Book Kury 08 PDF

Fear of Crime Punitivity
New Developments in Theory and Research

Helmut Kury (Editor)
Universittsverlag Brockmeyer
Bochum, 2008
Fear of Crime Punitivity

New Developments in Theory and Research
Helmut Kury (Editor)

Contents ................................................................................................................. V
Kerner, Hans-Jrgen
Foreword...............................................................................................................IX
Kury, Helmut
Introduction ....................................................................................................... XIII
Fear of Crime .................................................................................................. 1

Gray, Emily Jackson, Jonathan Farrall, Stephen
Researching Everyday Emotions: Towards a Multi-disciplinary
Investigation of the Fear of Crime.......................................................................... 3
Sessar, Klaus
Fear of Crime or Fear of Risk? Some Considerations Resulting from
Fear of Crime Studies and Their Political Implications ....................................... 25
Vanderveen, Gabry
How to Measure Fear of Crime in a Valid and Reliable Way.......................... 33
Kury, Helmut Obergfell-Fuchs, Joachim
Measuring the Fear of Crime. How Valid are the Results?.................................. 53
Kaal, Hendrien L. Vanderveen, Gabry N.G. van Oeveren, Moniek E.
Spatial Perception of Unsafety, Crime and Disorder ........................................... 85
Hirtenlehner, Helmut
Vulnerability Mediating the Perceived Risk Fear of Victimization
Linkage? Testing a Transactional Theory of Fear of Crime Using Data
from Austria........................................................................................................ 107
VI
Contents
Disorder, Social Anxieties and Fear of Crime. Exploring the Relationship
between Incivilities and Fear of Crime with a Special Focus on
Generalized Insecurities ..................................................................................... 127
Zarafonitou, Christina
Fear of Crime and Victimisation: The Greek Experience.................................. 159
Meko, Gorazd Kovo-Vukadin, Irma Muratbegovi, Elmedin
Social-Demographic and Social-Psychological Perspectives of Fear
of Crime in Slovenia, Croatia and Bosnia and Herzegovina.............................. 173
Punitivity ....................................................................................................... 197

Maruna, Shadd Matravers, Amanda King, Anna
Disowning Our Shadow: A Psychoanalytic Approach to Understanding
Punitive Public Attitudes.................................................................................... 199
Maruna, Shadd King, Anna
Public Opinion and Community Penalties ......................................................... 219
Green, David A.
Political Culture and Incentives to Penal Populism ........................................... 251
Kury, Helmut Obergfell-Fuchs, Joachim
Methodological Problems in Measuring Attitudes to Punishment
(Punitivity).......................................................................................................... 277
Obergfell-Fuchs, Joachim
Punitivity within the Criminal Justice System in Germany............................ 303
Serrano-Maillo, Alfonso Kury, Helmut
Insecurity Feelings and Punitivity: Relationship in a National Sample of
Adolescents and Young Adults in Spain ............................................................ 321
Hartnagel, Timothy F. Templeton, Laura
Perceptions, Emotions and Experiences of Crime: Effects on Attitudes
Toward Punishment in a Canadian Sample ....................................................... 349
Kossowska, Anna Rzeplinska, Irena Wozniakowska, Dagmara
Klaus, Witold
Criminal Policy Based on Fear of Crime Case of Poland .............................. 371
Contents
VII
Yoshida, Toshio
Problems Associated with Harsher Sanctioning. Trends in Returning to
more severe punishment in Japan....................................................................... 393
Which Sentencing Goals Do Victims of Crime in Austria Support?
Impressions from an Urban Victim Cohort ........................................................ 425
Elffers, Henk de Keijser, Jan W.
Different Perspectives, Different Gaps. Does the General Public Demand
a more Responsive Judge?.................................................................................. 447
De Keijser, Jan W. Elffers, Henk van de Bunt, Henk G.
Responsive but Misunderstood. Dutch Judges on Their Relation
to Society ............................................................................................................ 471
List of Authors ................................................................................................. 489
VIII
Contents
IX
Foreword
HANS-JRGEN KERNER
This is a very timely publication. The editor, Helmut Kury, has for decades been
one of the best known scholars in the fields of studying both, fear of crime and
punitivity either in the general population or among particular social strata or
societal sub-groups.
As chair of the organisation committee of the 6th Annual Conference of the
European Society of Criminology (ESC), taking place in August 2006 in Tbingen, Germany, and as then President of this scholarly association, I was already
highly content to see that Helmut Kury efficiently and effectively assembled
many people in well attended workshops on the issues which can be regarded in
conjunction under the umbrella perspective of peoples attitudes about the impact
of vaguely envisaged or concretely expected crimes on their physical and psychological well being, and their ideas of what could or should become the primordial means of responding to crime, either at large and/or on the individual
level.
In this volume Kury brings together and presents to the public an impressive
selection of updated and refined contributions to the Tbingen Conference workshops, which represent thought provoking recent research studies in different
countries, dealing with methodological problems, substantial results, and policy
implications alike.
Fear of Crime and Punitivity are to be considered as already highly complex phenomena in themselves. Their facets belong to divergent fields of scientific study and everyday life experience or common sense opinions alike. They
lead us from challenging questions of personality types to the psychology of
emotions, to life events and individual variations in coping with them, to small
group dynamics, to social-psychological structures and inter-group dynamics, to
anthropological basic beliefs about the human condition, to central moral convictions, and not the least to the politics of public outrage in cases of severe
crimes, just to name but a few of those facets.
Is fear of crime inseparably immersed in peoples fears of other menacing developments or events, including hard strokes of fate like losing a beloved partner
Foreword
or family member, and therefore always tainted with how individuals had been
able to cope with manifold past experiences and how they anticipate what will
occur to them individually, in their small group context or at societal level? Or is
fear of crime, to what small extent ever it might be influenced in detail by other
issues, eventually nevertheless a stark phenomenon of its own that can validly
be measured and used for different research, practice and policy endeavours? Is
there an interaction of the state of objective vulnerability of a person, her or his
personal assessment of this objective state, and general social anxiety on the one
hand, and socio-spatial particularities of the place of living or the broader environment on the other hand? Is fear of crime being more heavily influenced by the
perception of public decay and incivilities than by heavy real crime?
The contributors to this volume provide the reader with interesting substantial
information on those and related questions, but do not forget to point to still existing uncertainties and contradictions which need to be clarified by further careful empirical research and devoted theoretical deliberation.
The same remarks apply to the contributions to the second part of this volume.
Here the authors deal with questions and problems like: What lies behind punitive public attitudes in terms of personality, psychodynamics, emotions, life experiences, mindsets, and political affiliation? How far do different political cultures restrict or stir generalized emotions capable to nourish penal populism? Do
average people really almost always prefer harsh punishment of offenders, including juvenile offenders, or are they prepared and willing to differentiate depending on the circumstances? Do they differentiate independently via commonsense or do they need enlightened clarification by experts? Or taking it the other
way round: Do politicians playing the card of the urge to get tough with crime
and to get rid of mollycoddling criminals nothing but respond to strong and
deep seated public belief and sentiments, or do they instead exactly in the opposite instigate those beliefs and sentiments by their own activities, e.g. via
fuelling exaggerated debates in the public discourse, knowing well above all that
crime pays in the mass media and, by doing so, serving their own needs e. g.
with regard to election or re-election politics? Are victims of crime more inclined
towards harsh penalties as is the general population, or do they become perhaps
less fearful and even less punitive after having experienced concrete criminal acts
and offenders? What do people expect the criminal justice system to do in terms
of sanctioning offenders, and do they really demand a harsh sentencing practice
among judges? And what about judges themselves: Are they interested to learn
about the beliefs in and the demands of the population at large and, how precisely
do they know about the state of the nation so far and, eventually, how far are they
Foreword
XI
inclined and willing to respond in their everyday decision making and meting
out of sentences?
I wish the editor and all the contributors a wide spread and open-minded audience, and further success in unfolding the manifold aspects of these themes
which are pivotal for better knowledge and practice in criminology, victimology,
criminal justice, and crime and victim policy in our late modern societies.
XII
Foreword
XIII
Introduction
HELMUT KURY
Fear of crime and punitivity have sparked much criminological research, particularly in the industrial world and especially in Europe after the emergence of the
European Union and the collapse of the Soviet Union. Vast population movements, an ageing population, an increasing flow of goods from Russia and the
Far East together with the political and social realignment of European states after WWII all have contributed to the publics concern with these topics.
Such broad social changes are often accompanied by growing crime rates and
a vague sense of insecurity 1 . In the United States similar concerns have resulted
in a broad increase in punitivity toward offenders, which in turn has resulted in a
vast increase in Americas prison population. In Europe and in Japan many of the
same changes have occurred though not to the same degree as in the United
States.
The result has been a wide-ranging public discussion throughout the world regarding the negative effects of a growing crime rate, which have in the main increased the discomfort of the interested public. These discussions regarding
crime and punishment have invited public officials to take a more punitive approach toward crime, and in particular for criminologists to define more precisely
the concepts of punitivity and the fear of crime.
This reader includes essays devoted to the Fear of Crime and Punitivity as
presented at the sixth annual conference of the European Society of Criminology
in Tbingen, Germany from August 26th to the 29th in 2007. It focused on Understanding Crime: Structural and Developmental Dimensions and their Implications for Policy. Many of the papers given at the conference are published here
for the first time together with two additional papers (those by Maruna and King;
and Matravers and King).
See Thome, H., Birkel, C. (2007). Sozialer Wandel und Gewaltkriminalitt. Deutschland,
England und Schweden im Vergleich, 1950 bis 2000. (Social Change and Violent Criminality, A Comparison of Germany, England, and Sweden, From 1950 to 2000). Wiesbaden: VS Verlag fr Sozialwissenschaften.
XIV
Introduction
Part I: Fear of Crime

Gray, Jackson and Farrall analyze the social and psychological dimensions of
the fear of crime, and they suggest that its study requires an innovative, interdisciplinary approach. The authors present an overview of contemporary research
on the issue, including an analysis of the concepts and methodologies particularly
useful in its study. They suggest that the fear of crime is not a decisive factor in
everyday life. Current research, they suggest, tends to overestimate the significance of crime fear.
Sessar points out that the fear of crime became controversial only when registered crime rose dramatically roughly 30 years ago. The concept itself was a key
factor in the universal effort to sharpen criminal sanctions. The governing assumption here suggested that the fear of crime was a natural reaction to the
growth of crime. But he points out that the fear of crime has a political dimension as well as a scientific dimension. Fear of crime is a convenient, conservative
strategy to sway people toward greater severity.
Vanderveen shows how the concept can be defined more precisely and measured more reliably. He points to the multiple dimensions of the concept fear of
crime, and he suggests that each must be measured in its own right to arrive at a
complete understanding of the concept.
Kury and Obergfell-Fuchs examine how well the standardized definition of
the fear of crime assesses the concept. First, they measured the fear of crime
in the usual way using a randomly selected population. Second, they used the
standardized version in face-to-face interviews together with background information on their experience with crime. Their results suggest that about 50% of
the responses to such surveys do not reflect personal victimizations. Other unrelated fears and anxieties are expressed in surveys that deal with the fear of crime.
Kaal, Vanderveen, and van Oeveren discuss the relation of the fear of crime
to the nature of the neighborhood the subjects live within. They take care to define exactly the borders of the neighborhood, and they are especially interested in
the kinds of neighborhoods in which the fear of crime is greatest. It is important
that those interviewed have a clear understanding of both the area of their
neighborhood and the degree of threat in that area. Both issues are obviously crucial to the research question, as is the gender of the respondent.
Hirtenlehner points out that personal vulnerability is often used to explain
a greater fear of crime in such groups as women, the elderly, or in poor people.
But he proposes two additional concepts, coping strategies and support networks,
that may also affect a fear of crime. He looks at this question by comparing those
with coping strategies in fearful situations with others who have support net-
Introduction
XV
works. His study provides a precise estimate of each in shaping the fear of crime.
In a second paper Hirtenlehner examines the relationship between fear of
crime and the experience of incivilities in an Austrian neighborhood. He suggests
that the former, incivilities, also contributes to a sense of insecurity. Thus, many
factors contribute to a sense of insecurity, which in turn is an important factor in
the fear of crime.
Zarafonitou describes her research in Greece in 2004 regarding the links between a fear of crime and actual victimization. Based this research in Athens she
suggests that the fear of crime arises from a broader sense of vulnerability in society. She also reports that victims of crime feel much less safe than non-victims.
Clearly a fear of crime is related to ones experience.
Meko, Kovo-Vukadin, and Muratbegovi surveyed independently the fear
of crime in Slovenia, Croatia, and Bosnia-Herzegovina. Their results suggest that
the fear of crime is highest among women, people who are not physically fit, the
unemployed, and those who describe their environment as dangerous. They
based their research on the work of Van der Wurff et al. who measured the fear
of crime by using social psychological methodology. Their findings are explained in terms of conditions in their respective nations.
Part II: Punitivity
Maruna, Matravers, and King argue that punitivity cannot be attributed simply
to the fear of crime or even to an actual victimization. Much research has shown
that attitudes toward crime and sanctions express feelings of insecurity and
anxieties brought about by broad social changes. Psychoanalysis can relieve
these free-floating anxieties, but the fact that different cultures express different
levels of punitivity suggests that these anxieties are not inborn. Punitivity depends on several factors: the level of insecurity in the neighborhood; a broad social sense of insecurity; the social isolation and loneliness of modern life; and the
erosion of supporting social networks. More research in needed on all these issues.
Maruna and King analyze the following issues: (1) what do we know about
public opinion regarding punitivity and non-custodial sanctions; (2) how can we
understand the distinctive attitudes of the public regarding punitivity, and (3)
how can we encourage the publics support for non-custodial sanctions. In their
own research they find that the public is not as punitive as political figures assume. Despite the key importance of punitivity to most nations, careful research
has only just begun on punitivity and public opinion. The people are not of one
mind on this issue. When giving their opinions about specific punishments, they
XVI
Introduction
are less punitive than is often assumed. Moreover, little research has been carried
out regarding the publics non-punitive attitudes. The publics attitudes and sensibilities here are crucial if reform in sanctions is to be carried out.
Green asks the question, how does political culture affect punitivity? He explores two cases in which children murder another. The two cases (one in England and the other in Norway) were handled differently by the media and the police. In England the media offered the public a full discussion of the incident
complete with a demand for harsh sanctions, while the Norwegian press downplayed the incident as a rarity deserving of expert analysis rather than public
comment. Thus, the two distinct political cultures: one in England (a majoritarian
democracy) and the other in Norway (a consensus of expert opinion) play a large
part in shaping the way in which similar incidents are handled differently. More
careful research into the impact of political and social systems on punitivity is
needed.
Kury and Obergfell-Fuchs probe the measurement of punitivity. Immediately
after WWII punitivity in Germany was stable, but in the next few years it weakened noticeably. In the last 15 years however, the public has become increasingly
punitive particularly toward sex offenders and especially during political campaigns. The authors research shows that public opinion toward punishment is
vastly over estimated.
Obergfell-Fuchs looked at the question of punitivity in Germany by examining the sentences handed down in German courts over the last several decades.
He found a gradual trend toward sharper punishments in recent years.
Serrano-Mallo and Kury examine the links between the fear of crime, a
sense of insecurity, and punitive attitudes among Spanish citizens. A comparison
of Germany and Spain via the 2005 International Crime and Victimization Survey shows that Spaniards tend to be less punitive than Germans both in their attitudes toward punishment and in their courts.
Hartnagel and Templeton studied punitivity in Canada and found that Canadians clearly favored harsh punishments even though they enjoyed a relatively
low crime rate and relatively little fear of crime. They presented a statistical
model to explain their results and found that age, victimization, cognitive, and
emotional fears were all relevant to punitivity. They suggest that the emotional
dimension of punitivity deserves further attention.
Kossowska, Rzeplinska, Wozniakowska, and Klaus studied the situation in
Poland and found that both the rate of serious crimes and the fear of crime have
increased recently. Despite the fact that the issue has become politicized since the
collapse of the Soviet Regime and the opening of the western border neither pub-
Introduction
XVII
lic opinion nor prosecutors and judges support any change in laws or punishments.
Yoshida from Japan reports that punivity gathered support toward the end of
the 1990s. The penal laws were broadly strengthened in 2005 and the period during which an offender can be prosecuted was extended. These changes were justified as measures to curb the rise of crime in Japan. The growing crime rate,
however, can be understood partially at least as a greater willingness by the public to report offenses, and the increasing punitivity in Japan is defended as a
measure to deter further crime and to reassure the public that offenders are being
treated severely. Still, the prisons are overflowing with offenders and resocialization programs within the prisons has become more difficult.
Hirtenlehner maintains that punitivity depends upon public opinion. He studied the attitudes of victims who participated in the prosecution of their offenders.
To what extent are changes in legal punitivity tailored to the expectation of victims in the courtroom? Do victims of different kinds of crimes have different
views regarding punitivity? His research, which was based upon mailed questionnaires, suggests that victims are satisfied with rehabilitative or remedial actions by the court for offenders. Victims supported most sentencing policies, i.e.
punitive as well as rehabilitative measures, but to different degrees.
Elffers and Keijser surveyed both judges and citizens regarding how strongly
judges should consider the publics views of punishment. They broke the citizens
down into different groups, and they found that on the whole the public wants
judges to follow the law and not public opinion. Judges however, view the public
as interested in shaping the views of judges. Although leniency in judges is severely criticized by the public, the public seems to be saying they want fairness
in sentencing, not severity or leniency.
Keijser, Elffers, and van de Bunt interviewed 28 judges in different regions
in the Netherlands. They found that judges were confident they could judge each
case independently and prudently despite public opinion. They felt that the
public is usually unaware of the several dimensions of a case that influence the
punishment handed down by a judge. Nevertheless, public criticism of the courts
leniency has increased, and judges are concerned by their poor reputation in the
public.
This volume offers a broad range of studies from the whole world that deal with
the fear of crime and punitivity. I want to thank all the authors for their careful
work and their help in making this symposium a success. I also want to thank
Professor Hans-Jrgen Kerner, who organized this conference in 2006 in Tbin-
XVIII
Introduction
gen, for his generous support and his Foreword to this volume. Finally, I want to
thank the Universittsverlag and Dr. Brockmeyer for his excellent cooperation in
publishing this volume in a book series.
Helmut Kury
Freiburg, January 2008
FEAR OF CRIME
Researching Everyday Emotions: Towards a Multidisciplinary Investigation of the Fear of Crime
EMILY GRAY, JONATHAN JACKSON

AND STEPHEN FARRALL
The fear of crime has long been a high profile issue across Europe, America and
elsewhere. As Farrall et al. (2000:399) observe, the fear of crime is now one of
the most researched topics in contemporary criminology, with the risk of crime
being seen as one of the most pressing concerns affecting peoples quality of life
(Hale, 1996; Vanderveen, 2006). Yet technical problems have also dogged this
field of enquiry. Some criminologists believe that methodological limitations
have posed serious implications for the validity of the body of knowledge that
public policy relies upon (Skogan, 1981; Bernard, 1992; Farrall and Gadd, 2004;
Lee, 1999; 2001). Others argue that theoretical under-specification has restricted
the breadth and depth of definition and explanation, leaving us with a contested
and congested concept (e.g. Girling et al., 2000). Certainly, public concerns and
perceptions seem messier and more multi-faceted than current methods and concepts disclose. The fear of crime has social and psychological dimensions that
require interdisciplinary analysis and innovative methodological inroads yet the
vast majority of research in this area has lacked such ambition.
In this chapter we review, from an interdisciplinary perspective, literature on
everyday emotions that has so far gone unexplored by those criminologists interested in the fear of crime. There is a large and rapidly expanding body of psychological research on emotion (see Davidson et al., 2003). But only recently have
scholars begun to build an understanding about the emotions people experience
during the course of their day-to-day lives. One of the largest studies of real-life
emotions involved a series of surveys which focused on describing the probabilities of experiencing certain emotions in everyday life and on the socio-demographic and situational factors that influence these probabilities (Scherer et al.,
2002). The body of work on everyday emotions has also encompassed studies in
psychology (Frijda, 1986), a number of ethnographic studies concerned with
how emotions work (Katz, 1999) how emotions are managed across the settings of everyday life (Hochschild, 1983) and the development of a diverse soci-
Emily Gray, Jonathan Jackson and Stephen Farrall
ology of emotions (Kemper, 1990; Wouters, 1992; Williams and Bendelow,

1998). These contributions are noteworthy; not least because they offer a conceptual vocabulary that may well have much to offer criminological research, but
also because they address pertinent methodological questions concerning research on emotions and therefore place a significant emphasis on ensuring the
ecological and external validity of results.
These studies have not only provided rich and valuable data, but have also developed a constructive discussion of both methodological and theoretical issues.
Such issues are pertinent to the study the fear of crime since they may help us to
formulate a more comprehensive picture of what fear of crime actually is as a
lived experience. This chapter begins with a short review of the conceptual and
methodological problems within the fear of crime literature, before considering
the different perspectives on emotion which have emerged more recently. We
assess how a multi-disciplinary analysis might facilitate a more theoretically and
methodologically robust interpretative framework. Lastly, we outline how new
methodological techniques employed by researchers studying everyday emotions
might be employed to go back to basics, as it were, to assess what fear of crime
measures are actually measuring.
The utility and validity of fear of crime research

We make no attempt here to review the vast literature on fear of crime research,
since it has been covered comprehensively by previous writers (Hale, 1996;
Vanderveen, 2006). Nevertheless, we will highlight key concerns about the validity of the instruments used (Bernard, 1992; Fattah, 1993; Skogan, 1981; Farrall et
al., 1997) and some of the main issues surrounding the theoretical underspecification within this field of enquiry (Sparks, 1992; Girling et al., 2000).
Since the 1960s, large-scale victimisation surveys have given rise to data on
various offence-specific fears (burglary, rape, car theft etc.), perceived risk of
crime and experience of victimisation. Governments in America and later in the
UK and Europe were eager to fund research to assess the impact of crime on
communities, public perceptions and quality of life. One of the key research
conundrums to emerge from these data was the mismatch between officially
modelled likelihood statistics (self reported victimisation) and lay perceptions
of risk. On the basis of simple mean group differences, it was claimed that perceptions of fear of crime amongst certain groups such as women or older people
were out of proportion with their objective risk. Researchers soon took up the
challenge to resolve why lay perceptions might overlap objective risk, however,
these rationality debates quickly reached controversial heights and fear of crime
Towards a Multi-disciplinary Investigation of the Fear of Crime
research gave way to a broad critique of the theoretical and empirical tools used
to explore public perceptions of crime and fear.
Soon after the first sweep of the British Crime Survey (BCS) Maxfield (1984)
identified that operationalizing the concept of fear of crime from BCS data was laden
with technical difficulties. To begin with, studies tended to employ a narrow range of
feelings focusing on fear or worry, but never anger, frustration or even irritation
which may well be as common a reaction, not to mention as socially significant as
worry (Ditton et al., 1999). There has also been considerable criticism of the leading
nature of the question wording such as how safe do you feel walking alone in this
area after dark? or how worried are you about ?. These questions infer that this is
an issue about which respondents ought to have an opinion and can encourage respondents to provide answers which confirm the bias contained in the question
(Tourangeau et al., 2000).
Surveys typically ask respondents to reflect and recall emotions they are not
currently experiencing; their answers are assessed in contexts outside the realm
in which these feelings occur. Despite this, there has been little exploration of the
extent to which situational fears might be a symptom of structural or social circumstances. Furthermore, until recently we have not known the frequency or intensity of the emotions reported; questions do not ask whether fear is experienced
on a constant basis throughout the day, or whether it is concentrated at certain
times or in particular places. Perceptions of likelihood of victimisation produce
similar global estimates without providing contextual detail. As Walklate
(2007) notes, technical drawbacks have been particularly problematic for women,
specifically because questions usually cast fear as taking place outside the home
(how safe do you feel walking around in the dark ?). Pain (1991) stresses,
that while many women might relate to feeling worried about being alone in a
dark space, the context of fear for a small but significant number of women is
directly situated in the home in the company of their partner. A failure to recognise the role of the family domain in generating worries and perceptions of
victimisation will prevent a study from incorporating a key variable in the nature
of womens fear of crime.
These issues raise fundamental questions of how we interpret data in light of
the deeply intricate nature of many independent variables. How does one deal
with the various responses to a limited range of questions concerning feelings
(i.e. fear, worry), cognitions (perceived risk, safety assessments) and behaviours
which are very likely to represent different things to different people? A further
difficulty is the willingness (or reservation) of survey respondents to honestly
reveal emotions or vulnerability, something that may be more relevant to men
more than women (Stanko and Hobdell, 1993; Sutton and Farrall, 2005). Similarly, Tulloch et al. (1998) and Greve (1998) argue that fear of crime should be
conceptualised as a set of inter-related cognitions, emotions and behaviours, yet
this is very rarely done.
It has been suggested that the survey is simply too crude to effectively capture
the publics emotional reactions to crime. Smaller scale qualitative studies have
indeed been better able to demonstrate the more intricate and nuanced nature of
crime fears (Bannister 1993; Girling et al., 2000). As research on public perceptions has advanced we have discovered that lay perceptions are considerably
more complex and multi-faceted than previously thought. For example, the claim
that certain groups experience irrationally high levels of fear based on flawed
perceptions of risk has been deconstructed by new research evidence of repeat
or hidden victimisation (see Stanko, 1988). Feminist researchers employing
ethnographic and life-history research and female interviewers, were able to draw
attention to the hidden nature of crimes (such as domestic and sexual violence)
which were not only concealed from police statistics, but to large scale victimisation surveys. Such discoveries effectively put to rest the idea that women (or
other vulnerable groups) often reported fears in wild excess of their objective
risk. In sum, smaller scale, detailed studies have uncovered various new dynamics underpinning public emotions about crime and have been able to demonstrate
how gender, family, class, age, race and location variables interact and influence
an individuals perception of risk.
Research has also indicated that broad social and political contexts can shape
individuals perceptions of risk. Ferraro (1995) developed casual models in which
socio-cognitive variables, such as perceived risk, were shown to mediate the impact of other relevant socio-demographic factors. Similarly, Jackson (2004) notes
that an erosion of trust in the capacity of the local community to manage the risks
of crime impacts on how residents understand and relate to their environment. In
short, fear of crime revealed itself to be a considerably more complex and mysterious proxy of public perceptions; not only was it able to measure discrete experiences of crime-related worries, but it automatically linked into a whole host
of micro and macro-level factors. As Girling et al. conclude:
Researchers have begun to own up to diminishing confidence that they
know precisely what it is being measured or that they are all capturing indices of the same thing. Whatever it might be fear turns out to involve
multiple dimensions (Ferraro and LaGrange 1987; Killias 1990), vicarious
as well as direct experiences (Skogan and Maxfield 1981; Smith 1986) and
be open to varying conceptualizations as an expression of uneasiness
(Donnelly 1988), a judgement of government competence to deliver collec-
tive security (Taylor et al., 1986) and an expression of powerlessness and

uncertainty (Smith 1989). In other words it involves abstractions from experience and hence by definition their meanings and attributions as well as
the immediate impact of victimisation as such (2000:15).
Towards an interdisciplinary approach to fear of crime

Thus far, we have identified some of the key methodological and conceptual
problems within fear of crime research. Crucially, we believe there is potential to
improve the validity and utility of fear of crime research by drawing on related
disciplines, and one way to proceed is to utilise the growth in knowledge concerning everyday emotions. This discussion will hopefully be of interest to our
colleagues studying the fear of crime, but also to social scientists across the spectrum, who may appreciate an interdisciplinary enterprise which takes seriously
the role of theory, research and emotion in analysing crime, society and
psychology simultaneously.
Emotion has become an increasingly prominent issue in the humanities, social
sciences and psychology (see Davidson et al., 2003; Turner and Stets, 2005), and
there is now a large and expanding body of research on emotion emanating from
these disciplines. In the criminological arena we have been able to draw upon the
legacy of Durkheim and Elias who debated the relationship between human emotions and crime, punishment and social control. More recently, Karstedt (2002)
insists, there has been a return of emotions in criminal justice via the increasingly prominent voice of the victim, the use of restorative justice and increasingly emotionalised cultures of late modern societies (see Wouters, 1986; Williams, 2001). However, despite the apparent emotionality of crime and criminology, there are relatively few empirical studies that investigate the part that the
broad emotional palette has to play in influencing our responses to crime or even
the causation of crime or violence. De Haan and Loader (2002) further concur
that the emotions remain a somewhat peripheral topic within theoretical criminology. Indeed, despite a small number of studies, focusing for example on
shame (see Braithwaite, 1989), fear (see Hale, 1996) and anger or hate (Ditton et
al., 1999; Frijda, 1986; Gadd, 2006), in criminological research there is little to
aid the understanding of the wide ranging emotional experiences people often go
through in the course of their day-to-day lives. Nevertheless, everyday, offenders, victims and witnesses bring their emotions to the court, hate crimes and
domestic violence offences are committed, offenders are asked to express their
shame, the public fear crime and ingest information on the latest crime story
and impassioned demands for order are made by citizens and professionals alike.
As highlighted, the relative neglect of emotions in criminological work has not

been replicated in other fields. Not only have there been a number of works in
psychology (Frijda, 1986), a robust sociology of emotions has developed in
Britain, Europe and North America (Kemper, 1990; Wouters, 1992; Williams
and Bendelow, 1998) and a number of intriguing ethnographic studies concerned
simply with how emotions work (Katz, 1999) or the real-life occurrence of
emotions during everyday life have emerged (Hochschild, 1983).
One of the leading works of everyday emotional experience thus far is the collection of studies conducted by Scherer et al. (2004). Their research focused on
describing how regularly certain emotions occurred in everyday life, and on identifying demographic, personal and contextual factors that influence these experiences. Methodologically, this work drew attention to the ecological and external
validity of the results, and posed basic but pertinent questions relevant to exploring everyday emotions. In sum, these rich and diverse studies have considered
what emotions are, and were able to develop a sophisticated understanding of the
emotional experiences people typically go through in the course of their day-today lives. As such, an imaginative and significant corpus of work exists upon
which an analysis of crime fears and criminology might usefully draw. Moreover, it is very likely that criminology will be able to contribute in unique and
important ways to debates about the role and sway of emotions in crime and society. The following sections review some of the main issues highlighted by contemporary research into everyday emotions and consider key implications for
fear of crime research.
Research on everyday emotions a complex picture

As noted above, Scherer et al.s influential work examined the incidence of emotions in ordinary everyday life, the potential risk factors and the typical appraisal and reaction patterns. The study employed a population survey methodology in which data were canvassed from more than 1,000 German and French
speaking adults. There were two waves of the study, conducted four years apart.
Participants were asked to report an emotional event that happened yesterday as
well as to verbally label the experience. The extent of emotional complexity
among their respondents was expressed by the fact that respondents used 775
different words and phrases to describe their emotions on the previous day. In
fact, this work demonstrated that there are numerous ways in which people interpret and understand their emotional responses. Scherer et al. (2004) stated that
there were as many emotions as there are appraisal combinations or interpretations. Along these lines Ben-Zeve (2000) and Katz (2004) have also highlighted
the complexity and enormous diversity of human emotion in normal day-to-day

life.
Nevertheless, despite the vivid assortment of emotions reported by participants
in Scherer et al.s study, both sweeps, four years apart, produced very similar
distributions of emotional descriptions and frequencies. This work mapped a
wide array of emotional activity, but also revealed structural regularity. Moreover, individuals were better able to manage and adapt to significant world
events than one might have expected.
Ben-Zeve (2000) distinguishes between five main types of everyday affective
phenomena: emotions (such as fear, envy, anger, and guilt), sentiments or enduring emotions (enduring grief or love), moods (such as being cheerful, gloomy,
down), affective traits, (such as trait anger or shyness etc) and affective disorders
(such as depression and anxiety). Although there is considerable overlap between
the various types, the distinction between them is important as it represents the
possible reasons why an individual might respond to stimuli in a certain way;
someone who is fearful of crime when being followed at night is different to
someone who is arbitrarily scared of crime due to a psychiatric condition.
According to Turner and Stets (2005), everyday emotions are influenced by a
range of factors, all of which should be taken into consideration. Emotions involve certain elements. These include: 1) biological reaction of key body systems
such as the nervous system in the brain or hormonal influences (raised heart rate,
tears etc); 2) socially constructed cultural definitions and checks on what emotions should be experienced and expressed in a given situation (social desirability); 3) the application of common linguistic labels provided by culture and an
individuals semantic ability to describe an event or feeling; 4) the external expression of emotions through facial, vocal and/ or linguistic actions and 5) perceptions and appraisals of contextual objects or events. In short, being in a state
of fearing crime should be a considered multi-dimensional event. However, the
authors also suggest that not all of these elements need to be present for emotions
to exist. For example, people have unconscious emotional memories which trigger biological actions, which may later prompt physical cues to themselves and
others. At other times, individuals may repress their emotions with the result that
they do not experience the emotion or signal it to others at any level.
How often do we experience emotions?

Wilhelm (2001) and Myrteks (2004) work conclude that most people experience
some kind of emotion on a daily basis. Similarly, Scherer et al. (2004) found that
one in two individuals is likely to experience at least one strong emotion on any
10
given day. The most frequent emotions were happiness (9.1%) and anger (8.6%),
which reflect results in similar studies. The most pertinent results to those of us in
the fear of crime arena indicate the low frequency of basic emotions, thought to be
central to everyday thinking1 . Fear was a relatively rare emotion; 1.2% of the respondents reported experiencing fear on the previous day, although anxiety was
more common (6.5%). Scherer et al. noted that while it is possible that less intense
experiences of fear did occur, they could not have been consciously remembered or
perceived. They conclude: Serious fear situations are few and far between in the
normal course of events, (2004: 520). Indeed, Averill (2004) has also reflected that
fear is an infrequent emotion and rarely experienced outright. Similarly, other basic
emotions registered low frequencies: love was 0.8%, hate 0.2%, jealously 0.2%,
hope 0.1%, envy 0.1%, shame 0.1% and contempt in at less than one tenth of one
percent. While these emotions may be considered normal expressive fodder in
popular culture and to have a strong motivational force on our daily behaviour, they
may actually be much less common than we think. Other commentators have mused
that we may expect certain emotions to be more prevalent than they actually are because of their symbolism and significance;
Fear, love, hate, jealousy, hope, envy, shame and contempt are perceived to
be frequent since they appear to have tremendous impact upon our behaviour. While these emotions seem to shape our behaviour in many circumstances, the findings of the study under discussion (Scherer et al., 2004) indicate that they may have less impact than we think. Similarly aeroplane accidents are perceived to be more frequent than their actual occurrence because their immediate impact the death of so many people at one time is
so great and attracts large media attention. Accordingly we consider their
impact upon our life as more profound than it actually is (Ben-Zeve and
Revhon, 2004: 583).
Perhaps this low occurrence of basic emotions (including fear) signifies the
growing importance of other social and cultural emotional considerations. It has
been purported that basic emotions were some of the most important, both qualitatively and quantitatively due to their function in individual and evolutionary
development. However, the environment we now live in has posed new emotional stimuli, emotional processes particularly complex social, political and
philosophical ones. As Ben-Zeve comments; social comparative concerns have
become as crucial as the self-preservative biological concerns and cannot be re1
Although there are cultural differences in how emotions are expressed and interpreted, it has
been widely excepted that some emotions are basic and universal. Many writers (see Darwin, 1872; Emde, 1980; Turner, 1996) have agreed that happiness, fear, anger and sadness
are universal (Turner and Stets, 2005).
11
duced to them (2000:104). However, Scherer et al. remind us, that despite being
a rare event, the significance of fear as an emotion should not be underestimated.
Fear is described as a phylogenetically continuous emotion which produces
important biological, emotional responses essential for the maintenance of health
and the avoidance of imminent danger; the fact that it is apparently a relatively
rare event, at least in modern western democracies, does not detract from its important role in the emotion repertoire of humans (Scherer at al, 2004:557).
What does it mean to be fearful?

If fear is rare, what does it actually mean to think about crime and fear of crime
in everyday life? Similar to the approaches adopted in the above-mentioned studies, Gabriel and Greve (2003) argue that fear of crime research needs to utilise
approaches that are sensitive to the intricacies of fear as experienced by individuals. They argue that it is essential to distinguish conceptually between fear which
is due to a personal trait and a momentary affective state. There are clear and
important differences, the authors maintain, between a fear of crime that passes
quickly in response to very specific surroundings, and a general disposition or
trait of being afraid of becoming a victim of crime. Individuals who have dispositional fear consider more situations as being significant indicators of crime, are
more likely to experience fear in a given situation and probably more intensely.
The importance of this contribution is that the distinction demonstrates that fear
of crime can be experienced very differently in terms of individual relevance,
explanation and consequences. Along these lines, some people may experience
fear on a regular on-going basis on account of their individual characteristics and
psyche perhaps parochially described as nervous or vulnerable, while others
are more likely to respond to specific stimulus.
Furthermore, not all behaviours indicative of fear will have been prompted by
a fearful experience. For example, taking out insurance, locking ones car and
home may be an everyday precautionary action taken to minimise ones risk of
crime, but is not necessarily provoked by a fear-inciting situation. These sorts of
safety precautions are taken daily by the vast majority of people without an accompanying emotional theme-tune similar to decisions to take exercise, moderate alcohol intake or eat well to maintain general well-being and avoid the onset
of ill-health.
Emotional digestion and appraisal theory

In responding to Scherer et al.s study, Goldie (2004) makes the point that emotional responses do not always immediately follow an event. For example, if you
12
are robbed on the street, you may not feel a sense of anger until much later perhaps days or months after the episode when the full impact of the action has
had time to settle and digest in your mind 2 . As Goldie explains;
In much of our life the life of the mind the event is not so tightly
connected to the emotion that it elicits. We feel emotions in looking backwards at events that took place long ago in our own lives: nostalgia, grief,
sadness, regret and shame
Goldie (2004) argues that emotion-eliciting events can consist of different types
of incidence, they may be felt in the moment or following an event or crime for
example; they may be registered by the individuals unconscious mind, recalled
later from memory or even fabricated and inflated by imagination and reappraised at will. Accordingly, some experiences of fear of crime may be difficult
for a respondent to anchor in their memory or talk about accurately. Indeed,
Goldie (2004) highlights the subjective interpretation of thinking about facts and
events that shape our emotional responses. A central point in emotion research
and specifically appraisal theory is that the very same event may be considered
highly significant by one individual and irrelevant by another, depending on their
personal, cultural and biological characteristics at that point in time. One person
may become scared of crime because there are numerous teenagers hanging
around the local streets, whereas others may feel a sense of belonging to a local
community where young people are visible and playing outside. Even imagining
a situation of potential danger may elicit fear or worry in one person but not in
another (Warr, 1984).
Interestingly, there are numerous studies which have documented profound
discrepancies between peoples concurrent and retrospective reports of emotional
experiences (Gilbery and Ebert, 2002; Robinson and Clore 2002a, 2002b). Researchers have concluded that this is due, in part, to the highly fluctuating nature
of emotions over time, as well as from place to place (Brandstatter, 1983;
Fredrickson and Kahneman, 1993); the nature of memory (Scherer et al., 2004)
and the way respondents integrate subtle nuances of their experiences when they
are recounting historical emotional events (Kahneman, 1994). Exploring how
respondents recall knowledge, Robinson and Clore (2002a, 2002b) distinguish
between episodic and semantic knowledge; when people report their current or
very recent feelings they are accessing their episodic memory which is
grounded in the specifics of time and place. Conversely, reports of feelings experienced in the past (more than two weeks previously) are drawing on seman2
Post-traumatic stress disorder, which may indeed follow an incident of serious personal
violence, is not, by definition, felt until some time after the event.
13
tic knowledge. This information is conceptual in nature and draws upon peoples general beliefs related with this particular event. The actual experience may
not figure prominently in these semantic reports because the experience is no
longer accessible to the memory. Indeed, semantic knowledge allows respondents to characterise themselves, in general. This dynamic is supported by similar studies which have confirmed that people have fashioned beliefs and ideas
about themselves which can be divorced from experiences in their everyday life
(Klein, Babey and Sherman, 1997; Marsh and Yeung, 1998). Moreover, Robinson and Clore (2002a) maintain that individual beliefs about emotion are more
heavily influenced by semantic, rather than episodic knowledge. As such, when
surveys pose the question how worried are you about burglary the wording of
this question is more likely to illicit vague global summaries of intensity of
worry or fear. Warr (2000) suggests that these summaries represent future orientated anxiety rather than a summary of past episodes or current feelings of physical fear (see also Sacco, 2005). However, a subtle but nevertheless significant
rewording of this question which grounds the respondent into a specific time or
place, such as in the last two weeks have you ever felt worried about burglary
is likely to prompt time-framed dependent experiences.
Critically, fear of crime research has already encountered a divergence between self-reported levels of fear. Farrall et al. (2007) found that more general
estimates of fear of crime did not neatly map onto experiential time-focused accounts of crime-fears. Specifically this work found that some 88% of respondents
confirmed that they were worried (to some degree) about domestic burglary
when presented with a standard (semantic) survey question perhaps because
they have a vivid and accessible image of risk. Yet when asked an additional
time-limited question, 65% of the same sample reported that in the past 12
months that they had not experienced any events of worry about burglary. This
study concluded that the results point towards of two manifestations of fear one
based on expressive or semantic inspired fear, and the other more grounded in
the daily experience of everyday life. Fear of crime they maintain;
represents a continuum of feelings which are distributed along a spectrum
between two distinctly different ideal typical emotional reactions. At one
end, the most emotive aspect is the experience(s) of having felt fearful in a
specific situation. At the other, is a set of attitudes or opinions which are
brought forth when people are asked to discuss their feelings about crime.
This invocation of attitude which surveys provoke and measure we
shall refer to as the expressive dimension of the fear of crime (in contrast
to the experiential dimension) (2007:1).
14
Farrall and colleagues explored the idea that respondents who reported expressive fear, but who had not experienced any memorable crime fears in the past
year were comparable to the worried well in the health literature a concept
which describes people who express anxiety about their health despite not having
any appropriate symptoms (see Garfield, 1970; Lombardo, 2004). They suggest
that the responsibilisation of individuals to protect themselves from crime and
the high-profile circulation of images of crime in society have raised the general
salience and prominence of risk. They go onto suggest that the worried well
who are relatively protected from crime may, nevertheless feel at risk and obligated to protect themselves from the future victimisation. Moreover, crime is a
richly symbolic issue and public sensibilities about crime and fear of crime may
express their assessments of cohesion, social control and civility (see Jackson,
2004). Sparks et al. (2001) agree that peoples appraisal of fear of crime is deeply
connected to their unconscious interpretation of their environment and space and
that crime discourse is a particularly pertinent area in which to discover the impact of social and political change on peoples everyday experience of life.
New directions in everyday emotions research: methodological

issues
While work in the field of everyday emotions has opened up new ways of looking at emotional responses and provided a multidimensional view of our emotional life, it has also raised a series of methodological questions concerned with
the internal as well as the construct validity of such methods. If everyday emotions are revealed to be complex thought processes which individuals experience
in very personal and diverse ways, exploring the incidence and experience of the
fear of crime is also a multifaceted and methodologically challenging arena. Being alert to the issues identified here will not only increase the general validity of
any study, but also will better inform our understanding of how people understand, process and manage the fear of crime.
Evidently, we need to exercise greater sensitivity and seek more accuracy in
the way we understand and interpret survey reactions on the fear of crime. How
often do people actually feel fear are other less severe reactions more appropriate? How intense is this reaction and what are the implications for respondents
how do they cope with fear? If we were to let participants speak in their own
language would they use the term fear and if not, how would they describe their
feelings? Specifically, attempts to acquire knowledge about the frequencies of
emotions in daily life are confronted with conceptual problems, such as how
moods and unspecific emotional states can be distinguished from emotions. Even
15
though studies may ask respondents to only report incidences in which they felt
fear or worry, it is very likely that participants will also report what they perceive to be similar states, such as anxiety, dissatisfaction or mild concern for
example. Are we at risk of misinterpreting any reaction as a specifically fearful
one? Clearly, we need to pay close attention not just to the presence of these
emotions but also their intensity and frequency.
Interestingly, Katz advocates that as research strategies begin, researchers
should deliberately shed any assumptions about the topic, their field of work and
the theory or traditions surrounding these issues. He develops an alternative,
theoretically grounded research strategy, based on a theory of social ontology
that urges researchers to describe how social conduct is 1) created through symbolic interaction, 2) operationalised as a communicative action and 3) shaped by
corporeal processes. By developing a naturalistic social ontology one can guard
against artificial descriptions or meaningless data gathering. This process begins
with simply describing the phenomena as it is experienced by people in their everyday lives. Such an empirically and theoretically grounded description is the
first building block to theory development he stresses;
One initially asks not which theories the data validate and which they invalidate, but which ideas will best guide the description of social lie into
forms of data. Such a theory predicts nothing substantively differentiating
about the causes of the matters to be explained, but at the same time it does
predict something essential. The central claim is that unless the researcher
describes phenomena according to their nature, explanatory theories will
surely be wrong (Katz, 2002: 258).
In order to produce a more authentic picture of what fear of crime actually means
to respondents, data collection efforts need to minimise the potential for measurement errors in survey based studies. Methodologists have long been concerned with the validity of the results from surveys, since measurement errors
can easily affect the validity and reliability of the data. Cognitive interviewing
techniques were designed by an interdisciplinary team of methodologists and
psychologists in the 1980s for the very purpose of estimating the impact of
measurement errors. Specifically, the overall aim is to use cognitive theory to
understand how respondents perceive and interpret questions and to identify
potential problems that may arise in prospective survey questionnaires. For
example, an interviewer will probe the comprehension of the question; the ability
of the respondent to recall relevant information; the decision process involved in
reaching answers to questions and how these processes operate in the context of a
research process. Cognitive interviews are a useful means for pre-testing
questionnaires, particularly where the subject matter is sensitive or complex. As
16
such, cognitive interviewing techniques might be a particularly valuable method

in fear of crime research, usefully employed to allow researchers to go back to
basics and investigate what fear of crime tools are actually measuring.
This chapter has demonstrated that how we conceptualise fear and worry in
criminological circles and within popular culture requires thoughtful expansion.
Fear of crime has commonly been considered in exclusively negative terms, as a
social problem which impacts of peoples quality of life and sense of safety.
However, some degree of worrying about crime is both an inevitable and ordinary experience for most people. As Gladstone and Parker (2003) note, worry is
not a typically pathological process, it has functional properties, such as prompting the mental reminder to lock the house or check the burglar alarm. When people say they are worried about crime it is very likely that some of these respondents at least are describing a type of worry which is not detrimental to their well
being, but has motivational qualities which leads them to take common precautions, such as locking their car, taking insurance or purchasing security lights for
their property. These actions may indicate they are aware or alert to the possibility of becoming a victim, but do not necessarily mean these respondents are experiencing feelings of fear, anxiety or vulnerability. Indeed, it is clear that further
research is required to explore the multiple personalities of fear and what it
represents to people in the course of their daily lives.
Work in emotional research also suggests that respondents are more likely to
recall negative experiences. If a researcher asks a respondent about their last experience of fear or worry about crime they may only recall the most memorable
and possibly the most serious experience. For example, a respondent is more
likely to remember a time when they were concerned about being robbed, rather
than a time when they briefly thought something may have been stolen from their
car. Scherer et al. (2004) report that respondents have a tendency to associate the
term event in a survey with the most dramatic incidence they can recall. As
such, delicate internal thoughts, perhaps about less serious crimes or more subtle
worries are easily missed. Accordingly it is essential that 1) different crimes are
considered separately and 2) we assume that less intense experiences of fear of
crime are not consciously remembered and consequently have a reduced impact.
Everyday emotion research has also revealed that respondents are simply not
very good at recalling emotional responses for survey purposes. The recall of an
event or experience depends on its proximity and recent events are more accessible. One method for minimising response error is to employ a narrow time frame
in interview or survey questions which limits respondents to a specific reference
period which they will be able to recall more accurately. For example, in the last
17
12 months have you worried about? Nevertheless, retrievals are influenced by

a tendency to reconstruct events or experiences so as to make them consistent
with memories and subsequent interpretations. Respondents are unlikely to be
succinct or clear when describing their past emotional experiences in which numerous thoughts and feelings may have been intertwined. Consider this example:
a person may become worried that their car has been stolen when they cannot
locate it in an unfamiliar car park they find disorientating, he or she becomes
scared of not being able to get home late at night and then feels concerned about
the consequent financial implications as they search for their car. Eventually the
person is resigned to the fact that the car has been stolen and begins to think
about making alternative arrangements and later relieved when they realise they
were confused and it was parked on a different floor. In this situation a respondent has gone through a range of brief emotional encounters and asking them to
summarise this into one word or phrase is highly reductive. Indeed, much of the
difficulty in defining and studying emotion is due to the extraordinary changeability of the process, which can be hard to pin down into specific researchable
chunks. The use of multiple research methods including both qualitative and
quantitative techniques which allow respondents more freedom to define their
own and various emotional reactions to crime may help us to better understand
the properties of fear of crime and the consequent impact of such emotions on
individual perceptions.
The prolific work of Jack Katz provides imaginative methods for exploring
emotions as they occur in everyday life. Katz (2004: 611) argues that formalised
survey research ... functions as a surgical courier service for distant intellectual
audiences, hastily cutting out and neatly packaging experiences that have taken
subjects a messy lifetime to form. He describes how social phenomena can be
more usefully approached in four alternative ways: ethnographically, interactionally, diachronically, and with attention to corporeal practices. To be precise, his
study on how the emotions work (1999) uses in-depth reports and videotapes of
participants to investigate the emotional life of people in their daily routines. By
using video, he was able to capture the invisible features of events and emotional dilemmas, and as such was able to provide a unique glimpse into the hidden,
perhaps unexpected aspects of a persons life and thought processes. One chapter
from this work focuses on crying and he demonstrates the wide variety of
circumstances in which people cry. His examples demonstrate how crying might
be spontaneous, controlled or contrived and may express latent emotions as well
as immediate concerns. Similarly, exploring perceptions of disorder in Chicago,
Sampson and Raudenbush (2004) supplemented a neighborhood survey with a
18
systematic social observation. This involved their research team driving a sports
utility vehicle at five miles per hour down every street in a sample of Chicago
neighbourhoods. On both sides of the vehicle a video camera recorded captured
social activities and physical features, while a trained observer completed a log
for each block. By the end of the study, the research team had observed 23,816
blocks. A random sample of 15,141 logs and videos were then coded to measure
features of the streets, buildings, businesses, and social interaction. Sampson and
Raudenbush (2004) claim the experimental nature of their design was able to
generate a more sophisticated interpretation of how residents understand their
immediate environment by exploring the correlates for perceived disorder, while
controlling for reliably observed (videoed) disorder. Intriguingly, they concluded
that economic deprivation and racial diversity had the greatest influence on perceptions of disorder and that residents clearly supplement[ed] their knowledge
with prior beliefs informed by the racial stigmatization of modern urban ghettos
(2004:336).
Evidently, as long as it is possible to make the requisite observations, direct
observation is often a useful method for measuring behaviour in natural settings.
Indeed, naturalistic studies produce thick, detailed descriptions of the context,
stimuli and reach of human emotion in everyday life. In fear of crime research
this approach may be useful in identifying more precisely what fear, crime and
society actually mean to people. In so doing it would go beyond its narrow
parameters and would allow respondents to use their own words (and actions) to
describe what might be better classified as concern, panic, a passing thought,
worry, anger or frustration.
Conclusion
The fear of crime is a topical issue of social and political relevance. It has attracted a great deal of research. However, much of this work seems hampered by a
lack of clarity and ambition. It is even possible that standard approaches inadvertently exaggerate the fear of crime problem (Lee, 1999, Farrall and Gadd, 2004);
distort the nature of fear as it is experienced in everyday life (Gabriel and Greve,
2003; Farrall et al., 1997) and fail to recognise the functional as well as expressive aspects of the language of crime fears (Jackson, 2006). Throughout our recent work we have tried to develop the argument that researching emotional responses to crime involves important psychological and social considerations.
This chapter has given a flavour of some of our thinking in this regard. If we are
to continue with this endeavour, ways forward might include drawing upon the
research on the psychology of emotional self-report and risk, the sociology of
19
everyday emotions and public sensibilities to crime. Our own recent work has
sought to identify two manifestations of the fear of crime; one is intimately
grounded in the daily experience of everyday life, the other is better described as
more diffuse anxiety expressive fears regarding the cultural meaning of crime
(Farrall et al., 2006; 2007; Gray et al., 2006; Jackson et al., 2006). The fear of
crime is a fascinating social phenomenon that might be advanced by a multi-disciplinary dialogue capable of deepening our understanding of the wide variety of
emotions that resonate through our thoughts, perceptions and discussion of
crime. In so doing, we may invigorate some stagnant areas of theory and research
and open up new lines of substantive enquiry.
20
Bibliography
Averill, J. (2004). Everyday emotions: let me count the ways. Social Science Information. Vol 43 (4) 571-580.
Bannister, J. (1993). Locating fear: Environmental and ontological security. In H. Jones
(Ed.), Crime and the urban environment. Aldershot: Avebury.
Bernard, Y. (1992). North American and European Research on Fear of Crime, Applied
Psychology: An International Review, 41, 65-75.
Ben-Zeve, A. (2000). The Subtlety of Emotions. Cambridge, MA: MIT Press.
Ben-Zeev, A., & Revhon, N. (2004) Emotional complexity in everyday life. Social Science Information. Vol 43 (4) 581-589.
Braithwaite, J. (1989). Crime, Shame and Reintegration. Cambridge: Cambridge University Press.
Brandstatter, H. (1983). Emotional responses to other persons in everyday life situations. Journal of Personality and Social Psychology, 45, 871883.
Darwin, C. (1872). The expression of the emotions in man and animals. London: Oxford University Press.
Davidson, R.J., Scherer, K.R. & Goldsmith, H.H. (Eds.) (2003). Handbook of Affective
Sciences. New York: Oxford University Press.
De Haan, W. & Loader, I. (2002). On the Emotions of Crime, Punishment and Social
Control, Theoretical Criminology, 6, 3, 243-253.
Donnelly, P.G. (1988). Individual and Neighbourhood Influences on Fear of Crime.
Sociological Focus, 22(1), 69-85.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1999). Afraid or angry? Recalibrating the fear of crime. International Review of Victimology, 6, 83-99.
Emde, R.N. (1980). Levels of meaning in infant emotions: A biosocial view. In Collins,
W. A. (Ed.), Development of cognition, affect, and social relations: The Minnesota
symposium of child psychology. Hillsdale, NJ: Lawrence Erlbaum.
Farrall, S. (2004). Revisiting Crime Surveys: Emotional Responses Without Emotions, International Journal of Social Research Methodology, 7(2): 157-71.
Farrall, S., Bannister, J., Ditton, J. & Gilchrist, E. (1997). Questioning the Measurement of the Fear of Crime: Findings From A Major Methodological Study, British
Journal of Criminology, 37, 4: 657-78.
Farrall, S., Bannister, J., Ditton, J., & Gilchrist, E. (2000). Social psychology and the
fear of crime: Re-examining a speculative model, British Journal of Criminology,
40, 399-413.
Farrall, S. & Gadd, D. (2004). The Frequency of the Fear of Crime, British Journal of
Criminology, 44, 1: 127-132.
Farrall, S., Jackson, J. & Gray, E. (2006). Everyday Emotion and the Fear of Crime:
Preliminary Findings from Experience and Expression, (PDF) Experience and Expression in the Fear of Crime: Working Paper #1. Available at; http://www.
lse.ac.uk/collections/methodologyInstitute/pdf/JonJackson/E&E%20WP1.pdf
21
Farrall, S., Gray, E. & Jackson, J. (2007). Combining the Old and New Measures of
the Fear of Crime: Exploring the Worried Well, (PDF) Experience and Expression
in the Fear of Crime: Working Paper #4. Available at: http://www.lse.ac.uk/collections/methodologyInstitute/pdf/JonJackson/E&E%20WP4.pdf
Fattah, E.A. (1993). Research on fear of crime: Some common conceptual and measurement problems. In W. Bilsky, C. Pfeiffer, & P. Wetzels (Eds) Fear of crime and
criminal victimisation Stuttgart: Ferdinand Enke Verlag.
Ferraro, K.F. (1995). Fear of crime: Interpreting victimization risk. SUNY Press.
Ferraro, K.F. & LaGrange, R.L. (1987). The measurement of fear of crime. Sociological Inquiry, 57(1), 70-101.
Fredrickson, B.L., & Kahneman, D. (1993). Duration neglect in retrospective evaluations of affective episodes. Journal of Personality and Social Psychology, 65, 4555.
Frijda, N.H. (1986). The Emotions. Cambridge: Cambridge University Press.
Gabriel, U. & Greve, W. (2003). The psychology of fear of crime: conceptual and
methodological perspectives. British Journal of Criminology. Vol 43, 600-614.
Gadd, D. (2006). The role of recognition in the desistance process: a case study of a farright activist, Theoretical Criminology, 10(2): 179-202.
Garfield, S.R. (1970). The Delivery of Medical Care. Scientific American 222: 523.
Gilbert, D.T. & Ebert, J.E. (2002). Decisions and revisions: the affective forecasting of
changeable outcomes. Journal of Personality and Social Psychology, 82. 503-514.
Girling, E., Loader, I. & Sparks, R. (2000). Crime and Social Change in Middle England, Rotuledge, London.
Gladstone, G. & Parker, G., (2003). What's the use of worrying? Its function and its
dysfunction, Australian and New Zealand Journal of Psychiatry, 37 (3) 347-354.
Goldie, P. (2004). The life of the mind: commentary on Emotions in everyday life
Social Science Information. Vol 43 (4) 591-598.
Gray, E., Jackson, J. & Farrall, S. (2006).Reassessing the Fear of Crime: Frequencies
and Correlates of Old and New Measures, (PDF) Experience and Expression in the
Fear of Crime: Working Paper #3. Available at:
www.lse.ac.uk/collections/methodologyInstitute/pdf/JonJackson/E&E%20WP3.pdf
Greve, W. (1998). Fear of crime among the elderly: foresight, not fright. International
Review of Victimology, 5(3-4), 277-309.
Hale, C. (1996). Fear of Crime: A Review of the Literature, International Review of
Victimology, 4: 79-150.
Hochschild, A. (1983). The Managed Heart: The Commercialization of Human Feeling.
Berkeley, CA: University of California Press.
Jackson, J. (2004). Experience and Expression, British Journal of Criminology, 44:946- 66.
(2006). Introducing Fear of Crime to Risk Research, Risk Analysis, 26, 1, 253-264.
Jackson, J., Farrall, S. & Gray, E. (2006). The Provenance of Fear, (PDF) Experience
and Expression in the Fear of Crime: Working Paper #2. Available:
www.lse.ac.uk/collections/methodologyInstitute/pdf/JonJackson/E&E%20WP2.pdf
22
Kahneman, D. (1994). New challenges to the rationality assumption. Journal of Institutional and Theoretical Economics, 150, 1836.
Karstedt, S. (2002). Emotions and Criminal Justice Theoretical Criminology Vol.6 (3)
29-317.
Katz, J. (1999). How Emotions Work. Chicago, IL: University of Chicago Press.
(2002). Start Here: Social Ontology and Research Strategy, Theoretical Criminology,
6 (3): 255-278. 2002.
(2004). Everyday lives and extraordinary research methods. Social Science Information. Vol 43 (4) 609-619.
Kemper, T. (Ed.) (1990). Research Agendas in the Sociology of Emotions. New York:
SUNY Press.
Killias, M. (1990). Vulnerability: towards a better understanding of a key variable in the
genesis of fear of crime. Violence and Victims, 5, 97-108.
Klein, S.B., Babey, S.H. & Sherman, J.W. (1997). The functional independence of trait
and behavioural self-knowledge: Methodological considerations and new empirical
findings. Social Cognition, 15, 183203.
Lee, M. (1999). The fear of crime and self-governance: Towards a genealogy. The Australian and New Zealand Journal of Criminology, 32, 3, 227-246.
(2001). The genesis of fear of crime, Theoretical Criminology, 5, 467-485.
Lombardo, A.P. (2004). Anatomy of fear: Meads theory of the past. and the experience
of the HIV/AIDS worried well. Symbolic Interaction, 27, 531548
Marsh, H.W. & Yeung, A.S. (1998). Top-down, bottom-up, and horizontal models: The
direction of causality in multidimensional, hierarchical self-concept models. Journal
of Personality and Social Psychology, 75, 509527.
Maxfield, M. (1984). Fear of Crime in England and Wales, HMSO, London
Myrtek. M. (2004). Heart and Emotion. Ambulatory Monitoring Studies in Everyday
Life. Gottingen: Hogrefe and Huber.
Pain, R. (1991). Space, violence, and social control: integrating geographical and feminist analyses of womens fear of crime, Progress in Human Geography, 15: 415-31.
Robinson, M.D. & Clore, G.L. (2002a). Belief and Feeling: Evidence for an Accessibility Model of Emotional Self Report, Psychological Bulletin, 128: 934-960.
(2002b). Episodic and Semantic Knowledge in Emotional Self-report: Evidence for
Two Judgment Processes, Journal of Personality and Social Psychology, 83: 198215.
Sacco, V. (2005). When Crime Waves, London: Sage Publications.
Sampson, R.J. & Raudenbush. S.J. (2004). Seeing Disorder: Neighbourhood Stigma and
the Social Construction of Broken Windows. Social Psychology Quarterly 67: 319342.
Scherer, K.R., Wranik, T., Sangsue, J., Tran, V. & Scherer, U. (2004). Emotions in everyday life: probability of occurrence, risk factors, appraisal and reaction patterns. Social Science Information. Vol 43 (4) 499-570.
23
Skogan, W. (1981). Issues in the measurement of victimisation US Department of Justice). Washington, DC: US Government Printing Office.
Skogan, W. & Maxfield, M.G. (1981). Coping with Crime: Individual and Neighbourhood Reactions, Beverly Hills: Sage.
Smith, S. (1986). Crime, Space and Society. Cambridge: Cambridge University Press
(1989). Social Relations, Neighbourhood Structure and the Fear of Crime in Britain, In
Evans, D.J. & Herbert, D.T. (Eds.), The Geography of Crime. London, Routledge.
Stanko E. (1988). Hidden violence against women. In Maguire, M., & Pointing, J.,
(Eds) Victims of Crime. A New Deal? Milton Keynes: Open University Press.
Stanko, E. & Hobdell, K. (1993). Assault on Men: Masculinity and Male Victimization,
British Journal of Criminology, 33, 400-415.
Sparks, R. (1992). Reason and unreason in left realism: some problems in the constitution of the fear of crime, in: Matthews, R. & J.Young, J. (Eds) Issues in Realist Criminology, 119135. London: Sage.
Sparks, R., Girling, E. & Loader, I. (2001) Fear and Everyday Urban Lives, Urban
Studies, Vol. 38, No 56, 885898, 2001
Sutton, R. M. & Farrall, S. (2005). Gender, socially desirable responding and the fear of
crime: Are women really more anxious about crime? British Journal of Criminology,
45, 212-224.
Taylor, I., Evans, K. & Fraser, P. (1996). A Tale of Two Cities. London: Routledge.
Tourangeau, R., Rips, L. J. & Rasinski, K. (2000) The Psychology of Survey Response,
Cambridge. Cambridge University Press.
Tulloch, J. et al. (1998). Fear of Crime. Canberra: CCCRS.
Turner, J. H. (1996). The evolution of emotions in humans: A Darwinian- Durkheimian
analysis. Journal for the theory of social behaviour, 26, 1-33.
Turner, J.H. & Stets, J.E. (2005). The Sociology of Emotions. Cambridge: Cambridge
University Press.
Vanderveen, G. (2006). Interpreting Fear, Crime, Risk and Unsafety. Devon. Willan.
Walklate, S. (2007) Imagining the Victims of Crime. Berkshire. Open University.
Warr, M. (1984). Fear of Victimization: Why are Women and the Elderly More Afraid?,
Social Science Quarterly, 65, 681702.
(2000). Fear of crime in the United States: Avenues for research and policy. In Duffee, D. (Ed), Criminal Justice 2000. USA: National Institute of Justice.
Wilhelm, P. (2001). A Multilevel Approach to Analyze Ambulatory Assessment Data:
An Experiment of Family Members Emotional States in Daily Life, in Fahrenberg,
J., & Murtek, M., (Eds) Progress in Ambulatory Assessment: Computer Assisted
Psychological and Psychophysiological Methods in Monitoring and Field Studies.
Kirkland, WA: Hogrefe & Huber.
Williams, S. (2001). Emotions and Social Theory. London: Sage.
Williams, S. & Bendelow, G. (Eds.) (1998). Emotions in Social Life: Critical Themes
and Contemporary Issues. London: Routledge.
24
Wouters, C. (1986). Formalization and Informalization; Changing Tension Balances in

Civilizing Processes, Theory, Culture & Society 3(2): 1-18.
(1992). On Status Competition and Emotion Management: The Study of Emotions as
a New Field, Theory, Culture & Society 24(4): 699717.
25
Fear of Crime or Fear of Risk?

Some Considerations Resulting from Fear of Crime
Studies and Their Political Implications
KLAUS SESSAR
1. Fear of crime
Criminological investigations, like any other area of study in the social sciences,
try to find the truth (or at least some sort of true value). uninfluenced by the researcher and the way his or her observations are made. Attitudes such as punitiveness or fear of crime are commonly treated as being natural, in the sense that
their essence can be disclosed should appropriate methods be applied. That said,
the so-called standard question How safe do you feel walking alone in your
neighbourhood after dark has been highly criticised for not only failing to
measure fear of crime, but also for erroneously suggesting that fear of crime can,
indeed, be properly measured. Notwithstanding, a vast number of variations to
this standard question have been developed in an attempt to come closer to the
truth or reality. However, as is well known, different concepts, terms and methods usually produce different results. For example, according to the studies of
both Farrall and his colleagues in Great Britain (1997) and of Kury and his colleagues in Germany (2004). survey methods lead to a general over-counting of
fear when the answers are compared with those from qualitative interviews conducted with the same individuals a point that I will return to later.
For the moment it is important to note that the kind of observation, and
thereby the observer, are part of the process to find fear, not as a methodological
bias but as an intrinsic element of this process. Therefore, neither the researchers
personal interests nor their scientific background or skills are of any relevance,
the focus being rather on the political construction of the fear of crime problem
with the help of science. Dittons quote that there was no fear of crime in Britain until it was discovered in 1982 (Ditton et al. 1998: 10) may illustrate this
point. The issue first arose as crime began to increase in all Western countries. At

Paper presented at the 6th Annual Conference of the European Society of Criminology, Tbingen, 26 29 August 2006.
26
Klaus Sessar
the time, two main options were presented to deal with the problem. The first
option promoted the establishment of social programmes to support better education, the improvement of urban structures, anti-discrimination campaigns, etc.,
but above all else to tackle poverty. The second option was not a war on poverty
but rather a war on crime. This was to be brought about by increasing the power
of the police, by providing harsher sentences with eliminating tendencies or by
the application of zero tolerance measures to minor types of deviant behaviour.
To quote the then Senator Richard Nixon, that is, before his short presidency: If
the conviction rate were doubled in this country, it would do more to eliminate
crime in the future than a quadrupling of the funds for any governmental war on
poverty (quotation taken from a position paper called Toward Freedom from
Fear; see Harris 1969: 74). As is well known, law and order prevailed over social welfare in countering the perceived crime problem. And accordingly, fear of
crime became a similar target for criminal policy, that is, instruments to combat
crime also became instruments to combat the fear of crime.
There was and maybe still is a strong affinity between the political and the
criminological discourse regarding the almost uncontested connection of fear
with crime. The similarity was the basis of survey instruments used to confirm
this connection. The aforementioned standard question How safe do you feel ...
is one such instrument. It does not mention crime nor does it use the terms fear
or being afraid. Instead, it refers to darkness, to being alone and to walking
outside in the neighborhood. When combined, for some people these three items
are capable of sending a shiver down their spine even in the absence of any
threatening event. But the criminological interpretation of this kind of unsafety
was fear of crime: I feel unsafe because I am afraid of experiencing some sort
of violence, for example. (It should be mentioned that in former times unmarried
women were not allowed to walk alone outside their home after dark, see Ruhne
2002: 118; thus, the standard question has a gender-specific colouring from the
beginning.)
Another type of question closely linked to the first, is How likely do you
think the following crimes will happen to you in your neighbourhood in the near
future (or similarly worded). Again, if the expressed likelihood is high people
are said to be afraid of the streets, considering them to be dangerous. However,
should crime rates be low, motives other than fear should be assumed. For example, in a study that compared attitudes of East Germans with those of West Germans after the changes of 1989, 36% of women up to the age of 30 in major East
German cities felt that it was likely or very likely that they would be a victim of
rape (Boers et al. 1997); and in Krakow, only four years ago, 31% of 40 to 50
Fear of Crime or Fear of Risk? Some Considerations
27
year old women and 26% of 50 to 60 year old women considered it likely or very
likely that they would be raped. It is hard to believe that answers of this kind express fear (or worry, etc.); rather, they mirror some sort of uneasiness - or protest
- in view of the unbearable conditions in their communities or neighborhoods in
terms of social disorder (e.g., failure of collective efficacy) and the absence of
social control (Sessar 2007: 142).
Anyhow, both questions (How safe and How likely ) are mutually
supportive in that they assure each other a high degree of validity by means of
satisfying correlation coefficients. The problem however is that we still do not
know what we have measured when we have measured fear in this way.
This paper posits that largely independent of objects, persons or situations
which may induce all kinds of fear, fear of crime is first and foremost a construction; it is generated, not found. If we accept the social environment (the urban
space) as being formed and shaped by the people who have their existence there
(that is: the environment is not something given or natural). then it is not safe or
unsafe but is perceived as being safe or unsafe by the residents with their individual and social characteristics such as vulnerability or coping capabilities,
which in turn are linked to gender, age, social status or ethnicity (Jackson 2004).
Likewise the residents determine the impairment of their quality-of-life conditions. This is the very reason why the same objects, persons or neighborhoods are
considered to be unsafe or dangerous by one person but not by another; it is
much less why people have appropriate as opposed to inappropriate attitudes towards a dangerous thing.
It would be preferable, then, to understand fear of crime as a general concept
which is neither correct nor incorrect; rather, it is either useful or not useful, and
that is revealed by the results of the research (Skogan 1993: 131) a constructionist perspective.
2. Fear of risk
Insights of this kind must be treated carefully with regard to the political implications which have already been mentioned. Fear of crime is not only a scientific (a
criminological) issue, it is first of all a political one. Let us take the notion of risk
for a better understanding of the problem before coming back to fear risk in
terms of terrorism, transnational crime, mass unemployment, overpopulation,
migration, environmental degradation, nuclear contamination, etc. Risk societies
are mostly angst societies as we have learned from Ulrich Beck (1986; see also
Scott 2000); the problems of inequality in previous societies makes way for the
problems of anxiety and insecurity in modern societies. We are living in risk so-
28
Klaus Sessar
cieties much more than in crime societies. Very roughly: crime may generate fear
or anger mostly in a passing way, whereas risk evokes enduring feelings of insecurity whether we realise it or not. Crime is an issue of repression based on the
rule of law, that is: justice. Risk is a matter of prevention primarily based on
technological achievements, prevention which tends to stoke public unsafety for
its own expansion. Being unable to predict risk as it tells us only what we cannot do, not what we can do (OMalley 2004: 2). we are forced into substituting
our ignorance for a maximum degree of control to feign a sense of security (over
4 Millions CCTV machines in Great Britain did not prevent the London bombing
of 7 July, 2005). Control replaces knowledge, meaning: the less knowledge we
possess, the more control is thought to be needed, up to the point that risk prevention becomes an exclusionary strategy (OMalley 2004: 143). The most advanced control systems in terms of tactics, technology, data collection, statistics
and the like are continually being developed for better risk management, a development which is preferably left uninhibited by the rule of law or judicial surveillance. Control fears control.
3. Governing through fear of crime or fear of risk

The next step is to strip risk of any kind of objectivity. Risk says Mary Douglas, is not a thing, it is a way of thinking, and a highly artificial contrivance at
that (1992: 46). So far, risk is a construction like fear. Attitudes like fear of
crime and pessimistic perceptions of risk are thus excellent tools to govern a society for the sake of security or to exercise disciplinary power over a society for
the sake of its control. In fact, fearful people are willingly governable subjects in
that they are ready to make their social data, their attitudes, their feelings, their
body and their soul (Foucault) available to the controlling authorities. Therefore, what is necessary for these authorities is the fearing subject.
What is a fearing subject? To develop politics of fear, a large amount of data is
required, based on, among others, crime statistics and, in our case, on population
surveys with questions relating to fear of crime, to victimisation experiences, to
confidence in the political and the criminal justice system, to punitiveness, and
the like. To reiterate my initial remarks: the political discourse on the war on
crime and on fear of crime requires scientific information on these and other topics. And it will receive the information as far as the social sciences are able to
provide it. Murray Lee talks of a fear of crime feedback loop which means that
the interdependency between power and knowledge produce and intensify crime
fear and the research related to it; that research into victims produces and maintains the criminological concept of fear of crime quantitatively and discur-
29
sively; that this information operates to identify fear as a legitimate object of

governance or governmental regulation; that the techniques of regulation imagine
particular types of citizens fearing subjects; that these attempts to govern fear
of crime actually inform the citizenry that they are indeed fearful; that this sensitizes the citizenry to fear of crime; that the law and order lobby and populist
politicians use this supposed fearing population to justify a tougher approach on
crime, , and in doing so sensitize citizens to fear once again; and that this spurs
more research into fear of crime and so on (2001: 481). As one can see, it is
again the observer who produces his or her own observation, at least to a considerable extent.
4. Some recent developments

Evidence suggests that criminology is loosening the positivistic link between fear
of crime and crime. The following points provide first indications of these developments:
1. Criminology has become increasingly aware of the fact that the fear of
crime concept transcends the area of crime. Maybe crime fear is a sort of code or
a symbol representing social anxieties in a more general sense (Sessar 1998: 401;
Jackson 2004; Hirtenlehner 2006). Phrases such as insecurity with modern living, perception of disorder (incivility). and urban uneasiness are general
stances masked by crime fear (see Hale 1996: 84). Tim Hope and Richard Sparks
go even further by considering that sometimes the question of fear seems
chronically enmeshed with the dynamics of detraditionalisation and an accompanying sense of disruption of formerly settled moral and customary orders (2000:
5). And Hollway and Jefferson conclude from their studies that fear of crime is
an unconscious displacement of other fears which are far more intractable
(1997: 263).
As a consequence, it would be interesting to learn whether fear of risk is a
more qualified concept for further investigations; it encompasses or could encompass many of the modern anxieties including fear of crime and is less susceptible to seasonal fluctuations.
2. In recent years a number of studies have combined quantitative and qualitative methods to measure fear of crime more accurately. The same respondent is
interviewed twice, by way of a crime survey and, at a later date (weeks, months).
through an in-depth interview. It is desirable, though not always feasible, that the
second interviewer be aware of the answers from the first interview so that possible inconsistencies can immediately be discussed.
30
Klaus Sessar
To sum up the present results of such studies: surveys which operate mostly
with closed rather than open questions are turning sporadic and mostly passing
(sometimes even noncriminal) bad experiences into fear as a lasting attitude. The
surveys neglect contextual variables such as space, time and situation, thus ignoring processes. A given individual may have been fearful because of a mugging
two weeks ago near the train station and may still feel unsafe there, but this individual has not become a fearful person through this single event. In our studies
on Insecurities in European Cities (funded by the European Commission
within the 5th Framework Programme Key Action Improving the Socio-economic Knowledge Base; participants in the project were the cities of Amsterdam,
Budapest, Hamburg, Krakow, and Vienna) a 33 year old woman from Hamburg
seemed to be a fearful person in accordance with the questionnaire in that she
expected to fall victim in the near future to robbery, sexual molestation, theft, a
traffic accident as a pedestrian and a dog attack. She avoided city transportation
after dark, specific streets and places, and she sidestepped juvenile groups that
were hanging around. Finally, she was planning to move. No victimisation experience during the last three years was ticked. The qualitative interview revealed
one recurring experience that particularly worried her, namely, the existence of a
group of loitering juveniles and their dogs along her daily route to the kindergarten where she worked. They did not do anything to her, but she preferred to anticipate possible trouble and thus made a detour every day. There were no other
problems in her surroundings except that she found the place somewhat insecure
and noisy and not at all attractive. The plan to move resulted from the small
apartment which she had to share with her father and her brother. The impression
was that the diverse elements of fear as they were shown in the questionnaire
were condensed into one, which was then directed at this specific group of juveniles. One may call this fear, but again, it is restricted to a single, albeit repetitive, experience. For the rest of her grievances, annoyance, uneasiness or anger
may provide more suitable terms in interpreting our respondents feelings.
The common observation is that surveys in general produce higher levels of
fear, whereas in-depth interviews generate lower reported rates of fear. (The opposite is also true, see Kings study of 1992 concerning males who reported assault as a traumatic event only when interviewed qualitatively.) So far, fear, like
serious victimisation, is a rare event.
31
5. Conclusion
The main points of this contribution could be summarised as follows:
a. Fear of crime exists, of course, but it is less common than suggested by
means of surveys. Crime as a reason for fear is competing with signs of incivility in the community, with a worsening quality of life, with a lack of social
cohesion, with social decline, or with an incalculable future. Fear is part of a
more general, perhaps global feeling of insecurity, and is hardly discernible
from such feelings.
b. The way fear of crime is instigated and used by criminal policy makers to
get tough on crime and to legitimise mass surveillance for protecting fearridden societies is utterly unwarranted.
c. Maybe the concept of risk is the more seminal issue in criminology to study
peoples modern anxieties.
References
Beck, U. (1986). Risikogesellschaft. Auf dem Weg in eine andere Moderne. Frankfurt/M.: Suhrkamp (Trans.: Risk Society: Towards a New Modernity, 1992. London:
Sage).
Boers, K., Gutsche, G. & Sessar, K. (eds.) (1997). Sozialer Umbruch und Kriminalitt
in Deutschland. Opladen: Westdeutscher Verlag (the data are from the unpublished
annexed Codebook).
Ditton, J., Farrall, S., Bannister, J. & Gilchrist, E. (1999). Measuring fear of crime,
Criminal Justice Matters, 31: 10-12.
Douglas, M. (1992). Risk & Blame: Essays in Cultural Theory. London: Routledge.
Farrall, S, Bannister, J., Ditton, J. & Gilchrist, E. (1997). Questioning the measurement of the fear of crime: findings from a major methodological study, The British
Journal of Criminology, 37: 658-679.
Hale, C. (1996). Fear of Crime. A Review of the Literature, International Review of
Victimology, 4: 79-150.
Harris, R. (1969). The Fear of Crime. New York: Prager.
Hirtenlehner, H. (2006). Kriminalittsfurcht Ausdruck generalisierter ngste und
schwindender Gewissheiten?, Klner Zeitschrift fr Soziologie und Sozialpsychologie, 58: 307-331.
Hollway, W. & Jefferson, T. (1997). The risk society in an age of anxiety: situating fear
of crime, The British Journal of Sociology, 48: 255-264.
32
Klaus Sessar
Hope, T. & Sparks, R. (2000). Introduction, in Hope, T. & Sparks, R. (eds.). Crime,
risk & insecurity. Law & order in everyday life & political discourse. London, New
York: Routledge: 1-10.
Jackson, J. (2004). Experience and expression: Social and cultural significance in the
fear of crime, The British Journal of Criminology, 44: 946-966.
King, M.B. (1992). Male Sexual Assault in the Community, in Mezey, G. C. & King,
M. B. (eds.). Male Victims of Sexual Assault. Oxford: Oxford University Press.
Kury, H., Lichtblau, A., Neumaier, A. & Obergfell-Fuchs, J. (2004). Zur Validitt der
Erfassung von Kriminalittsfurcht, Soziale Probleme, 15: 141-165.
OMalley, P. (2004). Risk, Uncertainty & Government. London et al.: Glass House
Press.
Ruhne, R. (2002), RaumMachtGeschlecht. Annherung an ein machtvolles Wirkungsgefge zwischen Raum und Geschlecht am Beispiel von (Un)Sicherheiten im ffentlichen Raum Nachrichtenblatt zur Stadt- und Regionalsoziologie, 17: 107-121.
Scott, A. (2000). Risk Society or Angst Society? Two Views of Risk, Consciousness
and Community, in Adam, B., Beck, U. & Van Loon, J. (eds.). The Risk Society &
Beyond. Critical Issues for Social Theory. London: Sage: 33-46.
Sessar, K. (1998). Kriminalittseinstellungen: Von der Furcht zur Angst?, Kriminologie an der Schwelle zum 21. Jahrhundert. Festschrift fr Hans Joachim Schneider
zum 70. Geburtstag. Berlin, New York: de Gruyter: 399-414.
(2007). Die Angst vor dem Drauen. ber gemischte Gefhle in einer unwirtlichen
Welt, in Sessar, K., Stangl, W. & van Swaaningen, R. (eds.). Grostadtngste Anxious Cities. Untersuchungen zu Unsicherheitsgefhlen und Sicherheitspolitiken in
europischen Kommunen Studies on Feelings of Insecurity & Safety Policies in
European Societies. Wien: LIT-Verlag: 128-154.
Skogan, W.G. (1993). The Various Meanings of Fear, in Bilsky, W., Pfeiffer, C. &
Wetzels, P. (eds.). Fear of Crime & Criminal Victimisation. Stuttgart: Enke: 131140.
33
How to Measure Fear of Crime

in a Valid and Reliable Way
GABRY VANDERVEEN 1
Abstract
The traditional umbrella concept fear of crime is used as a term that actually
embraces all kinds of attitudinal knowledge regarding disapproved or unwanted
behaviour, crime and (perceived) safety. The term fear of crime itself suggests
something different than the several intertwined global and diffuse phenomena it
actually embraces. The way in which fear of crime is currently understood has
no inherent necessity at all.
In an extensive research project, the several prevailing usages of fear of
crime and its measurement have been analysed. The results point out the implicit
referents of the terms that are used in much research on fear of crime. The project has made these implicit referents more explicit and essentially pleads for the
complexity of the traditional fear of crime, which has resulted in two approaches of measurement; a traditional approach and a novel approach. In other
words, the research project not only criticised the inevitability of the meaning of
fear of crime and its measurement, but gives better alternatives for the status
quo as well.
First, the two basic problems regarding fear of crime are discussed, that is
the problematic and failing conceptualisation and operationalisation of fear of
crime. Next, advice and guidelines on how fear of crime is to be measured are
presented, in the form of general dos and donts. A recipe for the valid and
reliable measurement of fear of crime will be offered, which covers both the
ingredients that need to be included and the decisions that have to be made during the process.
This piece is based on Vanderveen (2006) and on a paper presented at the 2006 Conference
of the European Society of Criminology.
34
Gabry Vanderveen
1. Measuring an elephant
The so-called fear of crime is much like an elephant. That is, fear of crime can
be considered at least as complex as the elephant in the traditional story of the six
blind men who encounter an elephant. In this story, each man touches a different
part of the elephant, which leads to six different conclusions about what the elephant is like. For example, the man who feels the side says the elephant is like a
wall, but the man who feels the tusk is convinced the elephant is very like a spear
and the man touching the trunk thinks the elephant is like a snake. Each man assumes the whole elephant is like the part he perceives, but by touching only part
of the elephant, they fail to see the whole unique creature. The story illustrates
how different people can have distinctly different perceptions of the same thing,
how limited knowledge of reality can be, or even how truth or reality can be
many things. It is clear that, though the men may be partly right, they are wrong
as well. Not only are the men talking about different parts of the elephant, they
inadequately believe that by perceiving the one part which they touched, they
know the whole elephant. One man alone gives an incomplete description, and
their description would probably improve when they would have touched the
other parts of the elephant as well; that is, a description that incorporates the different parts, e.g. the trunk, the ear, the tail, as well as how they are joined together into the whole elephant, is a better description and closer to the truth. A
complex phenomenon cannot be described or pictured when only one or two
parts are studied in isolation, without reference to the context. Like an elephant,
the complex phenomenon fear of crime consists of different parts and dimensions as well. The traditional term fear of crime is used to refer to the diffuse
umbrella concept fear of crime, which actually embraces all kinds of attitudinal
knowledge regarding disapproved or unwanted behaviour, crime and (perceived)
safety, perceived risk, worry, fear of victimisation and so on (see Vanderveen
2006).
Contrary to the elephant, we do not know what the whole picture of fear of
crime looks like; of which and of how many parts it is constituted. Yet, when
examining the abundantly available literature it is clear that research on fear of
crime and findings is confusing, which begins with the very meaning of the concept. There seem to be several prevailing usages, and some of the disputes that
exist in this field may be of the blind-men-and-the-elephant kind, with the debating parties disagreeing on the implicit referents of the terms they use. For example, in studies in which fear of crime boils down to fear of becoming a victim
of a specific crime, the specific crime can be operationalised as common crimes
such as burglary, assault and vandalism (e.g. Sutton & Farrall 2005). Also, in
How to Measure Fear of Crime in a Valid and Reliable Way
35
many studies the whole complex phenomenon is reduced into one (possible) part
in isolation, and conclusions from these studies are gathered much like the blind
mens conclusions. In an extensive research project (reported in Vanderveen
2006), the focus has been on how the possible conceptual elements of fear of
crime have been constructed and how these are usually joined together. The
conceptualisation of fear of crime is similar to the question: what does an elephant look like, what characteristics are necessary in order for it to be referred to
as elephant.
Besides centring on conceptualisation, the research project investigated the
operationalisations as well. Several varieties of the blind-men-and-the-elephant
story exist; different origins have been put forward and the image of the elephant
that sometimes accompanies the printed story has several appearances. Clearly,
the whole elephant and the story can be presented in distinct ways, dependent on
the time period in which the image was created, the technique and the material
that has been used. Is the Japanese engraving from 1818 closer to the truth than
Galdones illustrations in pen and ink that accompany the story in verse by Saxe
(Atkinson & Stewart 1998; Saxe & Galdone 1963)? In other words, there is not a
single unique (re)presentation of a complex phenomenon; the presentation of reality is related to the characteristics of a time and place and relative to the perceptions of the scientific community as well. Yet, analogous to the different versions
and illustrations of the blind-men-and-the-elephant story, the various ways of
expression share some common characteristics regarding the notions or concepts
to which they refer. Moreover, all the different versions, pictures and drawings
are (re)presentations; they are models of reality. A fixed drawing of an elephant,
i.e. in a two-dimensional space, represents an elephant that lives alive and well in
Africa or Asia. All these drawings are representing reality and share some common features that point out the elephant, yet some of the pictures are said to be
more realistic than others. That is, a full colour photograph of an elephant in its
natural habitat is said to be more realistic than an abstract drawing or an elephant
that is drawn in a cartoon-like manner, yet they all remain models in the sense
that they are abstracted from reality. Thus, the research project was not only concerned about the conceptual elements of the common fear of crime and how
these are joined together, but was also concerned with the operationalisation of
the phenomenon, on the possible (numerical) representations of fear of crime.
The operationalisation and measurement of fear of crime is similar to the question: in what ways can an elephant be represented; what characteristics does the
(re)presentation, i.e. the operationalisation and instrument, need in order for it to
be referred to as elephant.
36
1.1
Gabry Vanderveen
Background of this paper: conceptualisation and operationalisation

of fear of crime
In this paper, the focus is on the practical outcomes of a larger research project:
the newly developed measurement instruments. This paragraph describes the
main research goals and questions of this interdisciplinary and multi-method research project (see Vanderveen 2006). The overarching objective has been to
contribute to clear conceptualisation and measurement in order to provide a better understanding of the phenomenon, by analysing, clarifying and improving the
conceptualisation and operationalisation of fear of crime and/or aspects thereof.
Thus, the main goals of the project were to:
x Give an overview of research on the topic and evaluate the current state of
affairs by hand focusing on familiar conceptualisations and operationalisations.
x Connect this field of research with other disciplines that have studied the
same or related phenomena and apply different perspectives, by making use
of other fields of research and by analysing recurring themes.
x Develop a more explicit and grounded conceptualisation of what has been
called fear of crime and aspects thereof, based on interdisciplinary conceptual analysis in which the socio-political nature of core concepts is taken into
account and by analysing the literature.
x Develop more explicitly grounded operationalisations and propose reliable
and more valid measurement instruments, based on two strategies, namely by
improving present instruments and by starting over and developing another
instrument.
Numerous research questions can be derived from these goals and the general
objective, unfortunately some restrictions, concerning both number and content,
had to be imposed. Therefore, the project was limited to the following main research questions, which were refined in more specific research questions in several studies:
x What is the current state of affairs, how has fear of crime been investigated
and what are the major flaws in this field of research?
x How did the concept, both the term and its meaning, as well as its measurement develop?
x How are core concepts related to one another, traditionally and especially
when making use of other disciplines and a multi-perspective conceptual
analysis?
x What meanings does fear of crime have?
x What constitutes a measurement instrument that is more founded on a conceptual analysis?
37
The overarching objective, the main research questions and goals lead to several general questions that were addressed, for example whether we can empirically and/or conceptually identify parts of fear of crime, whether the term fear of
crime is the most appropriate one, how fear of crime and possible parts thereof
are being measured and whether these current habits of measurement are the
most appropriate ones, and whether the results when measuring fear of crime
are the same as the results when only one part has been measured.
This paper is limited to the practical outcomes of the research project: the
newly developed measurement instruments. Thus, the focus is not so much on
the conceptual analyses. From those interdisciplinary analyses, it became clear
that the traditional umbrella concept fear of crime is used as a term that actually
embraces all kinds of attitudinal knowledge regarding disapproved or unwanted
behaviour, social norms and codes, crime and (perceived) safety. Much confusion arises from the lack of consistently applied and definite concepts (see also
Bilsky 1993; Hale 1996; Wilcox Rountree & Land 1996). Moreover, the term fear of
crime itself suggests something quite different than the several intertwined
global and diffuse phenomena it actually embraces. The two basic conditions of
fear of crime, that it concerns fear and crime, are assumptions, which are not
always met. The broad umbrella concept fear of crime embraces many more or
less distinct concepts, varying from the perception of the risk of becoming a victim of a burglary to concern that something will happen to the children, from the
fear that becoming a victim of rape evokes to the alarm that was installed.
When analysing the concepts history, the social and political context in which
the concept has and still is being used, it is obvious that fear of crime and its
meanings are not inevitable, but rather a product of historical events and social
forces. In other words, fear of crime need not have existed, or need not be at all
as it is; fear of crime is neither determined by the nature of things, nor is it inevitable. Fear of crime has been brought into existence and is shaped by and
therefore dependent on the social context. This social context could well have
been different, which would have brought about another type of fear of crime.
Both fear of crime research and the fear of crime concept are contingent in
nature; in other words, fear of crime is not a pre-discursive social fact (see Lee
1999, 2001; Vanderveen 2006). This means that the way in which fear of crime
is currently understood has no inherent necessity at all, but derives its meaning
from its historical roots and (politically) appealing features. The traditional notion of fear of crime is much more complex and many empirical studies do not
take that complexity into account when attempting to measure fear of crime.
The research project, on which this paper is based, intended to explicate the im-
38
Gabry Vanderveen
plicit referents of the terms that are used in much research on fear of crime and
essentially pleads for the complexity of the traditional fear of crime. This has
resulted in two approaches of measurement; a traditional approach and a novel
approach. The next paragraph will give an outline of this paper.
1.2
Outline of this paper
Analogous to the elephant, the so-called fear of crime is a complex phenomenon that consists of different parts. Empirical findings (reported in Vanderveen
2006) indicate that this traditional umbrella concept fear of crime consists of at
least five different (sub)constructs, of which the improved measurement instruments will be reviewed below. Clearly, a complex phenomenon as the traditional
fear of crime apparently is, cannot be described or pictured by referring to only
one or two parts, without any explicit reference. The term fear of crime itself
suggests something different than the several intertwined global and diffuse phenomena it actually embraces. The problems involved have been pointed out extensively and the inevitability of its meaning and measurement have been criticised (see for example chapter 4 in Vanderveen 2006, but also Jackson 2004,
2005). In other words, recent studies are essentially problematising fear of
crime.
In this paper, the results of two approaches of measurement; a traditional approach and a novel approach, are described. An outline of this paper is pictured
in Figure 1. First, the two basic problems that have been of primary interest are
recalled; the conceptualisation and measurement of fear of crime. Once the
problematic and failing conceptualisation and operationalisation are established,
the subsequent section offers advice and guidelines on how fear of crime is to
be measured in the form of general dos and donts. This is followed by a
concise review of the measurement instruments that have been developed and
which are listed in Table 1. Appendix A provides concrete examples of the instruments based on traditional instruments and Appendix B offers a concrete example of the instrument with the semantic differential format. Table 2 supplies
the recipe for the valid and reliable measurement of fear of crime, which covers
both the ingredients that need to be included and the decisions that have to be
made during the process.
39
Figure 1: Outline of this paper
Conceptualisation
and measurement
Two problems
Future research
Improve, but how ?
Defective
operationalisation
Unclear
conceptualisation
Dos &
donts
Interdisciplinary
Wellbeing
research:
start stealing
Improved
Validity of
instruments
differences
Recipe:
ingredients &
decisions
2. Two basic problems: concept and measurement

Research on fear of crime has two problematic features: flawed conceptualisation and operationalisation. While it is generally acknowledged that all measurement instruments need minimum levels of validity and reliability, it turns out that
the familiarity of these criteria has not been a guarantee for research of high quality; seemingly obvious and well-known qualities of research turn out to be not so
common after all, which is why the measurement instruments presented here
were improved and developed in the first place.
2.1
Unclear conceptualisation
Fear of crime is a diffuse term that embraces all kinds of (sub)concepts. These
(sub)concepts as well as their interrelations appear to be inconsistent, unclear and
therefore not very fruitful. Sometimes a concept is used as an explanation of a
phenomenon; other times it is an aspect of the same phenomenon, as in the case
of perception of risk and fear of crime. Despite common practice, it is definitely necessary to distinguish between different (sub)concepts, because the concept and the measuring instrument used influence the reported fear of crime. In
other words, one part of fear of crime or one part of the elephant does not necessarily tell us everything there is to know about the whole. A small trunk or a
small perception of risk does not mean that the whole elephant or the whole fear
of crime is small, neither empirically nor conceptually.
40
Gabry Vanderveen
The research project, briefly described before, has taken fear of crime as its
central core. Two chapters in Vanderveen (2006), one on the many meanings of
criminal victimisation and the other on risk have shown that fear of crime encompasses so much more than just plain fear of crime. The fear for ones
safety and the experience of safety, which is perceived to be threatened by an
expected, perceived or encountered dangerous other person, is connected with
attributions of responsibility, guilt and blame. Therefore, fear of crime exceeds
fear and exceeds crime likewise. The umbrella concept fear of crime actually
embraces all kinds of attitudinal knowledge regarding crime and (perceived)
safety or the perceived threat to ones own safety or the safety of a loved one because of the potential danger from others.
2.2
Defective operationalisation
In the introduction, the operationalisation and measurement of fear of crime

was said to be similar to the question: in what ways can an elephant be presented;
what characteristics does the (re)presentation, i.e. the operationalisation and instrument, need in order for it to be referred to as elephant. Vanderveen (2006)
discusses the ways in which fear of crime, or parts thereof traditionally have
been measured and exemplary traditional indicators were described. Findings
from the literature as well as the conclusions derived from own data argue that
the measurement of fear of crime is certainly not the same as the measurement
of a part thereof, i.e. as the measurement of a (sub)concept. Furthermore, the
practice of using only one item is problematic, as well as other current habits of
measurement in this field. The overall conclusion then is that the current habits
of measurement are definitely not the most appropriate ones, especially when
reliability and (construct) validity are concerned.
3. Improved conceptualisation and operationalisation

Once the problematic and failing conceptualisation and operationalisation have
been established, the next question is how these flaws can be resolved. Explicating the concept of interest and the reliable and valid measurement thereof is a
necessary condition without which it is neither possible to fruitfully discuss let
alone criticise fear of crime as social construct, the political uses of fear of
crime and its accompanying statistics, nor to influence fear of crime that leads
some people to adapt and constrain their behaviour by means of policy, or to
measure the effectiveness of these kind of interventions.
When something like fear of crime in fact is being measured, preference
should be given to the most proper indicators available, within the context of the
41
amount of time and money that the whole procedure may cost. Evidently, it is
better to measure aspects of fear of crime more properly in a smaller sample
than using only one item (e.g. how safe do you feel walking alone at night) in a
huge regionally representative sample. If fear of crime plays a role in present
society and is considered relevant for society and the government, the way
knowledge about it is acquired deserves a great deal of attention. So far, relatively little effort has been made to improve the survey contents, i.e. the instruments or items that supposedly measure fear of crime till now. The way in
which knowledge on fear of crime is acquired or fear of crime is measured,
needs and deserves improvement. Therefore, it is about time that a beginning is
made with the implementation of the improved measurement instruments that
were developed and are reviewed here.
Why is the term fear of crime still used here? Shouldnt the label fear of
crime be abandoned? On the one hand, one could consider to get rid of the term
fear of crime, since the label itself denotes something that is only a very small
part of the whole range of phenomena involved. It is not so much crime that is
of concern, but much more the potential of danger, the possibility of a violent
confrontation with the offender in person. Narrowing down fear of crime to the
fear of victimisation of legally sanctioned behaviour neglects too much. Though
the common approaches to the concept fear of crime with respect to the crimepart have the longest history, these common interpretations are not accurate and
lack validity (see Vanderveen 2006). The interpretation of events and its consequences are of importance, even in the case of legally defined criminal victimisation. This interpretation is grounded in attributions of deservedness, responsibility or culpability to the victim, which reflect schemas of the ideal types of victim
and perpetrator (see Vanderveen 2006). These schemas provide the information
on appropriate behaviour, attributions of deservedness, responsibility or culpability and feelings. Since this variety of phenomena is not reflected by the term fear
of crime, it would be best to abandon it.
On the other hand, the use of the term fear of crime can be continued, on a
couple of conditions. The reasons for not abandoning the term are mainly practical; a body of research and literature incorporate the term and apparently the term
is appealing to all kinds of professionals; it is evidently appealing to researchers,
the public, politicians and the media. The requirements that should be met in order to use the term least improperly or most appropriately are outlined below.
42
Gabry Vanderveen
4. Measuring fear of crime: guidelines

This section suggests what researchers in practice should do and should not do;
advices and guidelines that are all based on the studies as described in Vanderveen (2006). They contain the requirements that should be met in order to use
the term most appropriately. First, general advice is given on dos and donts,
followed by specific dos and donts, depending on the measurement strategy
that is preferred (traditional or novel). A summary of the improved traditional
instruments and the novel instruments are listed in Table 1. Appendix A provides
concrete examples of the instruments based on traditional instruments and Appendix B offers a concrete example of the instrument with the semantic differential format. Table 2 supplies the recipe for the valid and reliable measurement of
fear of crime, which covers both the ingredients that need to be included and
the decisions that have to be made during the process.
4.1
What to do and what not to do in general
This section offers general advice when measuring fear of crime. Not one clear,
uniform, let alone simple, conceptual model of fear of crime exists. Analyses of
the history of the concept and research on fear of crime and taking the socialpolitical nature of core concepts into account, problematises the simple, familiar,
traditional fear of crime. To put it simply: there is not one fear of crime. The
extensive analysis of three core concepts (fear, criminal victimization and risk)
that appear to be central to fear of crime argues for such a problematisation as
well. Fear of crime is an umbrella label that actually embraces several distinct
constructs. The measurement of only one of these constructs (by means of a set
of items) does not equal the measurement of another construct, nor does it equal
the measurement of the whole concept. The researcher has to choose between the
different constructs and select the construct or constructs of interest, while bearing in mind that it is better to measure one construct reliably than a bunch of constructs unreliably. The chosen construct(s) needs to be explicitly specified. The
researcher should consider the constructs that have been chosen thoroughly, since
the measurement of one aspect of fear of crime does not say that much about
other aspects. That is, conclusions derived from a study on only one part in isolation, like feelings of safety in the street or the perception of risk of becoming a
victim should not be generalised to the whole phenomenon like fear of crime.
Even when a (sub)concept such as the fear of becoming a victim related to a
range of crimes or events is measured extensively and reliably, the findings cannot be extended to other concepts. No attempt should be made to generalise the
findings to other (sub)concepts. In practice, this means that reference should be
43
made for example to the feelings of safety or the fear that an intimate other
will become a victim of crime, and not to fear of crime. Apart from being specific and being explicit about the construct or constructs s/he has chosen to
measure, these constructs should not be mixed up or used interchangeably with
other constructs.
Once the construct or constructs of interest have been selected, the instrument,
i.e. a series of items that measures a particular construct, needs to be chosen. The
series should consist of a set of at least three items to enhance reliability. Above
all, studies on fear of crime or the experience of and attitudes on safety, crime
and risk should not attempt to measure the (sub)concept of interest with only one
item. Using only one item seriously challenges both reliability and validity. A
complex phenomenon deserves more than just one item, such a phenomenon
simply cannot be described or pictured when only one part is studied by measuring this sole part by only one item. The common practice of using only one item
as indicator of fear of crime, needs to be ended.
Table 1:
Summary of improved measurement instruments
Construct of interest
Complete instrument
See Appendix A
Fear that other becomes victim of threat, physical abuse, robbery,
vandalism, a fight, sexual harassment, a burglary
Fear of becoming a victim of threat, physical abuse, robbery,
Fear of victimisation
a fight, sexual harassment, burglary
Seriousness of conseSeriousness of consequences when victim of violence, physical
quences of victimisation
abuse, threat, sexual harassment, sexual assault, burglary
Perception of risk on vic- Perception of risk on becoming victim of vandalism, physical
timisation
abuse, threat, burglary, robbery
Feeling safe when alone in street at night
Being at ease when alone in street at night
Feeling safe at night
Feeling safe when alone at home at night
Being at ease when alone at home at night
Being at ease when together in street at night
See Appendix B
Novel
Friendly unfriendly; pleasant unpleasant; good bad; risky
riskless; active passive; fast slow; steer let go; predictable
Meaning of concepts
unpredictable (per concept)
Nice nasty, pleasant unpleasant, desirable undesirable, sad
Evaluation of concepts
happy, friendly unfriendly and good bad (per concept)
Danger, unsafety and violence are rated on three pairs of adjecAttitude towards
tives each, namely friendly unfriendly, pleasant unpleasant
(un)safety; fear of crime
and good bad
Improved (traditional)
Fear of others victimisation
44
Gabry Vanderveen
Two measurement strategies are proposed here. The first is based on the traditional indicators of fear of crime and leads to the five most important (sub) constructs, namely the fear of others victimisation, the fear of victimisation, the seriousness of the consequences of victimisation, the perception of risk on victimisation and feelings of safety at night. Other constructs, such as behavioural constraints, are not recommended here. Using the improved Likert-type instruments
of these five constructs has the advantage of continuation, practicality and comparability. Besides that, the items are appealing and have, in that sense, face validity. The second measurement strategy proposed here is novel and touches on
ones attitude regarding crime and safety. These novel instruments take a semantic differential format. These new instruments are firmly embedded in the literature and empirical research on fear of crime and the perception of risk. The instrument that measures ones attitude is connected with the concepts derived
from these fields of research and has been created in conformance with generally
accepted tests of validity.
4.2
Improved instruments: concepts and their measurement
Despite their limitations, that are in part the well-known flaws of all research on
perceptions, attitudes and so forth, the proposed measurement instruments, either
based on a strategy of continuance or a fresh start, are certainly preferred over the
common traditional indicators. Table 1 lists the five distinct (sub)constructs that
need to be distinguished when one wants to continue with an instrument that is
most comparable with traditional items. These most important constructs, based
on an approach of continuance and on traditional indicators are the:
x fear of others victimisation
x fear of victimisation
x seriousness of the consequences of victimisation
x perception of risk on victimisation
x feelings of safety at night.
In accordance with the general advice on dos and donts described earlier,
choose one or more of the improved traditional constructs and verbalise these
constructs explicitly. When using two or more series of items, for example when
interested in both the fear of victimisation and the perception of risk on victimisation, apply a similar number (preferably at least five) of answer categories to
both sets. Also, make the range of answer categories symmetric; e.g. from entirely disagree to entirely agree. When indeed measuring more than one construct, it is recommended to use similar references to particular offences. That is,
when measuring both the fear of personal victimisation and the perception of risk
45
of becoming a victim of a specific type of crime, the crimes, like assault, burglary and so forth, that are referred to are preferably the same. When a short instrument, for example in a large-scale survey is preferred over more depth and
quality, e.g. in experimental settings or when good data quality is required, the
complete instrument can be reduced to a series of at least three items. Appendix
A provides some concrete examples of the improved instruments based on traditional instruments and an approach of continuance. Table 1 also lists the instrument that can be used when one wants to make a fresh start and use the validated
measurement instrument that employs a semantic differential format. When one
is primarily interested in the comparison between concepts, the selection of eight
pairs of adjectives performs well. When ones main concern is to measure the
attitude towards safety and crime, as another operationalisation of an aspect of
fear of crime, the instrument that consists of three concepts that are rated on
three pairs of adjectives each is recommended. Appendix B offers a concrete example of the latter instrument with the semantic differential format. In summary;
the novel approach leads to three different kinds of instruments that can be used
for varying purposes:
x portraits of concepts; comparing the meaning of concepts
x evaluation of concepts; comparing concepts on the evaluative/affective dimension (E)
x attitude measurement; attitude towards (un)safety.
The different instruments (see Table 1) still leave many options open: there are
still many decisions to make, which is evident in the description of the dos and
donts as well. All possibilities, instruments and choices to be made are concisely
summarised in Table 2, which provides the recipe that, when followed, gives the
best results. Six steps need to be taken, and three steps involve important decisions that have to be made, explicated and reasons should be advanced for and
against the chosen outcome. As noted previously, further research is needed, especially as far as the novel instruments are concerned. Though these instruments
are more conceptually embedded and make it possible to reveal new or other insights, it is still work in progress. Further research should focus on the multidimensional nature of fear of crime and the interrelationships between the concepts in different populations. While Table 2 provides a sort of recipe of measuring reliably and validly, the Appendices A and B offer concrete examples of actual
measurement instruments. These practical guidelines and suggestions make future research that employs these improved instruments actually possible. After
all, as the old proverb says, the proof of the pudding is in the eating. The next
paragraph offers some suggestions for future research, in which the improved
46
Gabry Vanderveen
measurement instruments might be fruitfully applied. When taking the findings

from the most recent studies into account, future studies cannot ignore the necessity of making choices and cannot use the label fear of crime when it is not
specified properly.
Table 2:
Recipe for the valid and reliable measurement of fear of crime
Step
1 Choose purpose & interest
2
3
4
5
Options & alternatives

Possible decision outcome
Comparability over time
Traditional instruments based on strategy of
Familiarity of instruments
continuance
Conceptual embeddedness
Novel instruments based on strategy of startAttitude measurement
ing over
Comparison of meaning or portraying concepts
No alternatives; choices are required
Acknowledge that fear of

crime is multidimensional,
which necessitates choices
Choose construct(s) of inter- Traditional
est
Denote the selected construct(s) explicitly

Choose the measurement
instrument that fits the chosen construct(s)

Seriousness of consequences of victimisation
Perception of risk on victimisation
Novel
Meaning of concepts
Attitude towards (un)safety
No alternatives; make a choice and give reasons for the decision made
Based on traditional indicators;

use at least 3 items (Table 1) per
construct:
See Appendix A and Appen- Fear of victimisation
Seriousness of consequences of
dix B
victimisation
Based on novel approach (see
Table 1):
Meaning of concepts: 8 items per
concept and # concepts
Use at least 3 items per construct of interest

Match the target crimes
x Choose format: statements or questions
x Apply a similar number of answer categories and at least five
Make the range of answer categories
symmetric
x
x
A number of relevant concepts

Items: friendly unfriendly; pleasant unpleasant; good bad; risky riskless; active
passive; fast slow; steer let go; predictable
unpredictable
Attitude towards (un)safety: 9 items Concepts danger, unsafety and violence rated
on 3 items (pairs of adjectives) each;
friendly/unfriendly, pleasant/unpleasant,
good/ bad
Evaluation of concepts: at least 3
A number of relevant concepts
bipolar scales of evaluative or affec- At least 3 items per concept from: nice/nasty,
tive dimension (E) per concept; #
pleasant/unpleasant, desirable/undesirable,
concepts
sad/happy, friendly/unfriendly and good/bad
Report findings and DO NOT No alternatives; report findings derived from the instrument that fitted the conovergeneralise
struct(s) of interest best. Be specific and explicit. Do not mix up the investigated
construct with (the) other constructs or use them interchangeably. Do not generalize findings to other constructs.
47
5. Conclusion
This paper has focused on the practical outcomes of a larger research project: the
newly developed and improved measurement instruments. The research project
has problematised traditional research on fear of crime. The problematic and
failing conceptualisation and operationalisation have been shown and suggestions for improvement have been put forward. The thesis demonstrated complexity of fear of crime; fear of crime encompasses much more than just plain
fear of crime. The fear for ones safety and the experience of safety, which is
perceived to be threatened by an expected, perceived or encountered dangerous
other person, is connected with attributions of responsibility, guilt and blame.
Therefore, fear of crime exceeds fear and exceeds crime likewise. However, the
research project does not only criticise the current state of affairs, but gives better
alternatives for the status quo as well. The complexity of the umbrella concept
has resulted in two approaches of measurement, one that is based on traditional
indicators and survey items, the other firmly embedded in the literature, connected with the concepts that are of importance and the instrument has been empirically validated.
The newly developed instruments that are presented here can be fruitful in practice; especially when the call for investment in quality and depth of the data is heard.
This also means that we might want to contain ourselves in employing excessively
technically advanced statistical modelling methods, especially when the data are of
the poor quality regarding the content. In other words, more of the same kind of research we were already doing does not have priority; new avenues of research have.
The instruments that were developed and presented here can contribute to that. Following Osgood (1998) in his plead for interdisciplinary research, new questions lead
to data that are to be collected that would otherwise not have been collected, which
will generate findings about new topics, leading to genuinely new knowledge, and
not only, or just, refinement of current knowledge.
So far, the newly developed measurement instruments, either based on traditional indicators or applying the semantic differential, have been used in several
samples. These opportunity and representative samples all consist of Dutch or
Dutch-speaking respondents. In the future, we hope that the instruments will be
used in English-speaking samples as well, or in other language communities.
This will enable more advanced nd necessary validation procedures (see Pauwels & Pleysier 2005).
48
Gabry Vanderveen
Appendix A
Improved Instruments (traditional)
In this appendix, concrete examples of the instruments (the sets of items) are presented. The first four sets contain mostly similar target crimes. When a researcher wants to measure the different constructs, these target crimes are preferably similar. Furthermore, the researcher can choose between statements and
questions; the example is in a question-format. It should be noted here that in a
study in which vignettes were used, the items that employed a statement-format
yielded higher responses than items in the question-format. In the example below, only the extreme answer categories are presented; the researcher can choose
whether to present all categories, i.e. not afraid at all, not afraid, not really afraid,
neutral, a little afraid, afraid and very much afraid when using a question-format.

An example of the instrument; i.e. the set of items, that measures the fear that a
significant other becomes a victim is presented below. Another decision to make
is on the significant other that is referred to; in an older subsample partner
(spouse) was used, in a student subsample friend was used. Another possibility
is referring to somebody you love.
How afraid are you that your friend will become a victim of
not afraid
at all
Threat?
Physical abuse?
Robbery?
Vandalism?
A fight?
Sexual harassment?
A burglary?
O
O
O
O
O
O
O
very much
afraid
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
An example of the instrument; i.e. the set of items that measures the fear that a
significant other becomes a victim is presented below. Please note that the target
crime vandalism has not been included in this series of items.
49
How afraid are you that you will become a victim of

not afraid
at all
Threat?
Physical abuse?
Robbery?
A fight?
Sexual harassment?
A burglary?
O
O
O
O
O
O
very much
afraid
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
Seriousness of consequences of victimisation

The inventory and the surveys that have actually been derived from this inventory did not refer to the same set of target crimes. The set of items that is presented here is based on the empirical analyses of the data from the inventory and
these particular surveys, which means that the target crimes do not mirror the
target crimes in the other three sets. The researcher should choose which target
crimes s/he wants to include.
How serious do you think the consequences will be when you happen to become
a victim of
very
large
Threat?
Physical abuse?
Violence?
Sexual assault?
Sexual harassment?
A burglary?
O
O
O
O
O
O
very
small
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O

The same issues regarding the choice of target crimes apply. Besides that, it
might be worthwhile to experiment with the suggested time frame, though the
next year is rather common.
How likely do you think it is that in the next year you will become a victim of
very
unlikely
Threat?
Physical abuse?
Robbery?
Vandalism?
A burglary?
O
O
O
O
O
very
likely
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
50
Gabry Vanderveen

As noted previously, the researcher can choose between statements and questions. The example below is in a statement-format.
Entirely
agree
I feel safe when Im walking alone

in the street at night
I am at ease when Im alone at home
at night
Im at ease when Im walking alone
in the street at night
I feel safe when Im alone at home
at night
I am at ease when Im walking together with somebody in the street
at night
Entirely
disagree
51
Appendix B
Final Instrument (semantic differential format)
Coincidentally, all three items in the original survey had the negative pole on the
left, the positive pole on the right. To minimise response-effects, the three pairs
of adjectives should vary, both the poles and ideally the order of the pairs as well.
Future research should investigate whether the order of the pairs and the concepts
has an independent effect on the responses given. A possible format of the instrument is as follows:
Unsafety
|..| good
friendly |.. |.. |.. |.. |.. |.. |..| unfriendly
unpleasant |.. |.. |.. |.. |.. |.. |..| pleasant
bad |.. |.. |.. |.. |.. |..
Danger
|..| unfriendly
unpleasant |.. |.. |.. |.. |.. |.. |..| pleasant
good |.. |.. |.. |.. |.. |.. |..| bad
friendly |.. |.. |.. |.. |.. |..
Violence
|..| friendly
bad |.. |.. |.. |.. |.. |.. |..| good
pleasant |.. |.. |.. |.. |.. |.. |..| unpleasant
unfriendly |.. |.. |.. |.. |.. |..
52
Gabry Vanderveen
Literature
Atkinson, D. & Stewart, L. (1998). Katsushika Hokusai (1760-1849): Life and Art of
Katsushika Hokusai. Retrieved from http://www.csuchico.edu/art/contrapposto/con
trapposto99/pages/essays/themefloating/hohusaimnt.html [18 May 2005].
Bilsky, W. (1993). Blanks and open questions in survey research on fear of crime. In:
W. Bilsky, C. Pfeiffer & P. Wetzels (Eds.), Fear of Crime and Criminal Victimization (pp. 9-19). Stuttgart: Enke.
Hale, C. (1996). Fear of crime: A review of the literature. International Review of Victimology, 4, 79-150.
Jackson, J. (2004). Experience and Expression: Social and Cultural Significance in the
Fear of Crime. British Journal of Criminology, 44 (6), 946-966.
Jackson, J. (2005). Validating New Measures of the Fear of Crime. International Journal of Social Research Methodology, 8 (4), 297-315.
Lee, M. (1999). The fear of crime and self-governance: Towards a genealogy. Australian & New Zealand Journal of Criminology, 32 (3), 227-246.
Lee, M. (2001). The genesis of fear of crime. Theoretical Criminology, 5 (4), 467-485.
Osgood, D.W. (1998). Interdisciplinary integration: Building criminology by stealing
from our friends. The Criminologist, 23 (4).
Pauwels, L. & Pleysier, S. (2005). Assessing cross-cultural validity of fear of crime
measures through comparisons between linguistic communities in Belgium. European Journal of Criminology, 2 (2), 139-159.
Saxe, J.G. & Galdone, P. (1963). The Blind Men and the Elephant. New York: Whittlesey House.
Sutton, R. M. & Farrall, S. (2005). Gender, socially desirable responding and the fear of
crime: Are women really more anxious about crime? British Journal of Criminology,
45 (2), 212-224.
Vanderveen, G.N.G. (2006). Interpreting Fear, Crime, Risk and Unsafety; Conceptualisation and measurement. The Hague: BJU.
Wilcox Rountree, P. & Land, K.C. (1996). Perceived risk versus fear of crime: Empirical evidence of conceptually distinct reactions in survey data. Social Forces, 74 (4),
1353-1376.
53
Measuring the Fear of Crime.

How Valid are the Results?
HELMUT KURY AND JOACHIM OBERGFELL-FUCHS
1. Introduction
In Germany, as in other countries, the fear of crime has become an important part
of criminological discussion, in particular in the field of crime policy. This trend
has been accentuated by the political events at the end of the eighties with the
German unification, the opening of the borders and the ensuing population migration. Based on survey results, the actual or putative fear of citizens of becoming a victim of crime is often used, particularly on the political front, as an argument for a tougher stance on crime. With the rise of victim studies in the second
half of the sixties, next to questions concerning the actual victimisation, the fear
of crime and other crime related aspects often became part of the inquiry, yet
only a restricted number of questions could be asked for each of these secondary
topics in order to avoid lengthy and thus impracticable research instruments.
Already in 1965 a standard indicator for assessing the fear of crime was developed in the US within the framework of the studies of the Law Enforcement
Assistant Administration. It was based on the question whether one was afraid to
walk alone in ones neighbourhood at night (see Kreuter 2002, p. 47). Opinion
research institutes, such as the German Institut fr Demoskopie Allenbach, later
adopted this form of measuring fear of crime. It was also incorporated in the German Allgemeine Bevlkerungsumfrage der Sozialwissenschaften (General population survey of social sciences, ALLBUS), as well as in numerous national and
international victim studies. We should note, however, that there is no longer a
unified standard indicator, but that there are by now several versions, in other
words, the question has been in part reformulated and, as a consequence, it has
become more difficult to compare the results. Kreuter (2002, p. 236), for example, cites the different versions of the standard indicator that were used in German speaking countries. Maxwell (1984) pointed out already shortly after the
54
H. Kury & J. Obergfell-Fuchs
first wave of British crime surveys that the research instruments used created
conceptual problems for measuring the fear of crime.
We also have to consider another important aspect: Fear (of crime) is a complex construct that cannot be made sufficiently operationalised with a single
question. Most researchers distinguish thus between affective, cognitive and
conative components (see Fattah and Sacco 1989; Gabriel and Greve 2003; Hale
1996; Boers 2003; Boers and Kurz 1997). Obergfell-Fuchs and Kury (1996), for
instance, were able in their research to extract two factors which they identified
as cognitive and emotional fear.
Yet there are still other problems with measuring fear of crime in standardized
surveys. As a rule, the fear of crime is assessed in the context of victim studies.
But if the persons questioned are first confronted with possible victimisations,
they may already be tuned in to fear once they become aware of all the dangers
they could have been exposed to. Questions concerning the fear of crime with a
set choice of answers can, against the background of media knowledge of
widespread insecurity in the population and of a real or imagined increase in
crime, lead respondents to indicate that they too live in fear or that they are
fearful. The results can also lead to wrong assessments because, as a rule, the
questions do not take into account how often the respondents experience this fear
and how it ranks compared to other insecurities (see Farrall 2004; Farrall and
Gadd 2004; Farrall et al. 2000). For an overview of the extensive literature on
fear of crime see Hale (1996) and Vanderveen (2006).
We can thus not be certain if what is presented by polls as fear of crime really
reflects the fear of citizens of becoming a victim of a criminal act or only a partial aspect thereof. The values obtained may also comprise to a large extent
other fears and insecurities (see Fattah 1993). Skogan (1993, p. 131) stressed
already in the early nineties that most researchers see it as a problem that there
is no clear consensus among researchers on what the concept of fear of crime
means or how it is best measured. There has been little change since. Comparisons of longitudinal studies on the evolution of the fear of crime suffer from the
fact that, given the different operationalization, formulation of the questionnaire
and sample selection, one cannot be too sure if the different results are due to a
real change or are influenced by methodological differences. That these differences can be considerable has been clearly shown by recent research (see Kury
1993; 1994a; 1994 b; 1995 a; 1995 b; Kury and Wrger 1993; Reuband and
Blasius 1996; Reuband 2002).
Measuring the Fear of Crime. How Valid are the Results?
55
2. The debate around the topic of fear of crime

Given these difficulties, it is a forgone conclusion that our knowledge about fear
of crime is at best limited. Yet the results from surveys concerning the fear of
Germans (see R+V Infocenter 2006 a) are eagerly picked up by the media and
often are presented as if these were solid facts. Fear that is only more or less present may thus be reinforced in sensitive and already insecure persons and give
rise to stronger feelings of insecurity. The public, but also the political class, tend
to assume that the fear of crime is a consequence of the (high) level of crime,
according to the motto: the higher the crime rate, the greater the fear of crime.
Such a link, if it exists at all, is at best marginal, because most citizens are illinformed about actual crime rates. Their information is mainly based on media
reporting which, as has been repeatedly pointed out, easily distorts the facts surrounding crime (see already Kerner and Feltes 1980; Roberts et al 2003; Beckett
and Sasson 2004). Countries with noticeably lower crime rates like Japan do not
show correspondingly lower fear of crime values (see already Ishii 1979).
This does not mean that there are no exceptions. If we have, for instance an
outbreak of serious crime in a region, such as a series of burglaries, and if this is
known, it cannot be excluded that the fear of becoming a victim of crime will
increase. But if the crime rate in a country or region rises or falls by a few percentage points, this will have at most a marginal effect on the insecurity levels of
the population, and only if the media stresses the change.
Given the way that the fear of crime is usually operationalised, we have to
assume that what is measured is rather a general feeling of insecurity than a specific fear of becoming a victim of crime. The standard item, for instance, inquires into the general fear of going out alone at night and does not specifically
address the fear of crime. Thus other fears could be touched upon such as the fear
of stray dogs or of having an accident.
Kreuter (2002) stresses rightly that the problem with the construct of fear of
crime could be linked to a cognitively weakly anchored attitude, meaning that
most respondents may have no or only a very vague perception of their personal fear of crime and are thus asked to answer items in the questionnaire about
which they have little concrete ideas and to which they have never given much
thought. As the most commonly used research instruments only require to check
a given value on a scale, most respondents will comply, even if they should truthfully acknowledge that they are unable to (precisely) answer the question. Unfortunately, this option is rarely given. It is obvious that information obtained in this
manner has lost some of its validity (see on this point Reuband 2000). This is one
56
of the reasons that explain the often contradictory results arrived at by different
studies.
We are thus left in doubt whether rising or decreasing levels of fear of crime
really indicate a change in the specific level of fear of crime in the population or
just a general level of insecurity which, because it was measured within the
theme of fear of crime, has been interpreted as such. Especially in the 1990ies,
after the fall of the Berlin Wall, we find constant references to an increase in the
fear of crime. This coincides with the time when the collapse of the Soviet Union
and the opening of the borders between countries lead to great political change
which inevitably transformed the life people were familiar with. As a matter of
fact there are few periods in history that have undergone similar societal transformation in such a short span of time. Against a background of rapidly changing
social values, life was perceived as increasingly risky in a world that had become
much more complicated. Old securities like a well paid job or an assured
pension collapsed or at least could no longer be taken for granted (see Beck
1986; Garland 1996; 2001). In this context, a rise in the feeling of insecurity is to
be expected, and with it a rise in the fear of crime, given how this fear is usually
measured.
In Germany the only longitudinal data available using the same research instruments over a period of time are provided by R+V Insurances, a general insurance company that has been conducting a yearly inquiry into fear since 1991
(R+V Infocenter 2006a). From 1994 on, children between the ages of 6 to 14
were also included in the survey and were given structured oral interviews with
closed questions. These inquiries involving children took place in 1994 and 1995
and from then on every second year, i.e. in 1997, 1999, 2001, 2003 and 2005
(R+V Infocenter 2006b).
Since 1991, every year a random sample of about 2 400 adult Germans (14
years and over) from both parts of the country are given a structured, personal
interview, usually around June. The main motivation of the R+V Insurances surveys is obviously not of a criminological nature but in assessing the general risks
of life, and among these also the risk of becoming a victim of crime. Unlike the
techniques used by victim studies, the question concerning the risk of crime is
put in a neutral context. This approach results in generally lower fear of crime
values as those obtained through victim studies.
The separate values for West and East Germans for the years 1991 through
2006 concerning the evaluation of 15 risk factors are summarized in figure 1. We
can see from the figure a clear increase in fear from 1991 to 1996 for both East
and West Germans. In West Germany the increase lasted till 1997. After this we
57
observe a decrease in West Germany until 2001, whereas the values remain relatively constant in the East. After 2001, there is a renewed increase, first in West
Germany, then also in the East until 2003. After this the values remain constant
on this relatively high level in both parts of the country but decrease significantly
in the West in 2006 (the last year for which data is available), whereas the values
in East Germany rather indicate an upward trend.
Figure 1 also indicates that in 1991 the fear levels were not only relatively low
in both parts of the country but also practically identical. Just one year later, the
level has become significantly higher in East Germany. Since then, the values in
the East have remained above those in the West but there has been a general and
noticeable increase in the level of the Fears of Germans from 1991 to 2006 in
both parts. There are no indications that the gap between the old and the new
Lnder is narrowing. Whereas one can observe an occasional decrease in fear
levels in West Germany the values remain consistently high in the East and tend
to go higher.
Figure1: The Germans live in great fear, Evolution of the fear index, comparison of West and East Germany (R+V Infocenter for Security and
Providence, The Fears of Germans 1991-2006).
East
West
means
great fear
The higher values for insecurities and fears coincide with the political turning
point in 1989 and the reunification of the two parts of Germany in 1990. Before
this, there were few differences between East and West. Even though the German
Democratic Republic had many problems, crime levels were clearly lower than
in the West (see Kerner 1997; Heide and Lautsch 1991). The fear of crime was
58
thus not an area of concern in the former GDR; the inhabitants felt safe and their
material existence was secure, even if supplies were not always abundant. Unemployment was no problem as the state took care of its citizens from the cradle
to the grave. After the reunification, the new Lnder went through a difficult
period of transition. They experienced problems of housing and work, and were
faced with a considerable rise in crime from its previously low levels (1/3 of
West German levels). Even if this rise was essentially an adaptation to Western
rates, it meant a tripling of recorded crime, a fact that could hardly be kept hidden from the population. In addition the citizens were faced with hitherto unknown forms of crime such as drug dealing and fraud. The decrease in fear of
crime in West Germany in 2006 coincides with an improved state of the economy and assurances from both politicians and business that the outlook was getting brighter. This positive evolution did not happen in most parts of Eastern
Germany.
Both the upward trend and the widening gap between the fear values in East
and West can thus be explained against the background of a different societal
evolution in the two parts of the country. After the initial euphoria caused by the
reunification, it soon became obvious in the early 1990s that this important societal and political change also brought with it considerable problems, especially of
an economic nature. In the following years, problems of pollution, terrorism, the
affordability of health care, pensions, the cost of care for older people, etc. increasingly became an issue. In addition, Germany was confronted with economic
problems leading to higher unemployment, especially in the East. These rapid
transformations, especially the dramatic way they were presented by the media,
could but lead to increased feelings of insecurity among the population.
If we look, on the basis of these data, at the evolution of the fear index for the
whole of Germany, restricted to the particular crimes the respondents were questioned about, we notice until 1993 a visible increase and, especially from 1997 on
an equally visible decrease. A comparison of the fear of crime curve with the
general fear curve shows thus great similarities, which seems to imply that the
general feeling of insecurity of the population has a direct influence on the fear
of crime values. In other words: what we call fear of crime is to a large extent the
expression of a general fear that has little to do with the specific fear of becoming
a victim of crime. If the general level of insecurity is on the increase, the fear of
crime will also show higher values, because the presence of other problems may
make it more difficult to cushion the possibility of becoming a victim of crime.
Even if we take into account possible problems with the validity of the data,
the R+V Insurances survey shows clearly that among the 15 fears listed the fear
59
of crime figures second last (figure 2), contrary to earlier years when it occupied
a prominent position. In other words, the risk of becoming a victim of crime is
not a major preoccupation of citizens both in East- and West Germany. They are
much more worried by other potential problems, such as the increase in the cost
of living, the distance of politicians from their electorate, unemployment in Germany, a slowdown of the economy or the risk of losing their job (see figure 2).
Only the fear of a marriage- or partnership break-up ranks after the fear of crime.
We notice also new fears, such as the fear of a terrorist attack or of a war involving German soldiers, which have only surfaced in the last few years when such
events have become concrete possibilities.
Figure 2: The people in East- and West Germany live in great fear of:
(Infocenter der R+V Versicherungen, Die ngste der Deutschen
2006)
82
74
67
in percent
60
72
59
West
68
67
58
47
63
50
East
59
49
47
43
47
41
53
35
47
40
41
28
39
27
32
24
27
D
ist
an
ce
o
fp
ol
iti
Co
st
of
liv
in
ci
g
a
ns
U
in
ne
cr
f
r
m
ea
o
m
pl
se
o
e
ym
Sl
le
c
ow
to
en
ra
do
ti
te
n
w
G
n
e
o
r
f
Pe
m
th
an
rs
ee
on
y
c
al
on
un
om
y
Ca emp
Te
re
lo
D
n
sio
ru
y
ca
m
gse
ns
en
or
in
th
t
al
r
ol
ou
co
d
S
g
ho
ag
D
er
h
ec
fo
io
la
e
re
us
r
d
e
as
di
ig
ill
ct
e
n
ne
io
in
im
ss
m
sta ns o
ig
nd
fo
ra
ar
tio
ne
d
'
n
s
of
ch
liv
il d
in
re
W
g
n
ar
in
w
ol
ith
d
a
G
T e ge
er
m
r
ro
an
r is
pa
m
Lo
r
t
i
ne
lin cipa
t io
es
si
n
n
ol
d
Ru
ag
pt
e
ur
e
Cr
of
im
pa
r tn e
er
sh
ip
16
We have to keep in mind that the German crime rates have been decreasing
slightly during the last couple of years (- 3.64 per cent in 2005 compared to 2004,
and again 1.37 per cent in 2006. See Bundeskriminalamt 2007). But these positive trends have been repressed by the media because of other more immediate
news and may have been hardly registered by the population at large. This interpretation finds confirmation in the fact that in 2006, compared to the year before,
virtually all fears included in the study showed lower levels except for the fear of
60
crime (+ 1 per cent), even if the objective situation had improved according to
official statistics. There are big differences how the various categories of fears
are presented by the media, and especially by politicians. Whereas the economic
situation, and in particular unemployment figures, are shown in a positive light to
highlight political achievements, crime is presented as a problem, sometimes
as an urgent problem that warrants additional measures, usually tougher laws
against certain categories of offenders such as sex offenders (see Rckert 2006;
Sack 2006; Brandenstein 2006; Obergfell-Fuchs 2006).
Crime rates and the fear of crime are two areas that up to now have not been
assessed with a high degree of accuracy. The results obtained show no clear correlation between the two.
3. Fear of crime and public policy

It is not only in Germany that the real or presumed fear of people of become a
victim of crime is used to justify the need to finally do something or to do
more to stop criminal activities. As mentioned above, the average citizen has little precise idea about actual crime rates but usually believes that they or on the
increase (see Roberts et al. 2003; Kury et al. 2002; see also the contributions in
this volume). Especially during election campaigns interior security will be presented as a problem that needs solving, usually through tougher measures
against (potential) delinquents. In the light of real or potential terrorist attacks we
find also in Germany an ongoing political debate on extending the rights of the
police to spy on computer data or to intensify street surveillance, despite the inherent risk to of civil rights and the protection of data (see Mller-Heidelberg et
al. 2007).
In Germany, as in other countries, the topic fear of crime was frequently instrumentalized to justify more stringent laws, higher sentences and a generally
tougher line of action against offenders. The focus was once more on the
young (males), foreigners and in particular on sex offenders. Even if in Germany
the number of registered sex offenders has in general declined over the last decades, sexual offences are perceived as a growing problem. This erroneous perception has probably much to do with the way the media play up particular incidents, especially cases involving the sexual abuse of children a phenomenon
that can be observed not only in Germany but also on an international level (see
Roberts et al. 2003).
The fear of crime has thus a considerable effect on shaping criminal policy. In
this context it is also worth pointing out the large sums people are willing to pay
for their own protection. The private security industry is one of the few branches
61
in Germany that has registered spectacular growth rates in the last few years (see
Obergfell-Fuchs 2000; DSD Der Sicherheitsdienst 2006). The projects for
community crime prevention that have been flourishing in Germany since the
early 1990s were also spurred by the real or imagined fear of crime. Crime prevention became an increasingly important topic, essentially as an answer to the
crime angst of the population as portrayed by the media (see Obergfell-Fuchs
2001). Given the impact of fear of crime figures, it is all the more important for
criminological research to produce dependable data that is based on precise theoretical concepts and reliable indicators for its measurement.
4. How reliable are the data concerning the fear of crime?

Recent studies into the fear of crime have shed additional doubt on the validity of
former research. In the following pages we will briefly touch upon a British
study and then present our own results.
4.1
Results from Great Britain
Already several years ago a British study has pointed out that the customary
standard polling techniques based on a standard, closed questionnaire have led to
a significant overrating of the fear of crime (see Farrall et al. 1997; 2000, Farrall
2004; Farrall and Gadd 2004; see also the article by Gray, Jackson and Farrall in
this volume).
In particular comparative qualitative studies, providing a more thorough analysis of the indicated fear values, have called into question the results from the
quantitative questionnaire surveys. Farrall et al. (1997, p. 658) criticise rightly
that from the outset the fear of crime as well as victimization experiences have
been almost exclusively the domain of quantitative studies, leading to results that
suggest that the fear of crime is an overriding social problem. It is only in the
last few years that these results have been increasingly questioned by more differentiated studies.
Farrall et al. (1997) have presented themselves a differentiated study on the
fear of crime for which they combined quantitative and qualitative research techniques. Lichtblau and Neumeier (2004) have later conducted a comparable study
in Germany (see below). In the British study of Farrall and colleagues (1997), the
researchers selected in October and November 1994 four urban districts in Glasgow according to the criteria: inner city versus suburb and lower versus higher
social strata among the inhabitants of each of the four neighbourhoods. 167 randomly selected persons (Random-walk-procedure) over 16 years of age were
given the customary standardized interview (standardized closed questionnaire)
62
and asked about their fear of crime, their fear of specific offences, their cognitive
risk assessment and their own victimisation experience. 64 of the respondents
were then given within a month an additional qualitative interview with open
questions during which they were asked again about their specific fears of crime,
how secure they felt in their neighbourhood, and about possible avoidance strategies if they showed high fear values.
The analysis of the data revealed considerable discrepancies between the results obtained from the quantitative and the qualitative approach and also between the different techniques used. In total, 114 substantial discrepancies were
found, 98 of which in relation to the two instruments used. 40 percent of these
discrepancies some of them considerable were caused by contradictions in
the answers obtained in the written and oral interviews. In the qualitative interviews the fear values were as a rule significantly lower than in the standardized
questionnaires. Farrell et al. (1997, p.669) quote as an example the case of a respondent who, in the written questionnaire, checked an intermediate value on a
five-point scale concerning the fear of vandalism, but said in the qualitative interview that he had no fear whatsoever of being vandalized. When asked about
this discrepancy, the respondent answered that he had understood the question in
the questionnaire as being purely hypothetical: would he be afraid if confronted
with vandalism? If this actually were the case, he would be somewhat afraid, but
as it had not happened he was at present not afraid. Similar errors in evaluation
were also noted for the area of offence specific fears. The authors (Farrall et al.
1997, p.3) conclude that closed inquiry systems as used by most polls, especially
by big scale surveys, consistently overestimate the fear of crime.
The British study also shows that qualitative inquiries tend to capture rather a
general fearfulness which, upon further questioning turns out to be based on very
specific and rare hypothetical occurrences such as meeting a group of young
drunk men while walking alone at night. Specific social and geographic contexts
are usually factored out in standardized polls. To conclude thus from a specific
situation the presence of a general fearfulness about crime seems more than
problematic. One may also ask if such fears, for example the fear of a child of a
dark cellar, are not even natural and, in the sense of a general protection mechanism, useful for survival. Hannah Arendt has justly pointed out that fear is indispensable for survival. It is hardly justified to extrapolate from such natural
fears the presence of a strong fear of crime and use it as an argument for introducing more incisive measures of crime prevention. Lichtblau and Neumeier
(2004, see below) mention the case of a young woman who describes her unease when she parks her car late at night near her house and has to walk from
63
the parking spot a short distance to the entrance along a well lit path that is bordered by bushes. When this young woman marks this unease in the questionnaire
as fear of crime one may wonder to what extent she can be generally classified as living in such a fear. But standardized quantitative inquires will normally
put it down as such because respondents will check the corresponding item for
lack of alternatives. Only that in this case almost all citizens will have to be classified as fearful of crime. It is perfectly natural and in no way exceptional for
people to feel uncomfortable in certain situation and even fear a possible victimisation.
Farrall et al (1997, p. 671) also point out that the wording of the question can
produce misleading results, for instance the word worry, because of its broad
spectrum (to alarm, frighten, scare, fret, upset, annoy, agonize), easily gives rise
to diverging interpretations. In German research literature we too find such expressions like fear of crime, feelings of insecurity, or fear of criminal acts
(see Kreuter 2002). As a rule no distinction is made between the terms fright
(Furcht) and fear (Angst). Numerous studies have shown the impact methodological differences can have on the outcome, such as the formulation of the question, the number, sequence and form of the set answer alternatives or the position
of the item in the questionnaire (see Kury 1994a, 1995a; Reuband 2002; Reuband
and Blasius 1996). For instance, the question: Which are, in your opinion, the
three most important problems in your part of town? received different answers
depending on where it was inserted into the questionnaire. The problem of crime
was mentioned 2.5 times more often when the question was put at the end of the
questionnaire, after several questions about possible personal victimisations for
different crimes than when it was put at the beginning (Kury and Wrger 2004).
In the first case, the respondents were already sensitised to the topic of crime by
specific references to possible victimisations, and consequently mentioned it
more frequently as a problem in their neighbourhood.
These various studies give a strong indication that the fear of crime values as
measured by standardised polls have to be interpreted with caution. The research
by Farrall et al. (1997) in particular has demonstrated clearly that classical victim studies tended and tend to overstate this fear to a considerable extent. The
researchers conclude (1997, p.676) that the result of these polls depend mainly
on the way the measurements are taken and less on how pronounced the fear of
crime actually is. In the afore mentioned study by Kury and Wrger (2004) the
authors were able to show that methodological differences could affect the result
of victim studies often more strongly than the fact of carrying out the inquiry at
different points in time. In short, there is much evidence to indicate that the fear
64
of crime has been hugely overestimated in such surveys and that the actual
values may be half of what is generally believed. Kreuters analysis of international studies and the results of his own calculations (2002) equally confirm this
hypothesis.
4.2
Our own research
4.2.1 Methodology
A German study conducted on similar lines as the one by Farrell et al. (1997)
arrives at comparable results. In a first step, Lichtblau and Neumaier (2004) developed a normal, standardised questionnaire for measuring the fear of crime and
victimisation in which they incorporated commonly used items such as the fear
of crime, the severity of victimisation, avoidance and protection techniques, general risks affecting the life of citizens, indications of social disorganisation (incivilities) and individual victimisations. The standard item for measuring the fear
of crime was introduced at the beginning of the questionnaire to avoid interference from subsequent victimisation questions. The same standard item was then
reintroduced in a slightly altered form (scale of 10 points for indicating the level
of fear) at the end of the questionnaire.
The wording of the standard item is based on the survey by Kreuter (2002) and is
couched in the following terms: Think only of your immediate neighbourhood,
i.e. of all the places that you can reach within a five-minute walk. How safe do you
feel, or would you feel, if you are out alone at night in these parts? Very safe,
rather safe, rather unsafe, very unsafe, dont know. The category dont know
was added to prevent that respondents without an opinion were forced to opt for one
of the other choices. By simply giving this option we can already expect lower fear
values compared to polls that do not include this category.
In addition to the standard item 6 vignettes were added describing situation
which, to a different extent, could induce fear (van der Wurff et al. 1989). The
vignettes covered the following situations: A. You are home alone late at night
when the doorbell rings. You dont expect anybody. B. You bring out the garbage at night and see two men standing by a car close by who start walking toward you. C. Early at night you lose your way to a party. A group of young people is following you and start making unpleasant remarks. D. In the afternoon
you are waiting for the bus next to your home. A group of 15 to 16 year olds enter the bus shelter and start spraying it with graffiti. E. You are about to go out at
night when the phone rings. There is no answer on the other side, all you hear is
deep, irregular breathing, and then the caller hangs up. F. You are driving
through an unfamiliar town and need to phone home to say that you will be late.
65
You stop at a coffee shop to find a phone. A group of motor cyclists are just
meeting there.
After describing each vignette the respondents were asked how secure they
would feel in such a situation on a scale corresponding to that of the standard
item. In addition, they had to assessed the probability of a negative outcome of
such a situation (five possibilities of answers: most unlikely, rather unlikely,
rather likely, very likely, and in addition dont know). Finally they were
asked whether they had experienced a similar situation before.
By taking an approach similar to Kreuters (2002, p. 49), we also attempted to
capture the specific fear of crime through an item asking about anxiety levels
caused by different events on a 4-point scale (not at all very worried). The
following eventualities were listed: injury in a traffic accident, to be mobbed,
to be beaten up and injured, burglary in ones own house or apartment, to be
attacked and robbed, car break-in, theft without violence, to be killed. For
female respondents the following situations were added: sexual harassment, sexual aggression, rape, to be beaten by a man or a male friend, to be sexually aggressed by a man, friend or partner. In a second item the respondents were asked
to assess the likelihood of finding themselves in such a situation and to indicate
the gravity of such incidents. Further questions concerned potential avoidanceand protection strategies as well as the level of insecurity caused by the general
risks of life based on the list of the R+V Insurances questionnaire. The questions
concerning personal victimisations were limited to the following offences: theft,
burglary, bodily harm and rape.
For the qualitative interviews a methodological guide was developed and pretested. The standardized questionnaire was also given a pre-trial run.
The main survey took place in the summer of 2003 in three neighbourhoods of
the city of Freiburg. The neighbourhoods were selected on the basis of two variables: 1. the actual crime rates and 2. the fear of crime of the inhabitants as
measured by previous comparable studies (see Obergfell-Fuchs and Kury 1995).
The three neighbourhoods had shown in earlier studies respectively high, medium and low levels for both variables (Weingarten: high fear of crime values,
medium crime rates; St. Georgen: low fear of crime values, high crime rates;
Sthlinger: medium fear of crime values, high crime rates). The respondents
were randomly selected (random-walk procedure) among the adult German
speaking population (18 years or older). The assignment of addresses to the two
interviewers (one male, one female) was also random. Before the actual survey,
the selected households received an introductory letter in which they were informed about the scope of the study and their random selection. They were then
66
encourage to participate in a research about the actual feeling of residing and

living in Freiburg. The target person in each household was the member who
was going to celebrate his or her birthday next. As a next step the questionnaire
was personally delivered and a collection date agreed upon. This procedure allowed to approach as closely as possible a random selection (see Schnell et al.
1999, p.266). When the questionnaire was collected, the target person was asked
whether he or she would be willing to take part in a personal interview at a later
date. In case of a positive answer, and only with the consent of the participant,
the address was marked on the otherwise anonymous questionnaire.
4.2
Random Sample
The total random sample included 293 individuals that were questioned. The response rate was 49.7 per cent which is an excellent result for the area compared
to other surveys, and it could only be achieved through a sustained effort to get in
touch with the target person, if necessary up to five times, if previous attempts
proved to be unsuccessful. Of the 590 (100 per cent) persons included in the initial sample, 86 (14.6 per cent) could never be contacted, 167 (28.3 per cent) refused to cooperate, 37 (6.3 per cent) did not return the questionnaire, and 7 (1.2
per cent) of questionnaires had to be eliminated because they were not filled in
properly (e.g. boxes were checked according to a fixed pattern). The response
rate of the three neighbourhoods fluctuated between 46.5 per cent and 55.0 per
cent (n between 94 and 100). 62.8 per cent (n = 184) of participants were women
and 37.2 per cent (n = 109) men. Women were thus overrepresented. The age
distribution was well spread over the different groups. 43.0 per cent of the respondents were married and living with their partner, 26.3 per cent were single,
13.7 percent were living with a partner without being married, 12.3 per cent were
divorced or separated and 4.8 per cent widowed. As for their educational level,
26.0 per cent had a university degree, 21.6 per cent had the German equivalent of
A-levels (Abitur), 23.3 per cent had attended the Realschule (comprehensive
school until grade 10), 20.2 per cent the Hauptschule (grades 7 to 9) and 5.1 per
cent had reached entry levels for attending a senior technical college.
In a second step, 30 persons were given a second, qualitative interview at a
distance of two months in form of open questions, which concerned in particular
the extent and reasons for the stated fear of crime. The two interviewers, a man
and a woman, were each in charge of 15 interviews. Based on the questionnaire
responses, 24 participants were chosen who had indicated relatively high levels
of insecurity, i.e. had relatively high fear of crime values, whereas 6 were chosen
for their low fear levels.
67
4.2.3 Results
4.2.3.1 The quantitative survey (questionnaire survey)
In the standard question at the beginning of the questionnaire, 23.9 per cent indicated that they felt rather unsafe (17.5 per cent) or very unsafe (6.4 per cent). The
others felt either rather safe (57.9 per cent) or very safe (18.2 per cent) in their
neighbourhood in Freiburg. Only 4.4 per cent (13) checked the category dont
know. Congruent with earlier studies the fear level in the three neighbourhoods
differed significantly. The highest values were registered in Weingarten, a part of
town with big apartment blocks (2.5 on a four-point scale: 1 = very safe 4 = very
unsafe). Sthlinger, a neighbourhood close to the station, obtained intermediate
values (2.1), and the lowest values went to St. Georgen, a rather rural neighbourhood on the outskirts of the town (1.8). The average fear value for the standard
indicator was 2.1.
The analysis of the 6 vignettes provided additional information about the general fear of crime. Contrary to the standard indicator, the questions are directed at
feelings of insecurity in a very specific context. Table 1 shows the mean values
(scale as above) obtained for each vignette.
Table 1: Mean value and standard deviation of feelings of security in six specific situations; five-point answer scale
Situation
Mean value
( )
Standard deviation
(s)
A. House door
B. Garbage container
C. Party
D. Graffiti
E. Telephone
F. Coffee shop
2.56 (2.35)
3.45 (3.27)
3.90 (3.77)
3.11 (2.30)
2.42 (1.96)
1.86 (2.31)
1.22 (1.29)
1.18 (1.23)
1.06 (1.08)
1.30 (1.14)
1.34 (1.25)
0.94 (1.32)
Dont know
n
%
12
4.1
22
7.5
26
8.9
21
7.2
13
4.4
15
5.1
Remarks: For comparison, the results from the study of van der Wurff et al. (1989) are indicated in between brackets.
The results are quite consistent with those obtained by van der Wurff et al., only
that the values of our study are slightly higher, with the exception of the coffee
shop vignette. Compared to the standard indicator, the fear values are higher for
the vignettes, again apart from the encounter in the coffee shop. In the three
neighbourhoods, the box dont know was checked by between 8.9 per cent (26)
and 4.1 per cent (12) of respondents. The every day relevance of the situations is
rather low. 39.5 per cent of respondents reported having gone through an experience similar to A (house door), 18.4 per cent to B (garbage container), 22.7 per
68
cent to C (party), 10.3 per cent to D (graffiti), 54.3 per cent to E (Telephone) and
25.0 per cent to F (coffee shop). The testing of the impact of the relevance factor
for daily life (i.e. the fact of having experienced such a situation) on general feelings of insecurity showed that the mean value for all six situations was higher if
the situation had not been experienced. The difference in mean value is however
only statistically significant for vignette B (garbage container) (t = 3,02; p<.01)
(see figure 3).
Figure 3: Mean value of the questions concerning insecurity (vignettes) in relation to everyday relevance (yes = such a situation has been experienced; no = such a situation has not been experienced)
4
3,5
3
2,5
2
1,5
1
House door
Garbage
container
Party
yes
Graffiti
T elephone
Coffee shop
no
A factor analysis of the 6 vignettes resulted in one factor only explaining 46.9 per
cent of the total variance. The factor loadings lie between the values .59 and .73.
A reliability analysis for the scale indicated item-total correlations between .43
and .57. The internal consistency (Cronbach Alpha) is .77.
To check the dimensionality of the individual items for fear of crime (general
and specific fear of crime, vignettes) a factor analysis was calculated which gave
the following results (table 2): Two factors emerged with eigenvalues of 5.4, respectively 1.8 which explain altogether 51.4 per cent of the variance (factor I:
38.5 per cent, factor II: 12.8 per cent). The factor loadings presented a relatively
unified picture in the sense that the offence specific items (victimisation risk) as
well as the item victim of a violent crime load on the first factor, whereas the
questions about feelings of insecurity of the vignettes load on the second. The
standard indicator also has to be attributed to the second factor, but presents with
.47 a comparatively low load. Its load of .42 for the first factor is only slightly
69
lower. Factor I covers thus worries over specific crime related risks. It is more
difficult to put a label on factor II. All items displaying the highest loads have in
common that there is no explicit reference to crime either in the wording of the
question or in the description of the situation. The vignettes, in particular, refer to
hypothetical and ambiguous situations. This factor covers thus rather a vague
feeling of insecurity in an ambivalent situation.
Table 2: Factor loadings and communalities of fear of crime items after
Varimax-Rotation
Victim of a violent crime
Standard indicator
House door
Garbage container
Party
Graffiti
Telephone
Coffee shop
Mobbing
Bodily harm
Burglary
Holdup and robbery
Theft
Murder
Factor I
Factor II
h2
.60
.42
.23
.11
.14
.04
.25
.17
.54
.79
.76
.87
.69
.67
.11
.47
.71
.74
.66
.59
.70
.64
.41
.19
.12
.26
.05
.21
.38
.40
.56
.56
.45
.35
.55
.43
.46
.66
.59
.82
.48
.50
The results also show clearly that the correlations between fear of crime and
demographic variables, such as sex, age or educational level depend to a large
extent on how the fear of crime is measured. The standard indicator tends to
measure significantly higher fear values for women as well as rising levels for
men as they grow older. It also shows a U-shaped distribution for women, with
high values for young and older women and considerably lower values for
women in the middle age range. As far as the specific fear of crime is concerned,
the age distribution remains fairly constant but the sex differences become less
pronounced.
The items for testing incivility were also brought into a factor analysis. The
Scree-test clearly indicates a two factor solution (see table 3). If we compare the
factor loads of the two extracted factors we see that the items loaded on the first
factor are youth, empty buildings, door to door selling, destruction,
drug addicts, drunks, graffiti, garbage, undisciplined drivers, foreign-
70
ers, abandoned car wrecks, and illegally parking drivers. We called this factor I: Social and physical adversities in ones home environment. The most
highly charged items on the second factor are social workers (few), alcohol/
drug counsellors (few), youth centres (few), xenophobia, recreational facilities (few), and sport clubs (few). With the exception of xenophobia all
items concern a lack of social institutions. Items that address problem groups
such as drug addicts and drunks show here if not the highest at least substantial
charges. We called this factor II Problem groups and lack of support. A complementary factor analysis that included the standard indicator showed that the
latter had to be subsumed under factor I (.63); for factor II it showed close to a
zero load (.-02).
Table 3: Factor loads und incivility items after Varimax-Rotation
Incivility-Items
Loitering youth
Abandoned buildings
Social workers (few)
Door to door selling
Vandalism
Drug addicts
Drunks
Alcohol-/drug counsellors (few)
Graffiti
Youth centres (few)
Garbage
Undisciplined drivers
Foreigners
Xenophobia
Recreational facilities (few)
Abandoned car wrecks
Sport clubs (few)
Illegally parking drivers
Factor I
.53
.55
.30
.67
.72
.57
.65
.41
.72
.19
.66
.43
.71
.39
.12
.55
.15
.58
Factor II
.41
.37
.76
.14
.42
.53
.48
.70
.14
.84
.36
.31
.03
.44
.77
.40
.73
.16
h2
.45
.44
.67
.46
.70
.60
.65
.66
.54
.73
.57
.28
.51
.35
.60
.46
.56
.36
There is a clear correlation (Pearson) of r = .42 (p < .001) between the standard
indicator and factor I Incivilities (loitering youth, abandoned buildings, vandalism, drug addicts, drunks, graffiti, garbage, foreigners, car wrecks, etc.), and a
lower but still significant correlation with factor II (lack of social workers, too
many drug addicts, not enough alcohol/drug counsellors, lack of youth centres,
xenophobia, insufficient recreation facilities, etc.) (Pearson correlation of r = .14;
p < .05). We were thus able to confirm that when the social and physical adversi-
71
ties of ones neighbourhood are perceived as a problem, this is often accompanied by increased feelings of insecurity and fear of crime.
A regression analysis with the variables which proved to be relevant in invariant comparisons regarding their influence on the fear of crime show the following result (see table 4): The factors which produced the strongest increase in fear
of crime levels were thus: insufficient illumination at night, perception of degradation, family status (higher fear levels for divorced and separated persons), the
level of education (the higher the level, the lower the fear of crime), sex (women
show higher fear levels than men), and a prior victimisation (victims tend to feel
slightly more insecure). There was no significant correlation for the other factors.
The two factors insufficient illumination and incivilities I, representing the
social and physical adversities of the neighbourhood, had the biggest explanatory
power. This confirms the hypothesis of Boers and Kurz (1997) that the incivility
approach is superior to, for instance, the victimisation approach (for a summary
see also Hale 1996).
Table 4: Representation of the different prediction variables measuring the
extent of fear of crime with their beta-coefficients (in ascending order
of importance) and the corresponding significance.
Prediction variable
Beta coefficient
Significance
Illumination at night
.25
.00
Incivility, Factor I
.25
.00
Family status
-.19
.00
Education
-.17
.00
Sex
-.12
.03
Victimisation
-.12
.03
Age
-.11
.14
Disability because of ill health
.11
.08
Present state of health
.09
.09
Incivility, Factor II
.09
.10
To be able to defend oneself
.06
.35
To be able to run away
-.05
.42
Length of residence in the neighbourhood
.04
.74
As shown above, 23.9 per cent of our Freiburg sample responds to the standard
indicator with more or less pronounced feelings of insecurity. As has been consistently demonstrated by the R+V Insurances surveys over the last 15 years, the
72
fear of crime, compared to other preoccupations, does not figure in first place but
usually in the lower third of the insecurity scale (see above). We also presented
the Freiburg respondents with the list of insecurity factors from the R+V Insurances questionnaire. Figure 4 shows the results.
Figure 4: Relative importance of the fear of crime compared to other worries
(proportion of rather or very worried respondents in percentage)
12,0
Act of violence
15,0
Traffic accident
Addiction (children)
16,6
Loss of apartment/house
16,9
very worried
somewhat worried
18,4
Left-wing extremism
22,4
Separation
26,5
Illness
28,0
Loneliness in old age
29,6
Unemployment
33,0
Foreigners
33,2
Islamism (radical Islam)
36,8
Standard of living in old age
37,9
War
44,3
Care case
46,5
Right-wing extremism
47,9
Destruction of environment
57,4
Pension
63,0
Cost of living
0
10
20
30
40
50
60
70
Our results show, as did the findings of the R+V Insurances surveys, that the fear
of crime takes a backseat compared to other preoccupations. 63.0 per cent of our
sample lists in first place their fear of not being able to cope with the cost of living, and 57.4 per cent are worried that their old age pension will no longer be
guaranteed. Only in last place, 12.0 per cent of the respondents mention the fear
of becoming a victim of a violent crime as a major concern.
The quantitative part of our inquiry produced thus results which are of methodological relevance for fear of crime research in general. The everyday rele-
73
vance of the vignettes, for instance, which describe concrete fear inducing situation is relatively small, except for vignette E (telephone) and A (front door). Less
than a quarter of the respondents had experienced a similar situation in the other
cases. By the same token, the lack of everyday relevance had an impact on the
results by increasing the fear values. The three vignettes that show the lowest
relevance in daily life (D Graffitti; B garbage can; C Party) also show the highest
fear values. It stands to reason that the less familiar a situation, the more fear inspiring it becomes.
A global factor analysis of the different fear of crime items did not allow the
conclusion that they represent a common construct. What we did find, however,
was a factor Worry about specific crime related dangers and a second factor
diffuse insecurity in ambivalent situations. The standard indicator can mainly
be attributed to the second factor but has also a significant correlation to the first.
Thus the findings of Reuband (2000) that the standard indicator in combination
with crime specific indicators loads one factor and is thus understood as a question relating to crime, has not be confirmed by our research. The item could not
be attributed conclusively to one single factor in any of our factor analyses. Our
study did also not confirm the hypothesis of Ferraro and La Grange (1987) that
the standard indicator basically measures personal risk assessment. One could
rather say that the standard indicator measures something of everything but
nothing properly, i.e. precisely.
As far as the individual fear of crime variables are concerned, the highest fear
values were obtained from the vignettes, with the exception of the coffee shop
incident. But relatively high fear values also resulted from using the standard
indicator. The findings implied that the more diffuse and hypothetical the question, the higher the measured fear values would be. The situation described in
the standard indicator is in point of fact for many respondents just as hypothetical
as the vignettes: most are hardly ever out alone at night. Depending on the measuring instrument used, very different conclusion may be arrived at about the level
of fear of crime among the population.
4.2.3.2 The qualitative part of the inquiry (interviews)

The willingness to participate in an additional interview was about equal in the
three neighbourhoods (between 65.7 per cent and 74.0 per cent).
Based on all relevant answers given to the open questions in the personal interview we tried to estimate the level of fear each respondent had indicated for
the standard item in the questionnaire. The evaluation was based on two points:
first, on the feeling of security indicated in the interview when out alone at night,
74
i.e. in the situation addressed by the standard item and, second, on the assessment
of the fear of crime that should ultimately be measured by the standard item. Using the random probes procedure (Schuman 1966), the level of correspondence
was measured by a point system based on preset criteria. Table 5 indicates the
results.
The results show that concerning the feeling of security in ones neighbourhood
in only 13 cases the answers given in the standardized questionnaire could be predicted correctly, and in only12 cases for the assessment of the fear of crime. In one
case each for both categories no prediction could be made because of the very vague
explanation given in the interview. This means that in 16, respectively 17 cases the
predictive value of the fear of crime levels at night alone, as measured by the standard item were incorrect, i.e. they did not agree with the explicit answers made in
the interviews, i.e. the answers did not correspond in over half of the cases. The general feeling of insecurity in the neighbourhood was overestimated in the standardized questionnaire by 15 of the 16 wrong predictions: in the oral interviews the respondents felt (considerably) safer, and the fear of crime was overestimated in all the
cases of wrong predictions in the written survey.
Table 5: Classification of the interviews in function of the evaluation criteria
Evaluation criteria
Clear explanation and precise prediction

Vague explanation und precise prediction
Explanation very vague, no prediction
possible
Clear explanation but wrong prediction
Vague explanation and wrong prediction
Respondent unable to give any kind of
explanation (dont know)
Points
Number of interviews
(Assessment of the
feeling of security in
the neighbourhood)
Number of
interviews
(Assessment of
fear of crime)
1
2
12
1
9
3
5
5
13
3
14
3
If the assessments made in the interviews are more valid than the answers given
to the closed questionnaire questions and much speaks for this view this
would mean that the questionnaire survey would erroneously classify every second respondent as feeling unsafe, respectively as a person with high fear of crime
levels. After the personal interviews only 2 of the 30 interviewees fell into the
category of persons with a high fear of crime. In 12 interviews, the answers to the
standard indicator contained no concrete reference to crime. Only in four inter-
75
views the answers referred exclusively to crime (concretely or vaguely) and not
to incivilities or other risk factors (such as stray dogs, darkness, the risk of falling
and hurting oneself).
The media seem to play a particularly important part in triggering thoughts of
victimisation. 11 interviewees indicated that local and regional media reporting
had shaped their views on crime and of the dangers associated with it. It needs to
be pointed out that, compared to the average crime figures for BadenWrttemberg, Freiburg has a relatively high official crime rate, sometimes ranking at the top a fact that has not been lost on the media. The citizens receive
thus a fair amount of crime related information through the media.
Both interviewers analysed the 30 interviews independently from each other
and estimated the values for the standard indicator in the questionnaire concerning insecurity and fear of crime. It was thus possible to calculate the assessor
concordance. To evaluate the interrater reliability, we calculated the correlation
of the two assessments. The correlation between the assessments of the interviewers is r = .90 (Kendall-Tau = .85) for the prediction of the insecurity values
that the respondents had checked for the standard indicator in the questionnaire.
For the prediction of the fear of crime values we calculated a correlation of r =
.89 (Kendall-Tau = .85). These values can be considered excellent and indicate a
high the degree of concordance between the assessments of the two interviewers.
The interviews highlight the following problems with assessing the fear of
crime solely based on the questionnaire data:
x
Respondents, who indicated in the questionnaire that they felt (very) unsafe,
often qualified this assessment during the personal interview. Filling the questionnaire made them tense and they felt prompted to indicate something.
During the personal interviews it became clear that the indicated fear really
stood for a certain unease or annoyance, for instance with the living conditions, the neighbourhood or perceived incivilities, but had little to do with real
fear of crime. The respondents knew from the media that crime was a problem and that people were scared, but as they did not have anything to say
of their own (non-attitude), they gave what they considered to be a normal answer and checked accordingly a smaller or greater degree of fear.
When questioned in the interview, some of the participants were much surprised that the interviewers should think they were afraid of crime.
x
This unease or annoyance is often linked to concrete, specific situations, such
as a nearby park being frequented by drunks or other conspicuous persons.
They are not particularly aggressive, leave other passers-by in peace, dont inspire fear, and are considered harmless by the interviewees, but they stand for
incivilities and are seen as an annoying degradation of the neighbourhood
76
that should be removed. The inhabitants are thus of the opinion that something should be done and that the town should intervene. The only way
open to them to express this in the questionnaire is to describe their feelings
as fear of crime. There was no other question that would have allowed them
to express their frustration. By simply asking more differentiated questions
about the problems that have been experienced and about desired changes
many that were fearful of crime might have migrated to these alternative
answers. The more differentiated the questions, the more multifaceted the
concept of fear of crime will become; as a consequence a greater number of
the fearful of crime will be classed in other categories, leaving fewer in the
group of the really fearful.
The respondents, who indicated also in the personal interview that they were
fearful, often limited this fear to certain narrowly defined situation, such as
walking at night alone from the parking lot in a high rise complex to the entrance door. Yet it would appear almost normal and by no means singular to
experience feelings of unease or even of fear in this situation. These respondents often do not qualify their behaviour, do not feel insecure in other situations and have never personally experienced a dangerous situation. Against
the background of such singular situations it seems difficult to assign those
persons a global fear level. But this is exactly what happens when the current evaluation techniques are applied to written surveys. Given this difficulty, it does not make sense to determine the fear of crime with a global indicator such as the standard question. Also interactional fear concepts stress the
domain specificity of fear (see Becker 1980).
As pointed out above, persons who indicated in the standardised questionnaire a
(very) high fear of crime, often interpret this as irritation, anger and frustration
about problems in the neighbourhood, especially about manifestations of neglect
(incivilities) that need to be addressed. They express this wish for change
through high fear values in order to give a clear signal to the town authorities that
something has to be done. As a rule, this has very little to do with real fear of
crime. In 19 of the 30 interviews the fear of crime was related to incivilities. 12
interviews showed a direct reference to crime, 8 a vague allusion, and in 7 the
fear was caused by other situations (fear of stray dogs, of an accident, etc.). In
only four cases the standard indicator referred solely to criminal acts (concrete or
vague) and not to incivilities or other aspects.
The results of both the written and oral survey show furthermore that the assessment of personal danger and the evaluation of the local environment depend to a
large extent on the social integration of the respondents in the neighbourhood
(for details see Lichtblau and Neumaier, 2004). Persons indicating high fear of
77
crime levels tended to be less integrated and vice versa. People who lived in the
same large apartment complex and thus in the same environment would sometimes make very different assessment of the safety of the neighbourhood and the
degree of incivilities (loitering youth, drunks, beggars, garbage, graffiti and other
signs of neglect) depending on their degree of social integration. Retiring persons, who have little contact with other tenants and live a rather isolated and
lonely life, seem to experience their environment as more threatening.
The personal interviews also show that fear of crime does not play a central
role in the lives of these people and clearly takes second place behind other problems (see also the results of R +V Insurances). It seems that some participants
dramatised their fear in the questionnaire in order to point out grievances on
the communal and political level and the annoyance they felt about it. It was, for
instance, not uncommon for respondents to list protective measures (such as
door- and window locks, etc.) which, as they admitted in the interview, did not
exist. This too confirms the doubtful validity of the information about fear of
crime and the steps taken for protection.
Our research confirms thus in a different geographical and social context the
findings of the British survey that the common standard indicator used to measure the fear of crime considerably overestimates the extent of this fear. In 57 per
cent of the cases, the fear values based on the standardised questionnaire were
too high compared to the more reliable information obtained through the interviews. Almost half of the respondents (46 per cent) were wrongly classified as
insecure, respectively as showing high fear values, even though the standard
question had been put in a neutral context. Sometimes higher fear values will be
obtained simply through the wording of answer alternatives. In their fear of crime
survey, Lichtblau and Neumaier (2004, p. 281) relied on a conservative approach: in case of doubt, the respondent were classified as experiencing fear of
crime, even if the subjective appraisal would have suggested a different categorisation but that could not be substantiated by corresponding concrete statements
during the interview. It can thus not be excluded that other questionnaire surveys
that rely exclusively on the standard question overestimate the fear of crime even
more. This could be especially the case in neighbourhoods where the incivility
levels are still higher than in Freiburg.
Overestimating the fear values by using the standard item can have several
causes, such as the use of a closed question with given answer categories. Respondents, who have no concrete opinions on the topic, also fall into this category. The
vaguer their own ideas (in the sense of a non attitude), the more they will be
guided by generally accessible information, such as media reporting on increased
78
feelings of insecurity among the German population (see Reuband 2000). We found,
for instance, in a large scale victimisation survey in Freiburg and Jena higher fear
values if the fear item was placed at the beginning of the questionnaire rather than at
the end. It seems that a general question about personal fear of crime will more
likely be answered affirmatively if put in a neutral context, without specific reference or background, than if the topic of victimisation has been given a structure and
rendered more concrete by specific questions about ones own experience. If the
respondents are first asked about possible personal victimisations, they are more
likely to realise that they have not had such a (serious) experience and will feel reassured. By the same token, if the respondents are asked at the beginning of the questionnaire about the problems of their own town (Freiburg), the topic of crime comes
up much less frequently than when the same question is put at the end. In the latter
case, the previous questions pertaining to possible victimisations for specific offences make the participants aware of the risk of crime with the effect that they mention it more frequently as a problem in their town. This shows again the strong impact methodology can have on survey results, an impact that is often neglected (see
Kury and Wrger 2004).
Another cause can be found in the semantic stimuli of the standard question
(Lichtblau and Neumaier 2004, p. 281). Formulations like at night, alone and
outside point to situations in which most people would feel more or less uncomfortable and unsafe. These expressions alone can provoke feelings of fear
and suggest that in such a situation everybody would feel unsafe or fearful.
What is induced and then polled is thus not a specific fear, but a generally expected sentiment. If anything is recorded at all, it is mainly a very natural fear
of crime, but not something exceptional.
Finally, the point must be made that the situation described in the standard
question is for many respondents purely hypothetical and is one in which they
find themselves very rarely, if at all. A majority of the persons polled indicated
that to be out alone at night was not part of their daily routine, quite irrespective
of whether they were afraid or not. For most respondents, no reduction in the
quality of life because of feelings of insecurity or of fear of crime could be made
out. But the survey of Lichtblau and Neumaier (2004) has also shown that that
absence of everyday relevance will most likely increase the fear values. As mentioned above, the two researchers introduced, in addition to the questionnaire,
several vignettes describing potentially fear inducing situations. The less familiar
the respondents were with the situations, the higher was the indicated level of
fear of crime. People who never take walks at night in their neighbourhood will
probably have higher fear of crime scores in the standard question than those
79
who regularly go out at night. The unknown is more fear inducing that what is
known from personal experience.
5. Discussion of the results

The studies that we have presented illustrate clearly the methodological shortcomings of many polls on the fear of crime. Especially the big scale victimisation surveys that included few questions directly pertaining to fear of crime have substantially overestimated the presence of this fear. This phenomenon has already been
pointed out by earlier studies. An analysis by Boers (1995) of three crime surveys
from the years 1991 and 1993 showed that higher fear values were obtained when
measured with the standard indicator than when the questions addressed crime specific fears. The overestimation was particularly noticeable in female respondents and
in persons from the new Lnder, i.e. in those that have generally higher fear of crime
levels. In comparing the extreme categories of the fear of crime items for female
respondents, the difference between very insecure (standard item) and very worried (item for specific fear of crime) varied between 15 per cent for robbery and 23
per cent for break and entry (Boers 1995, p. 13). In a further survey in 1995 the proportion of very insecure female respondents in East German cities was with 42 per
cent in part more twice as high as the different types of specific fear of crime
(Boers and Kurz 1997, p. 203). Our own research shows as well that other dangers
loom larger in the minds of citizens than that of becoming a victim of crime. Of the
18 risk factors listed at the beginning of the questionnaire (question 1), the respondents ranked the risk of becoming a victim of violent crime at the very bottom. It
was actually checked by only 12.0 per cent. This result contributes as well to qualify
the often exaggerated statements about individual fear of crime (see Lichtblau and
Neumaier 2004, p. 128).
All our results point to the fact that the fear of crime does not play the important role in the life of people as the polls, the media and the politicians have us
believe. The current overestimation of fear values in quantitative surveys becomes even more problematic if we take into account that the fear of crime was
and is a central driving force for changes in crime policy, especially for introducing more restrictive and tougher measures. This does not mean that the problem
of fear of crime should be trivialized, but if everybody is afraid, in the end
nobody is afraid, with the effect that little attention is given to the persons that
are truly afraid of crime and whose quality of life is affected by it. The ones who
would really need help are thus left without support.
The polling results and the resulting political demands can create the impression as if an attempt should be made to completely eradicate the fear of crime
80
and to create a world that is absolutely safe from crime. Other, much bigger risks,
such as traffic accidents, in which 5.091 persons lost their life in Germany in
2006 compared to 2.468 deaths through murder and manslaughter, are thus
pushed into the background. It is not impossible that this may be the reason why
some interest groups channel the discussion onto fear of crime.
6. Conclusion
The results of our study on fear of crime tally to a surprising degree with those of
the British study by Farrall et al. (1997), thus further corroborating their validity.
The magnitude of the fear people experience of becoming a victim of a (serious)
crime is clearly overrated by traditional surveys. Both studies make the point that
the actual fear levels are probably only half of those indicated in such polls.
Given the impact fear of crime has on criminal policy, exact information on its
dimension becomes all the more important. This implies a precise theoretical
conception of the construct and more reliable measurement of the fear values.
For this we need, in the first place, reliable and valid measuring instruments. The
standard indicator has not proved to be ideal for measuring the fear of crime. It
leads to a substantial overestimation of fear values and at best to a partial capture
of the construct. Recent results indicate that the fear of crime is no longer a major social problem (Box et al. 1988), at least not in Western Europe. Fear of
crime levels that are only apparently high, should therefore not be used to justify
a law and order policy. Regrettably, this is done again and again by means of
persuading citizens first that, they too, live in fear of crime.
Further research will have to develop precise instruments for measuring the
specific fear of crime and means to determine how far these values are influenced
by context variables, such as the position in the questionnaire. The fear of crime
has to be seen and weighted in the context of other risk- and danger assessments
in order to avoid the impression that it expresses the central feeling of insecurity of people. In the knowledge that fear of crime has different aspects it seems
questionable to use a single global fear value. Not only the extent of the fear, but
also its frequency and its everyday relevance have to be taken into consideration
(see Farrall and Gadd 2004). People experience every day many different emotions among which the fear of personal victimisation plays only a minor part.
Ultimately, it is the task of criminology to inform the public about the problems connected with assessing the fear of crime and the possible overestimation
of this fear. This same problem of validity does not only apply to the fear of
crime but also to other concepts, such as punitivity (see the contributions in this
volume). As the survey results can easily be politically misused, it is of great
81
importance that such studies are conducted and presented in a responsible manner. In the empirical social sciences, especially in psychology, the differentiated
use of methodology is an established practice. For many years numerous studies
have, for instance, pointed out the suggestibility of poll results. The emphasis on
methodology is still less developed in criminology, especially on the European
continent, probably because criminology has evolved here as a branch of law,
and social and behaviour sciences still take second rank. Empirical studies are
thus often headed by researchers who have not been trained in social science
methodology.
7. References
Beck, U. (1986). Risikogesellschaft. Auf dem Weg in eine andere Moderne. Frankfurt/M.: Suhrkamp (Englisch: Risk Society: Towards a new modernity. 1992. London: Sage).
Becker, P. (1980). Studien zur Psychologie der Angst. Weinheim und Basel: Beltz Verlag.
Beckett, K. & Sasson, T. (2004). The politics of injustice. Crime and punishment in America. Thousand Oaks et al: Sage.
Boers, K. (1995). Kriminalittseinstellungen und Opfererfahrungen. In: Kaiser, G., Jehle, J.-M. (Eds.), Kriminologische Opferforschung. Neue Perspektiven und Erkenntnisse. Teilband II: Verbrechensfurcht und Opferwerdung. Individualopfer und Verarbeitung von Opfererfahrungen, Heidelberg: Kriminalistik Verlag, 3-36.
(2003). Fear of violent crime. In: Heitmeyer, W., Hagan, J. (Eds.), International Handbook of Violence Research. Dordrecht: Kluwer Academic Publications, 1131-1150.
Boers, K. & Kurz, P. (1997). Kriminalittseinstellung, soziale Milieus und sozialer Umbruch. In: Boers, K., Gutsche, G, Sessar, K. (Eds.), Sozialer Umbruch und Kriminalitt in Deutschland. Opladen: Westdeutscher Verlag, 187-253.
Box, S., Hale, C. & Andrews, G. (1988). Explaining fear of crime. British Journal of
Criminology, 28, 340-356.
Brandenstein, M. (2006). Strafzweckerfllungen als abhngige Variable der Zeit. In:
Obergfell-Fuchs, J., Brandenstein, M. (Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift fr Helmut Kury zum 65. Geburtstag. Verlag
fr Polizeiwissenschaft: Frankfurt/M., 357-394.
Bundeskriminalamt (Ed.) (2007). Polizeiliche Kriminalstatistik Bundesrepublik Deutschland. Berichtsjahr 2006. Wiesbaden: Bundeskriminalamt.
DSD Der Sicherheitsdienst (2006). Security 2006: Der Sicherheitsmarkt in Zahlen Daten Fakten. DSD Der Sicherheitsdienst, 58(2-3), 35.
Farrall, S. (2004). Revisiting crime surveys: Emotional responses without emotions.
International Journal of Social Research Methodology 7, 157-171.
82
Farrall, S., Bannister, J., Ditton, J. & Gilchrist, E. (1997). Questioning the measurement of the fear of crime. Findings from a major methodological study. British
Journal of Criminology, 37, 658-679.
(2000). Social psychology and the fear of crime: Re-examining a speculative model.
British Journal of Criminology 40, 399-413.
Farrall, S. & Gadd, D. (2004). The frequency of the fear of crime. British Journal of
Criminology 44, 127-132.
Fattah, E. A. (1993). Some common conceptual and measurement problems. In: Bilsky,
W., Pfeiffer, C., Wetzels, P. (Eds.), Fear of crime and criminal victimization. Stuttgart: Enke, 45-70.
Fattah, E. A. & Sacco, V. F. (1989). Crime and victimization of the elderly. New York:
Springer.
Ferraro, K. F. & LaGrange, R. (1987). The measurement of fear of crime. Sociological
Inquiry, 57, 70-101.
Gabriel, U. & Greve, W. (2003). The psychology of fear of crime. Conceptual and
methodological perspectives. British Journal of Criminology 43, 600-614.
Garland, D. (1996). The limits of the sovereign state: Strategies of crime control in contemporary society. British Journal of Criminology 36, 445-471.
(2001). The culture of control: Crime and social order in contemporary society. Oxford: Clarendon Press.
Heide, F. von der & Lautsch, E. (1991). Entwicklung der Straftaten und der Aufklrungsquote in der DDR von 1985 bis 1989. Neue Justiz 45, 11-15.
Ishii, A. (1979). Die Opferbefragung in Tokyo. In: Kirchhoff, G.-F., Sessar, K. (Eds.),
Das Verbrechensopfer. Bochum, 133-157.
Kerner, H.-J. (1997). Kriminologische Forschung im sozialen Umbruch. Ein Zwischenresmee nach sechs Jahren deutsch-deutscher Kooperation. In: Boers, K., Gutsche,
G., Sessar, K. (Eds.), Sozialer Umbruch und Kriminalitt in Deutschland. Opladen:
Westdeutscher Verlag, 331-372.
Kerner, H.-J. & Feltes, T. (1980). Medien, Kriminalittsbild und ffentlichkeit. Einsichten und Probleme am Beispiel einer Analyse von Tageszeitungen. In: Kury, H.
(Ed.), Strafvollzug und ffentlichkeit. Freiburg: Rombach, 73-112.
Kreuter, F. (2002). Kriminalittsfurcht: Messung und methodische Probleme. Opladen:
Leske und Budrich.
Kury, H. (1993). Der Einfluss der Art der Datenerhebung auf die Ergebnisse von Umfragen erlutert am Beispiel einer Opferstudie. In: Kaiser, G., Kury, H. (Eds.),
Kriminologische Forschung in den 90er Jahren. Freiburg 1993, 2. Vol., 321-410.
(1994a). Zum Einfluss der Art der Datenerhebung auf die Ergebnisse von Umfragen.
Monatsschrift fr Kriminologie 77, 22-33.
(1994b). The influence of the specific formulation of questions on the results of victim studies. European Journal on Criminal Policy and Research 2, 48-68.
83
(1995a). Zur Bedeutung von Kriminalittsentwicklung und Viktimisierung fr die

Verbrechensfurcht. In: Kaiser, G., Jehle, J.-M. (Eds.), Kriminologische Opferforschung. Neue Perspektiven und Erkenntnisse. Teilband II: Verbrechensfurcht und
Opferwerdung. Individualopfer und Verarbeitung von Opfererfahrungen. Heidelberg:
Kriminalistik Verlag, 127-158.
(1995b). Wie restitutiv eingestellt ist die Bevlkerung? Zum Einflu der Frageformulierung auf die Ergebnisse von Opferstudien. Monatsschrift fr Kriminologie und
Strafrechtsreform 78, 84-98.
Kury, H. & Wrger, M. (1993). The Influence of the Type of Data Collection Method
on the Results of the Victim Surveys. A German Research Project. In: Alvazzi del
Frate, A., Zvekic, U., Dijk, J.J.M. van (Eds.), Understanding Crime. Experiences of
Crime and Crime Control. Acts of the International Conference. Rome 18.20. November 1992. Rom 1993, 137-152.
(2004). Zur Validitt von Umfrageergebnissen. Unpublished manuscript. Freiburg.
Kury, H., Obergfell-Fuchs, J. & Wrger, M. (2002). Strafeinstellungen. Ein Vergleich
zwischen Ost- und Westdeutschland. Freiburg: edition iuscrim.
Lichtblau, A. & Neumaier, A. (2004). Die Messung der Kriminalittsfurcht. Ein Versuch
der Kombination einer quantitativen und qualitativen Herangehensweise. Unpublished thesis, Psychologisches Institut der Universitt Freiburg.
Maxfield, M. (1984). Fear of crime in England and Wales. London: Home Office.
Mller-Heidelberg, T., Finckh, U., Steven, E., Assall, M., Micksch, J., Kaleck, W., Kutscha, M., Gssner, R. & Engelfried, U. (Eds.). Grundrechte-Report 2007. Zur Lage
der Brger- und Menschenrechte in Deutschland. Frankfurt/M.: Fischer Verlag.
Obergfell-Fuchs, J. (2000). Privatisierung von Aufgabenfeldern der Polizei. Wiesbaden:
Bundeskriminalamt.
(2001). Anstze und Strategien Kommunaler Kriminalprvention. Begleitforschung
im Pilotprojekt Kommunale Kriminalprvention in Baden-Wrttemberg anhand der
Stadt Freiburg im Breisgau. Freiburg: edition iuscrim.
(2006). Sinn und Unsinn kriminalprventiver Anstze bei Sexualstrafttern. In:
Obergfell-Fuchs, J., Brandenstein, M. (Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift fr Helmut Kury zum 65. Geburtstag. Verlag
fr Polizeiwissenschaft: Frankfurt/M., 599-632.
Obergfell-Fuchs, J. & Kury, H. (1995). Verbrechensfurcht und kommunale Kriminalprvention Analyse anhand der Bevlkerungsbefragung in den Projektstdten und der
bundesweiten reprsentativen Bevlkerungsumfrage. In: Feltes, T. (Ed.), Kommunale
Kriminalprvention in Baden-Wrttemberg. Holzkirchen: Felix Verlag, 31-68.
(1996). Sicherheitsgefhl und Persnlichkeit. Monatsschrift fr Kriminologie und
Strafrechtsreform, 2, 97-113.
Reuband, K.-H. (2000). Pseudo-Opinions in Bevlkerungsumfragen. Wie die Brger
fiktive Politiker beurteilen. ZA-Information 46, 26-38.
84
(2002). Frageformen, themenspezifische Sensibilitten und Antwortmuster: Wie Fragen in Statementform und in Form dichotomer Antwortvorgaben Antwortmuster beeinflussen. ZA-Information 51, 82-99.
Reuband, K.-H. & Blasius, J. (1996). Face-to-face, telefonische und postalische Befragungen. Klner Zeitschrift fr Soziologie und Sozialpsychologie 48, 296-318.
Roberts, J.V., Stalans, L.J., Indermaur, D. & Hough, M. (2003). Penal populism and
public opinion. Lessons from five countries. Oxford: Oxford University Press.
Rckert, S. (2006). Ab in den Knast. Die Zahl der Verbrechen sinkt, doch das Strafrecht
wird systematisch verschrft. Und immer mehr Menschen werden zu immer lngeren
Gefngnisstrafen verurteilt. DIE ZEIT, No. 22, May 24, 2006, 15-18.
R+V-Infocenter fr Sicherheit und Vorsorge (2007a). Die ngste der Deutschen 2006.
Frankfurt: R+V-Infocenter.
(www.ruv.de/index.htm?url=/presse/2_rv-infocenterfrvorsorgeundsicherheit/intro.htm).
(2007b). Die ngste der Kinder 2006. Frankfurt: R+V-Infocenter.
(www.ruv.de/index.htm?url=/presse/2_rv-infocenterfrvorsorgeundsicherheit/intro.htm)
Sack, F. (2006). Deutsche Kriminologie: auf eigenen (Sonder)Pfaden? Zur deutschen Diskussion der kriminalpolitischen Wende. In: Obergfell-Fuchs, J., Brandenstein, M. (Eds.),
Nationale und internationale Entwicklungen in der Kriminologie. Festschrift fr Helmut
Kury zum 65. Geburtstag. Verlag fr Polizeiwissenschaft: Frankfurt/M., 35-71.
Schnell, R., Hill, P. B. & Esser, E. (1999). Methoden der empirischen Sozialforschung.
Mnchen und Wien: Oldenbourg.
Schuman, H. (1966). The random probe: a technique for evaluating the validity of closed
questions. American Sociological Review, 31 (2), 218-222.
Skogan, W. G. (1993). The various meanings of fear. In: Bilsky, W., Pfeiffer, C., Wetzels, P. (Eds.), Fear of crime and criminal victimization. Stuttgart: Enke, 131-140.
Van der Wurff, A., Van Staalduinen, L. & Stringer, P. (1989). Fear of crime in residential environments: Testing a social psychological model. The Journal of Social Psychology, 129 (2), 141-160.
Vanderveen, G. (2006). Interpreting fear, crime, risk and unsafety. Devon: Willan.
85
Spatial Perception of Unsafety, Crime and Disorder
HENDRIEN L. KAAL, GABRY N.G. VANDERVEEN

AND MONIEK E. VAN OEVEREN
1. Introduction
1.1
Spatial aspects in research on fear of crime
In many surveys regarding fear of crime 1 questions are asked that contain some
form of reference to a spatial aspect to these feelings, perceptions and attitudes.
Questions that refer to ones area are by no means exceptional. Typical items
that are used in current surveys such as the International Crime Victimization
Survey (ICVS) are for example: How safe do you feel walking alone in your
area after dark? and Did you stay away from certain streets or places for reasons of safety ()? (Van Kesteren, Mayhew & Nieuwbeerta 2000). Another
large-scale survey, namely the Gallup Poll, uses the items: Is there more crime
in your area than there was a year ago or less? and Is there any area near where
you live that is, within a mile where you would be afraid to walk alone at
night?.
Also in several empirical studies the items used refer to the respondents
neighbourhood. A common item is for example: How safe do you feel your
neighborhood is compared to the rest of (Greenberg, 1980). In a survey on
fear of crime, social and physical disorder or incivilities, informal social control
and social disorganisation processes in general, a set of items was used that constituted a scale for fear of crime in the neighbourhood (ICPSR 8167 & ICPSR
2371: McPherson 1978; McPherson, Silloway & Frey 1983; Taylor 1995a,
1995b, 1996a, 1996b, 1997a, 1997b, 1999). The questionnaire was employed as a
telephone survey and contained the following introduction: We would like to
ask you some questions about crime in your neighbourhood. First, here are some
statements people have made about crime. For each one, please tell me if you
1
In this paper, the traditional term fear of crime is used to refer to the diffuse umbrella concept fear of crime, which actually embraces all kinds of attitudinal knowledge regarding
disapproved or unwanted behaviour, crime and (perceived) safety, perceived risk, worry,
fear of victimisation and so on (see Vanderveen 2006).
86
H. L. Kaal, G. N. G. Vanderveen & M. E. van Oeveren
think its mostly true or mostly false. Subsequently, respondents were asked to
respond to the five items; namely Im often a little worried that I will be the victim of a crime in my neighbourhood, I would not be afraid if a stranger stopped
me at night in my neighbourhood to ask for directions, I worry about the safety
of people close to me while they are in the neighbourhood, When I have to be
away from home for a long time, I worry that someone might try to break in and
finally When I hear footsteps behind me at night in my neighbourhood, it makes
me feel uneasy (page 4 resident questionnaire; ICPSR 8167 & ICPSR 2371).
Another study administered a survey that was developed to evaluate initiatives
of community policing. This survey included six items that referred to concern
about a specific crime (ICPSR 2800: McCoy 1997). The interviewer introduced
the items: each of the following statements indicate the degree of concern you
have that the problem might happen in your neighbourhood. Besides the six
items on concern, a rather familiar item was used, namely How serious a problem do you think crime is in your neighbourhood compared to other neighbourhoods in the U.S. Two other questions were posed about changes in personal
safety in the US and the neighbourhood following the format: In your
neighbourhood, do you think that personal safety is changing? with permitted
responses becoming safer, not changing and becoming less safe.
Clearly, all these items and surveys on fear of crime include questions that
contain some form of reference to a spatial aspect to these feelings, perceptions
and attitudes. The information gained from this type of research is often given a
spatial dimension, either by aggregating the data to city districts or postal code
areas, or by linking data on fear of crime to neighbourhood characteristics. However, hardly ever are any questions asked about the boundaries or size of the area
brought to mind by respondents. Neither is it questioned whether aggregating the
data we have gathered is justifiable if we do not know the boundaries of the area
people had in mind when they answered the questions asked in these surveys.
In other types of criminological studies, spatial dimensions are more commonly studied. In particular GIS is used both by academic researchers and crime
investigators as a means of depicting where crime hotspots may be found (e.g.
Bernasco & Luykx 2002; Ratcliffe & McCullagh 1998). Furthermore, observational research on social and physical characteristics of the neighbourhood gives
us an idea of what types of areas are considered to be unsafe by the public. From
these studies we know that criminal hotspots can be determined at various levels:
they can be determined at the level of countries, cities, neighbourhoods, street
sections, and even at the level of bus stops (Loukaitou-Sideris 1999). At the same
time we know that people feel unsafe at places that are dark, quiet, busy and so
87
on (for an overview see: Hale 1996; Vanderveen 2006). What we do not generally study is how the objective crime hotspots relate to our mental maps of unsafe
areas.
1.2
Alternative questions for research on fear of crime
From the above, two main research questions were developed. The first question
we wanted to answer is What are the boundaries of the areas brought to mind
when answering questions on safety and perceived crime? This should help gain
some insight into the question whether it is justifiable to aggregate survey data on
neighbourhood questions to large administrative areas. Two pilot studies aiming
to gain information on the size of the neighbourhood have come to our attention.
In 2001 Coulton et al. asked respondents in Ohio to draw what they believed
were the boundaries of their neighbourhood on a printed map. More recently the
Netherlands Ministry of Housing, Spatial Planning and the Environment
(VROM) did a small pilot in which they asked people to indicate on a digital map
what they considered to be their area and their neighbourhood (Wassenberg
et al., 2006). Both studies found that peoples own neighbourhood was considerably smaller than the administrative areas to which data are often aggregated,
although the neighbourhood in Ohio was considerably larger still (0,83 km2, SD
0,39 km2) than in the Netherlands, where the area was found to be on average
0,025 km2, and the neighbourhood only marginally larger (unspecified). The
VROM-report warned that aggregation to large administrative areas could lead to
the loss of data and incorrect assumptions of homogeneity. Coulton et al. (2001)
found that changes in borders could lead to significant changes in social indicators. We would like to consider further ways of measuring the size of the perceived own neighbourhood and the possible consequences for research.
The second question is concerned with hotspots of subjective unsafety. It is
clear that subjective safety must have hotspots as well, but little research has
been done. Questions we would like to see asked are: Where are the unsafe areas?, and Which areas are avoided?. Rather than trying to give answers to
these types of questions, this paper primarily aims to give some insight into the
ways in which information on the spatial perception of unsafety, crime and disorder can be gathered and displayed. It focuses on the methodology rather than
on the outcomes.
The next paragraph will describe what methods of research were considered
and chosen for this study. The same methods were applied to both research questions described above, although conclusions should be drawn for each question
separately. Our conclusions are drawn in paragraph 3, which describes our ex-
88
periences with the chosen research methods in three exploratory studies that were
conducted as part of this project. Paragraph 4 discusses the use of the street plans
for each of two research questions described above. Despite this paper having an
emphasis on methodology, it gives some examples of applications of this study
and future projects building on its findings in order to show what further paths
could be explored.
2. Methodology
2.1
Possible research methods
In order to gain more information about spatial perception, we can do a number

of things: we can ask respondents to describe what places they fear or have in
mind when describing crime and disorder; we can ask them to literally show us
what they consider to be their neighbourhood and the unsafe areas; or we can ask
them to draw their neighbourhood and the places they consider to be unsafe. This
last method is the one we chose for the studies we describe in this paper. It
seemed to make sense to give people a spatial task in order to gain spatial information. Also, it seemed that making respondents represent their ideas in a scaled
model rather than taking them for a walk and having them show us these things
might make it easier and less time-consuming to analyse the resulting data.
Asking people to represent their knowledge of their environment is known as
mental mapping. According to Kitchin (1994) mental mapping is the process in
which subjective, personal data is connected to graphic knowledge. This process
is best reflected by the use of graphic sketches. Horan (1999) reported that a
sketch map is an inferred product of a cognitive map, which is formed through
experiences and surroundings. In this way, sketch maps tell us what the drawer
experiences as important, to what extent the instructions were understood and
which experiences of subjective safety have been memorised. Thus sketch maps
and street plans can provide useful information on salient spatial perceptions.
2.2
Three exploratory studies
In three separate, exploratory, studies the idea of asking people to draw their
neighbourhood and the places they considered to be unsafe was experimented
with.
1) During a first-year student assignment respondents from two areas (an affluent area A adjacent to the city centre and a less affluent area B partly containing the city centre, see table 1) in Leiden (a medium-sized Dutch city with
around 120.000 inhabitants) were given a questionnaire on fear of crime and
perceived incidence of crime and incivilities in the neighbourhood. This
89
questionnaire was accompanied by a street plan of a large part of Leiden on

which people were asked to indicate what they considered to be their
neighbourhood and what places, if any, anywhere on the entire map they felt
to be unsafe. A total sample of 170 respondents was questioned.
2) The second study was part of a bachelor thesis (Van Oeveren et al., 2006).
For this thesis a number of small scale pilots were conducted trying to finetune methods. This was followed by a study amongst a small sample of respondents living in one block of houses surrounding a park in Leiden. In this
latter study, 23 respondents answered a questionnaire on fear of crime and
perceived incidence of crime and incivilities in the neighbourhood similar to
the one used in study 1. These respondents were asked to do a number of
things: they were asked to draw their own neighbourhood on a blank sheet of
paper. Subsequently, they were asked not only to draw places within their
neighbourhood which they felt were unsafe, but also places where they felt to
be at risk and places where they felt ill at ease. Later, they were given a street
plan of Leiden (the same one as was used in study 1) and again they were
asked to draw their neighbourhood, places which they felt were unsafe,
places where they felt to be at risk or places where they felt ill at ease.
3) The third study was part of a study funded by the local government of Leiden
on fear of crime in a certain area of town where it was planned to build a
new homeless project. This area was a subsection of area B in study 1. The
questionnaire, that was more elaborate but similar to the previous ones, was
supplemented with a street plan of the research area as defined by the city
council, which was considerably smaller than the street plan used in studies 1
and 2 (see table 1). On this street plan, respondents were again asked to indicate their neighbourhood and unsafe areas. A random sample of 271 respondents, drawn from the population census, was thus questioned.
3. Sketch maps or street plans?

When focusing on the methodology of measuring subjective safety with respect
to spatial references, several questions need to be answered. Is it manageable for
both researcher and respondent? How can you translate the obtained information
into analysable data? How are sketch maps related to street plans? Does the obtained information justify the effort of the research? Or, in other words: does the
use of these methods bear repeating in studies on fear of crime? The next sections briefly discuss the different methods, their possibilities and their (dis) advantages.
90
3.1
The sketch map method
Whether the sketch map task was manageable for the respondent was felt to depend on a few things. It appeared important to carefully consider the most suitable object for the sketch map. In one of the first pilots the respondents were
asked to draw the whole city first and subsequently indicate what they felt was
their neighbourhood. However, with a large area like this, it was found that some
respondents had problems maintaining an overview of what they were dealing
with. Also, when the drawing took too long to complete, respondents became
impatient and the possibility existed that drawings became less accurate and
fewer places were drawn. Therefore it seemed better to use the sketch map for
small areas only. Respondents got one A4-size sheet of paper, and were instructed to draw what they considered to be their neighbourhood. Within this
drawing they were then asked to indicate unsafe areas.
One of the advantages of using sketch maps is that drawing is universal. Everyone has the ability to draw and because of this it is less dependent on the language a respondent is used to. Therefore, this method is suitable for several
groups with different language skills or for groups with different ethnic backgrounds and native tongue. However, the study in which sketch maps were used
found that the results were nevertheless dependent on the drawing skills of the
respondent. There was a large difference between the accuracy with which the
respondents drew. For example, one respondent only drew eight lines to indicate
the four streets he considered to be his neighbourhood, whilst another respondent
drew shrubs, street corners and bus stops to indicate his neighbourhood. Furthermore, it was not always clear what the respondent meant by specific signs or
with the drawing in general. The level of detail seemed positively correlated to
the level of education of the respondent.
It was also found that it was very important that the instructions for the respondents were detailed enough to ensure that all respondents understood the
same by central terms such as unsafe and neighbourhood. In study 2, unsafe was defined as places where you feel unpleasant or less comfortable while
neighbourhood in this study was initially defined as the streets you consider to
be your direct environment 2 . Other definitions may have been more appropriate,
2
In the subsequent study (3) this definition of neighbourhood was dropped as it seemed to
refer more to your area than to your neighbourhood; instead people were asked to indicate what the considered to be their neighbourhood after they had answered a list of standard questions on incivilities and crime in their neighbourhood. When respondents were
asked what was meant by the term neighbourhood they were referred back to these questions and asked to indicate the area they had in mind when answering those questions.
91
but respondents at times felt at a loss if no further explanation of the terms used
could be given.
Obviously, it is advisable that a complementary questionnaire is added to the
sketch maps. For example, questions containing information about sexes, religious background, and the length of time that a person lived in Leiden are of interest. Social-demographic aspects influence the way a person sees his
neighbourhood. The person that lived in Leiden for more than 30 years had a
complete different perception about safety compared to the person who lived
there for only a year or less. The first heard more stories than the latter, knew the
place better and had more negative experiences, like burglary and presence of
gangs. Furthermore, the questionnaire was useful to collect information with a
possible clarification for certain characteristic aspects and patterns made in the
graphic part. Questions like: Can you explain why you find these places unsafe? yielded valuable information to be used when interpreting the sketch
maps.
The most important disadvantage of using sketch maps, at least for the researcher, is that it is hard to translate the results into analysable data. It was difficult to find a reliable method as the data could be interpreted in various ways.
First of all, there were several ways to assess the different drawings. For example, where a person drew only half a street, is this counted as one street or none?
And is the undefined object in the park a shrub or did the respondent mean to
draw something else? Also, what to do with places that are not drawn? When a
person did not indicate a street corner as unsafe for example, while five other
respondents did, had this something to do with the fact he really did not consider
it as unsafe? Or was it not in his perception of his environment, did it not belong
to his mental map of his environment? So, the difference between drawing styles
and indicated neighbourhoods made mutual comparison of sketch maps difficult,
albeit not impossible. In this study, several aspects were chosen on which the
maps could be scored. For example, the indicated unsafe places marked on the
sketch map were scored on the level of detail (very detailed, i.e. a gate, street
corner, bus stop etc.; medium detail, i.e. a whole street, station, shopping centre;
and little detail, i.e. a whole area, a collection of streets). Also, the sketch maps
were scored with respect to the vicinity of unsafe places to familiar places, being
defined as places which people visit at least once a week. The scoring of the
maps in this way turned out to lead to problems with the inter-rater reliability: the
interviewers did not always agree about the different degrees of detail, nor on the
vicinity of unsafe places to familiar places. The scoring method therefore needs
to be improved in order to lead to reliable results.
92
Despite the disadvantages, it was felt to be a benefit of this research method

that respondents are forced to think about their area more closely and that therefore more detail and depth is obtained in the data. Not only whole parks and districts were drawn, also little street corners and small alleys. Also, as they were
forced to think about their neighbourhood and unsafe places themselves, and
were in no way steered as could happen by the use of a pre-existing map, the information gathered with the sketch maps is felt to be a more valid representation
of the respondents actual mental map. Because of this free method of collecting
information, the data collected is much richer than when a person is asked five
standard questions, let alone just two items like the items used in the ICVS and
the Gallup Poll that were described before.
In conclusion, it was found that sketch maps can be useful indeed for research
in a small area with a small number of respondents who are extensively questioned. Obviously, depending on the research question the richness of the data
can be an advantage. However, the methods of data analysis need further attention in the future.
3.2
The street plan method
Some of the issues that were raised for the sketch maps also apply to the use of
street plans. The need for detailed instructions and the use of a complementary
questionnaire applies equally to research with street plans. Also, careful consideration of the scale of the map used is important when using street plans as was
the scale of the area to be drawn in a sketch map albeit for slightly different
reasons. The scale of the area drawn in sketch maps was important as it influenced the level of difficulty of the task for the respondent; the scale of the map
used in the street plan method is important as it was found to influence both the
size of the neighbourhood indicated by respondents (larger scale leads to smaller
neighbourhood) and the level of detail in the unsafe places identified (larger scale
leads to more detail in unsafe places) (see par. 4).
The use of street plans demanded less of the respondents skills than the use of
sketch maps. Whereas some people experienced drawing a sketch map as difficult, with street plans this was no problem. Although map-reading skills were
required, this never posed a problem in our studies. Most people felt that using a
street plan to indicate the size of their neighbourhood and to point out unsafe
places was a welcome change during a lengthy questionnaire. The easiness of the
task seemed to influence the quality of the data. The neighbourhoods indicated
on the street plans were a lot larger than when using sketch maps; this seemed to
be at least partially the result of the fact that drawing a larger neighbourhood re-
93
quired substantially more effort with the sketch map method3 and might thus be
avoided out of indolence.
The number and location of unsafe places indicated by respondents was also
no longer dependent on drawing skills when street plans were used. In general,
people drew more unsafe places on street plans than in sketch maps. This can be
partially attributed to the fact that it does not take much effort to draw on a street
plan. More important is the fact that the street plans often covered a larger area
than the sketch maps. Also, respondents were less likely to forget places, because
they were confronted with a map, which triggered their memory and their own
mental map(s) of the environment. This might improve the accuracy of the data,
although as we described before it could also be argued that such data are less
strongly related to the mental map(s) readily at their disposal. Although respondents tended to draw more unsafe places, these places also tended to be less detailed: they would no longer indicate one specific alleyway or street corner, but
whole areas. It seemed, however, that this loss of detail could at least partially be
remedied by using larger-scale street plans.
Data analysis using the street plan method can be done the same way as when
using sketch maps: number of unsafe places, levels of detail and the vicinity of
unsafe places to familiar places can be scored the same way. However, the street
plan provided more standardised data and did not leave as much room for interpretation by the researchers. This made analysis through scoring more reliable.
Moreover, the use of an identical map for each respondent made the use of GIS
possible, which made analyses both easier and more advanced. Through the use
of GIS-programmes the exact surface of the own neighbourhood could be calculated, and all unsafe places could be overlaid to obtain a map of hotspots of perceived unsafe areas (see par. 4).
In conclusion, although we lose some of the richness of the data obtained with
sketch maps when street plans are used as a basis instead, the task at hand is easier for the respondent, and perceptions of a larger area can be studied. The lower
level of detail in respondents answers seems to be remedied to some extent by
using a larger scale street plan. The analysis, finally, is made easier and more
sophisticated by the uniform basis used for every respondent. The use of street
plans therefore seems a useful addition to the usual (face-to-face) surveys on fear
of crime and perceived incivilities.
As was described in footnote 2, it was also felt that the instructions used with the sketch
maps led people to interpret the term neighbourhood more as own area, which tends to
be a smaller area (e.g. VROM, 2006).
94
4. The use of street plans in practice

To show the application and the dilemmas of the above research methods a number of examples will be given. One of the alternative questions we posed earlier
was What are the boundaries of the areas taken in mind when answering questions on safety and perceived crime? It was with this research question in mind
that we asked people to draw what area they considered to be their neighbourhood. This enabled us to compute the average size of the neighbourhood in the
various studies we did (par.4.1). It also enabled us to combine this information
with the answers on questions concerning crime and incivilities in the neighbourhood to see if we could discover perceived hotspots (par. 4.2). The second research question we put forward concerned the location of areas respondents perceived to be unsafe. The use of data on feelings of unsafety gathered with street
plans will be illustrated in paragraph 4.3.
4.1
Measuring neighbourhood size
For measuring the size of the perceived neighbourhood sketch maps seemed the
best way without hesitation. There is, however, an important choice to be made
concerning the scale of the map used. Our research findings indicate that a larger
scale yields smaller neighbourhood sizes. The question that remains to be answered is what scale produces the more valid results. We would love to give
some guidelines, but up to now it is unclear what these guidelines should be:
more research needs to be done first. The question as to what this means to the
common research practice of aggregating data based on neighbourhood questions
to large administrative areas is not easily answered either. A few of our findings
can possibly give an indication of whether problems can be expected.
It has already been mentioned that the perceived neighbourhood is much
smaller than the administrative areas data are often aggregated to. Not only did
we find that the scale of the map influenced the size of the perceived neighbourhood as reported, also different research areas yielded different results (Table 1):
study 1A and 1B used the same map, but the reported neighbourhood size differed significantly. Possible explanations for this difference can be found in the
fact that neighbourhood A is more homogenous in building style and socioeconomic status, has street names that all refer to mayors and professors, which
creates a sense of connection, and has clearer natural boundaries in the form of
main roads and waterways. We do not know whether there is a relationship with
the size of the administrative area in which the neighbourhood size is measured,
but it appears that the discrepancy between the perceived neighbourhood and the
area the data are aggregated to might differ between different types of areas.
Table 1:
95
Perceived neighbourhood size in two studies
Study
Map size
Scale
Research area
Size of research area
N
% men
av. age (sd)
Range neighbourhood size
1
A3
1:119.000
A
0,52 km2
60
43%
52 (14)
0,03-2,67 km2
B
0,53 km2
102
37%
40 (15)
0,004-1,71 km2
3
A4
1:74.000
C
0,89 km2
270
44%
38 (14)
0,003-0,78 km2
Average neighbourhood size

(sd)
0,67 km2
(0,50 km2)
0,24 km2
(0,30 km2)
0,12 km2
(0,14 km2)
Aggregating to a larger area might lead to data loss (see also par. 4.2), but if the
researcher is only interested in larger-scale data this in itself is not a problem. It
becomes a problem when the data do no longer reflect reality. This could be the
case when differences in perceived neighbourhood size (a) have any bearing on
the answers given on the survey questions, and (b) if differences in perceived
neighbourhood size are not independent of other relevant variables.
The neighbourhood questions as mentioned in paragraph 1.1 often have no absolute answers. For example, the questions concerning the respondents concern
that some form of criminality might happen in their neighbourhood have answers
ranging from very often to never. It is likely that the answers given by the
respondents depend on the size of the area they have in mind: if a very small area
is being brought to mind, often will have a different meaning than when a large
area is being brought to mind, which might to some extent neutralise the effect of
difference in perceived neighbourhood size. At the same time: if someone considers a very small area as his neighbourhood he might have to conclude that
street robbery never occurs, while if a larger area is considered it might occur
sometimes. It is therefore likely that the perceived neighbourhood size has an
impact on the answers given.
The argument that adding up all answers still gives a good indication of the
average score only holds firm when respondents are evenly spread over the research area, and neighbourhood size and location are evenly spread through the
research area and are not related to other relevant variables. The first issue is a
matter of research design, but is hardly ever checked. Figure 1 gives an indication of how often each location within the research area is considered their
neighbourhood by the respondents. Firstly, the figure shows that a large area surrounding the research areas is taken into consideration by at least some of the
96
respondents as well. This means that if a particularly criminal area is adjacent to

the research area this will influence the average scores on the neighbourhood
questions. The varying shades of grey in the figure also show that the chance of a
location being considered the respondents neighbourhood is not equally spread
within the research area. Again, whether the locations considered by most respondents are highly criminal or not will influence the average scores. An even
spread through the neighbourhood therefore seems by no means something that
can just be assumed.
Figure 1: Percentage of respondents that considered part of research areas
A and B as own neighbourhood (study 1)
97
To what extent are differences in perceived neighbourhood size related to

other relevant variables? As can be seen in table 1, the variation between individual respondents in perceived neighbourhood size is substantial. We found little evidence that this variation was related to the standard background variables
gathered in surveys. In area B, age was found to correlate positively with
neighbourhood size (r=0.24; p=.016), in area C it was found that women consider
a smaller area to be their neighbourhood than men (average 0.11 km2, SD 0.12
km2 compared to average 0.15 km2, SD 0.12 km2) (t(205)=2.18; p=.031). None of
the other tested variables (e.g. level of education, length of stay in area, main occupation, location of work) was found to relate to perceived neighbourhood size.
It is clear that the explanation of individual differences in perceived neighbourhood size needs further study.
Although this is no firm research finding, our interviewers got the impression
that respondents did not consider the same area with each neighbourhood question they answered. It was felt that with more serious types of crime respondents
had a larger area in mind than with less serious types of crime. They would for
example consider the area directly surrounding their home address when thinking
about something common like littering or graffiti, but when asked about something less common like street robbery or sexual assaults they would in their
minds circle their home address in ever larger loops until they found an occurrence of the offence at hand or until they really felt they could not consider this
their neighbourhood any more. We do not know how this influences the answer
patterns, but it could lead to overestimation of the less frequent sorts of crime.
Although we did not find a correlation between neighbourhood size and income on an individual level, we did find that the more affluent research area had
a larger average neighbourhood size than the mixed city centre research area. If
this is not a coincidence but a systematic difference, it would mean that the differences that are commonly found in experienced crime and incivilities are conservative estimates: if people would be forced to have comparably sized neighbourhoods in mind the differences would increase.
It is clear that a lot of research needs to be done before we know what the influence of perceived neighbourhood size is on research findings. One first little
step that we have planned to undertake is to conduct an experiment in which part
of the respondents are asked about neighbourhood size after answering a series of
neighbourhood questions, while the other part of the respondents is given a map
with the borders of their neighbourhood highlighted before answering the same
series of neighbourhood questions. If the answers on the neighbourhood questions between the two groups differ significantly, it is clear that the influence of
perceived neighbourhood size on survey results needs further attention.
98
4.2
Hotspots of perceived incivilities
To show that asking about neighbourhood size can be valuable in itself other than
to investigate whether common research practice is valid, we have created a map
that shows how knowledge about neighbourhood size could be used to circumvent aggregation to larger areas. In figure 2, it is shown how a hotspots map of
perceived incivilities can be created by combining in a grid the neighbourhoods
drawn by the respondents with the answers to their questions on nuisance from
dog littering. The procedure used here is simple: for each grid cell the average
nuisance score is calculated from all respondents that considered (part of) that
cell as their neighbourhood. The figure shows that this does not produce a homogenous picture. In this case two locations can be identified were dog littering
is considered to be more of a problem than in the rest of the research area: one
alongside a main traffic artery north of the research area, and one just below that
in the north-east corner of the research area where there is a large, uninhabited,
building site with a dog-toilet, a strip of land designated for dogs to relieve
themselves.
Figure 2: Hotspots of perceived nuisance from dog littering (study 3)
99
The advantage of this way of displaying the data is that for each grid cell only
those respondents who really consider this area as their neighbourhood are used.
It does not throw all people living in one artificially created administrative area
in a heap, and thus does more justice to reality. Aggregated data, on the other
hand, are often used for more sophisticated analyses, that might be more complicated to accomplish with these grid data. More practically, the grid method only
works when there is a sufficient number of respondents: each cell should be considered the neighbourhood of a minimum number of respondents to give a reliable result. We did not take account of this problem when creating figure 2 but
assumed this condition to be met due to the very large number of respondents in
this study. We are however aware that this respondent density is rarely matched
in standard survey research.
4.3
Hotspots of perceived unsafe areas
In the three studies that were conducted respondents were not only asked to draw
their neighbourhood, but also to draw the areas they felt were unsafe. Although
the data gathered this way need further elaborating upon, we do not want to deny
you some of the findings from this way of data gathering. Figures 3 to 5 show
some hotspot maps of perceived unsafe areas, constructed by combining the unsafe areas drawn by the respondents. The maps show some of the potential of this
research method. Figure 3 shows the unsafe areas in Leiden as perceived by the
people living in research area A, the affluent area. Darker colours imply that a
higher number of people considered this area to be unsafe. Figure 4 shows the
unsafe areas in Leiden as perceived by people living in research area B, a heterogeneous, partly city-centre area. When comparing figures 3 and 4 it can be seen
the shades in figure 3 are darker, which means that the unsafe places mentioned
by residents of the affluent area are mentioned by a larger proportion of the respondents. Furthermore, it seems that perceived unsafe areas of those living in
the (less affluent) city centre are more within their own neighbourhood than the
perceived unsafe areas of those living just outside the city centre in a more affluent area. The unsafe areas of the latter residents are more spread over the entire
city. A few hotspots are shared by both groups: the two dark areas to the left of
figure 3 are the central train station and a notorious entertainment district. The
large dark area at the right is a city district with a bad reputation. It is remarkable
that, although these areas appear in figure 4 as well, they are smaller and less intense in colour.
Figure 5 shows the differences in perceived unsafe areas between men and
women. The procedure used to create this map is simple: percentage of women
100
considering (part of) a grid cell as unsafe is deducted from the percentage of
men. Dark-grey areas are considered unsafe by more relatively men than women,
light-grey areas are considered unsafe by more women than men (at least 2 percent points difference). Areas that have no colour are perceived more or less
equally safe or unsafe by both men and women. This reveals some important differences in the patterns of perceived unsafety between men and women. More
men than women feel the areas just south-west and north of the central station
and the aforementioned entertainment area are unsafe. At the same time, more
women than men consider Leidens main shopping street, aforementioned disreputable city-district and a large woodlands recreational area in the south as unsafe. There are clear functional differences between the spaces considered unsafe
by men and women, which are suspected to correlate with different patterns
emerging from police statistics. Once again, it appears that fear of crime has different determinants for women than for men (Warr, 1987; Vanderveen, 2006).
The suspected correlation between crime patterns based on police statistics
and hotspots of perceived unsafety mentioned above has not yet been tested by
us. This is one of the things we aim to do in the future. Another correlation we
would like to examine is that between the hotspots of perceived disorder and incivilities described in the previous paragraph and the hotspots of perceived unsafe areas described in this paragraph. In general, we believe that because already
much research has been done on large-scale interrelations with socialdemographic variables as explanatory factors, the time has come to aspire to
more refinement and look at things at a more detailed level. The methods of both
data gathering and analysis illustrated in this paper may inspire such research.
Figure 3: Perceived unsafe places by residents of more affluent area
101
102
Figure 4: Perceived unsafe places by residents of less affluent area
Figure 5: Differences in perceived unsafe places by men and women in

percentages
103
104
References
Bernasco, W. & F. Luykx (2002). De ruimtelijke spreiding van woninginbraak; Een
analyse van Haagse buurten, Tijdschrift voor Criminologie, 44, 231246.
Coulton, C., J. Korbin, T. Chan & M. Su (2001). Mapping residents' perceptions of
neighborhood boundaries: a methodological note, American Journal of Community
Psychology, 29, 371-383.
The Gallup Poll [Online]. Retrieved from http://www.galluppoll.com/ [28 December
2006] Princeton, NJ: The Gallup Organization, Inc.
Greenberg, S. (1980). Characteristics of high and low crime neighborhoods in Atlanta,
1980 [Computer file]. ICPSR ed. Ann Arbor, MI: Inter-university Consortium for
Political and Social Research [producer and distributor], 1997.
Horan, M. (1999). What students see: sketch maps as tools for assessing knowledge of
libraries. The Journal of Academic Librarianship, 25 (3), 187-201.
Kesteren, J. van, Mayhew, P. & Nieuwbeerta, P. (2000). Criminal Victimisation in Seventeen Industrialised Countries: Key Findings from the 2000 International Crime
Victimisation Survey (Vol. 187). The Hague: WODC.
Kitchin, R.M. (1994) Cognitive maps: what are they and why study them? Journal of
Environmental Psychology, 14: 1-19.
Loukaitou-Sideris, A. (1999). Hot Spots of Bus Stop Crime. The Importance of Environmental Attributes. Journal of the American Planning Association, 65, 395-411.
McCoy, D.C. (1997). Evaluation of community policing initiatives in Jefferson County,
West Virginia, 1996-1997 [Computer file]. ICPSR version. Charles Town, WV: Focus (Free Our Citizens of Unhealthy Substances) [producer], 1997. Ann Arbor, MI:
Inter-university Consortium for Political and Social Research [distributor], 2000.
McPherson, M. (1978). Realities and perceptions of crime at the neighborhood level.
Victimology, 3(3/4), 319-328.
McPherson, M., Silloway, G. & Frey, D. (1983). Crime, fear, and control in neighborhood commercial centers: Minneapolis and St. Paul, 1970-1982 [Computer file]. 2nd
ICPSR version. Minneapolis, MN: Minnesota Crime Prevention Center, Inc. [producer], 1983. Ann Arbor, MI: Inter-university Consortium for Political and Social
Research [distributor], 1998.
Oeveren, M.E. van, G.I.G.L. Ellis, S. Gosepa, L.J. Branderhorst & C.M. Klauwer
(2006). Onveiligheidsbeleving in kaart gebracht: Een onderzoek naar de grafische
weergave van onveiligheidsbeleving. Unpublished bachelor thesis, University of
Leiden, Department of Criminology.
Ratcliffe, J. H. & McCullagh, M. J. (1998). Identifying repeat victimization with GIS.
British Journal of Criminology, 38(4), 651-662.
Taylor, R.B. (1995a). Impact of neighborhood structure, crime, and physical deterioration on residents and business personnel in Minneapolis-St.Paul, 1970-1982 [Computer file]. ICPSR version. Philadelphia, PA: Temple University [producer], 1995.
105
Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 1998.
(1995b). Responses to Disorder: Relative Impacts of Neighborhood Structure, Crime,
and Physical Deterioration on Residents and Business Personnel. Philadelphia: Department of Criminal Justice; Temple University.
(1996a). Impacts of specific incivilities on responses to crime and local commitment,
1979-1994: [Atlanta, Baltimore, Chicago, Minneapolis-St.Paul, and Seattle] [Computer file]. ICPSR version. Philadelphia, PA: Temple University [producer], 1996.
Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 1998.
(1996b). Neighborhood responses to disorder and local attachments: The systemic
model of attachment, social disorganization, and neighborhood use value. Sociological Forum, 11(1), 41-74.
(1997a). Social order and disorder of street blocks and neighborhoods: Ecology, microecology, and the systemic model of social disorganization. Journal of Research in
Crime and Delinquency, 34(1), 113-155.
(1997b). Crime, Grime, Fear, and Decline: A Longitudinal Look (Final Report).
Washington, DC: United States Department of Justice; National Institute of Justice.
(1999). Crime, Grime, Fear, and Decline: A Longitudinal Look. [Washington, DC]:
U.S. Dept. of Justice Office of Justice Programs, National Institute of Justice.
Vanderveen, G.N.G. (2006). Interpreting Fear, Crime, Risk and Unsafety; Conceptualisation and measurement. The Hague: Boom.
Warr, M. (1987). Fear of victimization and sensitivity to risk. Journal of Quantitative
Criminology, 3(1), 29-46.
Wassenberg, F.M. Arnoldus, R. Goetgeluk, F. Penninga & L. Reinders; Ministerie van
VROM (2006). Hoe breed is de buurt? Typologie van woonmilieus: herkenbaar,
bruikbaar en beschikbaar, VROM 5322/Januari 2006, Den Haag: VROM.
106
107
Vulnerability Mediating the Perceived Risk

Fear of Victimization Linkage?
Testing a Transactional Theory of Fear of Crime
Using Data from Austria
HELMUT HIRTENLEHNER
1. Introduction
Fear of crime is a topic which has its own tradition and cycles. Since the beginning of the 60s the feeling of security or insecurity concerning crime has been
the subject of many criminological studies. A large amount of the research done
uses the concept of vulnerability, mostly in order to explain the increased sense
of fear suffered by certain groups of the population, such as women, older people, people living in poverty, etc. (cf. Bals 2004; Carcach et al.1995; Killias
1990; Killias/Clerici 2000; Mawby 1988; Pantazis 2000; Skogan/Maxfield 1981).
Vulnerability is either defined as the ability to cope with a situation (meaning the
ability to deal successfully with a possibly dangerous situation connected with
crime) or as the ability to recover (meaning the ability to restore the physical,
mental, social and economic situation to what it was before the crime took place).
People who feel unable to protect themselves, either because they cannot run
fast, or lack the physical prowess to ward off attackers, or because they cannot
afford to protect their homes, or because it would take them longer to recover
from material or physical injuries, might be expected to fear crime more than
others (Hale 1996, p. 95).
According to Killias (1990), it is possible to distinguish between a physical
and a social dimension of vulnerability. The physical dimension emphasises the
physical ability to cope with a situation (by running away or defending oneself,
but also being able to recover if injured, and it also includes any physical reason
that someone could be more liable to attack eg. gender). In contrast, the social
dimension emphasises the perceived or expected amount of social support. It becomes clear that fear of crime can only be properly understood in the full context
of the risk of victimization, controllability and the expected consequences of
criminal acts (Killias 1990).
108
H. Hirtenlehner
The following study deals with vulnerability referring to competence in coping

with threatening situations, with the additional factor of perceived social support,
and then flows into a transactional model of the development of crime-related
feelings of insecurity. There is an interactive model of understanding widely accepted in German-speaking professional circles which is based on transactional
stress and emotion theories (Lazarus/Averill 1972; Lazarus/Folkman 1984). This
interactive model sees fear of victimization as the result of a disparity between
the risk of a specific crime and the chances of coping with that crime (Boers
1991; 1993; 2003; Boers/Kurz 1997; 2001). Fear of crime is said to arise through
the interaction of a primary appraisal of a potentially threatening situation and a
secondary appraisal of the persons own ability to deflect the danger, where it is
the discrepancy between the two which causes fear. The popularity of this model,
however, should not obscure the fact that it has only been insufficiently tested up
to now. There are few studies in which this interactive model has been explicitly
examined. When researchers deal with the topic, they generally do not inquire
about a persons subjective appraisal of their competence to deal with danger or
to cope with it. Instead of this, they attempt to determine coping ability indirectly
through socio-demographic characteristics such as age, gender, education and
income. The social resources that a person can mobilise in a potentially threatening situation are generally completely ignored. When the data is analysed, most
of the times evaluation strategies are used which missspecify the relationship
between risk and coping. Risk appraisal and social variables are generally introduced simultaneously in regression models without allowing for their interaction.
In this way the potential main effects can be isolated, but it is not possible to determine any combined effect of risk perception and vulnerability.
The aim of the following work is to conduct an empirical-statistical examination of the interactive model of fear of crime which takes the above concerns into
account. The test case is a survey of 558 persons living in the city of Linz, Upper
Austria, conducted in June 2004.
2. Conceptual clearings
Looking through the different papers on fear of crime it quickly becomes apparent that, in spite of all the research, the terms used are not clear. The term
fear of crime has so many meanings that it is tempting to say that there is a terminological confusion (Fattah 1993; Ferraro/LaGrange 1987; Hale 1996).
Based on the 3 component model of attitudes (Rosenberg/Hovland 1960), it
seems reasonable to view crime-related fear as a theoretical construct with three
dimensions: cognitive, affective and conative (Boers 2003; Fattah/Sacco 1989;
Testing a Transactional Theory of Fear of Crime Using Data from Austria
109
Gabriel/Greve 2003; Hale 1996). The cognitive component is a persons appraisal of risk, which is often measured as the subjective perception of how probable
it is to become a victim of a crime. The cognitive dimension also comprises
someones appraisal of his own ability to cope with danger (coping ability). The
affective component is the crime-related feeling of insecurity. Fear of crime in
the stricter sense describes an emotional response of dread and anxiety to crime
or symbols that a person associates with crime (Ferraro 1995, p. 23). The conative component refers to behavioural reactions such as avoidance and safety
seeking behaviours. The following work focuses on the affective component,
namely on the emotional reaction to crime and the symbols associated with it.
The relationship between the cognitive anticipation of risk and the emotional
fear of crime is often interpreted as causal. 1 Many studies which deal with risk
appraisal and fear of crime show the perception of victimization risk as a preceding determinant for the feeling of insecurity (Ferraro 1995; Giles-Sims 1984;
Jackson 2004; Lagrange/Ferraro 1989; LaGrange et al. 1992; Mesch 2000;
Smith/Torstensson 1997; Smith et al. 2001; Warr/Stafford 1983). Possible moderators of the risk-fear relationship are generally ignored. The conclusion is usually drawn that, with the anticipation of risk, a substantial determinant for the
feeling of insecurity has been found, and no further consideration is given to the
common discrepancy between the perceived risk and the feeling of safety which
can often be seen when the effect strength is only between .30 and .50.
3. Theoretical framework
3.1
Transactional Model of Fear of Crime
Klaus Boer has presented an interactive model for the understanding of crimerelated feelings of insecurity in many of his publications (1991; 1993; 2003;
Boers/Kurz 1997). Inspired by the transactional stress and emotion theory (Lazarus/Averill 1972; Lazarus/Folkman 1984) he rejected simplistic notions of a linear relationship between fear of crime and external factors such as crime reports,
experiences of victimization and incivilities. Lazarus and his colleagues assume
that fear is the result of two cognitive assessment processes: a primary appraisal
of the threat present in the environment and a secondary appraisal of personal
coping abilities. The conativeemotional result of these two processes of assessment can either be inactivity and helplessness accompanied by anxiety, or flight
and avoidance behaviour corresponding to fear, or it can be a defensive action
1
Objections to this procedure come from Gabriel/Greve (2003, p. 619 f) who explicitly reject
a causal order of the components of crime-related insecurity.
110
H. Hirtenlehner
coupled with anger. Only when there is a discrepancy between the perceived demands of the situation (the situation is perceived to be dangerous) and the resources available to deal with the situation (lack of confidence in problemsolving skills) do feelings of anxiety and fear arise. 2 Applied to fear of crime this
means that an understanding of the emotional reaction to crime is not possible
without taking the individual coping abilities into consideration. In this context
coping ability means the personal appraisal of the ability to deal with dangerous,
crime-related situations successfully. Risk appraisal was introduced as a counterpart to the appraisal of the situation. It deals with the perceived likelihood of becoming a victim of crime. Fear only arises when there is a discrepancy between
the threatening content of the situation (operationalised as risk appraisal) and the
available coping resources; meaning when insufficient coping ability can be assumed for a particular perception of risk.
The analytical differentiation between cognitive risk appraisal, affective fear
of crime and personal coping ability enables a considerably deeper and more detailed understanding of how crime-related feelings of insecurity are formed, in
contrast to the previous approaches (victimization perspective, media hypotheses,
disorder thesis, concepts of social inclusion, cf. Hale 1996); and, additionally,
existing research results can be integrated better. Work on relevant literature also
led Boers (2003) to the conclusion that experiences of victimization, media reports and signs of social disorganisation affect the perception of risk more than
they affect the fear of crime. Cognitive risk appraisal acts as a mediator between
the person and the environment which influences the level of fear only indirectly
through the personal coping abilities.
Whereas it has proved easy to collect information on risk appraisal, personal
coping abilities have been insufficiently surveyed up to now. Most empirical
studies attempt to collect data on coping abilities indirectly through using sociodemographic variables such as age, gender, education and income as indicators
(Boers 2003). A direct measurement of the subjective evaluation of personal coping abilities has been neglected in the majority of cases. For this reason, existing
findings on the interactive model are not really convincing. In fact, there are innumerable studies which document a connection between the level of fear and
the assessment of risk (Ferraro 1995; Giles-Sims 1984; Jackson 2004; LaGrange/Ferraro 1989; LaGrange et al. 1992; Mesch 2000; Smith/Torstensson
1997; Smith et al. 2001; Warr/Stafford 1983) or the socio-demographic variables
2
Although primary and secondary appraisal should be carefully separated analytically, in

practice both appraisal processes are simultaneous and influence each other (Lazarus/
Folkman 1984).
111
mentioned (Bennet/Flavin 1994; Carcach et al. 1995; Borooah/Carcach 1997;

Box et al. 1988; Ferraro 1995; LaGrange/Ferraro 1989; Mawby 1988; Pantazis
2000; Rountree/Land 1996) However, it is precisely the assessment of the personal ability to cope with danger which is generally neglected. Additionally, risk
appraisal and social variables are mostly introduced simultaneously in regression
models without allowing for their interaction. This means that the potential main
effects of these characteristics are isolated, but there is just no modelling of an
interactive effect between risk perception and coping indicators. The few studies
in which data on personal coping skills is collected explicitly often show the expected relationship between coping abilities and fear of crime, but generally ignore risk appraisal (Farrall et al. 2000; Killias/Clerici 2000; van der Wurff et al
1989) or do not specify any interaction effects (Bals 2004). On the whole it can
be assumed that the explanatory power of the interactive model has only been
examined roughly up to now.
An attempt to find a relationship between fear of crime, anticipated risk, and
coping abilities collected through direct questioning has been undertaken by
Klaus Boers and Peter Kurz (2001). They refrained from a causal examination of
the interactive model but used a multiple correspondence analysis to illuminate
the relationship patterns. Although the results of the correspondence analysis are
not explicit and convincing, the authors conclude that coping ability has an instrumental function, and therefore find that the interactive approach is confirmed.
In another study, Boers and Kurz (1997) attempted to deduce coping abilities
from milieu-related vulnerability. They based this on the assumption that in social milieus where there is a higher degree of physical, mental and social vulnerability, personal coping abilities are felt to be lower and therefore there is a
greater fear of crime. The different levels of fear observed between the social
milieus were also seen as conforming to the theory by the authors, but ultimately
showed no direct validation of the interactive explanatory model.
3.2
Self-efficacy and perceived collective efficacy as indicators of

vulnerability
Fear is influenced by the perception of risk, and this influence depends on the
capabilities for coping successfully with threatening situations involving crime.
Within these capabilities, it is necessary to differentiate between personal and
social resources (Skogan/Maxfeld 1981, p. 60 f.). Personal resources arise from
the individuals skills and abilities. The concept of self-efficacy presented by Albert Bandura in social sciences (1997) could be of help here. Self-efficacy describes the belief in a persons own abilities to cope successfully in situations not
112
H. Hirtenlehner
previously encountered. Perceived self-efficacy is defined as peoples judgements of their capabilities to organise and execute courses of action required to
attain designated types of performances (Bandura 1986, p. 391.) This refers to a
general confidence in a persons own problem-solving skills which can be effective also in situations where crime is involved. If someone has a basic confidence
in his own abilities, he also feels able to deal with situations in which protected
interests are compromised.
It is necessary to differentiate between general perceived self-efficacy and the
specific ability to deal with crime-related risk situations (defending oneself
against a physical attack, for instance). The latter could be called specific perceived self-efficacy, referring to someones judgement of his ability to deal with
a dangerous crime-related situation so that it is possible to get away unharmed,
either by fighting or fleeing.
As well as individual coping abilities, social resources, meaning external help
and support, may also be available to help cope with the danger. The expectation
of support in crisis situations can be directly connected to the concept of collective efficacy in communities which was developed by Sampson and colleagues
(1997; 1999; 2004). Collective efficacy describes the ability of a community to
carry out informal social control and in this way to uphold social order in public.
It arises from the integration and cohesion amongst residents in an area. It takes
form and becomes measurable when limited to the readiness shown by residents
of a neighbourhood to protect common goods such as safety and order. There is a
close connection between this and perceived social support in problem situations.
The belief that other people will come to help when accepted rules of behaviour
are broken in a neighbourhood provides an indicator of what support someone
can expect if a crisis should occur. The estimated willingness of other people in
the neighbourhood to intervene works as a proxy variable for the support which
someone can expect when dealing with a threatening situation. 3
The ground-breaking work of Lewis/Salem (1986) has already shown that citizens feelings
of security can only be adequately understood in the context of their confidence in the efficacy of local institutions.
113
4. Method
4.1
Data
The data in the following study was gathered in a survey of 558 citizens of the
city of Linz 4 in Austria, carried out in June 2004. The study was carried out by
sociology students during a course on empirical social research. A total of 600
persons over 20 were interviewed according to a quota sample using the characteristics age, sex and district as quota criteria. In order to avoid distortion through
interviewer effects and social desirability as far as possible, face to face interviewing was avoided, and a drop-off technique was preferred. The respondents
received a questionnaire to fill in by themselves, and this questionnaire was
picked up two days later by the same student who had delivered it. The final participation rate of 93 % was caused by one student dropping out during the collection phase. Nevertheless, the sample population measured by age, sex and district - is representative for the population of Linz.
4.2
Measurement
Crime-related feelings of insecurity: Fear of crime was measured as worry concerning specific named offences. The respondents were asked to assess their feelings of worry concerning the crimes listed in Table 1. Responses on a Likert-type
scale ranged from 1 (extremely worried) to 5 (not worried at all).
Table 1:
Fear of crime
Robbery
Assault/fight
Abuse
Burglary
Fraud
Theft and pickpocketry
Sexual harassment
Rape
M
3.56
3.56
3.43
3.17
3.59
3.17
3.70
3.74
SD
1.11
1.12
1.05
1.22
1.12
1.25
1.26
1.32
ri;t-i
.76
.82
.75
.70
.70
.74
.83
.81
Cronbachs = .93;
M = arithmetic mean; SD = standard deviation; r i,t-i = corrected item-total correlation.
Linz is the capital of Upper Austria, has approximately 200,000 inhabitants and is known
for its numerous industries and a good situation on the labour market. The official unemployment rate was 3.5% in June 2004 (Electronic Information System of the Labour Market
Service in Upper Austria, http://www.ams.or.at/neu/ooe/1400.htm, downloaded June 7,
2005).
114
H. Hirtenlehner
Risk appraisal: The respondents were asked to state how probable they considered it that they would be the victim of a range of selected crimes and annoyances within the next 12 months. A scale of 1 (very likely) to 4 (unlikely) was
provided to grade the answers.
Table 2:
Perceived risk
... being subject to abuse

... being hit and injured
... having something stolen
... being attacked and robbed
... having your home broken into
... being sexually harassed
... being raped or attacked sexually
M
2.62
3.39
2.84
3.22
3.01
3.49
3.59
SD
.83
.63
.74
.69
.76
.62
.60
ri;t-i
.43
.60
.65
.69
.51
.59
.58
Cronbachs = .83;
General self-efficacy: The basis for measuring confidence in general self-efficacy is

the inventory Dimensions of Self-Efficacy developed by Jerusalem and Schwarzer
(2004). Six items were selected from their sub-scale on general self-efficacy, which
show the expected ability to deal with various challenging situations. The items are
formulated as statements on the general subjective problem-solving abilities, which
can be assessed on a scale of 1 (completely true) to 4 (not true at all).
Table 3:
General self-efficacy
I can remain calm when facing difficulties because I can

rely on my coping abilities.
Thanks to my resourcefulness, I know how to handle
unforeseen events.
If someone opposes me, I can find ways and means to get
what I want.
I can usually handle whatever comes my way.
If I am in trouble, I can usually think of a solution.
I am confident that I could deal efficiently with unexpected events.
M
2.00
SD
.61
ri;t-i
.65
2.14
.61
.60
2.19
.63
.62
1.97
1.98
1.83
.64
.58
.57
.67
.74
.69
Cronbachs = .86;
Specific self-efficacy: The crime-related coping abilities were measured using the
example of being physically attacked by another person. Several different kinds
of reaction were provided for which the respondent could assess the chances of
being successful in coping with the attack. The chances of success could be
evaluated on a scale of 1 (very likely) to 5 (very unlikely).
Table 4:
115
Specific self-efficacy
Successful physical defence

Run away successfully
Defuse the situation by talking
Deter through confident appearance
M
2.97
2.95
3.02
2.82
SD
1.01
1.13
.96
1.03
ri;t-i
.65
.56
.53
.64
Cronbachs =.79;
Collective efficacy: The collective efficacy in residential areas was operationalised in a similar way to Sampson et al. (1997; 1999) by assessing the probability
that someone from the neighbourhood would intervene in cases of conflict and
disorderly conduct. The actual question was Imagine the following things happening in your neighbourhood. What do you think? Would someone ask the people involved to stop? The probability of intervention was evaluated on a scale of
1 (yes, of course), 2 (yes, probably), 3 (probably not) or 4 (definitely not).
Table 5:
Collective efficacy
Young people making noise

Damaging property intentionally
Hassling passers-by
Fighting between young people
Sexual attacks on women
M
2.33
2.19
2.56
2.68
2.18
SD
.87
.90
.89
.88
.99
ri;t-i
.62
.77
.77
.70
.67
Cronbachs = .88;
As well as the concepts important for the core area of the interactive model, a
number of control variables were used in the study. Incivilities were measured as
irritation concerning physical and social nuisances in the neighbourhood. The
respondents had to indicate how much of a problem for their area certain indications of disorder were, for instance litter on the streets, homeless people begging,
etc. (Cronbachs = .87). To measure the personal level of victimization the respondents were given a list with different kinds of offences and nuisances where
they had to state if that had happened to them in the last 12 months. Concerning
vicarious victimization, information about victims in the social environment was
collected. Using the same list of offences, the respondents could say whether any
of these offences had happened to anyone in their family or to anyone of their
friends and acquaintances. Age was measured in years, education was dichotomised on the Abitur level (school leaving exam qualifying for university entrance): lack of it was given the code 0; education at and above this level received the code 1. Men were designated by 0 and women by 1.
116
4.3
H. Hirtenlehner
Model specification and analysis
Figure 1 shows the theoretical model underlying the research. The interactive
model assumes that the effect of perceived risk on fear of crime is moderated by
different coping potentials (general and specific self-efficacy, perceived collective efficacy). If it appears that an anticipated crime can be successfully dealt
with through ones own coping resources, then the perceived risk should have no
effect on the feeling of safety. Conversely, if one has little confidence in ones
chances to master the situation successfully, then an increased awareness of risk
should enhance the fear of crime.
Figure 1: Theoretical Model
Collective
Efficacy
Perceived Risk
General
Self-Efficacy
Specific
Self-Efficacy
Fear of Crime
The statistical examination of the theoretical model took place through a series of
multiple regression analyses (Cohen et al. 2002). Complex inter-relations, as formulated in the core area of the interactive model, are not open to procedures of causal
analysis, which only take additive effects into account. They necessitate the expansion of conventional regression models to include interaction terms (Aiken/West
1991). Interaction terms specify the combined influence of two or more independent
variables on the target variable. Arithmetically they are the product of the predictors
concerned. To eliminate a multicollinearity of the product terms with the incoming
independent variables, a mean centring of the predictors involved followed by a
multiplication of the deviation scores to build the interaction terms is recommended.
The dependent variable and other predictors which may be included in the regression equation do not need to be centred, however.
117
The inclusion of interaction terms changes the interpretation of the isolated effects of the predictors in the interaction. It is no longer possible to speak of main
effects as they are generally understood. The first order regression coefficients
henceforth represent the conditional effect of a predictor at the zero point of the
other predictors (Aiken/West 1991, p. 37 ff). In the case of centred data, the zero
point marks the mean. The first order regression coefficient of the individual predictors included in the interaction then indicates the conditional effect of one
variable at the mean of the other variable.
The inclusion of interaction effects also results in the fact that the standardised
solutions which are automatically produced by statistical software are no longer
usable 5 (Aiken/West 1991, p. 43). To obtain undistorted standardised regression
coefficients for the core area of the interactive model too, the procedure recommended by Friedrich (1982) was used. First the initial data was standardised,
then the cross product of the standardised predictors was obtained and only after
this the standardised criterion variable was regressed on the predictors. The unstandardised solution of this analysis is the appropriate standardised solution of
equations with interaction terms (Aiken/West 1991, p. 44).
5. Results
Presented below is a series of linear regression analyses whose specification
represents the relationship postulated in the interactive model between the risk
appraisal and the coping resources on the one hand and the fear of crime on the
other. The core of the interactive approach is the assumption that there is interaction between risk and coping. At the centre of the regression models there is a
group of interaction effects which represent the interdependence between risk
perception (primary appraisal) and trust in available coping abilities (secondary
appraisal). For statistical reasons of model building, the first order effects of the
risk and coping variables in the interaction were additionally taken into account.
These individual effects of each predictor were of little importance when judging
the appropriateness of the theory 6 .
In the first step delict-unspecific regression analyses were calculated. The values
of items included in a latent concept were added up and if the concept was involved
in interactions the total score was centred by the mean of the scale. Table 6 shows
the results. The most important result is that no statistically significant interaction
effect could be seen. Neither the general nor the specific self-efficacy or the per5
6
The analysis of data was conducted with SPSS 12.0.

A pure main effect model was calculated for the sake of completeness, and also to assess the
direct influence of the risk and coping variables.
118
H. Hirtenlehner
ceived readiness of residents to intervene had any moderating effect on the influence
of perceived risk on the feeling of safety. Perceived risk proved to be a direct determinant whose effect did not depend on the assessment of coping abilities. The intensity of fear increases together with the perception of risk but, in contrast to the theory, it is not affected by the coping resources available.
Table 6:
Delict-unspecific regression analyses of fear of crime (standardised

regression coefficients)
Risk appraisal
General self-efficacy
Specific self-efficacy
Collective efficacy
Risk appraisal*general self-efficacy
Risk appraisal*specific self-efficacy
Risk appraisal*collective efficacy
Incivilities
Personal victimization
Vicarious victimization
Age
Women
Abitur level of education
Model fit
R2
*
**
***
-----
Model 1
.51***
.01
-.07
.02
.05
-.02
-.07
-------------------
Model 2
.47***
.02
-.06
.04
.07
-.03
-.06
.15***
-.09*
-.03
.02
-.01
-.07
Model 3
.45***
.03
-.07
.03
---------.15***
-.09*
-.03
.02
-.01
-.07
.27***
.30***
.29***
p .05;
p .01;
p .001;
Not included in the model.
Models:
Model 1= only the core area of the interactive model;
Model 2 = core area with added predictors;
Model 3= pure main effect model.
Figure 2 shows the conditional effect of risk anticipation on the feeling of security for different levels of coping. Reference points for all three coping variables,
taken simultaneously, were once, the respective mean (medium level of coping)
once, the mean decreased by one standard deviation (low level of coping) and
once, the mean increased by one standard deviation (high level of coping)
(Aiken/West 1991, p. 13). It is immediately obvious that the conditional regression lines run almost parallel to each other. This shows that almost identical riskfear relationships can be observed for the different coping levels a result that is
in direct contradiction to the interactive approach.
fear of crime
Figure 2
119
Conditional effects of risk perception on fear of crime for different

levels of coping
50
40
coping capacity
high
30
(B = 1.223)
medium
(B = 1.151)
low
(B = 1.079)
20
0
10
15
20
perceived risk
B = conditional unstandardised regression coefficient.
It may be that in some fields of crime there is a relationship which conforms to

the theory, but that this is obscured because the analysis mingles different forms
of crime. In a subsequent analysis a separate regression model was calculated for
each offence (table 7). The main result of these analyses was that no interdependency between risk and coping could be found even in offence-specific models.
The perceived risk affects the offence-specific level of fear and this is not affected by the coping capacities.
In all of these models it was possible to see direct effects of perceived risk. In
property offences (theft, burglary, robbery), the specific self-efficacy showed
some modest effects on the feeling of safety. An increase in the ability to deal
with dangerous situations means a slight decrease of fear, but this is completely
decoupled from the perception of victimization risk. In total, the fear of crime
proves to be a product of the anticipated risk, which is sometimes also influenced
by the individual coping skills. However, interdependency between risk appraisal
and coping abilities on the formation of crime-related fear cannot be proved with
this method.
120
Table 7:
H. Hirtenlehner
Delict-specific regression analyses of fear of crime (standardised

regression coefficients)
Abuse
Assault
Rape
Sexual
harassment
Theft
Burglary
Robbery
Risk appraisal
.46***
.49***
.37***
.35***
.32***
.43***
.45***
General selfefficacy
.05
.02
.04
-.04
-.00
-.01
.01
Specific selfefficacy
-.11*
-.11*
-.12*
-.09
-.07
-.07
-.07
Collective efficacy
.01
.03
-.05
-.02
.02
.05
.05
Risk appraisal*
general self-efficacy
.05
.01
.00
-.01
.03
.04
.02
Risk appraisal*
specific selfefficacy
.02
-.02
-.01
.01
.07
.03
.02
Risk appraisal*
collective efficacy
-.04
-.07
.00
.04
-.05
-.03
-.05
0.24**
*
.27***
.17***
.14***
.12***
.21***
Model fit
R2
.22***
* = p .05
** = p .01
*** = p .001.
6. Discussion
The results of this study cast doubt on the ability of a transactional theory to explain the development of crime-related feelings of insecurity. No support was
found for the central thesis of the interactive approach, which is that the influence of perceived risk on feelings of security is tempered by the appraisal of personal coping resources. No significant interaction was found between the risk
appraisal and the subjective perceived coping resources. Fear of crime grows as a
direct result of the expectations of risk; confidence in self-efficacy and perceived
social support play no intermediary role. In sum, these findings falsify the interactive model.
The reason for the lack of explanatory power of the interactive theory could lie
in several factors. We would like to exclude operational and statistical deficiencies. We have no doubts about the operationalisation of the central constructs or
in the chosen statistical model. Even if data on any one of the three coping dimensions was not collected in a valid way, the lack of any interdependency between risk and coping indicates that there are false assumptions in the interactive
121
model. A linear regression model with interaction terms added provides the closest corresponding evaluation procedure to the theory. The statistical requirements
for such a regression are fulfilled: the centred predictors are free of multicollinearity 7 , the dependant variable is practically normally distributed8 and no sign of
heteroscedasticity 9 or autocorrelation10 was observed. If instead of causal analyses a multiple correspondence analysis were used (as Boers/Kurz 1997; 2001), it
would mean weakening the core thesis of the theory and consequently reducing
its explanatory power. The transactional stress theory of Lazarus and colleagues
(Lazarus/Averill 1972; Lazarus/Folkman 1984), which is the basis for the interactive model of fear of crime, suggests that there is a causal dependence of emotions on cognitive judgement processes.
The survey approach itself seems more problematic. The level of abstraction
in data collected by survey may be too high to enable a situation-dependent relationship between cognition and emotions to be recognised. Risk appraisal, coping
assessment and feelings of safety are measured in isolation from their everyday
contexts, possibly further distorted by the atypical situation of filling in a questionnaire 11 . It is possible that there is some explanatory power present in the interactive model - for acute situations which stimulate primary and secondary appraisals - but that this explanatory power is lost through being separated from
everyday life and therefore cannot be apparent in the survey. Such a point of
view seems completely compatible with a transactional approach.
The question whether these results can be generalized or not remains unanswered. In international comparison, Austria has a relatively low level of crime,
low victimization rates, a modest level of inner-city problems and very low fear
of crime (Mayhew/van Dijk 1997; van Wilsem 2004; Sessar et al. 2004). Additionally, Austria still has a considerable degree of social security in comparison
to other countries. Although reorganization of the welfare state is also beginning
in Austria, the trust that Austrians have in institutional protection when they are
7
8
9
10
11
All tolerances exceed a value of .60.

A Kolmogorov-Smirnov test shows a significant deviation from the normal distribution
(p d .001); but the deviation is far from being a problem for the calculation, having a skewness of -.66 and a kurtosis of -.17.
Plots of the residuals against the expected values show no sign of heteroscedasticity.
The Durban-Watson test statistic is 1.95 (core area of delict-unspecific interactive model).
Farrall et al. (1997) point out that the common measurement of fear of crime in quantitative
survey research, ie. in the form of Likert scales, creates an entry point for feelings of insecurity which are only vaguely connected with crime, even if the measurement is not global but
related to specific named offences. These general insecurities can hinder the elucidation of
situational risk-fear relationships. For a better understanding of the state-trait-distinction of
fear of crime see Gabriel/Greve (2003).
122
H. Hirtenlehner
confronted with existential risk is still considerable (Hanak et al. 2004). Possibly,
the low crime rates and the trust in abstract systems (Giddens 1990) in Austria
work against any fear-stimulating effects that could arise from a lack of personal
coping abilities. This would mean that the Austrian population is not well enough
aware of fear and risk to allow the interactive model to work. This assumption
would fit well with the results obtained in this study. Measured according to the
standard indicator of fear of crime (How safe would you feel if you were alone
outside in your area after dark?), 22 % of the residents of Linz feel very or fairly
unsafe. 9 % think it is likely that they will be the victim of crime within the next
12 months 12 . This shows that the perceived risk is not high enough to explain
even the comparatively low level of crime-related feelings of insecurity. If the
level of perceived risk is too low, then any possible filtering through the coping
resources will also not have much explanatory power.
References
Aiken, L.S. & West, S.G. (1991). Multiple Regression: Testing and Interpreting Interactions. Thousand Oaks u.a.: Sage.
Bals, N. (2004). Kriminalitt als Stress Bedingungen der Entstehung von Kriminalittsfurcht. Soziale Probleme 15, 5476.
Bandura, A. (1986). Social Foundations of Thought and Action. A Social Cognitive
Theory. Englewood Cliffs: Prentice Hall.
(1997). Self-Efficacy. The Exercise of Control. New York: Freeman.
Bennett, R.B. & Flavin, J.M. (1994). Determinants of Fear of Crime: The Effects of
Cultural Setting. Justice Quarterly 11, 357381.
Boers, K. (1991). Kriminalittsfurcht. ber den Entstehungszusammenhang und die
Folgen eines sozialen Problems. Pfaffenweiler: Centauraus.
(1993). Kriminalittsfurcht. Ein Beitrag zum Verstndnis eines sozialen Problems.
MschrKrim 76, 6582.
(2003). Fear of Violent Crime, in: W. Heitmeyer & J. Hagan (Hrsg.), International
Handbook of Violence Research. Dordrecht: Kluwer Academic Publications. 1131
1150.
Boers, K. & Kurz, P. (1997). Kriminalittseinstellungen, soziale Milieus und sozialer
Umbruch, in: K. Boers, G. Gutsche & K. Sessar (Hrsg.), Sozialer Umbruch und Kriminalitt in Deutschland. Opladen: Westdeutscher Verlag. 187253.
12
These values were calculated as following: The values of the respondents on the seven items
measuring risk appraisal were added together; the sum was divided by the number of items
and rounded up or down to the next figure. The mean scores measure on the original scale.
123
(2001). Kriminalittsfurcht ohne Ende?, in: G. Albrecht, O. Backes & W. Khnel (Hrsg.),
Gewaltkriminalitt zwischen Mythos und Realitt. Frankfurt a.M.: Suhrkamp. 123144.
Borooah, V.K. & Carcach, C.A. (1997). Crime and Fear. Evidence from Australia. British Journal of Criminology 37, 635657.
Box, S., Hale, C. & Andrews, G. (1988). Explaining Fear of Crime. British Journal of
Criminology 28, 340356.
Carcach, C., Frampton, P., Thomas, K. & Cranich, M. (1995). Explaining Fear of Crime in
Queensland. Journal of Quantitative Criminology 11, 271 287.
Cohen, J., Cohen, P., West, S.G. & Aiken, L.S. (2002). Applied Multiple Regression /
Correlation Analysis for the Behavioral Sciences. Hillsdale: Erlbaum.
Farrall, S., Bannister, J., Ditton, J. & Gilchrist, E. (1997). Questioning the Measurement of the Fear of Crime. Findings from a Major Methodological Study. British
Journal of Criminology 37, 658679.
(2000). Social Psychology and the Fear of Crime. Re-Examining a Speculative
Model. British Journal of Criminology 40, 399413.
Fattah, E.A. (1993). Research on Fear of Crime: Some Common Conceptual and Measurement Problems, in: W. Bilsky, Ch. Pfeiffer & P. Wetzels, (Hrsg.), Fear of Crime
and Criminal Victimization. Stuttgart: Enke. 4570.
Fattah, E.A. & Sacco, V.F. (1989). Crime and Victimization of the Elderly. New York:
Springer.
Ferraro, K.L. (1995). Fear of Crime. Interpreting Victimization Risk. Albany: State
University of New York Press.
Ferraro, K.L. & LaGrange, R.L. (1987). The Measurement of Fear of Crime. Sociological Inquiry 57, 70101.
Friedrich, R.J. (1982). In Defense of Multiplicative Terms in Multiple Regression
Equations. American Journal of Political Science 26, 797833.
Gabriel, U. & Greve, W. (2003). The Psychology of Fear of Crime. Conceptual and
Methodological Perspectives. British Journal of Criminology, Vol. 43, pp. 600 614.
Giddens, A. (1990). The Consequences of Modernity. Cambridge: Polity Press.
Giles-Sims, J. (1984). A Multivariate Analysis of Perceived Likelihood of Victimization
and Degree of Worry About Crime Among Older People. Victimology 9, 222233.
Hale, C. (1996). Fear of Crime: A Review of the Literature. International Review of
Victimology 4, 79150.
Hanak, G., Karazman-Morawetz, I. & Stangl, W. (2004). Insecurities in European Cities
Local Report Vienna. Wien: Institut fr Rechts- und Kriminalsoziologie.
Jackson, J. (2004). Experience and Expression. Social and Cultural Significance in the
Fear of Crime. British Journal of Criminology 44, 946966.
Jerusalem, M. & Schwarzer, R. (2004). Dimensionen der Selbstwirksamkeit, in: A.
Glckner-Rist (Hrsg.), ZUMA-Informationssystem. Elektronisches Handbuch sozialwissenschaftlicher Erhebungsinstrumente. ZIS Version 8.00. Mannheim: ZUMA.
124
H. Hirtenlehner
Killias, M. (1990). Vulnerability: Towards a Better Understanding of a Key Variable in

the Genesis of Fear of Crime. Violence and Victims 5, 97 08.
Killias, M. & Clerici, Ch. (2000). Different Measures of Vulnerability in their Relation
to Different Dimensions of Fear of Crime. British Journal of Criminology 40, 437
450.
LaGrange, R.L. & Ferraro, K.L. (1989). Assessing Age and Gender Differences in Perceived Risk and Fear of Crime. Criminology 27, 697719.
LaGrange, R.L., Ferraro, K.L. & Supancic, M. (1992). Perceived Risk and Fear of
Crime: Role of Social and Physical Incivilities. Journal of Research in Crime and
Delinquency 29, 311334.
Lazarus, R.S. & Averill, J.R. (1972). Emotions and Cognition, in: C.S. Spielberger
(Hrsg.), Anxiety. Vol. 2. New York und London: Academic Press. 242283.
Lazarus, R.S. & Folkman, S. (1984). Stress, Appraisal and Coping. New York: Springer.
Lewis, D.A. & Salem, G. (1986). Fear of Crime: Incivility and the Production of a Social
Problem. New Brunswick und Oxford: Transaction.
Mawby, R.I. (1988). Age, Vulnerability and the Impact of Crime, in: M. Maguire &. S.
Pointing (Hrsg.), Victims of Crime: A New Deal? Milton Keynes: Open University
Press. 101111.
Mayhew, P. & van Dijk, J. (1997). Criminal Victimization in Eleven Industrialised
Countries. Key Findings from the 1996 International Crime Victims Survey. The
Hague: WODC.
Mesch, G.S. (2000). Perceptions of Risk, Lifestyle Activities, and Fear of Crime. Deviant Behavior 21, 4762.
Pantazis, C. (2000). Fear of Crime, Vulnerability and Poverty. British Journal of Criminology 40, 414436.
Rosenberg, M.J. & Hovland, C.I. (1960). Cognitive, Affective, and Behavioral Components of Attitudes, in: C.I. Hovland, M.J. Rosenberg, W.J. McGuire, R.P. Abelson &
J.W. Brehm, (Hrsg.), Attitude Organization and Change. New Haven: Yale University Press. 114.
Rountree, P.W. & Land, K.C. (1996). Perceived Risk versus Fear of Crime: Empirical
Evidence for Conceptually Distinct Reactions in Survey Data. Social Forces 74,
13531376.
Sampson, R.J. (2004). Neighborhood and Community: Collective Efficacy and Community Safety. New Economy 11, 106113.
Sampson, R.J., Raudenbush, S.W. & Earls, F. (1997). Neighborhoods and Violent Crime: A
Multilevel Study of Collective Efficacy. Science 277, 918924.
Sampson, R.J. & Raudenbush, S.W. (1999). Systematic Social Observation of Public
Spaces: A New Look at Disorder in Urban Neighborhoods. American Journal of Sociology 105, 603651.
Sessar, K., Herrmann, H., Keller, W., Weinrich, M. & Breckner, I. (2004). Insecurities
in European Cities. Crime-Related Fear Within the Context of New Anxieties and
125
Community-Based Crime Prevention. Hamburg: Department of Criminology of the

University of Hamburg.
Skogan, W. (1992). Disorder and Decline. Crime and the Spiral of Decay in American
Neighborhoods. Berkeley und Los Angeles: University of California Press.
Skogan, W. & Maxfield, M.G. (1981). Coping with Crime. Individual and Neighborhood
Reactions. Beverly Hills: Sage.
Smith, W.R. & Torstensson, M. (1997). Gender Differences in Risk Perception and Neutralizing Fear of Crime. British Journal of Criminology 37, 608634.
Smith, W.R., Torstensson, M. & Johansson, K. (2001). Perceived Risk and Fear of
Crime: Gender Differences in Contextual Sensitivity. International Review of Victimology 8, 159181.
Van der Wurff, A., van Staalduinen, L. & Stringer, P. (1989). Fear of Crime in Residential Environments: Testing a Social Psychological Model. Journal of Social Psychology 129, 141160.
Warr, M. & Stafford, M. (1983). Fear of Victimization: A Look at the Proximate Causes. Social Forces 61, 10331043.
Wilsem, J. van (2004). Criminal Victimization in Cross-National Perspective. An Analysis of Rates of Theft, Violence and Vandalism across 27 Countries. European
126
H. Hirtenlehner
127
Disorder, Social Anxieties and Fear of Crime.

Exploring the Relationship Between Incivilities and Fear of Crime
with a Special Focus on Generalized Insecurities1
HELMUT HIRTENLEHNER
Abstract
Many studies show a close relationship between perceptions of incivility and fear
of crime. The relationship is usually modelled within the outlines of a causal disorder theory, assuming that signs of incivility act as cues for norm erosion, victimisation risk and loss of control, which then lead to fear of crime. This view is
contrasted here with an approach that sees worries connected with incivility or
crime as part of a generalized syndrome of insecurity. Perceptions of incivility
and fear of crime are understood as parallel manifestations of an unspecific insecurity whose roots can be found in changes within late-modern societies. Both
conceptualizations of the incivility-fear-relationship were examined using survey
data from Austria. The results obtained from the comparison of competing structural equation models show that the data corresponds best with the generalized
insecurity approach.
Thanks are due to Professor Dr. Johann Bacher of the Institute of Sociology at the University of Linz (Austria) and Dr. Dietrich Oberwittler of the Max Planck Institute of Foreign
and International Penal Law in Freiburg (Germany) for providing many valuable suggestions after reading an earlier draft of this paper.
128
H. Hirtenlehner
1. Introduction
Many studies show that the perception of signs of social disorganisation socalled incivilities (Hunter 1978) or disorders (Skogan 1992) and fear of crime
are closely connected 2 (Carcach et al.1995; Covington/Taylor 1991; Gibson et al.
2002a; Hohage 2004; Jackson 2004; LaGrange et al.1992; Ldemann 2006;
Markowitz et al. 2001; McGarrel et al. 1997; Perkins/Taylor 1996; Robinson et
al. 2003; Ross/Jang 2000; Scheider et al. 2003; Snell 2001; Taylor/Covington
1993; Wikstrm/Dolmen 2001; Xu et al. 2005). However, there is considerably
less clarity concerning the exact form of the connection between disorder and
crime-related insecurity and the mechanisms behind the connection of these constructs. It is possible to differentiate roughly between two theoretical conceptualizations of the relationship. Social control theory oriented explanations presume that signs of disorder have the causal effect of increasing fear, and this is
thought to arise either because personal victimisation is expected or because a
lack of social control is perceived (Hunter 1978; Lewis/Salem 1986; Skogan
1992; Taylor 2001; Wilson/Kelling 1982). Approaches which are more strongly
inspired by macro-level social theory suggest that worries concerning incivilities
and fear of crime are the expression of a deeper, diffuse anxiety which arises
from transformations and changes inherent in late-modern societies (Ewald 2000;
Garofalo/Laub 1978; Hermann et al. 2003; Hollway/Jefferson 1997; 2000; Jackson 2004).
The aim of this study is to throw light on the pattern of relationships between
fear of crime and the perception of disorder. We will illuminate the intermediary
processes and mechanisms contributing to the confounding of emotions aroused
by incivility and fear of crime. Using survey data from an Austrian town we will
examine any concrete lines of connection which exist between incivilities and
fear of crime. Linear structural equation models (SEM) will be constructed for
both the disorder thesis, which is based on social control theory, and the generalization thesis, which sees both disorder-induced irritation and crime-related
fear as a manifestation of diffuse existential insecurity. These models will be
tested competitively.
The terms disorder and incivility are used synonymously in this paper, as are the terms fear,
anxiety and feelings of insecurity. For possible nuances of the meaning of the latter see
Walklate (1998).
Disorder, Social Anxieties and Fear of Crime
129
2. Conceptual clearings
Looking through the different papers on fear of crime it quickly becomes apparent that, in spite of all the research, the terms used are not clear. The term
fear of crime has so many meanings in professional circles that it is tempting to
say that there is a terminological confusion (Fattah 1993; Ferraro/LaGrange
1987; Hale 1996). As there is no consensus on what fear of crime is, it is necessary to clarify the terms used here.
Based on the three-component model of attitudes (Rosenberg/Hovland 1960),
it seems reasonable to view crime-related insecurity as a theoretical construct
with three dimensions: cognitive, affective and conative (Boers 2003; Fattah/
Sacco 1989; Gabriel/Greve 2003; Hale 1996). The cognitive component is a persons appraisal of risk, which is often measured as the subjective perception of
how probable it is to become a victim of a crime. The conative component refers
to behavioural reactions such as avoidance behaviours and security precautions.
The affective component is the crime-related feeling of insecurity. Fear of crime
in the stricter sense describes an emotional response of dread and anxiety to
crime or symbols that a person associates with crime (Ferraro 1995: 23). Two
points result from this. Firstly, fear of crime is specified as an affective sentiment
and carefully distinguished from cognitive risk anticipation. Secondly, it is no
longer limited to acts subject to criminal law, thus including the preliminary
stages of crime: annoyances and irritations which are often below the threshold
of criminalisation, but which people associate with crime (Wilson/Kelling 1982).
Such a definition emphasizes the inseparable connections between crime-related
insecurity and signs of disorder (Skogan 1992) resp. incivility (Hunter 1978).
Disorders or incivilities refer to the visible signs of destabilisation, loss of control
and neglect in the neighbourhood (graffiti, vandalism, garbage on the streets, loitering teens, etc.), which serve to create a feeling of personal insecurity 3 . What
these low level breaches of community standards (La Grange et al. 1992: 312)
all have in common is that they fall into the grey zone between delinquency and
obnoxiousness.
Skogan (1992) distinguishes between physical and social disorders. Physical disorders are
signs of neglect in buildings (e.g. neglected facades, broken windows). Social disorders refer to behaviour in public which is considered objectionable or unseemly (begging, public
intoxication). Social and physical incivilities can be differentiated in a similar way. There is
often a strong connection to be seen between the two phenomena (Lewis/Salem 1986; LaGrange et al. 1992; Ldemann 2006; Ross/Jang 2000; Xu et al. 2005).
130
H. Hirtenlehner
The relationship between the cognitive anticipation of risk and the emotional
fear of crime is often interpreted as causal 4 . Many studies which deal with risk
appraisal and fear of crime show the assessment of the risk of personal victimisation as a preceding determinant for the personal feeling of safety (Ferraro 1995;
Jackson 2004; LaGrange/Ferraro 1989; LaGrange et al. 1992; Mesch 2000;
Smith/Torstensson 1997; Smith et al. 2001). Objections to this procedure come
from Gabriel and Greve (2003: 619 f) who explicitly reject a causal ordering of
the components of crime-related insecurity.
3. Theoretical Considerations
3.1
Classical Disorder Approach
The classic disorder perspective sees fear of crime as resulting from the appearance of a neighbourhood. It assumes that the subjective feeling of safety depends
on the structure and state of the residential area (Hunter 1978; Lewis/Salem
1986; Skogan 1992; Taylor 2001; Wilson/Kelling 1982). The more tidy and
friendly the area appears, the less afraid the residents are of crime. There are presumed to be two connecting lines between incivility and fear of crime (Jackson
2004). The first connection emphasises the mediating role of perceived victimisation risk. A second causal chain, which is strongly influenced by social disorganisation theory (Shaw/McKay 1942; Sampson/Raudenbush 1999) focuses on
the importance of social capital and informal social control. Both arguments will
be outlined here.
The relationship between signs of incivility and fear of crime postulated in the
classic disorder thesis is complex and has many links. On the one hand, incivilities show a collapse of moral order in a community. It does not seem possible to
predict the behaviour of people one meets in the neighbourhood any longer.
When imponderableness increases, there is also an increase in the subjective perception of the risk of being victimised. Many signs of disorder, such as damaged
telephone boxes, walls with graffiti or used syringes lying on the pavement, suggest that criminal offences have been committed. These cues signal the presence
of crime and therefore influence peoples perception of risk, which results in an
increased fear of crime.
On the other hand, incivilities show that there is something wrong with the
neighbourhood. They suggest that the communitys capacity to regulate peoples
behaviour is impaired. Such an erosion of informal social control in the area be4
Sometimes individual vulnerability and coping skills are seen as moderators of the risk-fear
relationship. For further details see Killias/Clerici (2000).
131
comes subjectively perceptible as a feeling of not having any control over what
happens in the neighbourhood and not being able to influence any upheavals
which may occur. The impression that the community is powerless and helpless
shatters the residents feeling of security and feeds their fear of crime. From this
point of view, fear of crime appears to be the expression of unease people have
concerning the state of their close neighbourhood.
Signs of unpleasant changes in the neighbourhood do not necessarily cause
fear of crime. Lewis and Salem (1986) stress the moderating effect of local social
capital. The term social capital (Coleman 1990; Putnam 2000) refers to the
strength of social cohesion and the sum of the relationships in the neighbourhood
which are characterized by mutual trust, acceptance, recognition and respect.
This social capital is the basis for cooperation in enforcing the norms of orderly
behaviour in public. As Robert Putnam (2000: 312) states: the presence of social
capital individuals connected to one another through trusting networks and
common values allows for the enforcement of positive standards. Cohesive,
well integrated neighbourhoods can temper the fear-producing effects of external
signs of social disorganisation (Ross/Jang 2000; Ross et al. 2001). Thereby it is
possibly not the number of individual social relationships had by each resident
interpreting the situation that matters, but the impression that the neighbourhood
is intact, that social control in the district still functions. If the residents of a
neighbourhood keep an eye on each others houses and on what goes on in the
streets, then the commuter who comes home late at night and who knows hardly
anyone in the neighbourhood also benefits (Putnam 2000: 20). This commuter
can also have confidence in the local institutions and thereby be protected against
fear of crime, without actually having any increased contact to other residents.
The positive externalities of intact local networks for non-members could be one
explanation for the partly contradictory results concerning the relationship between social integration and fear of crime, as well as differences in the measurement of social integration (Gibson et al. 2002a; Hale 1996) 5 .
Collective efficacy, introduced into scientific discussion by Sampson and colleagues (1997; 1999), is being increasingly used as a measure for the extent of
social control in a community. Collective efficacy captures the link between
cohesion especially working trust and shared expectations for action (Sampson 2004: 108). It is manifested in the shared expectations of the residents of a
neighbourhood that other residents will intervene when particular rules of behav5
Sampson et al (1999) illustrate the positive externalities of intact networks of social control
by demonstrating that areas with little collective efficacy benefit from adjoining neighbourhoods with successful social control.
132
H. Hirtenlehner
iour are broken in public. This concept of collective efficacy is particularly popular in social-ecological analysis of crime (Duncan et al. 2003; Morenoff et al.
2001; Sampson et al. 1997; Sampson/Raudenbush 1999; Simons et al. 2005).
Collective efficacy has seldom been used as a predictor of fear of crime (eg Gibson et al. 2002a; Ldemann 2006; Xu et al. 2005), although using it that way
would seem to be obvious. Taylor (2002), however, remarked correctly that it is
primarily a question of the terms used. Related concepts of informal social control have already been used in research into fear of crime, although admittedly
without using the term collective efficacy (cf Taylor et al. 1984; Wikstrm/Dolmn 2001).
Here it is necessary to stress that the incivility-fear of crime relationship is
considered to be causal in the usual disorder thesis, with incivility preceding
and (...) influencing fear of crime (La Grange et al. 1992: 314). Farrall and
colleagues (1997) point out that the usual measurement of crime-related fear in
the form of Likert-type scales can be a personnel door for more general fears
even when the fear of specific named offences is measured and not just formless
fear. Decoupling the experience of insecurity from the situational contexts in
which fear really arises in favour of a generalized assessment of personal feelings
of safety leaves room for insecurities which are only vaguely connected with
crime 6 . Similarly, Jackson (2004: 961) reports that in surveys there are more
people who claim to be afraid of crime than have actually experienced situations
in which they were afraid. This means there are people who say they are afraid of
crime but who seldom feel such fear. Jackson interprets this divergence as an
indication for an expressive component of fear of crime which enables concerns
about broader social issues (2004: 962) to be expressed through crime. Kury et
al. (2004) conducted a methodological study in Germany that came to similar
conclusions. The frequently observed incivility-fear connection therefore opens
to a wide spectrum of modern insecurities. This consideration leads us to an alternative understanding of the connection between irritation caused by incivility
and fear of crime.
3.2
Disorder in the Context of Generalized Insecurity
In their diagnoses of the times many sociologists focus on the far reaching
changes in late-modern societies and the risks and insecurities they have brought
(Baumann 1997; 1998; Beck 1992; Giddens 1990; 1991; Vail et al. 1999). They
view the developed societies of the West primarily through the perspective of
6
In this connection, the comprehensive discussion of the state-trait-distinction by Gabriel/

Greve (2003) is enlightening.
133
change and call attention to fundamental transformations which are taking place
at a speed never before known in history. The political, economic, social, cultural
and ecological changes usually discussed under headings such as globalisation,
flexibilisation, and detraditionalisation do not take place without having an effect
on peoples sensitivities. The dismantling of the social security system, the deregulation of the labour market, the erosion of normal biographies, the pluralisation of life forms and other similar kinds of reorganisation of institutions which
were previously considered to be stable show that traditional certainties are disposable. Life plans and future perspectives, orientation patterns and action
knowledge, they appear all suddenly challenged. This loss of traditional certainties can be the basis for complex fears and worries. We can assume that this removal of certainty, which can be felt in all spheres of life, promotes a general
feeling of unease, in which the various risks and insecurities lose their uniqueness and blend into a generalized threat 7 (Herrmann et al. 2003).
The starting point for considering what we have called the generalization thesis is the understanding that fear of crime in social reality does not occur as a
phenomenon which can be clearly differentiated from other fears (Eve/BrownEve 1984; Ewald 2000; Garofalo/Laub 1978; Girling et al. 2000; Hermann et al.
2003; Hollway/Jefferson 1997; 2000; Jackson 2004; Taylor et al. 1996). More
often than not, fear of crime appears in connection with other social and existential fears. One could speak of an amalgam of unease, where various fears overlap
and permeate each other. In this way, fear of crime can be seen as a facet of a
wider general feeling of insecurity. Crime-related fears are taken out of their insulated reference to events close to criminal law and to social upheaval in the
immediate environment, and set into the wider field of societal problems. The
generalization thesis overcomes the concentration which the disorder approach
has on the unease caused by unwelcome changes of the neighbourhood, and is
open to more diffuse feelings of insecurity which are fed by the fundamental
transformations in late-modern societies.
From such a perspective, fear of crime appears to be the expression of an unspecific insecurity which has its basis in diffuse existential fears and an abstract
fear of the future. These generalized fears are projected onto crime, which then
appears as the tangible embodiment of the risks inherent in transformation, which
are otherwise difficult to grasp. The intra-psychological use of this projection can
7
Whereas current social theorists tend to presume that a pronounced feeling of insecurity is a
characteristic of rapidly changing late-modern societies, Pearson (1983) emphasises in his
cultural history of fear that fear of crime and violence is not a new phenomenon, and the
causes have always been sought in social change.
134
H. Hirtenlehner
be seen in a division of fear into portions (Hermann et al. 2003; based on Baumann 1999) or in a defence against anxiety (Hollway/Jefferson 1997; based on
Giddens 1991). By breaking down the diffuse anxieties into specific problems,
they become nameable and communicable, workable and sometimes even manageable. Someone who is afraid of being attacked on the street can put his fear
into words and manage his risk through his lifestyle, his avoidance behaviours
and his safety precautions. A reduction of complexity takes place (Luhmann
1995), and this creates capacity to act and confidence in the possibility to solve
problems in a world that is basically seen as unsafe.
While, technically seen, fear of crime is an indicator of an abstract anxiety, an
analysis which is more strongly influenced by sociology will see fear of crime as
a metaphor for a societal insecurity which uses crime as a code to articulate itself
(Girling et al. 2000: 170; Jackson 2004: 963; Taylor 1995: 1568). It is not always
crime which is meant when crime is spoken about! The anxieties measured under
the name of fear of crime may have more to do with general fear of the future
and existential insecurity than with the specific fear of being the victim of a
crime. This is comparable with Jackson (2004) who distinguishes between experienced fear and expressive fear. Whereas experienced fear embraces crime
and a range of things that have come to be associated with crime (960) and
therefore sticks to the classic disorder perspective, expressive fear emphasizes
concerns about broader social issues (962). In this light, fear of crime appears
to be a code for social and global changes which are difficult to name, and which
seem threatening and inscrutable for the individual.
The state of research concerning the connection between crime-related fears
and existential and social anxieties is scanty and unsatisfactory. A range of qualitative studies illustrate the practically inextricable interweaving of fear of crime
and worry concerning changes in the life-world (Girling et al. 2000; Hollway/Jefferson 2000; Stangl 1996; Taylor et al. 1996). An inter-relation between
the perception of anomie and fear of crime appears enlightening. It has repeatedly been shown that higher values of anomie are associated with increased fear
of crime (Gomme 1986; Reuband 1999; Sacco 1985). The findings on regional
inequalities in the distribution of fear of crime in Germany after the unification
speak for the assumption that crime-related fear is brought about by social
change. The political and social upheaval in the east was accompanied by an increased level of fear of crime, compared to the western districts, which was in no
relation to the distribution of registered crime (Bilsky 1995; Ewald 2000;
Kury/Ferdinand 1998). Although crime has risen in the east after the unification,
it has not risen above the level in the west. The difference in the prevalence of
135
fear can therefore not come from a difference in crime, but at best comes from a
divergence in the development of crime, which is faster in the east. 8
4. Method
4.1
Data
The data basis for this study is a survey of 558 residents of the Austrian city of
Linz conducted in June 2004. The survey was performed by students of sociology during a practicum on empirical social research. A total of 600 people over
the age of 20 were surveyed according to a quota sample plan with the characteristics age, sex, and residential area as quota criteria. The final response rate of
93% was due to the dropping out of a student during the survey phase. Table 1
shows that the sample according to age, sex and residential area is nevertheless representative for Linz 9 .
Table 1: Socio-demographic profile of the respondents and the urban population
Respondents
Persons
Percent
Age in years
20 29
30 39
40 49
50 59
60 69
70 and older
Sex
Male
Female
Area of residence
Inner city
South of the Danube
North of the Danube
Residents of Linz
Persons
Percent
91
121
111
82
66
84
16 %
22 %
20 %
15 %
12 %
15 %
23.625
31.488
26.172
23.594
18.554
25.278
16 %
21 %
18 %
16 %
12 %
17 %
256
302
46 %
54 %
86.686
96.818
47 %
53 %
185
196
174
33 %
36 %
31 %
65.254
71.712
46.538
36 %
39 %
25 %
Source: Census 2001. From: Magistrat Linz, Amt fr Stadtforschung 2003: Linz 2003. Fakten, Bilder,
Graphiken. Linz: CD-ROM.
8
An alternative explanation of the higher fear of crime in the new federal states stresses the
role of the media after the unification. The former citizens of the GDR were now confronted
with sensation-oriented reports on crime put together according to the rules of dramatic
presentation. The increased number of reports and the different way of reporting on crime
created the feeling that there was in fact an explosion of crime in the country (Reuband
2000). This attempt at explanation must be contrasted with the findings of research into the
impact of media, which suggest that crime reporting has a rather low effect on citizens feeling of security (Boers 2003; Ditton et al. 2004; Hale 1996).
Linz is the capital of Upper Austria, has about 200,000 inhabitants, and is known for many
types of industry and the good labour market situation. The unemployment rate was 3.5% in
June 2004 (Electronic data system of the Upper Austrian Labour Market Service http:www.
ams.or.at/neu/ooe/1400.htm, downloaded 7 June 2005).
136
H. Hirtenlehner
So that distortion through interviewer bias and social desirability could be

avoided as far as possible, face-to-face interviewing was avoided, and a drop-off
technique was preferred. The respondents received a questionnaire to fill in by
themselves, and this questionnaire was picked up two days later by the same student who had delivered it.
An examination of representativity concerning the level of crime-related fears
can be done by comparing the results of our survey with a 2004 mail survey
based on a randomly drawn sample of 21.000 residents of Linz (Stadtforschung
Linz 2005). 22 % of our respondents and 18 % of the randomly selected residents
report to feel somewhat or very unsafe in their residential area. The comparison
speaks for the quality of our sample.
4.2
Measurement
Incivilities: The measurement of perceived incivility combines physical and social sources of irritation. The respondents had to assess how much of a problem
for their area the nuisances given in Table 2 were. The possible answers ranged
from 1 (a great problem) to 4 (no problem). In such an operationalisation the frequency of the nuisances and their worrying effect presumably work together in a
different ratio for every person. The inclusion of the quantity dimension could
enable the incivilities to gradually gain distance from other facets of insecurity. If
the respondents sentiments connected to disorder do not fit so easily into a general insecurity as other fears do, the reason may be in the cognitive element
which is stronger in this scale than in the others. But if the indicators of disorder
thus collected fit into a generalized feeling of insecurity, there is strong proof that
sentiments aroused by incivilities are influenced by abstract insecurities.
Table 2:
Variable
V8
V9
V10
V11
V12
V13
V14
V15
Items used to measure incivilities

Dirt and litter on the streets
Graffitti on walls
Desolate buildings
Groups of loitering teens
Drunks on the street
Vandalism
Harassment of passers-by
Homeless and beggars on the streets
Cronbachs D = .87
M = arithmetic mean;
SD = standard deviation;
ri,t-i = corrected item-total correlation.
M
2.99
3.35
3.51
3.26
3.35
3.28
3.54
3.63
SD
.84
.78
.67
.81
.78
.83
.73
.65
ri;t-i
.56
.61
.55
.65
.66
.71
.69
.59
137
Fear of crime: Fear of crime was measured as worry concerning specific named
offences and annoyances. The respondents were asked to assess their feelings of
worry regarding the crimes listed in Table 3. Responses on a Likert-type scale
ranged from 1 (extremely worried) to 5 (not worried at all).
Table 3:
Variable
V22
V23
V24
V25
V26
V27
V28
V29
Items used to measure fear of crime

Robbery
Assault
Abuse
Burglary
Fraud
Theft
Sexual harassment
Rape
M
3.56
3.56
3.43
3.17
3.59
3.17
3.70
3.74
SD
1.11
1.12
1.05
1.22
1.12
1.25
1.26
1.32
ri;t-i
.76
.82
.75
.70
.70
.74
.83
.81
Cronbachs D = .93;
M = arithmetic mean; SD = standard deviation; ri,t-i = corrected item-total correlation.
Risk appraisal: The respondents were asked to state how probable they considered it that they would be the victim of a range of selected crimes and annoyances within the next 12 months. A scale of 1 (very likely) to 4 (unlikely) was
provided to grade the answers.
Table 4:
Items used to measure risk appraisal
Variable
V31
V32
V33
V34
V35
V36
V37
... being subject to abuse

... being assaulted and injured
... having something stolen
... being attacked and robbed
... having their home broken into
... being subject to sexual harassment
... being raped
SD
ri;t-i
2.62
3.39
2.84
3.22
3.01
3.49
3.59
.83
.63
.74
.69
.76
.62
.60
.43
.60
.65
.69
.51
.59
.58
Cronbachs D = .83;
Interpersonal trust: Interpersonal trust refers to the trust of the residents in the
neighbourhood. It is one component of the local social capital, and is expressed
in the awareness of cohesion and solidarity. It was measured using four statements that could be assessed on a scale ranging from 1 (completely true) to 4 (not
true at all).
138
Table 5:
Variable
V3
V4
V5
V6
H. Hirtenlehner
Items used to measure interpersonal trust.

I trust my neighbours.
Im friendly with my neighbours.
I can rely on people.
I have a lot of friends in the neighbourhood.
M
1.87
2.02
2.09
2.11
SD
.75
.90
.83
.93
ri;t-i
.52
.69
.74
.56
Cronbachs D = .81;
Collective efficacy: The perceived collective efficacy was operationalised in a

similar way to Sampson and colleagues (1997; 1999) by assessing the probability
that when a certain kind of attack or public misbehaviour occurred, someone
from the neighbourhood would intervene. The actual question was Imagine the
following things happening in your neighbourhood. What do you think? Would
someone ask the people involved to stop?. The probability of intervention was
evaluated on a scale of 1 (yes, of course), 2 (yes, probably), 3 (probably not) or 4
(definitely not).
Table 6:
Variable
V16
V17
V18
V19
V20
Items used to measure collective efficacy

Young people making noise
Damaging property intentionally
Hassling passers-by
Fighting between young people
Sexual attacks on women
M
2.33
2.19
2.56
2.68
2.18
SD
.87
.90
.89
.88
.99
ri;t-i
.62
.77
.77
.70
.67
Cronbachs D = .88;
Social fears and personal fears: Social fears refer to the unease felt concerning economic and ecological risks which arise from globalisation, the reduced ability of
national state structures to regulate matters, and the deconstruction of the welfare
state. Personal fears are more concerned with personal matters, especially worries of
what effects any negative development could have on ones own future. Social and
personal fears were measured using a common battery of questions. The respondents
were asked to use the already well-known five-step answering scale to state how
worried they felt about a range of risks and dangers. Tables 7 and 8 show the indicators used10 . It should be noted that the items for the measurement of social and personal fears were not given consecutively, but mixed.
10
One more item was included in the measurement of personal fears which was intended to
find out about worries concerning losing a job (V44). Explorative data analysis suggested
that any item dealing with potential unemployment should be removed. The fear that I will
Table 7:
Variable
V40
V41
V43
V45
V47
V48
139
Items used to measure social fears

Increased taxes
Natural catastrophes
Loss of pension
Economic crisis
Increased prices
Destruction of the environment
M
2.36
3.18
2.86
3.07
2.40
2.31
SD
1.13
1.16
1.35
1.11
1.17
1.06
ri;t-i
.62
.52
.48
.62
.68
.55
Cronbachs D = .81;
Table 8:
Variable
V42
V46
V49
V50
Items used to measure personal fears

Serious illness
Loss of friends and acquaintances
Foreigners moving in
Decrease in standard of living
M
3.37
3.87
2.67
3.19
SD
1.18
1.10
1.41
1.19
ri;t-i
.53
.43
.50
.60
Cronbachs D = .72;
4.3
Models
Data analysis was conducted using linear structural equation models (SEM)
(Byrne 2001). For both a causal model, taken from the classic disorder perspective, and a factor model, compatible with the generalization thesis, SEM were
formulated, which were first assessed individually and then compared with each
other regarding their explanatory power.
4.3.1 Measurement models

Measurement models are the lowest level of linear SEM. Whereas structural
models show relationships between the latent variables which are of interest,
measurement models specify the relationship between the theoretical constructs
and their empirical indicators. As the variables used to measure the theoretical
constructs have already been introduced, it is not necessary to give a comprehensive description of the models here. A graphic representation of the measurement
models used can be seen in figures 3 and 4. However, the measurement error correlations in the model should be explained more exactly, as this is not shown in
the figures. The literature recommends that caution should be exercised when
become unemployed is not relevant for a large part of the Austrian population (pensioners,
housewives who do not intend to take up an occupation, officials with permanent positions)
and does therefore not fit into a general factor.
140
H. Hirtenlehner
releasing correlated error measurements and only to accept residual correlations

when there is a good reason for it 11 (Byrne 2001: 134). The reasons are given
individually for this.
Correlated error terms are basically only accepted between indicators of the
same theoretical construct, and then only when particular analogies of the item
content suggest this. In fear of crime it can be assumed that the information given
by the respondents concerning property offences have more in common than is
expressed by the general fear of crime factor. Property offences are not only
combined with fear but also more with anger than is the case with violent offences
(Boers 2003; Ditton et al. 1999a,b). This additional emotional colouration should
generate a joint variance that cannot be completely reproduced by the factor
model. The residual correlation between fear of being raped and the fear of sexual harassment expresses the sex-specific importance of sexual offences. It represents the realisation that sexual attack is only an issue for female respondents
and then particularly for younger women (Ferraro 1996; Fisher/Sloan 2003; Jefferson/Hollway 2000). Women in general show a higher fear of crime than men
(Hale 1996; LaGrange/Ferraro 1989; Smith/Torstensson 1997). Whereas for the
most offences there appears to be a gradual difference between men and women,
concerning sexual offences there is a structural break which cannot be completely reflected by the measurement model used. As these objections are also
valid for the assessment of personal victimisation risks, correlated error terms
were specified for the perceived probability of being the victim of sexual harassment or rape.
Among the incivilities, the items groups of loitering teens and drunks on
the street seem to be disproportionately linked. Young people hanging out
aimlessly in streets, parks or other public places often consume alcohol and are
drunk as a result. If the problem perceptions of the respondents refer to the same
group of people, there will be an additional source of joint variance that must be
taken into consideration by specifying a measurement error correlation.
Social anxieties refer to the unease concerning a wide range of risks of which
some are more homogenous than others. Fears that taxes and duties will increase
and that everything will get more expensive seem to be closely connected, as the
second is a logical result of the first. The concerns that natural catastrophes are going to come and the environment is gradually being destroyed are both connected
to ecological dangers. In both cases the item formulation indicates a common back-
11
The unlimited release of correlated error terms would favour atheoretical model-fitting and
endanger model identification.
141
ground variable, which is not specified in the factor model, and therefore can only
be adequately reproduced by residual correlations.
4.3.2 Classical Disorder Model.

Figure 1 shows the structural model taken from the classic disorder approach
which has been augmented with elements from the social disorganisation theory.
Many signs of disorder represent the direct effects of criminal acts be it only
the often given example of broken windows (Wilson/Kelling 1982) or the debris
of drug use. Other incivilities are somewhere in the grey area between crime and
non-crime (for example graffiti or vulgar behaviour). As a preliminary stage or
consequence of crime, incivilities are open to interpretation as a sign that crime
exists. For this reason it is presumed that incivilities increase the residents perception of risk, and consequently lead to fear of crime.
Fig. 1:
Classical Disorder Model

Classical Disorder Model
Neighborhood
Cohesion
Collective
Efficacy
Fear of Crime
Disorder
Risk
Perception
Over and above this, signs of disorder act as cues for an erosion of informal social control in the area. Litter on the streets, dilapidated building facades or prostitutes soliciting in public are all signs of a neglected environment which can
negatively affect the confidence people have in their neighbours. This reduction
of interpersonal trust undermines the expectations that other community members will actively support the community. Confidence declines in the readiness of
neighbours to intervene should the standards of acceptable behaviour be broken.
The assumption is that signs of incivility wear away interpersonal trust and in
142
H. Hirtenlehner
this way impair the perceived collective efficacy 12 . The lack of expectation of
help and support in potentially dangerous situations is specified as a condition for
the development of crime-related fears.
The question remains open if there are further connections between incivility
and fear of crime in addition to the two connecting lines sketched. The dashed
pathway from incivility to fear symbolises a direct effect that was specified to
catch additional relationships between incivility and crime-related feelings of
insecurity 13 . At first the conventional disorder model is estimated without such a
direct effect, as the theory does not require it. The next step is to estimate a
model augmented by a direct incivility-fear effect. It will be necessary to test
whether the introduction of such an effect leads to a significant improvement of
the model. The existence of an incivility-fear of crime relationship that cannot be
explained by perceived risk or perceived lack of control will certainly leave room
for interpretations. These interpretations are anyway compatible with the assumption that there is an abstract anxiety which is divided down to concerns connected with crime and disorder in the neighbourhood among others.
4.3.3 Generalized Insecurity Model

To test the generalization thesis, a SEM in the form of a second-order confirmatory factor analysis (CFA) (Byrne 2001: 120 ff) was specified (figure 2). The
starting point and hierarchic centre of the model structure is a diffuse insecurity
located on the secondorder factor. This insecurity is understood theoretically as
a result of social transitions, the origins of which are not further investigated
here. The generalized insecurity is broken down into four first-order factors: social fears, personal fears, incivilities and fear of crime. This model is based on
the assumption that doubts about the stability of a neighbourhood are at least
partly to be understood as a kind of abstract anxiety. The conceptualization of
concerns connected to incivility as manifestation of an abstract feeling of insecurity as is fear of crime conforms to the opinion of Sampson and Raudenbush
(1999: 608) who see disorders and crime as being merely opposite ends of a seriousness continuum. If signs of incivility are indications that administrative laws
12
13
The discussion whether interpersonal trust and readiness to intervene form two different
theoretical constructs or whether they should be combined as one concept is artificial. There
are good reasons both for the separation and for the fusion, so that both sides have their supporters (Gibson et al. 2002a,b; Sampson et al. 1997; Taylor 2002). The assumption of a
close connection between solidarity and readiness to intervene, which is modelled causally
here for reasons of analytical clarity and theoretical stringency, is decisive for our purpose.
The incivility-fear covariance that is not covered by other links must be modelled causally
(as a regression path), as correlations can only be specified between exogenous variables in
SEM (Byrne 2001).
143
are being broken or minor offences are being committed, then they can also serve
as a projection area for existential fears in the same way as serious crime.
The arrangement of the elements in a factor structure implies that the various
insecurities are interwoven, but their covariance comes primarily from their
common roots in a more abstract unease. Between the areas of fear there is no
causal order, but they can all be traced back to an insecurity of a higher order,
and this is why they appear more or less in bundles.
Fig. 2:
Generalized Insecurity Model

Generalized Insecurity Model
Generalized
Insecurity
Social
Fears
Personal
Fears
Disorder
Fear of Crime
It should be mentioned here that a covariation of the residual terms was specified
between social fears and personal fears. Social fears and personal fears have
more in common than can be seen from the general model. Firstly, they represent
two fields of anxiety which are related conceptually but which must be separated
analytically. An explorative factor analysis (Child 2006) can extract two separate
factors 14 , which are correlated with .47 after an oblique-angled Oblimin-rotation.
Additionally, the empirical indicators of the two theoretical constructs were
given together in a common battery of questions, which is why they may reflect
the same response sets. Conceptual similarity and synchronous response sets are
represented mathematically in a correlation of the residual terms.
14
The eigenvalues of the factors are 4.42 and 1.08.
144
H. Hirtenlehner
5. Results
The statistical analysis was conducted with Amos 3.6. (Arbuckle 1997). The
maximum likelihood algorithm was used for model estimation.
5.1
Classical Disorder Approach
The SEM derived from the classical disorder approach is only partially supported
by the data. The examination criteria used for the assessment of the overall
model indicate that the disorder approach is basically compatible with the data,
but are far from providing a complete reproduction of the empirical correlation
structure 15 . The goodness-of-fit index of .85 must be considered unsatisfactory.
The comparative fit index meets exactly the acceptance threshold of .90. The root
mean square error of approximation falls below the upper limit of .08. The ratio
of F2 to degrees of freedom is just under the threshold level of 3. Although the
F2-test shows significant deviation between the empirical and the estimated covariance matrix, because of its dependence on the sample size this does not necessarily imply an insufficient model fit (Byrne 2001: 81) 16 .
The postulated relationships between disorder and fear of crime are only partially
confirmed. In total, 28 % of the variance in fear of crime could be explained. The
assumption that signs of disorder have an effect on the perception of risk was correct. Seeing external signs of social disorganisation causes an increase in perceived
victimisation risk, and this increased awareness of risk creates fear of crime. Incivilities seem really to be interpreted as cues for or symbols of crime, which indicate that
victimisation is possible, and in this way produce fear of crime. Additionally, observing incivilities undermines the trust in neighbours and thus weakens the social
cohesion in an area. If interpersonal trust is reduced, residents do not believe their
neighbours will intervene when the generally accepted standards of behaviour are
broken. Such an erosion of collective efficacy, however, has no consequences for
15
16
The threshold values for assessing the goodness-of-fit are taken from Byrne (2001: 81 ff).
With regard to the specified measurement models it should be noted that all the items used
are acceptable indicators of the theoretical constructs to be modelled. All the items load on
the allocated factor at least to a level of .50 and are significant with p .001. There were no
signs of cross-loadings of individual items on several factors. Further, the most important
measurement error correlations should be commented on. In the area of the emotions arising
from property offences, residual correlations of between .38 and .42 occur. We may well
conclude that information given by respondents concerning property crime have in fact
more in common than is expressed in the general fear of crime factor. The assumption that
anger is here an additional component cannot be rejected. Even closer is the residual correlation between rape and sexual harassment: .72 in risk appraisal and .82 in fear. The covariation of the error terms shows that the items on sexual offences have more in common than
the model presumes. The background to this additional covariation could be the genderspecific meaning of sexual delinquency.
145
the crime-related anxieties of the inhabitants of Linz. The level of crime-related fear
is independent of the perceived capacity of the community to control peoples behaviour. That the perception of informal social control is irrelevant for the residents
feeling of safety could have to do with the deeply-rooted confidence in statutory
institutions and authorities which has grown from the strong tradition of the welfare
state in Austria (Esping-Anderson 1990). Findings from Vienna show that low fear
of crime the residents there have comes mainly from a trust in abstract systems
(Giddens 1990) and only to a lesser extent from social cohesion and social capital
(Hanak et al. 2004).
Fig. 3:
V15
Classical Disorder Model (Standardized Coefficients)

V14
.75
V13
V12
V11
.79 .66
.61
V10
.57
.67
V9
.63
V8
.58
Disorder
-.26
V3
.61
.79
Trust
.89
.39
.59
V4
V5
V6
.34
.68
V31
V32
Collective Efficacy
.50
.62
.76
.80
V34
.60
.56
V35
.84
.72
V33
V36
V16
.82
Risk Assessment
.52
.77
V17
V18
V19
V20
-.00
.53
V37
Fear of Crime
.83
V22
V23
.91
.84
V24
.63
V25
.77
.62 .69
V26
V27
.78
V28
V29
F2 = 1343.65; d.f. = 453; p = .000; F2/df = 2.97; GFI = .85; CFI = .90; RMSEA = .06
The point of view taken by Lewis and Salem (1986) differs slightly from the relationship structure tested here. They question the linearity of the connection be-
146
H. Hirtenlehner
tween incivility and fear of crime and postulate a buffering effect of local social
control. Only when there is a lack of confidence in the ability of a community to
regulate peoples behaviour does perceiving signs of disorder result in feeling
doubtful about safety. As long as the neighbourhood is considered to have the
power to enforce the common norms, the feeling of security shall be resistant to
incivilities. This raises the question whether there is an interactive impact of disorder and social control. To test whether such an interactive effect exists, the
sample was divided into two groups along the median-dichotomized collective
efficacy: a low collective efficacy group and a high collective efficacy
group 17 . Models consisting of only a regression path from incivility to fear of
crime were estimated for both groups simultaneously 18 . A first estimation took
place with freely varying disorder-fear effects between the groups. In a second
estimation, invariance of the structure coefficients between the groups was postulated as an added restriction. A comparison of the fit of the two multiple group
analyses using a F2-difference test shows no significant difference between the
two models (F2diff = 2.14; d.f.diff = 1; p > .05) 19 . This means that extending the
classical disorder model with different incivility-fear relationships according to
the extent of perceived collective efficacy does not bring any improvement to the
model. The assumption that there is an interactive effect of disorder and informal
social control on peoples feeling of safety can therefore not be confirmed.
As there is no connection between collective efficacy resp. informal social
control and fear of crime, one of the core assumptions of the classical disorder
theory is shaken. Further doubts about the (sole) explanatory power of a conventional disorder approach occur with the removal of model restrictions. In a new
analysis, a disorder model was estimated which had been extended by an additional regression path running directly from incivility to fear of crime. Such a
model enables a significantly better reproduction of the data. If a hierarchical
model comparison is undertaken that is only the relationship structure between
the variables is changed, but not the number of variables then a F2-difference
test can be used to test whether the introduction of an additional effect brings a
significant improvement of the model. This is here the case (F2diff = 13.44;
17
18
19
A total score was calculated from the items measuring collective efficacy, and the median of
this score was used to divide the sample population into two halves.
The measurement models (factor loadings, error terms and error correlations) were fixed at
the values of the previous model.
With regard to the contents there is a decline in the structural coefficients which is contrary
to the postulated direction. In the model with the free varying disorder-fear effects, the standardised regression coefficient in the high collective efficacy group ( = .42; CR = 6.61)
is even higher than in the low collective efficacy group ( = .23; CR = 3.10).
147
d.f.diff = 1; p .001) 20 . A moderate positive direct effect of incivilities on fear of

crime is shown ( = .19; CR = 3.70). This means, for a start, only that there must
be another mechanism of connection between disorder-related irritation and
crime-related fear, one which is beyond risk appraisal and perceived lack of control. It says nothing about the origin of the joint dispersion of incivility sentiments and fear of crime, which is not reproduced by the disorder model initially
specified. One has to use interpretations inspired by theory here. Possibly, the
erosion of moral order in the neighbourhood, visible in the form of incivilities,
leads straight to a crime-related feeling of insecurity in the residents. This kind of
connection between incivility and fear also seems compatible with the idea of a
diffuse existential insecurity which is transported by both incivilities and crime.
With this, we can begin reach an understanding of the connections between fear
of crime and irritations arising from incivilities which is not based on the conventional disorder approach.
5.2
Generalized insecurity approach
The CFA model derived from the generalization thesis reproduces the data very
well. All the goodness-of-fit statistics show satisfactory values. The goodness-of-fit
index and the comparative fit index exceed the threshold value of .90. The root mean
square error of approximation is less than the upper limit of .08. The ratio of F2 to
degrees of freedom remains lower than the rejection threshold of 3. Indeed, the F2
test shows significant deviations of the empirical and the reproduced covariance
structures, but as a result of its dependency on the sample size this does not necessarily mean there is an unsatisfactory model fit (Byrne 2001: 81)21 .
The relationship between the first-order factors and the second-order factor is
of particular interest. The sub-dimensions of insecurity are significantly
(p .001) 22 and substantially loaded on the diffuse insecurity of the second order. On closer consideration it becomes noticeable that the social fears cannot
contribute so much to abstract anxieties as fear of crime, personal fears and fears
concerning incivilities. The loading of social anxieties on the diffuse secondorder insecurities is shown to be somewhat lower than that of the other insecurities, but at .44 it is still significant. In the same way the variance in social anxie20
21
22
The various statistics of model evaluation change only marginally: F/d.f. = 2.94; GFI = .85;
CFI = .90; RMSEA = .06.
A look at the details of the first-order measurement models shows that the items used are
acceptable indicators of the assumed insecurities. All the factor loadings are greater than .50
and significant with p .001. There was no evidence of cross-loadings of individual items
on more than one factor.
CRmin = 5.94.
148
H. Hirtenlehner
ties that is explained by the generalized insecurity (.20) is somewhat lower than
in the crime-related fears (.41), the personal fears (.41) and the disorders (.26).
The reason for the slightly worse reproduction of social fears although still to
be seen as adequate on the whole could lie in the fact that the indicators used
here do not focus so sharply on individual involvement, as they do for fear of
crime, personal fears and incivilities. Seeing the signs of dwindling stability in
ones own neighbourhood every day, having ones own integrity or ones own
possessions threatened, and fearing a negative development for ones own future
could contribute more to the development of individual feelings of insecurity
than the unease felt concerning general risks (increased taxes, economic crises,
natural catastrophes, etc) which must first be divided down to ones life-world. It
may be merely the different focuses of social fears (indirect involvement) and
personal fears (direct involvement) which could be responsible for the lack of
unidimensionality of the two constructs in the exploratory factor analysis 23 .
Fig. 4:
Generalized Insecurity Model (Standardized Coefficients)
V22
V23
.84
V24
V25
.84
.92
V26
V27
.64
.64
V28
.69
.78
V29
.78
Fear of Crime
V40
V41
.62
.61
V43
.56
.64
Social
Fears
.72
V45
.44
.75
V47
.53
Generalized
Insecurity
V48
.64
V42
.60
.51
V46
.51
Personal
Fears
.62
V49
.79
V50
Disorders
.60
V8
23
.64
V9
.58
V10
.68
V11
.66
V12
.76
.79
V13
V14
.63
V15
F2 = 663.70; d.f. = 287;

p = .000; F2/df = 2.31;
GFI = .91;
CFI = .95;
RMSEA = .05
As expected, the residual terms of social fears and personal fears are close correlated
(r = .84). The reasons for this are probably the conceptual similarity of the two constructs
and their synchronous measurement in a common battery of questions.
5.3
149
Model Comparison
The previous comments on the explanatory power of the two competing SEM
suggest that the generalized insecurity model shows a better fit to the data than
the classical disorder model. If two models are not hierarchically related to one
another, the exact comparison of the models must be carried out using information criteria of model fit. Table 9 shows the results of the comparative assessment
of the generalized insecurity model and the classical disorder model. The comparison of the goodness-of-fit was conducted using four information criteria
(AIC, CAIC, BIC, BCC). Smaller values indicate a better fit to the data.
Table 9:
Model Comparison
Information criteria:
AIC
CAIC
BIC
BCC
Variance explained in fear of crime:
SMC
Generalized
Insecurity Model
Classical
Disorder Model
791.70
1126.70
1271.22
798.87
1493.65
1882.15
2067.08
1504.68
.41
.28
SMC = squared multiple correlation.
The findings of the model comparison support the view that fear of crime has a
metaphorical meaning. The CFA model derived from the generalization thesis
shows a better fit than the classical disorder model. All four information criteria
support this view. Furthermore, the variance of fear of crime is better reproduced
by the generalized insecurity model than by the conventional disorder model.
Whereas the generalized insecurity model explains 41 % of the variance in fear
of crime, the traditional disorder model can only explain 28 % of the variance in
crime-related fears. It may then be concluded that understanding fear of crime as
an embodiment of an abstract unspecific feeling of insecurity more closely approaches the covariance structures observed than the assumption that crimerelated fears are brought about by the appearance and state of the neighbourhood.
6. Discussion
The findings of the research presented here are suitable to fertilize international
research into fear of crime in two ways. Firstly, they show, using quantitative
data, that fear of crime is a projection of more general anxieties. Fear of crime is
inseparable from other forms of insecurity. It represents one component of a generalized insecurity, whose origins can be found, according to many sociological
150
H. Hirtenlehner
diagnoses of the times, in the political, economic and social changes in latemodern societies. Secondly, the findings detect a mechanism of confounding between incivility and crime-related fears which has not been much studied up to
now. Fear of crime and feelings of incivility appear to be parallel manifestations
of a generalized syndrome of insecurity which can only be understood in the light
of social change. The often seen, inextricable interweaving between incivility
and crime-related fear seems to be caused by the fact that both are subject to projections and ciphers for diffuse transformation-related fears of the future. The
classical disorder thesis cannot be ignored completely, but the Austrian figures
did not confirm a connection between perceived deficiencies in informal social
control and fear of crime. The level of fear of crime here is independent of the
collective efficacy of the community. In contrast, however, even in urban Austria, signs of disorder do generate an anticipation of risk of criminal victimisation
which is then accompanied by fear of crime.
In summary, the data presented here give more support to the assumption that
fear of crime and its associated symbols (incivilities) is brought about by abstract
insecurities than they do promote a pure causal model hypothesizing that perceptions of disorder create fear of crime. That offences and obnoxious behaviours
below the threshold of criminalisation also mark one component of fear of crime
is not disputed here it was shown that signs of incivility cause doubts concerning the safety of the neighbourhood but our findings suggest we should agree
with Jackson (2004) who wrote that articulated fear of crime is made up of both
experienced fear and expressive fear. Fear of crime unites an emotional reaction
to symbols of danger in the immediate environment with a general unease concerning the world. The analyses carried out here suggest that the expressive component prevails. A factor model derived from a background of generalized insecurities shows a better fit to the data and explains the subjectively articulated fear
of crime better than a conventional disorder model. Fear of crime seems to be
less the result of observed upheavals in the neighbourhood and more a condensation of a diffuse insecurity, which stems from the changes and transformations of
late-modern society. The connection between incivilities and fear of crime also
appears in a new light. The assumption that there is a causal ordering between the
two constructs cannot be completely rejected24 , but should be extended by an
additional pattern of connections. Both crime and incivilities act as a metaphor
24
The assumption of a causal ordering can be based on the chain of impact running from disorders via risk appraisal to fear of crime. This could be countered by pointing out that the
cognitive risk assessment cannot be sharply distinguished from emotional insecurity and
therefore this connection, too, only reflects the common embedding of incivility sentiments
and fear of crime in a generalized insecurity syndrome.
151
through which a general unease with the changes in late-modern societies and
their institutions can be expressed.
Finally, some possible points of criticism should be discussed. Firstly, there is
undoubtedly the question of whether the findings in Austria are transferable to
other countries. Here, there are two key points to consider the first being that
the lower importance of experienced fear could be caused by the generally low
risk of crime in Austria. Compared internationally, Austria has a low crime rate,
low numbers of victimisation, and a low level of inner city problems (Mayhew/
van Djik 1997; van Wilsem 2004; Sessar et al. 2004). Additionally, Austria still
has a considerable degree of social security when compared to many other countries. Although the welfare state is also being reconstructed in Austria, the population has not yet lost confidence in institutional protection against existential
risks (Hanak et al. 2004). The question whether this trust in public institutions,
which still exists in many ways, weakens or strengthens the connections between
crime-related fear and other forms of insecurity can only be answered after additional research. For example, the irrelevance of the collective efficacy of a
neighbourhood for the crime-related feelings of security could be a result of the
complete confidence of the residents in the state and its authorities (Walklate
1998). Findings from Vienna show that the positive feeling of security the residents enjoy mainly stems from a trust in abstract systems (Giddens 1990) and
less from social ties or social networks (Hanak et al. 2004). A range of qualitative studies from Great Britain indicate that in other countries, too, crime is a cipher for a general unease which is difficult to grasp (Girling et al. 2000; Hollway/Jefferson 2000; Taylor et al. 1996).
Further objections refer to the internal validity of the study. A shortcoming of
the study is certainly that it is only based on cross-sectional data. The resulting
problems for causal analysis mainly concern the structural model derived from
the classical disorder theory and less the CFA carried out to test the generalization thesis. Nevertheless, a longitudinal study would be preferable.
Additionally, it cannot be ruled out that the connection of emotions aroused by
incivility and crime with more abstract fears may have been brought about by our
research method. Doing quantitative survey research means that feelings of insecurity, which vary according to the particular situation, are taken out of their specific context in daily life 25 and replaced by someones general assessment of his
own anxiousness, and this opens up a surface on which to project other feelings
of insecurity. Such an objection can, of course, be countered by the findings
gained from qualitative research showing that fear of crime can be fully under25
See Gabriel/Greve (2003) on state-trait-distinction.
152
H. Hirtenlehner
stood only in the context of a general unease concerning urban life and the development of society (Girling et al. 2000; Hollway/Jefferson 2000; Stangl 1996;
Taylor et al. 1996).
Regarding the classical disorder thesis, it would be desirable to have measurements of the extent of disorder in an area that are independent of the individual perceptions of the residents, for example, by using systematic social observation (Sampson/Raudenbush 1999). Further, it must be considered that our measurement of incivilities includes both aspects of how frequently observations were
made and also how intense the concern is. That this measure of perceived disorder fits seamlessly into a concept of generalized insecurity is interpreted by us as
being particularly strong evidence that irritations connected to incivilities are influenced by diffuse insecurities. We can assume that a more direct measurement
of the affective component of incivility-related irritations will show an even
closer association with other insecurities.
153
References
Arbuckle, J. (1997). Amos users guide. Version 3.6. Chicago: Smallwaters.
Bauman, Z. (1997). Postmodernity and its discontents. Cambridge: Polity Press.
(1998). Globalization: The human consequences. Cambridge: Polity Press.
(1999). In search of politics. Cambridge: Polity Press.
Beck, U. (1992). Risk society: Towards a new modernity. London: Sage.
Bilsky, W. (1995). Die Bedeutung von Furcht vor Kriminalitt in Ost und West. Monatsschrift fr Kriminologie und Strafrechtsreform 79, 357372.
Boers, K. (2003). Fear of violent crime. In W. Heitmeyer and J. Hagan (eds.) International handbook of violence research, 11311150. Dordrecht: Kluwer.
Byrne, B. (2001). Structural equation modeling with AMOS. Mahwah: Lawrence Erlbaum Associates.
Carcach, C., Frampton, P., Thomas, K. & Cranich, M. (1995). Explaining fear of crime
in Queensland. Journal of Quantitative Criminology 11, 271287.
Child, D. (2006). Essentials of factor analysis. London: Continuum.
Coleman, J. (1990). Foundations of social theory. Cambridge: Havard University Press.
Covington, J. & Taylor, R. (1991). Fear of crime in urban residential neighborhoods:
Implications of between- and within-neighborhoods sources for current models. Sociological Quarterly 32, 231249.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1999). Afraid or angry? Recalibrating the fear of crime. International Review of Victimology 6, 8399.
Ditton, J., Farrall, S., Bannister, J., Gilchrist, E. & Pease, K. (1999). Reactions to victimisation: Why has anger been ignored?. Crime Prevention and Community Safety
1, 3754.
Ditton, J., Chadee, D., Farrall, S., Gilchrist, E. & Bannister, J. (2004). From imitation
to intimidation. A note on the curious and changing relationship between the media,
crime and fear of crime. British Journal of Criminology 44, 595610.
Duncan, T., Duncan, S., Hayrettin, O., Strycker, L. & Hix-Small, H. (2003). A multilevel contextual model of neighborhood collective efficacy. American Journal of
Community Psychology 32, 245252.
Esping-Anderson, G. (1990). The three worlds of welfare capitalism. Cambridge: Polity
Press.
Ewald, U. (2000). Criminal victimisation and social adaption in modernity. Fear of crime and risk perception in the new Germany. In T. Hope and R. Sparks (eds.) Crime,
risk, and insecurity. Law and order in everyday life and political discourse, 166199.
London: Routledge.
154
H. Hirtenlehner
Eve, R. & Brown-Eve, S. (1984). The effects of powerlessness, fear of social change,
and social integration on fear of crime among the elderly. Victimology 9, 290295.
Fattah, E. (1993). Research on fear of crime: Some common conceptual and measurement problems. In W. Bilsky, C. Pfeiffer & P. Wetzels (eds.) Fear of crime and criminal victimization, 4570. Stuttgart: Enke.
Fattah, E. & Sacco, V. (1989). Crime and victimization of the elderly. New York:
Springer.
Ferraro, K. (1995). Fear of crime. Interpreting victimization risk. New York: SUNY
Press.
(1996). Womens fear of victimization: Shadow of sexual assault?. Social Forces 75,
667690.
Ferraro, K. & LaGrange, R. (1987). The measurement of fear of crime. Sociological
Inquiry 57, 70101.
Fisher, B. & Sloan, J. (2003). Unraveling the fear of victimization among college
women: Is the shadow of sexual assault hypothesis supported?. Justice Quarterly
20, 633656.
Gabriel, U. & Greve, W. (2003). The psychology of fear of crime. Conceptual and
methodological perspectives. British Journal of Criminology 43, 600614.
Gibson, C., Zhao, J., Lovrich, N. & Gaffney, M. (2002a). Social integration, individual
perceptions of collective efficacy, and fear of crime in three cities. Justice Quarterly
19, 537564.
Gibson, C., Zhao, J. & Lovrich, N. (2002b). Sociological measurement confusion, paradigmatic imperfection, and etiological nirvana: Striking a pragmatic balance in pursuing science. Justice Quarterly 19, 793807.
Giddens, A. (1990). The consequences of modernity. Cambridge: Polity Press.
(1991). Modernity and self-identity. Cambridge: Polity Press.
Girling, E., Loader, I. & Sparks, R. (2000). Crime and social change in middle England.
Questions of order in an English town. London: Routlege.
Gomme, I. (1986). Fear of crime amongst Canadians: A multi-variate analysis. Journal
of Criminal Justice 30, 249258.
Hale, C. (1996). Fear of crime. A review of the literature. International Review of Victimology 4, 79150.
Hanak, G., Karazman-Morawetz, I. & Stangl, W. (2004). Insecurities in European cities
local report Vienna. Vienna: Institute for the Sociology of Law and Criminology.
Herrmann, H., Sessar, K. & Weinrich, M. (2003). Unsicherheit in der Moderne am
Beispiel der Grostadt. In W. Stangl & G. Hanak (eds.) Innere Sicherheiten. Jahrbuch fr Rechts- und Kriminalsoziologie '02, 251286. Baden-Baden: Nomos.
155
Hohage, C. (2004). Incivilities und Kriminalittsfurcht. Soziale Probleme 15, 7795.

Hollway, W. & Jefferson, T. (2000). The role of anxiety in fear of crime. In T. Hope and
R. Sparks (eds.) Crime, risk, and insecurity. Law and order in everyday life and political discourse, 3149. London: Routledge.
(1997). The risk society in an age of anxiety: Situating fear of crime. British Journal
of Criminology 48, 255266.
Hunter, A. (1978). Symbols of incivility. Evanston: Northwestern University Press.
Jackson, J. (2004). Experience and expression. Social and cultural significance in the
fear of crime. British Journal of Criminology 44, 946966.
Killias, M. & Clerici, C. (2000). Different measures of vulnerability in their relation to
different dimensions of fear of crime. British Journal of Criminology 40, 437450.
Kury, H. & Ferdinand, T. (1998). The victims experience and fear of crime. International Review of Victimology 5, 93140.
Kury, H., Lichtblau, A., Neumaier, A. & Obergfell-Fuchs, J. (2004). Zur Validitt der
Erfassung von Kriminalittsfurcht. Soziale Probleme 15, 141165.
LaGrange, R. & Ferraro, K. (1989). Assessing age and gender differences in perceived
risk and fear of crime. Criminology 27, 697719.
LaGrange, R., Ferraro, K. & Supancic, M. (1992). Perceived risk and fear of crime:
Role of social and physical incivilities. Journal of Research in Crime and Delinquency 29, 311334.
Lewis, D. & Salem, G. (1986). Fear of crime: Incivility and the production of a social
problem. New Brunswick: Transaction Books.
Ldemann, C. (2006). Kriminalittsfurcht im urbanen Raum. Eine Mehrebenenanalyse
zu individuellen und sozialrumlichen Determinanten verschiedener Dimensionen
von Kriminalittsfurcht. Klner Zeitschrift fr Soziologie und Sozialpsychologie 58,
285306.
Luhmann, N. (1995). Social systems. Stanford: Stanford University Press.
Markowitz, F., Bellair, P., Liska, A. & Liu, J. (2001). Extending social disorganization
theory: Modeling the relationships between cohesion, disorder, and fear. Criminology 39, 293319.
Mayhew, P. & van Dijk, J. (1997). Criminal victimisation in eleven industrialised countries. Key findings from the 1996 International Crime Victims Survey. The Hague:
WODC.
McGarrell, E., Giacomazzi, A. & Thurman, Q. (1997). Neighborhood disorder, integration, and the fear of crime. Justice Quarterly 14, 479497.
Mesch, G. (2000). Perceptions of risk, lifestyle activities, and fear of crime. Deviant
Behavior 21, 4762.
Morenoff, J., Sampson, R. & Raudenbush, S. (2001). Neighborhood inequality, collective efficacy, and the spatial dynamics of urban violence. Criminology 39, 517560.
156
H. Hirtenlehner
Pearson, G. (1983). Hooligan. A history of respectable fears. London: Macmillan Press.

Perkins, D. & Taylor, R. (1996). Ecological assessments of community disorder: Their
relationship to fear of crime and theoretical implications. American Journal of Community Psychology 24, 63107.
Putnam, R. (2000). Bowling alone. The collapse and revival of American community.
New York: Simon & Schuster.
Reuband, K.H. (1999). Kriminalittsfurcht. Stabilitt und Wandel. Neue Kriminalpolitik
11/2, 1520.
(2000). Kriminalitt als Thema ostdeutscher Massenmedien vor und nach der Wende. Eine Analyse Dresdner Tageszeitungen 1988 1994. Kriminologisches Journal
32, 4355.
Robinson, J., Lawton, B., Taylor, R. & Perkins, D. (2003). Multilevel longitudinal impacts of incivilities: Fear of crime, expected safety, and block satisfaction. Journal of
Quantitative Criminology 19, 237274.
Rosenberg, M. & Hovland, C. (1960). Cognitive, affective, and behavioral components
of attitudes. In C. Hovland, M. Rosenberg, W. McGuire, R. Abelson & J. Brehm
(eds.) Attitude organization and change, 1-14. New Haven: Yale University Press.
Ross, C. & Jang, S. (2000). Neighborhood disorder, fear, and mistrust: The buffering
role of social ties with neighbors. American Journal of Community Psychology 28,
401420.
Ross, C., Mirowsky, J. & Pribesh, S. (2001). Powerlessness and the amplification of
threat: Neighborhood disadvantage, disorder, and mistrust. American Sociological
Review 66, 568591.
Sacco, V. (1985). Perceptions of crime and anomic adaptions. Crimecare Journal 1, 86108.
Sampson, R. (2004). Neighborhood and community: Collective efficacy and community
safety. New Economy 11, 106113.
Sampson, R., Morenoff, J. & Earls, F. (1999). Beyond social capital: Spatial dynamics
of collective efficacy for children. American Sociological Review 64, 633660.
Sampson, R. & Raudenbush, S. (1999). Systematic social observation of public spaces:
A new look at disorder in urban neighborhoods. American Journal of Sociology 105,
603651.
Sampson, R., Raudenbush, S. & Earls, F. (1997). Neighborhoods and violent crime: A
multilevel study of collective efficacy. Science 277, 918924.
Scheider, M., Rowell, T. & Bezdikian, V. (2003). The impact of citizen perceptions of
community policing on fear of crime: Findings from twelve cities. Police Quarterly
6, 363386.
Sessar, K., Herrmann, H., Keller, W., Weinrich, M. & Breckner, I. (2004). Insecurities
in European cities. Crime-related fear within the context of new anxieties and community-based crime prevention. Hamburg: Department of Criminology of the University of Hamburg.
157
Shaw, C. & McKay, H. (1942). Juvenile delinquency and urban areas. Chicago: University of Chicago Press.
Simons, R., Gordon-Simons, L., Harbin-Burt, C., Brody, G. & Cutrona, C. (2005). Collective efficacy, authoritative parenting and delinquency: A longitudinal test of a
model integrating community- and family level processes. Criminology 43, 989
1029.
Skogan, W. (1992). Disorder and decline. Crime and the spiral of decay in American
neighborhoods. Berkeley: University of California Press.
Smith, W. & Torstensson, M. (1997). Gender differences in risk perception and neutralizing fear of crime. British Journal of Criminology 37, 608634.
Smith, W., Torstensson, M. & Johansson, K. (2001). Perceived risk and fear of crime:
Gender differences in contextual sensitivity. International Review of Victimology 8,
159181.
Snell, C. (2001). Neighborhood structure, crime, and fear of crime. Testing Bursik and
Grasmicks neighborhood control theory. New York: LFB Scholarly Publishing.
Stangl, W. (1996). Wien Sichere Stadt Ein bewohnerzentriertes Prventionsprojekt. Kriminologisches Journal 28, 4868.
Stadtforschung Linz (2005). Analyse Brgerbefragung 2004. Linz: Amt fr Stadtforschung.
Taylor, I. (1995). Book review Fear of crime Interpreting victimization risk, by K.F.
Ferraro. Urban Studies 32, 15671568.
Taylor, I., Evans, K. & Fraser, P. (1996). A tale of two cities. Global change, local feeling and everyday life in the north of England. London: Routledge.
Taylor, R. (2001). Breaking away from broken windows: Baltimore evidence and implications for the nationwide fight against crime, grime, fear and decline. New York:
Westview Press.
(2002). Fear of crime, social ties, and collective efficacy: Maybe masquerading
measurement, maybe dja vu all over again. Justice Quarterly 19, 773792.
Taylor, R. & Covington, J. (1993). Community structural change and fear of crime. Social Problems 40, 374397.
Taylor, R., Gottfredson, S. & Brower, S. (1984). Block crime and fear: Local social ties
and territorial functioning. Journal of Research in Crime and Delinquency 21, 303
331.
Vail, J., Wheelock, J. & Hill, M. (1999). Insecure times: Living with insecurity in contemporary society. London: Routledge.
Walklate, S. (1998). Excavating the fear of crime: Fear, anxiety or trust?. Theoretical
Wickstrm, P.O. & Dolmn, L. (2001). Urbanisation, neighbourhood social integration,
informal social control, minor social disorder, victimisation and fear of crime. International Review of Victimology 8, 121140.
158
H. Hirtenlehner
Wilsem, J. van (2004). Criminal victimization in cross-national perspective. An analysis

of rates of theft, violence and vandalism across 27 countries. European Journal of
Wilson, J. & Kelling, G. (1982). The police and neighborhood safety: Broken windows.
Atlantic Monthly 127, 2938.
Xu, Y., Fiedler, M. & Flaming, K. (2005). Discovering the impact of community policing: The broken windows thesis, collective efficacy, and citizens judgment. Journal
of Research in Crime and Delinquency 42, 147186.
159
Fear of Crime and Victimisation:

The Greek Experience
CHRISTINA ZARAFONITOU
1. Introduction
In contemporary western societies there is a tendency to identify personal and
collective insecurities with the fear of crime. The feeling of insecurity related to
crime is not limited to the perception that crime is so much a real and very serious threat, as to affect the management of daily life on a personal level (Killias
2001: 399). Rather, it reflects citizens general fears that are directly linked to
quality of life as well as doubts concerning the ability of relevant authorities to
offer effective protection (Forum Europen pour la Scurit Urbaine 1996: 19).
This tendency leads to overestimation of the size of the phenomenon of criminality and to spurious conclusions concerning linked threats, since fear of crime
does not stem exclusively from personal experience, but also from the assimilation of the experiences of others, formulated by various information conduits.
The fear of crime is also fitting into broader narratives concerning anxieties
about the way society is today (Lupton and Tulloch 1999: 521).
In any event, the experimentally verified difficulties of linking insecurity with
crime have reinforced controversy concerning its conceptual determination. A
basic distinction that was established relatively early and has significantly contributed in studying the phenomenon is between fear of crime (peur du crime)
and concern (proccupation). In the first case, fear is a rational or irrational
state of alarm engered by the belief that one is in danger of criminal victimization (McLaughlin and Muncie 2006: 164). However, this apperception of threats
is formed on the basis of vulnerability 1 attributed by respondents to themselves
or those close to them. In the second case, insecurity is more generalized and
focuses on criminality as a social problem and not as a personal condition
(Robert and Pottier 2004: 218). More than likely, it is this form of insecurity that
1
For the determinative role of this variable see Killias (2001); Killias and Clerici (2000);
Box, Hale, Andrews (1988); Taylor and Hale (1986).
160
Christina Zarafonitou
is being referred to by most respondents of relevant research and polls that say
they fear they will be victimized, expressing their general social concerns
through the symbolically dense concept of crime 2 .
One of the paradoxes that scientific experience seeks to clarify is the great
difference between the low victimisation of certain categories of people (such as
women and the elderly) and their especially great fear of crime. Relative to this
issue, Steven Balkin had already posited in 1979, the crime occurrences depend
on both the amount of criminality in ones environment and the adjustments one
makes in avoiding it. It is this ex ante criminality upon which fears of crime and
safety are based not the rate of crime occurences (Balkin 1979: 344). From
this point of view, certain persons, although exhibiting a high risk of victimisation, are not victimized since they dont expose themselves to dangers.
The expression of similar feelings is dependent upon a series of factors of differing origins, such as the elements contributing to the quality of life of the inhabitants of an area, their trust in the penal system, but also on their broader
socio-ideological perceptions. Shaking up this trust of citizens (very often) reflects the perception of a more general inability of the state and therefore challenges the state itself.
Within this framework, and under the influence of the significant role played
by the media (Cario 2004) as well as the political exploitation of criminality related issues, large sections of citizens develop especially punitive and nontolerant positions 3 vis--vis specific population groups, which because of age,
ethnic origins, and economic conditions incur the greatest distrust (Zarafonitou
2004). As Robert and Pottier (2004: 218) characteristically state the anxiety that
is related to security correlates to punitiveness and xenophobia: these three dimensions form a sort of solid behavioral syndrome.
2. The role of victimisation

Research findings are not homogeneous concerning the relationship between past
victimisation experience and the feeling of fear and insecurity, while distinguishing according to crime type. Thus, the research by Skogan (1987) on the victimisation of 1738 persons in two American cities and its evolution over the course
of a twelve month period, assessed the intensity of this feeling after each victimi2
According to the distinction made by Jackson (2004: 963) between expressive and experienced fear: in the first case the fear of crime is approached as an expression of related
concerns, funnelled through this concept of crime, while in the second case as much as
summed expressions of threat and vulnerability.
See also Kury, Obergfell-Fuchs, Woessner and Wuerger (2002); Zvekic (1997).
Fear of Crime and Victimisation: The Greek Experience
161
sation. The victimisation survey undertaken in two countries (Texas, USA and
Baden-Wurttemberg, Germany) representing two different western cultures (Arnold and Teske 1988), found out that victimisation was significantly associated
with fear of crime in Texas. According to those data, Texans were almost twice
as likely to have been the victim of a crime during the previous year and this
may, in part, contribute to increased levels of fear of crime. Nevertheless, it was
not evaluated as crucial variable in their predictive model (Ibid: 378).
Research, in general, confirms the correlation between fear of crime and victimisation, noting, however, that this relationship is not a strong one (Quann and
Hung 2002: 313), and explaining this conclusion through the mitigation of the
emotion caused by victimisation, and therefore the mitigation of the relevant fear
(Box, Hale and Andrews 1988) and the de-dramatization of criminality (Killias
2001: 400).
Within this framework, three basic explanations emerge attempting to delineate this complex relationship (Box, Hale and Andrews 1988: 352): a) victims
may take more precautions and so become less fearful 4 , b) some neutralize the
negative effects of being victimized, and so it becomes less salient to them and c)
others simply allow the experience to atrophy as time passes. However, the picture is different in the event that this relationship is examined within an environment with a high incivilities index. In this context and relative to the above,
victimisation increases the fear of crime 5 .
Furthermore, this relationship is differentiated by the effects of other factors,
such as the type of crime. Research by Killias, conducted in Zurich in 1998 and
1999, linked fear of crime of the inhabitants of certain areas with frequent victimisation in crimes against the person that took place near their homes (Killias
2001: 124-5, 405). However, research results do not converge on this point, either, since the 1989-2000 ICVS noted that victims of a household offence were
slightly more fearful of crime than victims of an offence against the person
(Quann and Hung 2002: 313). This unexpected conclusion is interpreted by
researchers as the result of the great probability that the victim and perpetrator
4
Killias (2001: 402) also maintains that self protection and self-restraining measures that are
taken after a first victimisation decrease fear of crime and therefore account for the negative
correlation between this fear and the victimisation experience.
Within this framework, the main explanations expressed include the difficulties faced by
victims in taking self-protection measures that are seen as effectively dealing with dangers
and threats connected to the relevant areas, while the process of neutralization and deadening of the negative repercussions of their experience as victims, because of their constant
contact with signs of environmental disorder, that not only remind of them of their victimization, but also of a possible repeat (Box, Hale and Andrews 1988: 352).
162
may know each other in cases of crimes against the person. This creates greater
conditions for rationalization on the part of the victim. This in turn engenders
less fear relative to household victimisation where the invader is a stranger and
the attack is more likely to be planned and with intent (Ibid) 6 .
3. The Greek experience

3.1
Victimization and feelings of (un)safety
This connection between fear of crime and victimisation, clearly and steadily
come out of Greek research (Zarafonitou 2000; 2002). The latest such study was
undertaken in three municipalities of the greater Athens metropolitan area, in the
Spring of 2004 7 . The sample comprised 450 persons8 selected on the basis of
residence (150 inhabitants from each area) 9 . The study saw the distribution, doorto-door, of a 39-question questionnaire that was completed in the presence of
field researchers.
The examination of victimisation in this study followed international victimisation research methodology (Kesteren, Mayhew and Nieuwbeerta 2000: 80, 83).
Two questions were posed that dealt with respondents fears of walking within
their home municipality and their fear of being home alone10 , as well as a ques6
10
Where the aforementioned explanation by Newhart, Smith and Hill, Victimisation and fear
of crime, Criminal Justice and Behavior, 18(2), 1991: 217-239, is cited.
The data for this study come from a survey entitled Insecurity, fear of crime and attitudes
of the inhabitants of Athens to criminal phenomenon (Zarafonitou 2004b).
Men represented 51.3% of the sample and women 48.7%. As concerns age the sample distribution was as follows: 19.3% were aged 15-24, 26.9% were 25-34, 22.2% were 35-44,
14.6% were 45-54 and 16.9% were over 55. 53.5% were married with children and 52.1%
had a medium education. As concerns occupations, 40.9% were private or public employees, 24.9% were freelance professionals, 10% were students, 9.3% were pensioners, 6.9%
were housewives, 4.5% were entrepreneurs and 3.5% were unemployed. Finally, 69.7%
owned their own homes (something trued for Greeks overall), and two thirds had lived in
the area for over five years.
Questionnaires were distributed to representative residents on the basis of address in such a
manner as to cover the entire area. This method was realized in the following stages: An initial stratification was conducted based on existing administrative subdivisions using maps of
the area. These subdivisions were further arranged in ten zones in a second phase. In a third
phase, fifteen questionnaires there were further distributed . This methodology was augmented with on the spot observation of the areas in order to further map out their distinctive characteristics. The SPSS statistical software package was used to process the data
along with additional multiple-regression analysis.
The questions were posed thus: How safe do you feel walking alone in your area after
dark? and how safe do you feel when you are at home alone after dark? Available answers were: very safe, fairly safe, a bit unsafe, very unsafe and in certain cases they were
just: safe unsafe.
163
tion concerning their perception of a possible future victimisation 11 . The question

relating to (un)safety walking in the street was chosen from among the three for
correlation with personal experience of victimisation, since it expresses fear of
crime more directly, as a perception of the threat of criminal attack, as mentioned.
According to the data above, it was found that the percentage of those feeling
unsafe was greater among victims than among non-victims (72.8% vs 47.5%,
Table 1). Respectively, responders with no victimisation experience felt almost
twice safer than victims (52.5% vs. 27.2%).
Table 1: Victimisation and feelings of (un)safety
Victims/No Victims
Victims
No Victims
Total
Victimisation and feelings of (un)safety

Safe
Unsafe
25 27,20%
67 72,80%
187 52,50%
169 47,50%
212 47,30%
236 52,70%
Total
92
356
448
x2: .000
Multiple-regression analysis also reached the same conclusions. According to

this, those that have fallen victim to crime feel insecurity almost three times as
much (2.942) as the rest (Appendix, Table A). Previous research data give a
similar (arafonitou 2002: 105), the multivariate analysis of which came to the
conclusion that, during the year preceding the study, the victimisation process
almost doubles (87%) the rate of additional fear possibility (Tseloni 2002: 189).
The role of indirect victimisation in the intensity of the fear of crime is also
proven to be significant. Also according to the Greek research findings (2004), it
was found that those having a former experience of indirect victimisation felt
more unsafe than those that did not claim something similar (61.1% vs. 43.6%,
Table 2) and therefore, security levels were higher among the latter relative to
those registering indirect victimisation (56.4% vs. 38.9%). Within the framework
of the above research, the influence of indirect victimisation seems less important
than of the aforementioned direct experience.
11
The question was: How likely do you believe the possibility of becoming the victim of a
criminal act in the immediate future? and the answers were: very likely or likely enough,
little or not at all likely.
164
Table 2: Indirect Victimisation and Feelings of (un)safety

Indirect victims/
No victims
Indirect victims
No victims
Total
Indirect victimisation and feelings of (un)safety

Safe
Unsafe
Total
86 38,90%
135 61,10%
221
123 56,40%
95 43,60%
219
209 47,60%
230 52,40%
439
x2: .000
The examination of the hypothesis on the role of self-protection measures, possibly, taken by victims to mitigate fear, was conducted through the question what
has changed in your everyday life after the victimisation of yourself, or some acquaintance (Table 3).
Table 3: Changes in your life after the direct of indirect victimisation
Measures of safety in their houses (locks, alarm etc)

Moving in another area
Avoidance of some places
Carrying weapons (knife, gun, spray)
General unsafety
Improvement of relations with neighbours
No change
Total
Changes in your life after the direct

or indirect victimisation
83
23,30%
6
1,70%
51
14,30%
16
4,50%
112
31,40%
21
5,60%
68
19,10%
357 100,00%
From the registered answers, it was ascertained that more than half took absolutely no measures and answered either that they feel generally insecure
(31.4%), or nothing has changed (19.1%), while 23.3% made reference to security measures taken at home (locks, alarms, etc) and 14.3% answered that they
avoid certain areas 12 . This image seems to confirm the relevant hypothesis that in
part explains the positive correlation between victimisation and a fear of crime as
a result also of not taking special measures to avoid the possibility of subsequent
victimisation (Killias 2001: 402). However, it is more likely that a more general
unsafety and worry, as well as dissatisfaction for services rendered by the
state (both in terms of protection from crime, and areas that influence quality of
life, in general) is expressed through fear of crime.
12
In the hierarchy of households without security measures (alarms, neighborhood watch,

security locks) in the European victimisation study, Athens received fifth place with 12%
(average:11%). Madrid, Edinburgh, Paris, and Copenhagen (25%, 23%, 22%, 15%) were in
the first places, while London and Budapest were in the last positions (2%).The citizens of
the last two cities seem to place greater emphasis on issues of self-protection, (Nyiri 2005).
3.2
165
Neighbourhood environment perception and quality of life
The tendency of combining quantitative with qualitative methodological approaches in researching the complex phenomenon of fear of crime seems to be
developing in this direction 13 . One such approach focuses on the social meaning of the notions incivility and social cohesion(Jackson 2004: 960). Within
this framework, it has been found that worry about crime is formulated by a series of subjective parameters, such as the psychological perception of vulnerability, the general social attitudes, and the perception of everyday risk 14 . Approaches of this nature, in any event, highlight the significant role of information
concerning crime, which in large urban centers is almost exclusively dominated
by the mass media.
Within this framework, those that registered unsafety in moving about at night
in their neighborhood, explain this feeling as the result of the existence of a lot of
foreigners, inadequate police patrolling, and deserted and badly-lit areas (23.7%,
22.9%, 15.2%). A significant factor that emerges is the lack of social cohesion
since in 20% of responses insecurity was given as the result of the indifference of
neighbors (9.6%) and the indifference of passers-by in the event of a criminal
attack (10.4%, Table 4).
Table 4: Reasons of unsafety*
Lot of foreigners
Inadequate police patrolling
Deserted and badly-lit areas
Indifference of passers-by in the event of a criminal attack
Indifference of neighbours
Slums
Rumors of crime occurrence
Abandoned buildings
Many homeless
Dense circulation
Night clubs
Large crowds
Total
Reasons of unsafety
317
23,70%
307
22,90%
203
15,20%
139
10,40%
128
9,60%
63
4,70%
54
4,00%
46
3,40%
30
2,20%
25
1,90%
14
1,00%
14
1,00%
1340
100,00%
*v = 438 (multiple answers)

13
14
See indicatively Lupton and Tulloch (1999), which examines cultural representations and
the various levels of symbolism that contribute to the formulation of fear of crime.
For example, within the framework of this research, it was found out that persons with more
authoritarian views on law and order were more prone to perceive disorder in their environment and more easily linked it to consensual and social cohesion problems, and degradation of social structures and unofficial social checks (Jackson 2004: 960).
166
All of the above refer to the quality of life issue that emerged in this latest Greek
survey, as satisfaction from services in health, education, public transportation,
the environment, etc. 15 According to the findings from the study, quality of life
emerges as an important factor in insecurity, since 76.8% of those reporting fear
were also unsatisfied by the quality of life in their municipality, while among the
safe respondents the relevant figure was 58.0% (Table 5).
Table 5: Satisfaction from the quality of life in the area of domicile
Satisfaction from the quality of life in the area of domicile
Satisfied
Not satisfied
89
123
Safe
41,99%
58,01%
55
182
Unsafe
23,20%
76,80%
Total
212
100,00%
237
100,00%
x : ,000
This finding is also borne out through multiple-regression analysis, according to

which those that are unsatisfied by the quality of life in their residential area feel
twice as (2.042) unsafe as those claiming satisfied (Appendix, Table A). The inhabitants of Athens indeed ranked as their greatest problems (Table 6) drugs
(25.5%), immigrants (21.2%), unemployment (19.9%), while criminality, in general, was in fourth place (13.2%) 16 .
15
16
The question concerning environment, that was included in the European Victimisation
Survey (EU ICS), is relevant to environmental disorder. This question examines the stance
of citizens of European capitals on the basis of certain characteristics that delineate the notion of deprived area, meaning: youth on the streets, homeless persons, beggars, littering,
graffiti, vandalism and public intoxication (Hideg and Manchin 2005). Research data rate
inhabitants of Athens first in negative assessments with 86%, along with inhabitants of Budapest. These are followed by the inhabitants of Brussels and Paris (84%), while the inhabitants of Lisbon are in last place (and therefore have the most positive image) (56%). This
stance is correlated to (in)security.
The view is posed that criminality emerges as a main social problem when social concerns
for unemployment subside, while in the opposite case, when unemployment takes on a
dominant position, insecurity concerning criminality is overtaken (Robert and Pottier 2004:
237). Of course, its much more likely that criminality in the responses of inhabitants of
Athens is linked to drugs and therefore is ranked first. Its also not impossible that there is a
linkage between immigrants and criminality, as well as with unemployment, something
which has come out of other research.
167
Table 6: Most important social problems in their domicile area

(multiple answers)
Most important social problems in their domicile area
Drugs
Immigrants
Unemployment
Crime
Lack of sanitys structures
Environmental problems
Lack of educations structures
Total
3.3
297
247
232
154
96
69
68
1163
25,50%
21,20%
19,90%
13,20%
8,20%
5,90%
5,80%
100,00%
Attitudes to the police
Finally, for the unity of the syllogism it is imperative to refer to the significant
role played in expressions of insecurity and fear of crime, by the trust of citizens
to the penal justice system. The first studies in explaining fear of crime noted the
decisive role that the presence of police can play, especially if the force is willing, effective, and acceptable by the community (Box, Hale and Andrews 1988:
353). This role becomes even more important in modern urban environments,
because of the absence of informal social control and the loosening of social
bonds. Within this framework, the police in the minds of citizens is an organization in the service of the local population and as such satisfaction from police
services constitutes a logical criterion for its assessment (Killias 2001: 429).
In this way, the finding that those that feel more intense fear are those that are
also most dissatisfied with the work of the police and who seek greater policing
(Zvekic 1997: 8).
This image is also verified by Greek research results. More specifically, the
2004 study, in general, registers a negative assessment of the work of the police
(Table 7) and almost of respondents assess police work as not very or not at all
effective (71.8%). This assessment, however, becomes even more negative when
answers come from victims (75.8%) or those feeling insecurity (77.6%)17 .
17
A corresponding ranking also comes out of international studies. The average satisfaction of
inhabitants of European capitals with police work in the field of criminality is generally very
high (72%) as is borne out by data from the aforementioned european victimisation survey.
The inhabitants of Helsinki and Edinburgh are most satisfied with percentages reaching
90%, while those least satisfied are the inhabitants of Athens and Rome with 52% and 60%
respectively (Nyiri 2005). This study also registers the great dissatisfaction of those that
claimed victimisation within the last five years with almost a third (29%) believing that the
police does a fairly or a very poor job in dealing with crime (20% for non-victims). The
findings of the European crime survey especially highlight the great dissatisfaction with
treatment by police of victims that proceeded with crime reporting to police.
168
Table 7: Evaluation of police job in their domicile area

Police is ...
Evaluation of police job in their domicile area

Total
Unsafe
Victims
Effective
Ineffective
122 28,20%
310 71,80%
53 22,40%
184 77,60%
22 24,20%
69 75,80%
Total
432
237
91
100%
100%
100%
x2: .017; x2: .344
It is obvious, that the above significantly influence the general tendency for reporting crime to police, something also verified by international victimisation
crime survey. According to these data (Kesteren, Mayhew and Nieuwbeerta
2000: 64), the average reporting rates for the six representative crimes has been
within a narrow range of 51% to 53% for the countries taking part at least three
times. Its worth noting that the rate of reporting is higher among Athens residents that took part in the 2004 Greek survey (64.5%), despite their much more
negative image of police effectiveness.
4. Conclusions
According to the aforementioned Greek experience, a stable relation between
fear of crime and victimisation is established. In fact, multiple-regression analysis pointed out that those that have fallen victim to crime feel insecurity almost
three times as much as the rest. A similar correlation is verified in the case of
indirect victimisation. The hypothesis, according to the aforementioned, of not
taking self-protection measures by a significant number of victims can only
partly account for this positive correlation. t is more likely that a more general
unsafety and worry, as well as dissatisfaction for services rendered by the
state is expressed through fear of crime. Within this framework, quality of life
emerges as a significant factor, as this is determined through services offered and
the level of satisfaction from environmental and social conditions in the area of
residence. This image is also borne out through multiple-regression analysis, according to which those that are unsatisfied by the quality of life in their residential area feel twice as unsafe as those claiming satisfied.
Inhabitants of Athens explain their feelings of insecurity as the result of the
existence of a lot of foreigners, inadequate police patrolling, and deserted and
badly-lit areas. A significant factor that emerges is also the lack of social cohesion. Athenians rank as most significant social problems drugs, immigrants, unemployment, and criminality, in general.
169
Greek research findings verify the standing correlation between lack of satisfaction with police and feelings of insecurity. More specifically, the 2004 study,
in general, registers a negative assessment of the work of the police. This assessment, however, becomes even more negative when answers come from victims or those feeling insecurity. Its worth noting, however, that the reporting
rates by victims to police is greater among Athens residents in comparison to
international research findings, despite the much more negative image Athenians
have of police efficacy.
In any event, and despite whatever vagueness in the conceptual content of insecurity and the difficulties of a spherical explanation of the phenomenon, insecurity has taken on global dimensions with indisputable effects on a personal and
societal level. Thus persons feeling unsafety are led to self imposed restrictions
in their daily lives downgrading their quality of life, exacerbating conflicts
among citizens, and reinforcing stigmatizing stereotypes (racism, xenophobia)
that lead to the marginalization of whole population groups and also whole
neighborhoods. Citizens punitiveness is reinforced, which acts in a disorienting
manner towards the application of a rational social/preventive criminal policy.
Also, personal initiatives in self-protection measures are affected to a significant
degree with, often, dubious results.
The intensification of punitiveness that has been observed over the last few
years on the part of a great percentage of citizens in Europe and the USA is not,
however, exclusively due to their fear of crime but also to a series of other factors. Among these are: the level of information of citizens and the media, political conservativism, social problems, religion, etc (Kury and Obergfell-Fuchs
1998: 42). In any case, both insecurity and punitiveness on the part of the public
are favorite themes for journalistic and political exploitation, offering, at the
same time, favourable conditions for the emergence of theories of the people
(Walter 2002: 1305).
It is obvious that the insecurity, either in the form of an immediate fear of
crime, or in the form of anxiety over criminality as a social problem, significantly
affects crime policy. Social pressure for taking harsher measures to deal with the
phenomenon of crime, as well as a generalized tendency to politically exploit
insecurity, have as a logical continuation the reinforcement of repression and the
extension of the envelope for the penal intervention by the state. Within this
framework, penal justice is now used as a tool for managing social problems,
thus contributing to their exacerbation and expansion.
170
Appendix
Table A: Multiple-regression analysis
Independent variables
Dependent variable
Odds ratio (OR) (95% C.I)
Domicile area
Centre
D1
D2
0,425 (0,244-0,742)
0,809 (0,462-1,416)
Reference category
25-34
35-44
45-54
55+
15-24
2,067 (0,909-4,696)
2,087 (0,853-5,104)
1,472 (0,559-3,873)
0,593 (0,173-2,036)
Reference category
Age
Education
High school graduate
University/College gratuated
Has had no schooling/has had primary education
Gender
Men
Women
Profession
Free trader/Entrepreneur
Pensioner/housewife
Student undergratuated
Unemployed
Employee
Victimisation
No
Yes
Sources of information
Radio
Newspapers
Social environment
TV
Satisfaction from the quality of life in their domicile area
Yes
No
Main social problems
Other social problems
Unemployment
Assesment of police work
Effective
Ineffective
Fellow lodger
Family
Friends
Alone
1,582 (0,736-3,402)
2,160 (0,911-5,123)
Reference category
4,794 (2,928-7,851)
Reference category
1,130 (0,626-2,040)
1,096 (0,399-3,008)
2,063 (0,754-5,640)
1,990 (0,517-7,659)
Reference category
2,942 (1,622-5,338)
Reference category
0,727 (0,202-2,618)
1,766 (0,686-4,546)
2,541 (0,584-11,052)
Reference category
2,042 (1,219-3,421)
Reference category
1,330 (0,295-6,000)
Reference category
1,242 (0,744-2,073)
Reference category
0,780 (0,402-1,514)
0,777 (0,205-2,936)
Reference category
171
References
Arnold, H. and Teske, R. ( 1988), Factors related to fear of crime. A comparison of the
Federal Republic of Germany and the United States, in G. Kaiser and I. Geissler
(Eds), Crime and Criminal Justice. Criminological research in the 2nd decade at the
Max Planck Institute in Freiburg, Freiburg, 355-384.
Balkin, St. (1979), Victimization rates, safety and fear of crime, Social Problems, vol.
26, 343-357.
Box, St., Hale, C. and Andrews, G. (1988), Explaining fear of crime, the British Journal
of Criminology, vol. 28, 340-356.
Cario, R. (2004), Mdias et inscurit: entre droit dinformer et illusions scuritaires,
Recueil Dalloz, vol. 2, 75-80.
Forum Europen pour la Scurit Urbaine (1996), Outils pour l'action, Paris.
Hideg, G. and Manchin, R. (2005), Environment and safety in european capitals, Research paper, EUICS Consortium.
Jackson, J. (2004), Experience and expression. Social and cultural significance in the
fear of crime, the British Journal of Criminology, Vol. 44, 946-966.
Kesteren, J.v., Mayhew, P. and Nieuwbeerta, P. (2000), Criminal Victimisation in Seventeen Industrialised Countries, NSCR.
Killias, M. and Clerici, Ch. (2000) Different measures of vulnerability in their relation to
different dimensions of fear of crime, The British Journal of Criminology, 40, 437-450.
Killias, M. (2001), Prcis de criminologie, Staempfli Editions SA Berne.
Kury, H. and Obergfell-Fuchs, J. (1998), Public attitudes towards sanctioning, in H.J.
Albrecht and H. Kury (Eds), Research on crime and criminal justice at the Max
Planck Institute, Freiburg, 41-49.
Kury, H., Obergfell-Fuchs, J., Woessner, G. and Wuerger, M. (2002), Fear of crime and
victimization, in P. Nieuwbeerta (Ed.), Crime victimization in comparative perspective, Boom Juridische uitgevers, Den Haag, 317-334.
Lupton, D. and Tulloch, J. (1999), Theorizing fear of crime: beyond the rational/
irrational opposition, The British Journal of Criminology, Vol. 50, no 3, 507-523.
McLaughlin, E. and Muncie, J. (2006), The Sage dictionary of Criminology, Sage Publ.
Nyiri, Z. (2005), Crime and security in european capitals, European Union International
Crime Survey (EU ICS) Consortium, Research papers.
Quann, N. and Hung, K. (2002), Victimization experience and the fear of crime. A
cross-national study, in P. Nieuwbeerta (Ed.), Crime victimization in comparative
perspective. Results from the International Crime Victims Survey, 1989-2000,
NSCR, BJU, Den Haag, 301-316.
Robert, Ph. and Pottier, M-L. (2004), Les proccupations scuritaires : une mutation?,
Revue franaise de sociologie, 211-242.
Skogan, W.G. (1987), The impact of victimisation on fear, in Crime and Delinquency,
vol. 33, 135-154.
172
Taylor, R. and Hale, M. (1986), Testing alternative models of fear of crime, The Journal
of criminal law and criminology, vol. 77, 151-189.
Tseloni, . (2002), Fear of walking alone in the neighbourhood alone: A multivariate
and regression analysis, in Ch. arafonitou, The Fear of Crime, Studies in European
Legal Science, A. Sakkoulas publications, (in greek and english), 157-201.
Walter, M. (2002), Prevention of Crime in Local Society-Where will it Lead to, (translated in Greek by A. Pitsela), Penal Justice, vol. 12, 1304-1306.
arafonitou, Ch. (2000), Empirical study of the fear of crime in a municipality of Athens, in Daskalaki et al. (Ed.), Offenders and victims threshold of 21st century, National Centre of Social Research, Athens, 511-519 (in greek).
(2002), The Fear of Crime, Studies in European Legal Science, A. Sakkoulas publications, (in greek and english).
Zarafonitou, Ch. (2004), Insecurity and extention of social control: penalization of
incivilities and disorder, Poenicos Logos, vol. 4, 2049-2059 (in greek).
Zarafonitou, Ch. (2004), Insecurity, fear of crime and attitudes of the inhabitants of
Athens to criminal phenomenon, umpublished research, PanteionUniversity, Athens.
Zvekic, U. (1997), Les attitudes des victimes envers la police et la punitivit: rsultats
des sondages internationaux de victimisation, Revue Internationale de Criminologie
et de Police Technique, vol. 1, 3-16.
173
Social-Demographic and Social-Psychological

Perspectives of Fear of Crime in Slovenia,
Croatia and Bosnia & Herzegovina
GORAZD MEKO, IRMA KOVO-VUKADIN

AND ELMEDIN MURATBEGOVI
Abstract
Research and theorising about the fear of crime has, in the main, been dominated
by researchers who have relied upon sociological or socio-demographic variables
to account for variations in fear levels. Whilst this body of work has contributed
greatly to our understanding of the fear of crime, we are still far from a full understanding of this important and most corrosive aspect of contemporary society.
This paper first presents three independent evaluations of fear of crime. The data,
collected in Slovenia, Scotland and Holland, sheds further light on both the social
psychological model proposed and the wider study of fear of crime. These
evaluations served as a background for the present study of fear of crime in Slovenia in 2001 and Croatia and Bosnia and Herzegovina in 2002. The results show
higher level of fear of crime in women, people who perceive themselves as less
physically fit, the unemployed, those who expose themselves by walking alone in
dark and those who perceive streets and woods as sources of danger. In addition,
higher level of fear of crime can be attributed to respondents who consider themselves a potential victim due to the jealousy of others or being attractive to a potential criminal in other ways, incapable of chasing of a potential assailant, and
perceiving places as dangerous (criminalisable space).
174
G. Meko, I. Kovo-Vukadin & E. Muratbegovi
1. Introduction
The fear of crime is nowadays one of the most researched topics in contemporary
international criminology. In the United Kingdom and United States, crime surveys have expanded rapidly since the late 1960s. For example, the British Crime
Survey now biennially interviews in the region of 10,000 residents of England
and Wales. This survey has shed light on attitudes to policing, victimisation, perceptions of risk and peoples fear of crime. Crime surveys are conducted, not
only in Western European countries, but also in Central and Eastern European
countries (Hatalak, del Frate, Zveki 1998; Zveki 1998; Kury 2001).
In Slovenia, three major crime and victimisation surveys have been undertaken to date. The first in 1992, the second in 1996/97 (Pavlovi 1998) and the
third in 2000/2001 (National Bureau of Statistics, Internet, Umek 2004.). The
fourth study on victims of crime (ICVS) is ready to be conducted in Slovenia in
2004/05.
In Croatia, two major crime and victimisation studies (ICVS) have been conducted so far (eparovi and Turkovi 1998; Gruszczynska 2001: 77-89). The
last study on fear of crime in Croatia was conducted by Ivii, aki and Franc
(2004).
In Bosnia and Herzegovina, only one crime and victimisation survey has been
undertaken to date (2001) by researchers from the University of Lausanne (within United Nations Mission of International Police Task Forces in Bosnia and
Herzegovina) (Muratbegovi, 2003).
The findings of the UK and USA research are now well known. A plethora of
studies have concluded that the fear of crime impinges upon the well-being of a
large proportion of the population. Some have even gone as far as to suggest that
the fear of crime is now a larger problem than crime itself (Hale 1992; Bennett
1990; Warr 1984.). Chambers and Tombs (1984: 29) reviewing the 1982 British
Crime Survey (Scotland) reported that "more than half of the respondents (58%)
said that at some time in the past they had been concerned about the possibility of
being a victim of crime".
The fear of crime impacts the quality of life. People tend to adjust their behaviour according to their fear of crime level in a way of not engaging in some activities or doing some (for example buying a gun for a protection of themselves
or their property). The fear of crime also influences personal effectiveness and
productivity in everyday life.
As well as receiving much attention at an empirical level, many have attempted
to explain the fear of crime. These efforts have tended to be dominated by researchers influenced by sociological insights. Thus such variables as age, gender, house-
Fear of Crime in Slovenia, Croatia and Bosnia & Herzegovina
175
hold income, friendship networks, length of residence, earlier victimization experiences and so on have been suggested as key in explaining fear of crime. This approach is also called socio-demographic model of fear of crime. Such models have
indeed been found to be of utility in explaining the fear of crime.
However, by concentrating on sociological variables, researchers have largely
ignored social psychological and psychological factors that may be important in
explaining the fear of crime (Van der Wurff et al. 1989: 142 and 1986; Farrall et
al. 2000.). This domination of the literature by sociologically informed theorising
has ignored important processes occurring at the individual level. In this article
we discuss further analyses of the only social psychological model of the fear of
crime (Van der Wurff et al. 1989). The aim of the current article is to shed light,
in addition to socio-demographic dimensions, also on the relationship between
social psychological factors and the fear of crime, and as such to develop further
the analysis undertaken by Van der Wurff et al. (1989) and Farrall et al. (2000)
and our own analyses on the basis of international data.
This article is constructed in the following fashion. We start by outlining the
social psychological model proposed by Van der Wurff et al. Following this we
describe the methodologies employed in each of the surveys (and the countries in
which these surveys were conducted in the past). We then discuss the question
wording employed in the surveys. Having assessed the data in terms of its suitability for replicating the earlier study, we present findings of the studies based
on the same methodology in the past. We further present results of our present
study on fear of crime in Slovenia, Croatia and Bosnia and Herzegovina, using
both models socio-demographic and social psychological (a holistic model).
We conclude the article with a discussion of what these findings mean in specific
social environments of Slovenia, Croatia, and Bosnia and Herzegovina.
1.1
The Socio-demographic and Social Psychology Model
The socio-demographic model consists of the following variables: age, gender,

household income, friendship networks, length of residence, earlier victimization
experiences and so on. The social psychological model that we shall discuss in
this article was originally proposed and tested by Van der Wurff et al. (1989)
using data collected in the Netherlands. Van der Wurff et al. develop a social
psychological model based on the assumption that the fear of crime is associated
with four social psychological components. These four they describe as follows
(1989: 144-5) 1 :
1
We discuss the original operationalisation of these components and our minor changes to
them in the Appendix.
176
The Attractivity [component 2 ] is intended to refer to the extent to which

people see themselves or their possessions as an attractive target or victim
for criminal activities. It involves the attribution of a characteristic to oneself and ones possessions. One thinks, for example, of the peculiar sensation one may have when walking on the street with a great deal of money.
Another example would be the fear of burglary, which may be experienced
if one keeps valuable articles in the house.
The Evil Intent [component] relates to the wrongdoers role in the phenomenon. It is represented by the extent to which a person attributes criminal intentions to another individual or particular group. Thus, one may be
afraid of having ones pocket picked the moment one sees a gypsy. Or one
can experience fear as a result of a feeling that society is in moral decay and
a conviction that present-day youth are prepared to commit murder for a
paltry sum of money.
The Power [component] refers to the degree of self-assurance and feeling of control that a person has with respect to possible threat or assault by
another. In principle it is a question of two related sub-factors: ones own
power and the power of the other. The first of these relates to a persons
confidence in his 3 own efficacy. This need not be directly related to the
dangers of crime, of course. Feelings of self-assurance, control, and confidence in meeting the challenges of life will by generalisation tend to lower a
persons sensitivity to feelings of threat. Almost anything can contribute to
the feeling of ones own power, from a good family relationship to an optimistic temperament.
The power of the other is the wrongdoers side of the coin. It concerns
characteristics attributed to potential criminals, such as their strength, agility, resources, and general ability to carry out their criminal intentions. A
comparison of ones own power with power of other determines whether a
person faces confrontations with that other with confidence or not. Thus, the
idea that even the smallest thief goes about carrying weapons can lead to
feelings of uneasiness or fear, if one has no compensating power of ones
own.
Criminalisable Space [component] is the fourth and final [component].
Whereas the first [component] refers to the potential victim, the second to
the potential wrongdoer, and the third to both of these parties, the last [component] has to do with the situation in which a crime may take place. The
emphasis is on characteristics of place and time and on the presence of others, It is a question of the extent to which a situation lends itself to criminal
activities in the eyes of a possible victim of how much the situation facilitates crime or the criminal. A criminalisable situation might, for example,
2
Van der Wurff et al. use the word factor. We use component so as not to cause confusion
when discussing factor analysis in future sections.
Sexism in original text.
177
include walking at night through a poorly lit pedestrian subway or through a

dark wood, although estimates of criminalisability for any one situation can
naturally vary between individuals. The interest here lies in the extent to
which people have a general tendency to heed the criminalisability of the
situations into which they venture.
As is evident from outlining their model, Van der Wurff et al.s primary consideration in the construction of the model was the perception of the individual. Van
der Wurff et al. (1989: 143-4) are careful to note that there is no causal ordering
inherent in their model. That is to say that these components are merely associated with fear of crime. The fear of crime may exert an influence upon these
components and vice versa they may exert an influence upon it. A further feature
of this model is that it operationalises Youngs (1988 and 1992) observation that
the research on the fear of crime has ignored the perpetrators of crimes, but might
include them in future models.
1.2
Van der Wurff et al.s Results
The first study of this kind was conducted by Van der Wurff et al. (1989: 155)
who found that the social psychological model explained about 24% of the variance of the fear of crime. This is more variance in fear of crime levels than is
usually explained. A previous replication of this model (Farrall et al. 1999) finds
similar results, although they report a significant gender bias in the model and
propose that socio-demographic variables can be added to the model to increase
its power still further.
2. The Past Studies: Holland Scotland Slovenia

The replication of any study is a generally worthwhile endeavour. Replication
allows for an assessment of the both the extent to which one can generalise the
findings of the model to different populations, and the validity of the instrument
used. In this article we present the initial results reported by van der Wurff et al.
with those of Farrall et al. and with a similar survey undertaken in Slovenia
(Meko and Farrall 1999). The authors of previous research were concerned primarily with the utility and operation of the social psychological model, and for
the sake of brevity did not concern themselves with either socio-demographic
variables or the confounding issue of gender.
2.1
Question Wording
We are fortunate that Van der Wurff et al. reproduced translations (into English)
of the items upon which they relied. These English translations were then used
178
largely unchanged for the Scottish survey, and translated into Slovene for the
first Slovenian survey (1998). For the Slovenian survey, the present authors corresponded on this issue so as to ensure that the meanings of the words used were
as comparable as was possible. In order to fully assess its utility, the questions
used to produce the social psychological model developed by Van der Wurff et
al. were employed in the crime surveys outlined above. All questions referred to
here are reproduced in the Appendix (along with notes concerning any recodings
performed). Unlike Van der Wurff et al., who relied upon a stratified survey of
110 respondents in each of four sampling points, the Slovenian and Scottish surveys employed simple random sampling techniques.
The measure of the fear of crime employed by Van der Wurff et al. relies upon
vignettes, which outline six situations. Vignettes (one of which is reproduced
below the rest are reproduced in the Appendix) have the benefit of providing
the measurement of complex phenomena in social contexts.
Vignette Four: To a party
Youve been invited to a party in a neighbourhood you dont really know.
Early that evening you set out by bus. When you get off you still have a long
way to walk. Suddenly you notice that youve lost your way. A group of
youths is following you and begins to make unpleasant remarks at you.
Each of these vignettes are followed by questions on associated feelings of unsafety. Of these questions, only one for each vignette, referred to as the unsafety question is employed in further analysis. These six unsafety questions
are summated and used as the measure of the fear of crime (i.e. as the dependent variable). Each of the vignettes was employed in accordance with the procedures outlined by Van der Wurff et al. (1989: 148).
The four components of the social psychological model (Attractivity, Power,
Evil Intent and Criminalizable Space.) were measured through the use of eight
questions (two per component). In the original article these eight questions were
left unnamed, however, we have employed the same titles as given to these variables by Farrall et al. These questions are also reproduced in the Appendix.
2.2
Previous Analysis of the Data Sets
Van der Wurff et al. go to considerable lengths to examine their data set. They
test the six unsafety items employed for their distribution of answers, their reli-
179
ability and their uniformity. In order to assess the extent to which our data set is
comparable to that of Van der Wurff et al.s, Meko and Farralls (1999) commenced their analysis by repeating these tests in order to establish the generalisability of the model to other populations. Tables 1 and 2 reproduce their results
for the six Unsafety questions for each vignette (Meko and Farrall, 1999) in Slovenia, Scotland and Holland.
Table 1:
Degree of Safety in 6 Situations

Mean
SD
Slovenia
Scotland
Holland Slovenia
Scotland
Holland
Doorbell
3.31
3.05
2.35
1.01
1.28
1.29
Car
2.41
2.53
3.27
0.92
1.15
1.23
To a party
2.05
1.73
3.77
0.86
0.86
1.08
Bus stop
2.76
2.53
2.30
0.86
1.14
1.14
Telephone
3.27
3.18
1.96
1.01
1.19
1.25
Caf
3.24
3.25
2.31
0.95
1.19
1.32
Situation:
Ns = Slovenia (443), Scotland (482-485) and Holland (440).
Table 2:
Summary of Factors Loadings

Slovenia
Scotland
Holland
Doorbell
.76
.66
.66
Car
.77
.54
.71
To a party
.69
.29
.63
Bus stop
.71
.51
.61
Telephone
.71
.71
.47
Caf
.67
.56
.36
Situation:
Ns = Slovenia (443), Scotland (482-485), and Holland (440). Maximum Likelihood

Extraction used. KMO statistic: (Slovenia) .835 (Scotland) .779.
As with the earlier study, both the Slovenian and Scottish data sets produced just
one factor, suggesting unidimensionality in the data set. Along with the similar
Alpha coefficients (higher than 0,70), this suggests that the Slovenian and Scottish data sets enable us to undertake a robust replication of the earlier study.
180
We come now to the presentation of our findings with regard to the replication
of the social psychological model.
Two results stand out as being of particular interest the relative differences
in the adjusted R-Square figures produced and the similarity of the models produced when only significant values are considered. The R-Square values range
considerably from over 35% (Slovenia) to under half of that (17%, Scotland).
The exact meaning of this is hard to fathom. It could be due to some of the differences observed earlier between the Slovenian and Scottish samples or alternatively due to some unanticipated influence of question wordings.
The second, more interesting observation concerns the elements which enter
the three examinations of the model. In all three of the examinations undertaken,
the two elements of Criminalisable Space enter the models and at the highest
levels of significance (p > .01 and .001). In addition to this, in all but one of the
models the two elements of Power also enter the model. The other two elements of the model (Attractivity and Evil Intent) fare less well, and their entry
to the model is ambiguous or uniformly non-existent. Van der Wurff et al. (1989:
155) report a similar finding in their regression analyses.
The variable, which we named Obstruction in the Criminalisable Space
component is partly interesting, as it is entered in the same direction for each
model at the highest level of significance (in short, thinking that one would have
ones path blocked, is associated with feeling unsafe). The direction of second
element of the Criminalisable Space component (Safe Route), is less clear,
but for two of the three models, taking a safe route is associated with feeling
less safe. In Scotland, taking a safe route is associated with feeling safer, but this
is at a lower level of significance. The elements of Power (Attacker and
Rows) act in a less uniform manner, but generalisations are still possible.
Thinking that one can chase off an attacker is associated with feeling more safe
(Slovenia and Scotland), whilst steering clear of rows has the same effect in
Scotland, but not Holland (See Meko and Farrall 1999).
This suggests a number of points to us. First of all anxieties about crime are
perhaps reducible to features of the relationship between the individual and the
person they imagine as their assailant. Anxieties are observed to be associated
with having ones route blocked, but mitigated if one feels one can chase off an
assailant. Hence anxieties appear to be related to imagined vulnerability and perceptions of ones physical strength relative to the assailant. Similarly, exhibiting
a degree of consciousness about ones personal safety (making sure one takes a
safe route home), is related to greater levels of fear suggesting that the fearful
take precautions.
181
Following the studies conducted in Holland, Scotland and Slovenia we decided to conduct the another study in Slovenia (2001), Croatia (2002) and Bosnia
and Herzegovina (2002), which were before 1991 republics of Yugoslavia and
faced different social developments in the nature and extent of crime (especially
violence related to war and post-war conditions). Therefore, we assume that different levels and sources of fear of crime can be attributed to each of the studied
countries.
3. The Present Study: Slovenia, Croatia and Bosnia & Herzegovina

We used the questionnaire previously used by Van der Wurff (1989) and Farrall
et al. (1997). The questionnaire was first translated into Slovene by Meko and
Umek (1998) and tested for its applicability. The same procedure was conducted
by Meko and Kovo (1999) for the use of the questionnaire in Croatia in 1998.
The questionnaire was later on translated into Bosnian language (which does not
differ much from Croatian). Question wording is the same in all three questionnaires and the meaning of statements and coding is the same. The authors have
thoroughly checked for the meaning of each item in the questionnaire regarding
its connotative and denotative meaning.
The Slovenian data come from a survey of people living in Slovenia. Sixty
percent of the sample comes from the capital city of Ljubljana. The interviews
were conducted in spring and autumn 2001 by the students of third year of criminal justice studies at the Faculty of Criminal Justice and Security in Ljubljana.
The Croatian data come from a random sample of people living in the capital
city of Croatia Zagreb. The data were collected in autumn 2002 by interviewing
inhabitants. The interviews with respondents were conducted by trained students
of Police College in Zagreb.
The Bosnian data come from a survey of people living in Canton Sarajevo the
biggest and social and economic most developed region in Bosnia and Herzegovina. A random stratified sample of 1000 residents, 200 from each of five urban
municipalities (Old Town, Center, New Sarajevo, New Town and Ilida). People were randomly selected by their address and interviewed face to face by the
trained students of 3rd year of graduate program at Faculty of Criminal Justice
Science University of Sarajevo. The survey was undertaken during September
December 2002. Characteristics of the respondents included into the samples are
presented in table 3.
182
Table 3:
Characteristics of studied samples in Slovenia, Croatia and B&H

(N= 3491)
Demographics
Slovenia
n
Age
Adolescents (15-20)
334
Young adults (21-30)
556
Middle aged adults (31-55)
702
Early seniors (56-70)
115
Seniors (71-90)
51
Gender
Female
933
Male
827
Educational level
Primary school finished
231
High school finished
1126
Vocational univ. degree
159
University degree
244
Homeowner
Yes
1161
No
599
Time living at the present address
Less than 1 year
125
1 to 2 years
95
3 to 5 years
138
6 to 10 years
223
More than 11 years
1179
Living with
Parents, family
1645
Alone
113
Financial ability
Low
832
Average
398
High
530
Total number of respondents
1760
Croatia
Bosnia & Herzeg.

n
%
19,0%
31,6%
39,9%
6,5%
2,9%
467
128
108
25
4
63,8%
17,5%
14,8%
3,4%
0,5%
79
291
397
212
21
7,9%
29,1%
39,7%
21,2%
2,1%
53,0%
47,0%
410
322
56,0%
44,0%
496
503
49,6%
50,4%
13,1%
64,0%
9,0%
13,9%
11
589
42
90
1,5%
80,5%
5,7%
12,3%
110
552
159
179
11,0%
55,2%
15,9%
19,9%
65,96%
34,04%
577
155
78,8%
21,2%
637
363
63,7%
36,3%
7,1%
5,4%
7,8%
12,7%
67,0%
51
31
71
81
498
7,0%
4,2%
9,7%
11,1%
68%
49
36
130
162
623
4,9%
3,6%
13,0%
16,2%
62,3%
93,6%
6,4%
705
27
96,3%
3,7%
815
185
81,5%
18,5%
47,3%
22,6%
30,1%
251
177
304
34,3%
24,2%
41,5%
436
214
349
43,6%
21,4%
34,9%
732
1000
KMO statistics (sampling adequacy coefficient):

Slovenia 0.844, Croatia 0.826 and Bosnia & Herzegovina 0.880.
4. Results
4.1
Comparison of socio-demographic and social psychological

variables
Socio-demographic variables of fear of crime were studied in other studies (Pavlovi 1992; 1996; Zveki 1998; Umek and Meko 1999; Meko and Kovo
1999). For the purpose of this article, we include socio-demographic and socio-
Table 4:
183
Socio-demographic model of fear of crime comparisons

n
Slovenia
%
Croatia
n
Bosnia & Herzeg.

n
%
Victimization in last year ***

Yes
176
No
1584
10,0%
90,0%
55
677
7,5%
92,5%
138
862
13,8%
86,2%
Speed running***
Slow
Average
Fast
495
722
542
28,1%
41,0%
30,8%
122
310
300
16,7%
42,3%
41,0%
439
289
272
43,9%
28,9%
27,2%
Health in last year***

Poor
193
Average
488
Very good
1078
11,0%
27,7%
61,3%
55
124
553
7,5%
16,9%
75,5%
209
207
584
20,9%
20,7%
58,4%
Chat to people locally***

Never
331
Sometimes
1429
Often
0
18,8%
81,2%
0,0%
26
269
437
3,6%
36,7%
59,7%
50
410
540
5,0%
41,0%
54,0%
Having friends locally***

No
510
Some
1250
Many
0
29,0%
71,0%
0,0%
41
263
428
5,6%
35,9%
585%
91
464
445
9,1%
46,4%
44,5%
Walk after dark***

Never
Sometimes
Often
917
843
0
52,1%
47,9%
0,0%
92
321
319
12,6%
34,9%
43,6%
283
424
293
28,3%
42,4%
29,3%
Streets unsafe***
Yes
No
1299
461
73,8%
26,2%
477
255
65,2%
34,8%
626
374
62,6%
37,4%
Shops unsafe***
Yes
No
192
1565
10,9%
89,1%
50
682
6,8%
93,2%
159
841
15,9%
84,1%
Woods unsafe***
Yes
No
853
904
48,5%
51,5%
411
321
56,1%
43,9%
738
262
73,8%
26,2%
Chi-square used for comparison of the samples.

Level of significance:
*
p < .05,
** p < .01,
*** p < .001.
psychological variables used in Farralls et al. (1996) study on fear of crime. We

will first start with the presentation of results of socio-demographic model, which
has its limits because it does not focus on real crime situations but is focused
on demographic characteristics of potentially fearful people, people at risk or
184
victimised people. Social-psychological model includes dimensions of fear of

crime and specific situations, which can be easily imagined by the majority of
people and connected to their feelings of insecurity and fear. Socio-demographic
model focuses on socio-demographic determinism, while social psychological
includes a variety of interactions in peoples living environments.
The results presented in table 4 show that the highest level of victimisation in
the last year is typical for Bosnian respondents (13,8%). Individuals assessment
on ones high physical ability (speed running) is the most positive in Croats
(41,0%), what might be due to the fact that Croatian respondents were the
youngest. Health problems in last year are the least significant for the Croats.
Regarding local social networks (chat to people locally and having friends locally) the Croats seem to be the most locally connected people in comparison to
the Slovenes who seem to be quite distant to the people living in their neighbourhoods. Exposure to potential victimisation (walk after dark) is the most typical
for Bosnian respondents while Slovenes seem to expose at night themselves
less than Croats and Bosnians. Perception of streets as unsafe is the most typical
for Slovenes. Considering shops as unsafe is the most typical in Bosnia. Woods
as place/space where people feel unsafe are considered as a threat in Bosnia
(73,8%) and Croatia (56,1%), followed by Slovenia (48,5%).
We transformed 5-point Likert scales into nominal variables (e.g. I agree, I
dont know, I disagree) in social psychological variables for the purpose of this
comparison. The results are presented in table 5.
The results in table 5 show that only obstruction item (of criminalisable space
component) does not result in statistical significant differences in compared samples. All other components of social-psychological aspects of fear of crime differ
significantly. The results imply that respondents in Bosnia and Herzegovina assess themselves as possible targets of criminal activity more often than respondents from Croatia and Slovenia. Respondents from Slovenia emphasise jealousy
as a possible motive for criminal activity against them less than respondents from
Croatia and Bosnia. A general distrust towards strangers is the most expressed by
Slovenian respondents. Distrust to specific people living in respondents neighbourhoods is the most typical for Bosnians, followed by Croats and Slovenes.
Slovenes seem to be most sure of themselves of being capable of chasing of a
potential assailant and avoiding conflicts stay clear of rows. There are no differences regarding individuals imagination that someone would obstruct his or
her path. Respondents from Slovenia and Bosnia are more inclined to reconsideration a safe route when going out than Croatian respondents.
Table 5:
185
Comparison of social psychological indicators of fear of crime

Slovenia
Croatia
Bosnia & Herzeg.

n
%
Target***
Agree
Dont know
Disagree
494
515
749
28,1%
29,3%
42,6%
212
279
241
29,0%
38,1%
32,9%
507
250
243
50,7%
25,0%
24,3%
Jealousy***
Agree
Dont know
Disagree
247
302
1211
14,0%
17,2%
68,8%
123
199
410
16,8%
27,2%
56,0%
233
297
470
23,3%
29,7%
47,0%
Trust***
Agree
Dont know
Disagree
281
495
984
16,0%
28,1%
55,9%
252
274
206
34,4%
37,4%
28,1%
323
246
431
32,3%
24,6%
43,1%
Distrust***
Agree
Dont know
Disagree
854
456
450
48,5%
25,9%
25,6%
369
196
167
50.4%
26,8%
22,8%
640
222
138
64,0%
22,2%
13,8%
Attacker***
Agree
Dont know
Disagree
529
531
698
30,1%
30,2%
39,7%
100
141
491
13,7%
19,3%
67,1%
164
162
674
16,4%
16,2%
67,4%
Rows***
Agree
Dont know
Disagree
1279
282
199
72,7%
16,0%
11,9%
377
191
164
51,5%
26,1%
22,4%
548
254
198
54,8%
34,9%
19,8%
ATTRACTIVITY+
EVIL INTENT
POWER
CRIMINALISABLE SPACE
Obstruction (n.s.)
Agree
Dont know
Disagree
618
391
751
35,1%
22,2%
42,7%
272
189
271
37,2%
25,8%
37,0%
380
225
395
38,0%
22,5%
39,5%
Safe Route***
Agree
Dont know
Disagree
577
453
791
32,8%
25,7%
44,9%
201
189
365
27,5%
25,8%
49,9%
379
293
392
37,9%
29,3%
39,2%
+ See appendix for the description of items. Chi-square used for comparison of the three samples. Level of significance: * p < 0.05, ** p < 0.01, ***p<0.001.
186
4.2
Comparison of selected fear of crime issues
Statistically significant differences among samples (2, p < .05). Among variables
of the socio-demographic (utmost left column) and socio-psychological model
(utmost right column) coefficient of correlation was estimated (statistically
significant values stated). Numbers inside tables represent percentage of participants with specific level of fear expressed.
Table 6:

Slovenia
Level of fear
Low
High
Self-estimation of heath
Bad health
Good health
Bad health
Good health
Bad health
Good health
Bad health
Good health
35
46
62
55
37
45
36
46
39
25
19
15
40
33
43
30
Gender
m
f
m
f
m
f
44
56
60
40
59
41
48
52
30
70
37
63
Participants with high

level of fear
Average proportion (%)
Croatia
0,12
0,07
0,07
0,14
Low
High
33
34
56
23
22
41
40
50
38
27
16
37
51
35
31
27
21
78
58
42
55
45
66
34
29
71
34
66
0,09
0,23
0,16
0,13
Bosnia &
Herzeg.
Low High
Variable
19
27
64
35
29
46
27
42
65
44
25
36
54
33
42
36
35
65
65
34
60
40
67
33
40
60
40
60
Evil intent
(trust)
Crim. space
(obstruction)
Crim. space
(safe route)
0,18
0,24
0,21
0,17
Attractivity
(target)
Evil intent
(trust)
Crim. space
(obstruction)
Crim. space
(safe route)
28
29
51
Attractivity
(target)
14
17
23
Attractivity
(jealousy)
16
34
32
Evil intent
(trust)
48
50
64
Evil intent
(distrust)
30
14
16
Power (attacker)
73
51
55
Power
(rows)
35
37
38
Crim. space
(obstruction)
33
27
38
Crim. space
(safe route)
35
32
39
4.3
187
Factor and regression analyses
We conducted factor and regression analysis in the same manner as Farrall et al.
(1996), and Meko and Farrall (1999). Such analyses allow us to compare our
results with other studies conducted in other countries using the same methodology. First step we made was factor analysis of six fear of crime situations presented in tables 7 and 8. We saved the factor fear of crime for further regression analysis (see table 8).
Table 7:

Mean
SD
Slovenia
Croatia
B&H
Slovenia
Croatia
B&H
Doorbell
3.31
3.40
3.12
1.01
.96
1.15
Car
2.41
2.62
2.44
0.92
.91
1.03
To a party
2.05
2.01
2.21
0.86
.86
1.09
Bus stop
2.76
3.14
2.66
0.86
1.03
1.11
Telephone
3.27
3.59
3.29
1.01
1.02
1.19
Caf
3.24
3.36
2.48
0.95
1.06
1.09
SITUATION:
Ns = Slovenia (1760),
Croatia (732) and
Bosnia & Herzegovina (1000).
Table 8:
Summary of Factors Loadings Slovenia Croatia Bosnia

& Herzegovina
Slovenia
Croatia
Bosnia &
Herzeg.
Doorbell
.706
.647
.774
Car
.707
.738
.798
To a party
.680
.569
.709
Bus stop
.686
.694
.788
Telephone
.680
.618
.745
Caf
.608
.656
.616
SITUATION:
Maximum Likelihood Extraction used.

Slovenia: 45.93% of variance explained;
Croatia: 43.02% of variance explained and
Bosnia and Herzegovina: 54.89% with the factor we named fear of crime.
We saved factor scores for further regression analyses. The results of regression
analysis are presented in table 9.
188
Table 9:
Social Demographic and Social Psychological Models Regression

Analysis - Slovenia 2001, Croatia 2002 and Bosnia and Herzegovina
2002
Variable/Country
Slovenia
Beta
Croatia
b
SEb
Beta
-,046
Bosnia & Herzegovina

b
SEb
Beta
SEb
-,160
,431
,035
,002
,002
Socio-demographic Model
Age
,022
,001
,002
Gender
-,259*** -,142
-,050
-,142*** -,260
,063
-,227*** -,427
,053
Victimisation in last year
-,009
-,028
-,009
,032
,111
,100
-,044
-,121
,066
Time living in area
,020
,001
,002
-,043
-,095
,069
,029
,002
,002
Speed running
,075**
,060
,075
,013
,012
,030
,099**
,070
,025
Financial resources
,062
,041
,016
,055
,035
,019
,134***
,087
,019
Health in last year
,009
,007
,023
,006
,006
,030
,072*
,053
,025
Educational level
,004
,004
,027
-,085*
-,111
,045
-,011
-,012
,027
Household composition
,014
,060
,098
,043
,206
,144
-,041
-,080
,054
Work activity
,061*
,128
,055
-,034
-,109
,112
-,075**
-,142
,052
Chat to people
-,008
-,009
-,032
,013
,011
,029
,075*
,059
,027
Have friends locally
,017
,016
,028
-,020
-,015
,027
-,056
-,042
,026
Walk after dark
,178***
,156
,024
,143***
,103
,022
,080*
,055
,020
Streets unsafe
-,084***
-,174
,049
-,091**
-,175
,058
-,044
-,085
,049
Shops unsafe
-,035
-,099
,066
-,028
-,099
,106
-,060*
-,154
,065
Woods unsafe
-,099***
-,179
,044
-,081
-,148
,058
-,088*** -,189
,053
Social-Psychological Model
Attractivity
Target
,085***
,058
,017
,113***
,082
,022
,071**
,048
,018
Jealousy
,068*
,005
,018
,027
,020
,023
,036
,025
,018
Trust
-,095
-,076
,019
-,179*** -,140
,026
-,087**
-,056
,017
Distrust
-,006
-,004
,017
-,001
-,001
,023
-,010
-,008
,019
Attacker
-,108*** - ,077
,019
-,082*
-,064
,024
,044
,032
,018
Rows
,035
,018
,009
,007
,022
,019
,013
,018
Evil Intent
Power
,027
Criminalisable Space
Obstruction
,247***
,162
,017
,215***
,144
,022
,219***
,138
,018
Safe Route
,094***
,064
,017
,136***
,092
,021
,112
,072
,018
R-Square (adj.)
43,0
Constant included in models.

* = p<.05
** = p<.01
*** = p<.001.
42,7
45,1
189
Table 9 shows that four components of fear of crime can be taken into consideration: gender, frequency of walking outside after dark, self-perception of a potential victim/target and the frequency of thinking about the possibility of obstruction in all three studied countries. Perception of ones physical ability, unemployment (retirement), perception of streets and woods as unsafe places, considering other people as jealous, ability of chasing of a potential assailant, and
making sure that a route taken is safe, are related to fear of crime in the Slovenian sample. Fear of crime, in addition to common variables, is related to the
educational level, unsafe streets, low level of trusting people, ability of chasing
of a potential assailant, and making sure that a route taken is safe. Fear of crime
in Bosnia is, in addition to common variables, related to the perception of ones
physical ability, financial resources, unemployment (retirement), local social
networks, shops unsafe and level of distrust.
We decided to control for gender because in all three countries gender differences showed to be significant. The reason for doing so are the results of previous studies on fear of crime (Umek and Meko 1998; Meko and Kovo 1999),
which implied that fear of crime is more severely expressed by women (also
Warr 1984.), especially because of patriarchal relations in certain societies and a
greater vulnerability of females in interpersonal, often not reported violent
crimes.
5. Discussion
Generally speaking, the results of the regression analysis imply higher level of
fear of crime in women, people who perceive themselves as less physically fit,
the unemployed, those who expose themselves by walking alone in dark and
those who perceive streets and woods as sources of danger. In addition sociodemographic variables, peoples fears appear to be related to perceptions of oneself relative to an attacker who approaches one outside of the confines of ones
home. Imagining oneself to be physically stronger than this person reduces fear.
These findings have implications for the original model, theoretical work in this
field and crime surveys.
The relation between variables estimation of health and fear of crime reveals
that persons with a high self-estimation of health have lower level of fear (stands
for all three samples, the difference is statistically significant and noticeable).
Slovenian participants also think that a potential perpetrator could be quite dangerous and they are not prone to take risks. This is not valid for other two samples. Relationship between fear of crime and health leads us to the conclusion
that persons with a high estimation of health are more self-protective, not only
considering their own health but also in avoiding any risk behaviour.
190
Difference between male and female respondents is revealed in the variables,

which are related to a more direct personal threat (i.e. particularly in some of the
components of the social psychological model: When Im on my way home, I
sometimes imagine that someone would obstruct my path, When I have to go out
somewhere, I make sure that I take a safe route and I think Im capable of chasing
of a potential assailant.). The results for female respondents show a higher level of
fear, the differences come up to 25 % (the proportion of persons with a highest level
of fear within gender) and they are all statistically significant in all three samples.
Results reveal the fact that the highest discrepancy between male and female respondents appears at the Slovenian sample, especially at the variable power attacker (I think Im capable of chasing of a potential assailant).
The results of the social psychological model lead us to the conclusion that
about one third of respondents in all three countries express higher level of fear.
Findings are in accordance with results abroad (Houghes 1998). The highest percentage is noticed among participants from Bosnia and Herzegovina (39 % in
average, 35 % in Slovenia and 32 % in Croatia). In Slovenia there is no clear picture about differences among different age groups. It seems that the highest proportion of persons with higher level of fear is among adolescents (small difference and statistically not significant). In Croatia and especially in Bosnia and
Herzegovina quite different situation is detected higher level of fear is noticed
mainly among older participants (55 years of age and above). These findings are
expected and they can be explained by cognitive processes, which are typical for
age groups mentioned the young and the elderly are aware that they are sometimes exposed to some kind of criminal threat (or they believe so). Adolescents
stay out late in the night and seniors are aware of their diminished physical ability to resist a possible perpetrator.
Social psychological model reveals that fear of crime is lower among participants with a higher socio-economic status and a higher education level. In Slovenian sample the correlation is quite low but significant ( = 0,1 to 0,2; p < .02).
In other two samples the correlation coefficient is lower (around 0,01) and
mostly statistically non significant. It is assumed that well situated participants
have no difficulties to ensure themselves satisfactory level of safety. They can
choose a safe living environment and provide themselves with a better technical
equipment or security service.
In all three samples the most threatening circumstances are presented in vignettes
A Car (68 78 % of participants assumes feelings of being jeopardized), To a
party (28 42 %) and A Doorbell (48 57 %). A plausible reason for these results is the fact that these events are quite possible to happen in real life.
191
The results of the study imply that specific social circumstances impact the
level of insecurity and crime related fears of the respondents. It is often stated
that crime and recently fear of crime are related to the same factors of poverty,
post-war conditions (land mines, consequences of post-traumatic stress syndrome), social disorganisation, bad housing, migration, discrimination, organised
crime, etc. Fear of crime issues can be closely connected to the existential fear
and fear of loosing a certain level of the quality of life and ones perception of a
potential assailant is often related to a concept of stranger danger and the negligence of danger within (e.g. domestic violence).
5.1
Concluding remarks
The present examinations were conducted by different researchers in different

countries. Despite this example of researcher, data set and population triangulation, discernable patterns do emerge which do not differ much from the studies in
Western Europe. This is understandable because of globalisation and some significant crimes which have an international importance (e.g. 9/11 in the USA,
which caused a new wave of efforts for a safer world).
Fear of crime is insofar a very important concept as there is a strong influence
on political decisions. We find more and more punitiveness in western countries
(see Roberts et al. 2003.) in the last years, often based on the discussion of fear of
crime. As Farrall et al. (1997) could show for Great Britain and Kury et al.
(2004) for Germany the regular measurement of fear of crime overestimates the
fear level that is we need a better operationalised and more valid measurement of
this important concept.
Fear of crime is an important element of penal populism in a way that politicians recognise expressed fear of crime generated by the media, well publicized
cases and changes in some legal norms (often without sufficient expert consultation and pressure of so called moral entrepreneurs). Their declared goal is citizens protection from victimization but the hidden goal is quite frequently the
political gain (election votes).
At a theoretical level, this suggests to us that more attention needs to be given
to the relationship between those that report being fearful and those that the fearful imagine will attack them. Or, in other words, theories, which aim to account
for fear levels need to address the relationship between an individuals beliefs
and stereotypes about crime and their levels of anxiety.
Comparison of the three cultures implies that measuring of fear of crime is
also related to other issues, which would deserve a special research attention (i.e.
masculinity, self-attribution of being fearless or fearful, resignation, self-esteem,
defence mechanisms of denial/projection/rationalisation) and quality of life is-
192
sues (i.e. living in a people friendly environment) in a contemporary society

characterised by uncertainty and culture of fear (Furedi 1998). The modern fears
of being excluded, ugly, unhealthy, fat, unfit, financially incapable, rejected at
work by ones chief, etc. presented by Furedi (ibid.) also call for studying their
relation to the fear of crime. The assessments of sources of fear of crime also imply that contemporary concepts of crime and criminality issues are related to
stereotypes and prejudice, which should be overcome.
Fear of crime issues are a political matter because of fearful people and politicians who abuse fear of immigrants, the poor, the southerners, the easterners and
other marginalised and stigmatised social groups and attribute the criminality to
them. The media also create ones social construction of crime related reality.
The medias influence on ones understanding of attribution of criminality to
specific individuals, circumstances and possible victimisation deserve additional
research attention.
The present study implies the fact that higher levels of wealth, health and quality of life are related to less fear of crime. Care for peoples basic needs and quality of life (especially health) are issues closely related to the reduction of all
kinds of existential fears, among which fear of crime manifests ones attribution
of dangers which threaten his or her wellbeing. These issues are, generally speaking, related to the issues of contemporary thoughts on safety and security in our
communities. Meko (2004: 740) found out that representatives of local communities in Slovenia attributed the greater importance to the following factors of
crime prevention, which supposedly should solve their crime and fear of crime
problems. The factors of preventative measures are ranked as follows: (1) social
prevention (training for parents, solving social problems, student friendly school,
competent school teachers, leisure activities available, responsibility of business
owners), (2) self-protective measures (availability of information on crime prevention, self-protective measures), (3) formal social control (police punishment
of criminals) and (4) private social control (private security guards at school,
private security industry).
It is evident that the respondents in the present and Mekos (2004) study
call for something what we know as a welfare state, but social reality in the
most of countries of contemporary western civilisation reinforce liberal capitalist
model, which works on the free market principles and especially competition,
which pushes all incapable to the social margins where they can easily experience all kinds of fears and real victimisations. So called countries in transition
are facing turbulent changes and it is difficult to manage such a situation without
learning new survival skills.
193
Appendix
The Vignettes employed:

Doorbell
One evening youre at home on your own. Its late. The doorbell rings, but youre not
expecting anyone.
Car
One evening you go out to put the dustbin out. A short way up the street you see two
men walking around a parked car. When they see you looking at them, they begin to
walk toward you.
To a party
Youve been invited to a party in a neighbourhood you dont really know. Early that
evening you set out by bus. When you get off you still have a long way to walk. Suddenly you notice that youve lost your way. A group of youths is following you and
begins to make unpleasant remarks at you.
Bus stop
One afternoon youre standing at the bus stop nearest home, when a group of 15-16
year olds comes along. They begin kicking the bus stop and daubing graffiti on the
bus shelter.
Telephone
Youve going out one evening. Youre ready and just about to leave when the phone
rings. You answer, giving your name. But at the other end you hear only irregular
breathing. You ask whos there. They hang up.
Caf
Youre travelling through a town where youve never been before. You have to ring
home to say youll be late getting back. Because you cant find a telephone box, you go
into a caf to ring from there. It turns out to be where a group of bikers meets.
Each of these scenarios was followed by the Unsafety Question How unsafe would
you feel in such a situation? [response codes: 1 = very unsafe, 2 = quite unsafe, 3 =
dont know, 4 = quite safe, 5 = very safe]. It is the summation of the answer to this
question after these six scenarios that is used as the measure of the fear of crime (i.e. as
the dependent variable).
194
The Components of the Social Psychological Model

Could you tell me whether you agree or disagree with the following statements
[response codes: 1 = agree strongly,
2 = agree,
3 = dont know,
4 = disagree,
5 = disagree strongly].
ATTRACTIVITY
Target:
I think that people who are up to no good are likely to target especially on me and my
possessions.
Jealousy:
I think that people are jealous of me.
POWER
Attacker:
I think Im capable of chasing of a potential assailant.
Rows:
I generally stay clear of rows.
EVIL INTENT
Trust:
I generally trust strangers.
Distrust:
I distrust particular people in my surroundings.
Obstruction: When Im on my way home, I sometimes imagine that someone would
obstruct my path.
Safe Route:
When I have to go out somewhere, I make sure that I take a safe route.
195
References
Bennett, T. (1990). Tackling Fear of Crime, Home Office Research Bulletin, 28, 14-19.
Chambers, G. & Tombs, J. (1984). The British Crime Survey Scotland. Scottish Office
Central Research Unit.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1996). Fear of Crime: Conceptual
Development. Field Testing and Empirical Confirmation: Final Report to the ESRC.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1999). Angry or Afraid? Recalibrating The Fear of Crime, International Review of Victimology.
Journal of Criminology 37, 658-679.
Farrall, S., Ditton, J., Bannister, J. & Gilchrist, E. (2000). Social Psychology and the
Fear of Crime: Re-Examining a Speculative Model. British Journal of Criminology.
Furedi, F. (1998). Culture of Fear: Risk-taking and the Morality of Low Expectations.
Cassel: London.
Grusczynska, B. (2001). Victimization, Crime and Social Factors in Post-Socialist
Countries. In: Kury, H. (Ed.), International Comparison of Crime and Victimization.
The ICVS. Willowdale: de Sitter Publications
Hale, C. (1992). Fear of Crime: A Review of The Literature. Report Prepared for the
Metropolitan Police Service Working Party on Fear of Crime, University of Kent,
Canterbury.
(1996). Fear of Crime: A Review of the Literature. International Review of Victimology, 4, 79-150.
Hatalak, O., Del Frate, A. & Zveki, U., (Eds.) (1998). The International Crime Victim
Survey in Countries in Transition. Rome: UNICRI.
Houghes, G. (1998). Understanding Crime Prevention Social Control, Risk and Late
Modernity. Buckingham: Open University Press.
Ivii, I., aki, V. & Franc, R. (2004). Fear of crime in Croatia: its extent and predictors. Presentation at the ECS conference, Amsterdam, August. 2004.
Kury, H. (Ed.)(2001). International Comparison of Crime and Victimization. The ICVS.
Willowdale: de Sitter Publications.
Meko, G. (2004). Local Safety Councils in Slovenia A Story on Attempts to Make
Communities Responsible for Solving Crime and Safety Problems. In: Meko, G.,
Pagon, M., Dobovek, B. (Eds.), Policing in Central and Eastern Europe Dilemmas
of Contemporary Criminal Justice. Ljubljana: Faculty of Criminal Justice.
Meko, G. & Farrall, S. (1999). The social psychology of the fear of crime: a comparison of Slovenian, Scottish and Dutch local crime surveys. Hrvatska revija za rehabilitacijska istraivanja, 35, No. 2, 141-149.
Meko, G. & Kovo, I. (1999). Strah pred kriminaliteto v Sloveniji in Hrvaki. Varstvoslovje,
1, 2, 26-30.
196
Meko, G. & Areh, I. (2003). Strah pred kriminaliteto v urbanih okoljih. Revija za
kriminalistiko in kriminologijo, 54, 3, 256-264.
Meko, G. & Pavlovi, Z. (1998). Strah pred kriminaliteto izzivi za raziskovanje.
Revija za kriminalistiko in kriminologijo, 49, 4, 346-353.
Meko, G. & Umek, P. (1999). Znailnosti socialnopsiholokega in demografskega
modela strahu pred kriminaliteto. Revija za kriminalistiko in kriminologijo, 50, 2.
Muratbegopbvi, E. (2003). Policija u prevenciji kriminaliteta na razini lokalne zajednice Kanton Sarajevo, studija sluaja. Sarajevo: Fakultet kriminalistikih nauka.
Pavlovi, Z. (1998). Mednarodna anketa o viktimizaciji 1997. Intitut za kriminologijo
pri Pravni fakulteti v Ljubljani, raziskava t. 118, Ljubljana.
Roberts, J.V., Stalans, L.J., Indermaur, D. & Hough, M. (2003). Penal Populism and
Public Opinion. Lessons from Five Countries. Oxford: Oxford University Press.
eparovi, Z. & Turkovi, K. (1998). Zagreb (Croatia). In: Hatalak, O., Del Frate, A.,
Zveki, U. (Eds.), The International Crime Victim Survey in Countries in Transition.
Rome: UNICRI.
Umek, P. (2004). Viktimizacijska tudija 2001. In: Meko, G. (Ed.), Prepreevanje kriminalitete teorija, praksa in dileme. Ljubljana: Intitut za kriminologijo.
Van der Wurff, A., Stringer, P. & Timmer, F. (1986). Feelings of unsafety in residential
Surroundings. In: Canter, D., Jesuino, C. Soczka, L. & Stephenson, G. (Eds.), Environmental Social Psychology, Dordrecht: Kluwer.
Van der Wurff, A., Van Staalduinen, L. & Stringer, P. (1989). Fear of Crime in Residential Environments: Testing a Social Psychological Model. Journal of Social Psychology, 129 (2), 141-160.
Viktimizacijska tudija (2001). Ljubljana, Urad za Statistiko Republike Slovenije (Bureau of Statistics of the Republic of Slovenia). www.stat.si. Accessed December 13,
2004.
Warr, M. (1984). Fear of Victimisation: Why are Women and the elderly more afraid?
Social Science Quarterly, 65, 681-702.
Young, J. (1988). Risks of crime and Fear of crime: A Realist Critique of Survey based
Assumptions. In: Maguire, M., Ponting, J. (Eds.), Victims of Crime: A New Deal,
OUP, Milton Keynes.
(1992). Ten Points of Realism. In: Young, J., Matthews, R. (Eds.), Rethinking Criminology: The Realist Debate. Sage: London.
Zveki, U. (1998). Criminal Victimisation in Countries in Transition. Rome: UNICRI.
173
Social-Demographic and Social-Psychological

Perspectives of Fear of Crime in Slovenia,
Croatia and Bosnia & Herzegovina
GORAZD MEKO, IRMA KOVO-VUKADIN

AND ELMEDIN MURATBEGOVI
Abstract
Research and theorising about the fear of crime has, in the main, been dominated
by researchers who have relied upon sociological or socio-demographic variables
to account for variations in fear levels. Whilst this body of work has contributed
greatly to our understanding of the fear of crime, we are still far from a full understanding of this important and most corrosive aspect of contemporary society.
This paper first presents three independent evaluations of fear of crime. The data,
collected in Slovenia, Scotland and Holland, sheds further light on both the social
psychological model proposed and the wider study of fear of crime. These
evaluations served as a background for the present study of fear of crime in Slovenia in 2001 and Croatia and Bosnia and Herzegovina in 2002. The results show
higher level of fear of crime in women, people who perceive themselves as less
physically fit, the unemployed, those who expose themselves by walking alone in
dark and those who perceive streets and woods as sources of danger. In addition,
higher level of fear of crime can be attributed to respondents who consider themselves a potential victim due to the jealousy of others or being attractive to a potential criminal in other ways, incapable of chasing of a potential assailant, and
perceiving places as dangerous (criminalisable space).
174
1. Introduction
The fear of crime is nowadays one of the most researched topics in contemporary
international criminology. In the United Kingdom and United States, crime surveys have expanded rapidly since the late 1960s. For example, the British Crime
Survey now biennially interviews in the region of 10,000 residents of England
and Wales. This survey has shed light on attitudes to policing, victimisation, perceptions of risk and peoples fear of crime. Crime surveys are conducted, not
only in Western European countries, but also in Central and Eastern European
countries (Hatalak, del Frate, Zveki 1998; Zveki 1998; Kury 2001).
In Slovenia, three major crime and victimisation surveys have been undertaken to date. The first in 1992, the second in 1996/97 (Pavlovi 1998) and the
third in 2000/2001 (National Bureau of Statistics, Internet, Umek 2004.). The
fourth study on victims of crime (ICVS) is ready to be conducted in Slovenia in
2004/05.
In Croatia, two major crime and victimisation studies (ICVS) have been conducted so far (eparovi and Turkovi 1998; Gruszczynska 2001: 77-89). The
last study on fear of crime in Croatia was conducted by Ivii, aki and Franc
(2004).
In Bosnia and Herzegovina, only one crime and victimisation survey has been
undertaken to date (2001) by researchers from the University of Lausanne (within United Nations Mission of International Police Task Forces in Bosnia and
Herzegovina) (Muratbegovi, 2003).
The findings of the UK and USA research are now well known. A plethora of
studies have concluded that the fear of crime impinges upon the well-being of a
large proportion of the population. Some have even gone as far as to suggest that
the fear of crime is now a larger problem than crime itself (Hale 1992; Bennett
1990; Warr 1984.). Chambers and Tombs (1984: 29) reviewing the 1982 British
Crime Survey (Scotland) reported that "more than half of the respondents (58%)
said that at some time in the past they had been concerned about the possibility of
being a victim of crime".
The fear of crime impacts the quality of life. People tend to adjust their behaviour according to their fear of crime level in a way of not engaging in some activities or doing some (for example buying a gun for a protection of themselves
or their property). The fear of crime also influences personal effectiveness and
productivity in everyday life.
As well as receiving much attention at an empirical level, many have attempted
to explain the fear of crime. These efforts have tended to be dominated by researchers influenced by sociological insights. Thus such variables as age, gender, house-
175
hold income, friendship networks, length of residence, earlier victimization experiences and so on have been suggested as key in explaining fear of crime. This approach is also called socio-demographic model of fear of crime. Such models have
indeed been found to be of utility in explaining the fear of crime.
However, by concentrating on sociological variables, researchers have largely
ignored social psychological and psychological factors that may be important in
explaining the fear of crime (Van der Wurff et al. 1989: 142 and 1986; Farrall et
al. 2000.). This domination of the literature by sociologically informed theorising
has ignored important processes occurring at the individual level. In this article
we discuss further analyses of the only social psychological model of the fear of
crime (Van der Wurff et al. 1989). The aim of the current article is to shed light,
in addition to socio-demographic dimensions, also on the relationship between
social psychological factors and the fear of crime, and as such to develop further
the analysis undertaken by Van der Wurff et al. (1989) and Farrall et al. (2000)
and our own analyses on the basis of international data.
This article is constructed in the following fashion. We start by outlining the
social psychological model proposed by Van der Wurff et al. Following this we
describe the methodologies employed in each of the surveys (and the countries in
which these surveys were conducted in the past). We then discuss the question
wording employed in the surveys. Having assessed the data in terms of its suitability for replicating the earlier study, we present findings of the studies based
on the same methodology in the past. We further present results of our present
study on fear of crime in Slovenia, Croatia and Bosnia and Herzegovina, using
both models socio-demographic and social psychological (a holistic model).
We conclude the article with a discussion of what these findings mean in specific
social environments of Slovenia, Croatia, and Bosnia and Herzegovina.
1.1
The Socio-demographic and Social Psychology Model
The socio-demographic model consists of the following variables: age, gender,

household income, friendship networks, length of residence, earlier victimization
experiences and so on. The social psychological model that we shall discuss in
this article was originally proposed and tested by Van der Wurff et al. (1989)
using data collected in the Netherlands. Van der Wurff et al. develop a social
psychological model based on the assumption that the fear of crime is associated
with four social psychological components. These four they describe as follows
(1989: 144-5) 1 :
1
We discuss the original operationalisation of these components and our minor changes to
them in the Appendix.
176
The Attractivity [component 2 ] is intended to refer to the extent to which

people see themselves or their possessions as an attractive target or victim
for criminal activities. It involves the attribution of a characteristic to oneself and ones possessions. One thinks, for example, of the peculiar sensation one may have when walking on the street with a great deal of money.
Another example would be the fear of burglary, which may be experienced
if one keeps valuable articles in the house.
The Evil Intent [component] relates to the wrongdoers role in the phenomenon. It is represented by the extent to which a person attributes criminal intentions to another individual or particular group. Thus, one may be
afraid of having ones pocket picked the moment one sees a gypsy. Or one
can experience fear as a result of a feeling that society is in moral decay and
a conviction that present-day youth are prepared to commit murder for a
paltry sum of money.
The Power [component] refers to the degree of self-assurance and feeling of control that a person has with respect to possible threat or assault by
another. In principle it is a question of two related sub-factors: ones own
power and the power of the other. The first of these relates to a persons
confidence in his 3 own efficacy. This need not be directly related to the
dangers of crime, of course. Feelings of self-assurance, control, and confidence in meeting the challenges of life will by generalisation tend to lower a
persons sensitivity to feelings of threat. Almost anything can contribute to
the feeling of ones own power, from a good family relationship to an optimistic temperament.
The power of the other is the wrongdoers side of the coin. It concerns
characteristics attributed to potential criminals, such as their strength, agility, resources, and general ability to carry out their criminal intentions. A
comparison of ones own power with power of other determines whether a
person faces confrontations with that other with confidence or not. Thus, the
idea that even the smallest thief goes about carrying weapons can lead to
feelings of uneasiness or fear, if one has no compensating power of ones
own.
Criminalisable Space [component] is the fourth and final [component].
Whereas the first [component] refers to the potential victim, the second to
the potential wrongdoer, and the third to both of these parties, the last [component] has to do with the situation in which a crime may take place. The
emphasis is on characteristics of place and time and on the presence of others, It is a question of the extent to which a situation lends itself to criminal
activities in the eyes of a possible victim of how much the situation facilitates crime or the criminal. A criminalisable situation might, for example,
2
Van der Wurff et al. use the word factor. We use component so as not to cause confusion
when discussing factor analysis in future sections.
Sexism in original text.
177
include walking at night through a poorly lit pedestrian subway or through a

dark wood, although estimates of criminalisability for any one situation can
naturally vary between individuals. The interest here lies in the extent to
which people have a general tendency to heed the criminalisability of the
situations into which they venture.
As is evident from outlining their model, Van der Wurff et al.s primary consideration in the construction of the model was the perception of the individual. Van
der Wurff et al. (1989: 143-4) are careful to note that there is no causal ordering
inherent in their model. That is to say that these components are merely associated with fear of crime. The fear of crime may exert an influence upon these
components and vice versa they may exert an influence upon it. A further feature
of this model is that it operationalises Youngs (1988 and 1992) observation that
the research on the fear of crime has ignored the perpetrators of crimes, but might
include them in future models.
1.2
Van der Wurff et al.s Results
The first study of this kind was conducted by Van der Wurff et al. (1989: 155)
who found that the social psychological model explained about 24% of the variance of the fear of crime. This is more variance in fear of crime levels than is
usually explained. A previous replication of this model (Farrall et al. 1999) finds
similar results, although they report a significant gender bias in the model and
propose that socio-demographic variables can be added to the model to increase
its power still further.
2. The Past Studies: Holland Scotland Slovenia

The replication of any study is a generally worthwhile endeavour. Replication
allows for an assessment of the both the extent to which one can generalise the
findings of the model to different populations, and the validity of the instrument
used. In this article we present the initial results reported by van der Wurff et al.
with those of Farrall et al. and with a similar survey undertaken in Slovenia
(Meko and Farrall 1999). The authors of previous research were concerned primarily with the utility and operation of the social psychological model, and for
the sake of brevity did not concern themselves with either socio-demographic
variables or the confounding issue of gender.
2.1
Question Wording
We are fortunate that Van der Wurff et al. reproduced translations (into English)
of the items upon which they relied. These English translations were then used
178
largely unchanged for the Scottish survey, and translated into Slovene for the
first Slovenian survey (1998). For the Slovenian survey, the present authors corresponded on this issue so as to ensure that the meanings of the words used were
as comparable as was possible. In order to fully assess its utility, the questions
used to produce the social psychological model developed by Van der Wurff et
al. were employed in the crime surveys outlined above. All questions referred to
here are reproduced in the Appendix (along with notes concerning any recodings
performed). Unlike Van der Wurff et al., who relied upon a stratified survey of
110 respondents in each of four sampling points, the Slovenian and Scottish surveys employed simple random sampling techniques.
The measure of the fear of crime employed by Van der Wurff et al. relies upon
vignettes, which outline six situations. Vignettes (one of which is reproduced
below the rest are reproduced in the Appendix) have the benefit of providing
the measurement of complex phenomena in social contexts.
Vignette Four: To a party
Youve been invited to a party in a neighbourhood you dont really know.
Early that evening you set out by bus. When you get off you still have a long
way to walk. Suddenly you notice that youve lost your way. A group of
youths is following you and begins to make unpleasant remarks at you.
Each of these vignettes are followed by questions on associated feelings of unsafety. Of these questions, only one for each vignette, referred to as the unsafety question is employed in further analysis. These six unsafety questions
are summated and used as the measure of the fear of crime (i.e. as the dependent variable). Each of the vignettes was employed in accordance with the procedures outlined by Van der Wurff et al. (1989: 148).
The four components of the social psychological model (Attractivity, Power,
Evil Intent and Criminalizable Space.) were measured through the use of eight
questions (two per component). In the original article these eight questions were
left unnamed, however, we have employed the same titles as given to these variables by Farrall et al. These questions are also reproduced in the Appendix.
2.2
Previous Analysis of the Data Sets
Van der Wurff et al. go to considerable lengths to examine their data set. They
test the six unsafety items employed for their distribution of answers, their reli-
179
ability and their uniformity. In order to assess the extent to which our data set is
comparable to that of Van der Wurff et al.s, Meko and Farralls (1999) commenced their analysis by repeating these tests in order to establish the generalisability of the model to other populations. Tables 1 and 2 reproduce their results
for the six Unsafety questions for each vignette (Meko and Farrall, 1999) in Slovenia, Scotland and Holland.
Table 1:

Mean
SD
Slovenia
Scotland
Holland Slovenia
Scotland
Holland
Doorbell
3.31
3.05
2.35
1.01
1.28
1.29
Car
2.41
2.53
3.27
0.92
1.15
1.23
To a party
2.05
1.73
3.77
0.86
0.86
1.08
Bus stop
2.76
2.53
2.30
0.86
1.14
1.14
Telephone
3.27
3.18
1.96
1.01
1.19
1.25
Caf
3.24
3.25
2.31
0.95
1.19
1.32
Situation:
Ns = Slovenia (443), Scotland (482-485) and Holland (440).
Table 2:
Summary of Factors Loadings

Slovenia
Scotland
Holland
Doorbell
.76
.66
.66
Car
.77
.54
.71
To a party
.69
.29
.63
Bus stop
.71
.51
.61
Telephone
.71
.71
.47
Caf
.67
.56
.36
Situation:
Ns = Slovenia (443), Scotland (482-485), and Holland (440). Maximum Likelihood

Extraction used. KMO statistic: (Slovenia) .835 (Scotland) .779.
As with the earlier study, both the Slovenian and Scottish data sets produced just
one factor, suggesting unidimensionality in the data set. Along with the similar
Alpha coefficients (higher than 0,70), this suggests that the Slovenian and Scottish data sets enable us to undertake a robust replication of the earlier study.
180
We come now to the presentation of our findings with regard to the replication
of the social psychological model.
Two results stand out as being of particular interest the relative differences
in the adjusted R-Square figures produced and the similarity of the models produced when only significant values are considered. The R-Square values range
considerably from over 35% (Slovenia) to under half of that (17%, Scotland).
The exact meaning of this is hard to fathom. It could be due to some of the differences observed earlier between the Slovenian and Scottish samples or alternatively due to some unanticipated influence of question wordings.
The second, more interesting observation concerns the elements which enter
the three examinations of the model. In all three of the examinations undertaken,
the two elements of Criminalisable Space enter the models and at the highest
levels of significance (p > .01 and .001). In addition to this, in all but one of the
models the two elements of Power also enter the model. The other two elements of the model (Attractivity and Evil Intent) fare less well, and their entry
to the model is ambiguous or uniformly non-existent. Van der Wurff et al. (1989:
155) report a similar finding in their regression analyses.
The variable, which we named Obstruction in the Criminalisable Space
component is partly interesting, as it is entered in the same direction for each
model at the highest level of significance (in short, thinking that one would have
ones path blocked, is associated with feeling unsafe). The direction of second
element of the Criminalisable Space component (Safe Route), is less clear,
but for two of the three models, taking a safe route is associated with feeling
less safe. In Scotland, taking a safe route is associated with feeling safer, but this
is at a lower level of significance. The elements of Power (Attacker and
Rows) act in a less uniform manner, but generalisations are still possible.
Thinking that one can chase off an attacker is associated with feeling more safe
(Slovenia and Scotland), whilst steering clear of rows has the same effect in
Scotland, but not Holland (See Meko and Farrall 1999).
This suggests a number of points to us. First of all anxieties about crime are
perhaps reducible to features of the relationship between the individual and the
person they imagine as their assailant. Anxieties are observed to be associated
with having ones route blocked, but mitigated if one feels one can chase off an
assailant. Hence anxieties appear to be related to imagined vulnerability and perceptions of ones physical strength relative to the assailant. Similarly, exhibiting
a degree of consciousness about ones personal safety (making sure one takes a
safe route home), is related to greater levels of fear suggesting that the fearful
take precautions.
181
Following the studies conducted in Holland, Scotland and Slovenia we decided to conduct the another study in Slovenia (2001), Croatia (2002) and Bosnia
and Herzegovina (2002), which were before 1991 republics of Yugoslavia and
faced different social developments in the nature and extent of crime (especially
violence related to war and post-war conditions). Therefore, we assume that different levels and sources of fear of crime can be attributed to each of the studied
countries.
3. The Present Study: Slovenia, Croatia and Bosnia & Herzegovina

We used the questionnaire previously used by Van der Wurff (1989) and Farrall
et al. (1997). The questionnaire was first translated into Slovene by Meko and
Umek (1998) and tested for its applicability. The same procedure was conducted
by Meko and Kovo (1999) for the use of the questionnaire in Croatia in 1998.
The questionnaire was later on translated into Bosnian language (which does not
differ much from Croatian). Question wording is the same in all three questionnaires and the meaning of statements and coding is the same. The authors have
thoroughly checked for the meaning of each item in the questionnaire regarding
its connotative and denotative meaning.
The Slovenian data come from a survey of people living in Slovenia. Sixty
percent of the sample comes from the capital city of Ljubljana. The interviews
were conducted in spring and autumn 2001 by the students of third year of criminal justice studies at the Faculty of Criminal Justice and Security in Ljubljana.
The Croatian data come from a random sample of people living in the capital
city of Croatia Zagreb. The data were collected in autumn 2002 by interviewing
inhabitants. The interviews with respondents were conducted by trained students
of Police College in Zagreb.
The Bosnian data come from a survey of people living in Canton Sarajevo the
biggest and social and economic most developed region in Bosnia and Herzegovina. A random stratified sample of 1000 residents, 200 from each of five urban
municipalities (Old Town, Center, New Sarajevo, New Town and Ilida). People were randomly selected by their address and interviewed face to face by the
trained students of 3rd year of graduate program at Faculty of Criminal Justice
Science University of Sarajevo. The survey was undertaken during September
December 2002. Characteristics of the respondents included into the samples are
presented in table 3.
182
Table 3:
Characteristics of studied samples in Slovenia, Croatia and B&H

(N= 3491)
Demographics
Slovenia
n
Age
Adolescents (15-20)
334
Young adults (21-30)
556
Middle aged adults (31-55)
702
Early seniors (56-70)
115
Seniors (71-90)
51
Gender
Female
933
Male
827
Educational level
Primary school finished
231
High school finished
1126
Vocational univ. degree
159
University degree
244
Homeowner
Yes
1161
No
599
Time living at the present address
Less than 1 year
125
1 to 2 years
95
3 to 5 years
138
6 to 10 years
223
More than 11 years
1179
Living with
Parents, family
1645
Alone
113
Financial ability
Low
832
Average
398
High
530
Total number of respondents
1760
Croatia
Bosnia & Herzeg.

n
%
19,0%
31,6%
39,9%
6,5%
2,9%
467
128
108
25
4
63,8%
17,5%
14,8%
3,4%
0,5%
79
291
397
212
21
7,9%
29,1%
39,7%
21,2%
2,1%
53,0%
47,0%
410
322
56,0%
44,0%
496
503
49,6%
50,4%
13,1%
64,0%
9,0%
13,9%
11
589
42
90
1,5%
80,5%
5,7%
12,3%
110
552
159
179
11,0%
55,2%
15,9%
19,9%
65,96%
34,04%
577
155
78,8%
21,2%
637
363
63,7%
36,3%
7,1%
5,4%
7,8%
12,7%
67,0%
51
31
71
81
498
7,0%
4,2%
9,7%
11,1%
68%
49
36
130
162
623
4,9%
3,6%
13,0%
16,2%
62,3%
93,6%
6,4%
705
27
96,3%
3,7%
815
185
81,5%
18,5%
47,3%
22,6%
30,1%
251
177
304
34,3%
24,2%
41,5%
436
214
349
43,6%
21,4%
34,9%
732
1000
KMO statistics (sampling adequacy coefficient):

Slovenia 0.844, Croatia 0.826 and Bosnia & Herzegovina 0.880.
4. Results
4.1
Comparison of socio-demographic and social psychological

variables
Socio-demographic variables of fear of crime were studied in other studies (Pavlovi 1992; 1996; Zveki 1998; Umek and Meko 1999; Meko and Kovo
1999). For the purpose of this article, we include socio-demographic and socio-
Table 4:
183
Socio-demographic model of fear of crime comparisons

n
Slovenia
%
Croatia
n
Bosnia & Herzeg.

n
%
Victimization in last year ***

Yes
176
No
1584
10,0%
90,0%
55
677
7,5%
92,5%
138
862
13,8%
86,2%
Speed running***
Slow
Average
Fast
495
722
542
28,1%
41,0%
30,8%
122
310
300
16,7%
42,3%
41,0%
439
289
272
43,9%
28,9%
27,2%
Health in last year***

Poor
193
Average
488
Very good
1078
11,0%
27,7%
61,3%
55
124
553
7,5%
16,9%
75,5%
209
207
584
20,9%
20,7%
58,4%
Chat to people locally***

Never
331
Sometimes
1429
Often
0
18,8%
81,2%
0,0%
26
269
437
3,6%
36,7%
59,7%
50
410
540
5,0%
41,0%
54,0%
Having friends locally***

No
510
Some
1250
Many
0
29,0%
71,0%
0,0%
41
263
428
5,6%
35,9%
585%
91
464
445
9,1%
46,4%
44,5%
Walk after dark***

Never
Sometimes
Often
917
843
0
52,1%
47,9%
0,0%
92
321
319
12,6%
34,9%
43,6%
283
424
293
28,3%
42,4%
29,3%
Streets unsafe***
Yes
No
1299
461
73,8%
26,2%
477
255
65,2%
34,8%
626
374
62,6%
37,4%
Shops unsafe***
Yes
No
192
1565
10,9%
89,1%
50
682
6,8%
93,2%
159
841
15,9%
84,1%
Woods unsafe***
Yes
No
853
904
48,5%
51,5%
411
321
56,1%
43,9%
738
262
73,8%
26,2%
Chi-square used for comparison of the samples.

Level of significance:
*
p < .05,
** p < .01,
*** p < .001.
psychological variables used in Farralls et al. (1996) study on fear of crime. We

will first start with the presentation of results of socio-demographic model, which
has its limits because it does not focus on real crime situations but is focused
on demographic characteristics of potentially fearful people, people at risk or
184
victimised people. Social-psychological model includes dimensions of fear of

crime and specific situations, which can be easily imagined by the majority of
people and connected to their feelings of insecurity and fear. Socio-demographic
model focuses on socio-demographic determinism, while social psychological
includes a variety of interactions in peoples living environments.
The results presented in table 4 show that the highest level of victimisation in
the last year is typical for Bosnian respondents (13,8%). Individuals assessment
on ones high physical ability (speed running) is the most positive in Croats
(41,0%), what might be due to the fact that Croatian respondents were the
youngest. Health problems in last year are the least significant for the Croats.
Regarding local social networks (chat to people locally and having friends locally) the Croats seem to be the most locally connected people in comparison to
the Slovenes who seem to be quite distant to the people living in their neighbourhoods. Exposure to potential victimisation (walk after dark) is the most typical
for Bosnian respondents while Slovenes seem to expose at night themselves
less than Croats and Bosnians. Perception of streets as unsafe is the most typical
for Slovenes. Considering shops as unsafe is the most typical in Bosnia. Woods
as place/space where people feel unsafe are considered as a threat in Bosnia
(73,8%) and Croatia (56,1%), followed by Slovenia (48,5%).
We transformed 5-point Likert scales into nominal variables (e.g. I agree, I
dont know, I disagree) in social psychological variables for the purpose of this
comparison. The results are presented in table 5.
The results in table 5 show that only obstruction item (of criminalisable space
component) does not result in statistical significant differences in compared samples. All other components of social-psychological aspects of fear of crime differ
significantly. The results imply that respondents in Bosnia and Herzegovina assess themselves as possible targets of criminal activity more often than respondents from Croatia and Slovenia. Respondents from Slovenia emphasise jealousy
as a possible motive for criminal activity against them less than respondents from
Croatia and Bosnia. A general distrust towards strangers is the most expressed by
Slovenian respondents. Distrust to specific people living in respondents neighbourhoods is the most typical for Bosnians, followed by Croats and Slovenes.
Slovenes seem to be most sure of themselves of being capable of chasing of a
potential assailant and avoiding conflicts stay clear of rows. There are no differences regarding individuals imagination that someone would obstruct his or
her path. Respondents from Slovenia and Bosnia are more inclined to reconsideration a safe route when going out than Croatian respondents.
Table 5:
185
Comparison of social psychological indicators of fear of crime

Slovenia
Croatia
Bosnia & Herzeg.

n
%
Target***
Agree
Dont know
Disagree
494
515
749
28,1%
29,3%
42,6%
212
279
241
29,0%
38,1%
32,9%
507
250
243
50,7%
25,0%
24,3%
Jealousy***
Agree
Dont know
Disagree
247
302
1211
14,0%
17,2%
68,8%
123
199
410
16,8%
27,2%
56,0%
233
297
470
23,3%
29,7%
47,0%
Trust***
Agree
Dont know
Disagree
281
495
984
16,0%
28,1%
55,9%
252
274
206
34,4%
37,4%
28,1%
323
246
431
32,3%
24,6%
43,1%
Distrust***
Agree
Dont know
Disagree
854
456
450
48,5%
25,9%
25,6%
369
196
167
50.4%
26,8%
22,8%
640
222
138
64,0%
22,2%
13,8%
Attacker***
Agree
Dont know
Disagree
529
531
698
30,1%
30,2%
39,7%
100
141
491
13,7%
19,3%
67,1%
164
162
674
16,4%
16,2%
67,4%
Rows***
Agree
Dont know
Disagree
1279
282
199
72,7%
16,0%
11,9%
377
191
164
51,5%
26,1%
22,4%
548
254
198
54,8%
34,9%
19,8%
ATTRACTIVITY+
EVIL INTENT
POWER
Obstruction (n.s.)
Agree
Dont know
Disagree
618
391
751
35,1%
22,2%
42,7%
272
189
271
37,2%
25,8%
37,0%
380
225
395
38,0%
22,5%
39,5%
Safe Route***
Agree
Dont know
Disagree
577
453
791
32,8%
25,7%
44,9%
201
189
365
27,5%
25,8%
49,9%
379
293
392
37,9%
29,3%
39,2%
+ See appendix for the description of items. Chi-square used for comparison of the three samples. Level of significance: * p < 0.05, ** p < 0.01, ***p<0.001.
186
4.2
Statistically significant differences among samples (2, p < .05). Among variables
of the socio-demographic (utmost left column) and socio-psychological model
(utmost right column) coefficient of correlation was estimated (statistically
significant values stated). Numbers inside tables represent percentage of participants with specific level of fear expressed.
Table 6:

Slovenia
Level of fear
Low
High
Self-estimation of heath
Bad health
Good health
Bad health
Good health
Bad health
Good health
Bad health
Good health
35
46
62
55
37
45
36
46
39
25
19
15
40
33
43
30
Gender
m
f
m
f
m
f
44
56
60
40
59
41
48
52
30
70
37
63
Participants with high

level of fear
Average proportion (%)
Croatia
0,12
0,07
0,07
0,14
Low
High
33
34
56
23
22
41
40
50
38
27
16
37
51
35
31
27
21
78
58
42
55
45
66
34
29
71
34
66
0,09
0,23
0,16
0,13
Bosnia &
Herzeg.
Low High
Variable
19
27
64
35
29
46
27
42
65
44
25
36
54
33
42
36
35
65
65
34
60
40
67
33
40
60
40
60
Evil intent
(trust)
Crim. space
(obstruction)
Crim. space
(safe route)
0,18
0,24
0,21
0,17
Attractivity
(target)
Evil intent
(trust)
Crim. space
(obstruction)
Crim. space
(safe route)
28
29
51
Attractivity
(target)
14
17
23
Attractivity
(jealousy)
16
34
32
Evil intent
(trust)
48
50
64
Evil intent
(distrust)
30
14
16
Power (attacker)
73
51
55
Power
(rows)
35
37
38
Crim. space
(obstruction)
33
27
38
Crim. space
(safe route)
35
32
39
4.3
187
Factor and regression analyses
We conducted factor and regression analysis in the same manner as Farrall et al.
(1996), and Meko and Farrall (1999). Such analyses allow us to compare our
results with other studies conducted in other countries using the same methodology. First step we made was factor analysis of six fear of crime situations presented in tables 7 and 8. We saved the factor fear of crime for further regression analysis (see table 8).
Table 7:

Mean
SD
Slovenia
Croatia
B&H
Slovenia
Croatia
B&H
Doorbell
3.31
3.40
3.12
1.01
.96
1.15
Car
2.41
2.62
2.44
0.92
.91
1.03
To a party
2.05
2.01
2.21
0.86
.86
1.09
Bus stop
2.76
3.14
2.66
0.86
1.03
1.11
Telephone
3.27
3.59
3.29
1.01
1.02
1.19
Caf
3.24
3.36
2.48
0.95
1.06
1.09
SITUATION:
Ns = Slovenia (1760),
Croatia (732) and
Bosnia & Herzegovina (1000).
Table 8:
Summary of Factors Loadings Slovenia Croatia Bosnia

& Herzegovina
Slovenia
Croatia
Bosnia &
Herzeg.
Doorbell
.706
.647
.774
Car
.707
.738
.798
To a party
.680
.569
.709
Bus stop
.686
.694
.788
Telephone
.680
.618
.745
Caf
.608
.656
.616
SITUATION:
Maximum Likelihood Extraction used.

Slovenia: 45.93% of variance explained;
Croatia: 43.02% of variance explained and
Bosnia and Herzegovina: 54.89% with the factor we named fear of crime.
We saved factor scores for further regression analyses. The results of regression
analysis are presented in table 9.
188
Table 9:
Social Demographic and Social Psychological Models Regression

Analysis - Slovenia 2001, Croatia 2002 and Bosnia and Herzegovina
2002
Variable/Country
Slovenia
Beta
Croatia
b
SEb
Beta
-,046
Bosnia & Herzegovina

b
SEb
Beta
SEb
-,160
,431
,035
,002
,002
Socio-demographic Model
Age
,022
,001
,002
Gender
-,259*** -,142
-,050
-,142*** -,260
,063
-,227*** -,427
,053
Victimisation in last year
-,009
-,028
-,009
,032
,111
,100
-,044
-,121
,066
Time living in area
,020
,001
,002
-,043
-,095
,069
,029
,002
,002
Speed running
,075**
,060
,075
,013
,012
,030
,099**
,070
,025
Financial resources
,062
,041
,016
,055
,035
,019
,134***
,087
,019
Health in last year
,009
,007
,023
,006
,006
,030
,072*
,053
,025
Educational level
,004
,004
,027
-,085*
-,111
,045
-,011
-,012
,027
Household composition
,014
,060
,098
,043
,206
,144
-,041
-,080
,054
Work activity
,061*
,128
,055
-,034
-,109
,112
-,075**
-,142
,052
Chat to people
-,008
-,009
-,032
,013
,011
,029
,075*
,059
,027
Have friends locally
,017
,016
,028
-,020
-,015
,027
-,056
-,042
,026
Walk after dark
,178***
,156
,024
,143***
,103
,022
,080*
,055
,020
Streets unsafe
-,084***
-,174
,049
-,091**
-,175
,058
-,044
-,085
,049
Shops unsafe
-,035
-,099
,066
-,028
-,099
,106
-,060*
-,154
,065
Woods unsafe
-,099***
-,179
,044
-,081
-,148
,058
-,088*** -,189
,053
Social-Psychological Model
Attractivity
Target
,085***
,058
,017
,113***
,082
,022
,071**
,048
,018
Jealousy
,068*
,005
,018
,027
,020
,023
,036
,025
,018
Trust
-,095
-,076
,019
-,179*** -,140
,026
-,087**
-,056
,017
Distrust
-,006
-,004
,017
-,001
-,001
,023
-,010
-,008
,019
Attacker
-,108*** - ,077
,019
-,082*
-,064
,024
,044
,032
,018
Rows
,035
,018
,009
,007
,022
,019
,013
,018
Evil Intent
Power
,027
Criminalisable Space
Obstruction
,247***
,162
,017
,215***
,144
,022
,219***
,138
,018
Safe Route
,094***
,064
,017
,136***
,092
,021
,112
,072
,018
R-Square (adj.)
43,0
Constant included in models.

* = p<.05
** = p<.01
*** = p<.001.
42,7
45,1
189
Table 9 shows that four components of fear of crime can be taken into consideration: gender, frequency of walking outside after dark, self-perception of a potential victim/target and the frequency of thinking about the possibility of obstruction in all three studied countries. Perception of ones physical ability, unemployment (retirement), perception of streets and woods as unsafe places, considering other people as jealous, ability of chasing of a potential assailant, and
making sure that a route taken is safe, are related to fear of crime in the Slovenian sample. Fear of crime, in addition to common variables, is related to the
educational level, unsafe streets, low level of trusting people, ability of chasing
of a potential assailant, and making sure that a route taken is safe. Fear of crime
in Bosnia is, in addition to common variables, related to the perception of ones
physical ability, financial resources, unemployment (retirement), local social
networks, shops unsafe and level of distrust.
We decided to control for gender because in all three countries gender differences showed to be significant. The reason for doing so are the results of previous studies on fear of crime (Umek and Meko 1998; Meko and Kovo 1999),
which implied that fear of crime is more severely expressed by women (also
Warr 1984.), especially because of patriarchal relations in certain societies and a
greater vulnerability of females in interpersonal, often not reported violent
crimes.
5. Discussion
Generally speaking, the results of the regression analysis imply higher level of
fear of crime in women, people who perceive themselves as less physically fit,
the unemployed, those who expose themselves by walking alone in dark and
those who perceive streets and woods as sources of danger. In addition sociodemographic variables, peoples fears appear to be related to perceptions of oneself relative to an attacker who approaches one outside of the confines of ones
home. Imagining oneself to be physically stronger than this person reduces fear.
These findings have implications for the original model, theoretical work in this
field and crime surveys.
The relation between variables estimation of health and fear of crime reveals
that persons with a high self-estimation of health have lower level of fear (stands
for all three samples, the difference is statistically significant and noticeable).
Slovenian participants also think that a potential perpetrator could be quite dangerous and they are not prone to take risks. This is not valid for other two samples. Relationship between fear of crime and health leads us to the conclusion
that persons with a high estimation of health are more self-protective, not only
considering their own health but also in avoiding any risk behaviour.
190
Difference between male and female respondents is revealed in the variables,

which are related to a more direct personal threat (i.e. particularly in some of the
components of the social psychological model: When Im on my way home, I
sometimes imagine that someone would obstruct my path, When I have to go out
somewhere, I make sure that I take a safe route and I think Im capable of chasing
of a potential assailant.). The results for female respondents show a higher level of
fear, the differences come up to 25 % (the proportion of persons with a highest level
of fear within gender) and they are all statistically significant in all three samples.
Results reveal the fact that the highest discrepancy between male and female respondents appears at the Slovenian sample, especially at the variable power attacker (I think Im capable of chasing of a potential assailant).
The results of the social psychological model lead us to the conclusion that
about one third of respondents in all three countries express higher level of fear.
Findings are in accordance with results abroad (Houghes 1998). The highest percentage is noticed among participants from Bosnia and Herzegovina (39 % in
average, 35 % in Slovenia and 32 % in Croatia). In Slovenia there is no clear picture about differences among different age groups. It seems that the highest proportion of persons with higher level of fear is among adolescents (small difference and statistically not significant). In Croatia and especially in Bosnia and
Herzegovina quite different situation is detected higher level of fear is noticed
mainly among older participants (55 years of age and above). These findings are
expected and they can be explained by cognitive processes, which are typical for
age groups mentioned the young and the elderly are aware that they are sometimes exposed to some kind of criminal threat (or they believe so). Adolescents
stay out late in the night and seniors are aware of their diminished physical ability to resist a possible perpetrator.
Social psychological model reveals that fear of crime is lower among participants with a higher socio-economic status and a higher education level. In Slovenian sample the correlation is quite low but significant ( = 0,1 to 0,2; p < .02).
In other two samples the correlation coefficient is lower (around 0,01) and
mostly statistically non significant. It is assumed that well situated participants
have no difficulties to ensure themselves satisfactory level of safety. They can
choose a safe living environment and provide themselves with a better technical
equipment or security service.
In all three samples the most threatening circumstances are presented in vignettes
A Car (68 78 % of participants assumes feelings of being jeopardized), To a
party (28 42 %) and A Doorbell (48 57 %). A plausible reason for these results is the fact that these events are quite possible to happen in real life.
191
The results of the study imply that specific social circumstances impact the
level of insecurity and crime related fears of the respondents. It is often stated
that crime and recently fear of crime are related to the same factors of poverty,
post-war conditions (land mines, consequences of post-traumatic stress syndrome), social disorganisation, bad housing, migration, discrimination, organised
crime, etc. Fear of crime issues can be closely connected to the existential fear
and fear of loosing a certain level of the quality of life and ones perception of a
potential assailant is often related to a concept of stranger danger and the negligence of danger within (e.g. domestic violence).
5.1
Concluding remarks
The present examinations were conducted by different researchers in different

countries. Despite this example of researcher, data set and population triangulation, discernable patterns do emerge which do not differ much from the studies in
Western Europe. This is understandable because of globalisation and some significant crimes which have an international importance (e.g. 9/11 in the USA,
which caused a new wave of efforts for a safer world).
Fear of crime is insofar a very important concept as there is a strong influence
on political decisions. We find more and more punitiveness in western countries
(see Roberts et al. 2003.) in the last years, often based on the discussion of fear of
crime. As Farrall et al. (1997) could show for Great Britain and Kury et al.
(2004) for Germany the regular measurement of fear of crime overestimates the
fear level that is we need a better operationalised and more valid measurement of
this important concept.
Fear of crime is an important element of penal populism in a way that politicians recognise expressed fear of crime generated by the media, well publicized
cases and changes in some legal norms (often without sufficient expert consultation and pressure of so called moral entrepreneurs). Their declared goal is citizens protection from victimization but the hidden goal is quite frequently the
political gain (election votes).
At a theoretical level, this suggests to us that more attention needs to be given
to the relationship between those that report being fearful and those that the fearful imagine will attack them. Or, in other words, theories, which aim to account
for fear levels need to address the relationship between an individuals beliefs
and stereotypes about crime and their levels of anxiety.
Comparison of the three cultures implies that measuring of fear of crime is
also related to other issues, which would deserve a special research attention (i.e.
masculinity, self-attribution of being fearless or fearful, resignation, self-esteem,
defence mechanisms of denial/projection/rationalisation) and quality of life is-
192
sues (i.e. living in a people friendly environment) in a contemporary society

characterised by uncertainty and culture of fear (Furedi 1998). The modern fears
of being excluded, ugly, unhealthy, fat, unfit, financially incapable, rejected at
work by ones chief, etc. presented by Furedi (ibid.) also call for studying their
relation to the fear of crime. The assessments of sources of fear of crime also imply that contemporary concepts of crime and criminality issues are related to
stereotypes and prejudice, which should be overcome.
Fear of crime issues are a political matter because of fearful people and politicians who abuse fear of immigrants, the poor, the southerners, the easterners and
other marginalised and stigmatised social groups and attribute the criminality to
them. The media also create ones social construction of crime related reality.
The medias influence on ones understanding of attribution of criminality to
specific individuals, circumstances and possible victimisation deserve additional
research attention.
The present study implies the fact that higher levels of wealth, health and quality of life are related to less fear of crime. Care for peoples basic needs and quality of life (especially health) are issues closely related to the reduction of all
kinds of existential fears, among which fear of crime manifests ones attribution
of dangers which threaten his or her wellbeing. These issues are, generally speaking, related to the issues of contemporary thoughts on safety and security in our
communities. Meko (2004: 740) found out that representatives of local communities in Slovenia attributed the greater importance to the following factors of
crime prevention, which supposedly should solve their crime and fear of crime
problems. The factors of preventative measures are ranked as follows: (1) social
prevention (training for parents, solving social problems, student friendly school,
competent school teachers, leisure activities available, responsibility of business
owners), (2) self-protective measures (availability of information on crime prevention, self-protective measures), (3) formal social control (police punishment
of criminals) and (4) private social control (private security guards at school,
private security industry).
It is evident that the respondents in the present and Mekos (2004) study
call for something what we know as a welfare state, but social reality in the
most of countries of contemporary western civilisation reinforce liberal capitalist
model, which works on the free market principles and especially competition,
which pushes all incapable to the social margins where they can easily experience all kinds of fears and real victimisations. So called countries in transition
are facing turbulent changes and it is difficult to manage such a situation without
learning new survival skills.
193
Appendix
The Vignettes employed:

Doorbell
One evening youre at home on your own. Its late. The doorbell rings, but youre not
expecting anyone.
Car
One evening you go out to put the dustbin out. A short way up the street you see two
men walking around a parked car. When they see you looking at them, they begin to
walk toward you.
To a party
Youve been invited to a party in a neighbourhood you dont really know. Early that
evening you set out by bus. When you get off you still have a long way to walk. Suddenly you notice that youve lost your way. A group of youths is following you and
begins to make unpleasant remarks at you.
Bus stop
One afternoon youre standing at the bus stop nearest home, when a group of 15-16
year olds comes along. They begin kicking the bus stop and daubing graffiti on the
bus shelter.
Telephone
Youve going out one evening. Youre ready and just about to leave when the phone
rings. You answer, giving your name. But at the other end you hear only irregular
breathing. You ask whos there. They hang up.
Caf
Youre travelling through a town where youve never been before. You have to ring
home to say youll be late getting back. Because you cant find a telephone box, you go
into a caf to ring from there. It turns out to be where a group of bikers meets.
Each of these scenarios was followed by the Unsafety Question How unsafe would
you feel in such a situation? [response codes: 1 = very unsafe, 2 = quite unsafe, 3 =
dont know, 4 = quite safe, 5 = very safe]. It is the summation of the answer to this
question after these six scenarios that is used as the measure of the fear of crime (i.e. as
the dependent variable).
194
The Components of the Social Psychological Model

Could you tell me whether you agree or disagree with the following statements
[response codes: 1 = agree strongly,
2 = agree,
3 = dont know,
4 = disagree,
5 = disagree strongly].
ATTRACTIVITY
Target:
I think that people who are up to no good are likely to target especially on me and my
possessions.
Jealousy:
I think that people are jealous of me.
POWER
Attacker:
I think Im capable of chasing of a potential assailant.
Rows:
I generally stay clear of rows.
EVIL INTENT
Trust:
I generally trust strangers.
Distrust:
I distrust particular people in my surroundings.
Obstruction: When Im on my way home, I sometimes imagine that someone would
obstruct my path.
Safe Route:
When I have to go out somewhere, I make sure that I take a safe route.
195
References
Bennett, T. (1990). Tackling Fear of Crime, Home Office Research Bulletin, 28, 14-19.
Chambers, G. & Tombs, J. (1984). The British Crime Survey Scotland. Scottish Office
Central Research Unit.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1996). Fear of Crime: Conceptual
Development. Field Testing and Empirical Confirmation: Final Report to the ESRC.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1999). Angry or Afraid? Recalibrating The Fear of Crime, International Review of Victimology.
Journal of Criminology 37, 658-679.
Farrall, S., Ditton, J., Bannister, J. & Gilchrist, E. (2000). Social Psychology and the
Fear of Crime: Re-Examining a Speculative Model. British Journal of Criminology.
Furedi, F. (1998). Culture of Fear: Risk-taking and the Morality of Low Expectations.
Cassel: London.
Grusczynska, B. (2001). Victimization, Crime and Social Factors in Post-Socialist
Countries. In: Kury, H. (Ed.), International Comparison of Crime and Victimization.
The ICVS. Willowdale: de Sitter Publications
Hale, C. (1992). Fear of Crime: A Review of The Literature. Report Prepared for the
Metropolitan Police Service Working Party on Fear of Crime, University of Kent,
Canterbury.
(1996). Fear of Crime: A Review of the Literature. International Review of Victimology, 4, 79-150.
Hatalak, O., Del Frate, A. & Zveki, U., (Eds.) (1998). The International Crime Victim
Survey in Countries in Transition. Rome: UNICRI.
Houghes, G. (1998). Understanding Crime Prevention Social Control, Risk and Late
Modernity. Buckingham: Open University Press.
Ivii, I., aki, V. & Franc, R. (2004). Fear of crime in Croatia: its extent and predictors. Presentation at the ECS conference, Amsterdam, August. 2004.
Kury, H. (Ed.)(2001). International Comparison of Crime and Victimization. The ICVS.
Willowdale: de Sitter Publications.
Meko, G. (2004). Local Safety Councils in Slovenia A Story on Attempts to Make
Communities Responsible for Solving Crime and Safety Problems. In: Meko, G.,
Pagon, M., Dobovek, B. (Eds.), Policing in Central and Eastern Europe Dilemmas
of Contemporary Criminal Justice. Ljubljana: Faculty of Criminal Justice.
Meko, G. & Farrall, S. (1999). The social psychology of the fear of crime: a comparison of Slovenian, Scottish and Dutch local crime surveys. Hrvatska revija za rehabilitacijska istraivanja, 35, No. 2, 141-149.
Meko, G. & Kovo, I. (1999). Strah pred kriminaliteto v Sloveniji in Hrvaki. Varstvoslovje,
1, 2, 26-30.
196
Meko, G. & Areh, I. (2003). Strah pred kriminaliteto v urbanih okoljih. Revija za
kriminalistiko in kriminologijo, 54, 3, 256-264.
Meko, G. & Pavlovi, Z. (1998). Strah pred kriminaliteto izzivi za raziskovanje.
Revija za kriminalistiko in kriminologijo, 49, 4, 346-353.
Meko, G. & Umek, P. (1999). Znailnosti socialnopsiholokega in demografskega
modela strahu pred kriminaliteto. Revija za kriminalistiko in kriminologijo, 50, 2.
Muratbegopbvi, E. (2003). Policija u prevenciji kriminaliteta na razini lokalne zajednice Kanton Sarajevo, studija sluaja. Sarajevo: Fakultet kriminalistikih nauka.
Pavlovi, Z. (1998). Mednarodna anketa o viktimizaciji 1997. Intitut za kriminologijo
pri Pravni fakulteti v Ljubljani, raziskava t. 118, Ljubljana.
Roberts, J.V., Stalans, L.J., Indermaur, D. & Hough, M. (2003). Penal Populism and
Public Opinion. Lessons from Five Countries. Oxford: Oxford University Press.
eparovi, Z. & Turkovi, K. (1998). Zagreb (Croatia). In: Hatalak, O., Del Frate, A.,
Zveki, U. (Eds.), The International Crime Victim Survey in Countries in Transition.
Rome: UNICRI.
Umek, P. (2004). Viktimizacijska tudija 2001. In: Meko, G. (Ed.), Prepreevanje kriminalitete teorija, praksa in dileme. Ljubljana: Intitut za kriminologijo.
Van der Wurff, A., Stringer, P. & Timmer, F. (1986). Feelings of unsafety in residential
Surroundings. In: Canter, D., Jesuino, C. Soczka, L. & Stephenson, G. (Eds.), Environmental Social Psychology, Dordrecht: Kluwer.
Van der Wurff, A., Van Staalduinen, L. & Stringer, P. (1989). Fear of Crime in Residential Environments: Testing a Social Psychological Model. Journal of Social Psychology, 129 (2), 141-160.
Viktimizacijska tudija (2001). Ljubljana, Urad za Statistiko Republike Slovenije (Bureau of Statistics of the Republic of Slovenia). www.stat.si. Accessed December 13,
2004.
Warr, M. (1984). Fear of Victimisation: Why are Women and the elderly more afraid?
Social Science Quarterly, 65, 681-702.
Young, J. (1988). Risks of crime and Fear of crime: A Realist Critique of Survey based
Assumptions. In: Maguire, M., Ponting, J. (Eds.), Victims of Crime: A New Deal,
OUP, Milton Keynes.
(1992). Ten Points of Realism. In: Young, J., Matthews, R. (Eds.), Rethinking Criminology: The Realist Debate. Sage: London.
Zveki, U. (1998). Criminal Victimisation in Countries in Transition. Rome: UNICRI.
197
PUNITIVITY
198
199
Disowning Our Shadow:

A Psychoanalytic Approach to Understanding
Punitive Public Attitudes1
SHADD MARUNA, AMANDA MATRAVERS

AND ANNA KING
Abstract
Punitive public attitudes cannot be easily explained by pointing to instrumental
concerns (e.g., fear of crime, personal victimization, or real or perceived levels of
crime). Instead, numerous observers have suggested that public punitiveness is
more a symptom of free-floating anxieties and insecurities resulting from social
change than it is a rational response to crime problems. We argue that these public concerns might be better understood by drawing on the insights of psychoanalytic theory, and we review relevant theoretical work to that effect.
This chapter first appeared in Deviant Behavior, 25, 277-299. Reprinted with permission.
The authors wish to thank Bruce Arrigo, David Garland, Loraine Gelsthorpe, Nicole Hahn
Rafter, Claire Valier, and Veronique Voruz for their comments and criticisms.
200
Shadd Maruna, Amanda Matravers and Anna King
As criminal justice policies are increasingly said to be advanced to win votes

without much regard for their effects (Roberts, Stalans, Indermauer & Hough
2003), research that attempts to better understand the nature and origin of punitive public attitudes is badly needed (see e.g., Tyler & Boeckmann 1997). Instead, there is a tendency in some criminal justice research to take the existence
of the punitive public almost for granted and assume that public hostility to
offenders is natural or innate 2 . In Punishment & Democracy, for instance, Zimring, Hawkins & Kamin (2001) argue that public punitiveness has probably been
a constant in US history (what changed was the level of victimization and opportunities for populist political decision-making). Like most other observers, the
authors therefore seem to accept popular punitivism as a fact of nature and
therefore offer no real explanation for it (Greenberg 2002: 246).
Yet, comparative and historical research contradicts this image of punitive attitudes as a constant. Research suggests that there are significant cross-cultural
differences in punitive attitudes (Braithwaite 1989; Karstedt 2003; Mayhew &
Van Kesteren 2002). That is, crimes of similar seriousness and frequency do not
seem to provoke parallel public reactions across different cultures and historical
epochs. Sutherland and Cressey (1974: 336) write:
There has been no constant desire to make all criminals suffer and the
system used for inflicting suffering on (criminals) has changed from time to
time. The punitive reaction to lawbreaking has not been present in all societies (in fact it) varies from time to time even within a given society. A
theory which precisely explains or accounts for all of these variations has
not been developed.
Additionally, there seem to be important individual-level differences in punitive

attitudes within a single cultural context (see the survey and opinion-poll analyses in Cullen, Cullen & Wozniak 1988; Roberts and Hough 2002; Warr 1995).
Apparently, the execution of a Death Row inmate can evoke disgust and tears
from one person while triggering celebration and cheers from another.
These differences in emotive reactions to crime and punishment may hold
some important clues to better understanding the psychosocial dynamics of
vengeance and forgiveness. They also may have implications for criminal justice
reform efforts. Garland (1990: 62), for instance, argues that reformers tendency
to ignore public punitiveness or dismiss these views as a form of false consciousness is counterproductive. If such sentiments do exist, and give support
2
Indeed, this argument is made explicit in theoretical work in evolutionary psychology (e.g.,
Fehr and Gachter, 2002), in which support for the punishment of wrongdoers is considered
an almost universal human trait, crucial to the evolution of civilization.
A Psychoanalytic Approach to Understanding Punitive Public Attitudes
201
to current penal practices, then penal reformers will have to address themselves
directly to popular feelings if they intend to produce real change.
According to Tyler & Boeckmann (1997), explanations for punitive attitudes
broadly fall under two main categories: instrumental and expressive explanations. Instrumental arguments suggest that punitive attitudes are likely to result
when individuals feel a personal threat to themselves or their communities. People are fed up. Like Charles Bronson in Death Wish, they are normal, goodnatured individuals who get pushed too far by the crime and disorder around
them. Hence, punitive attitudes should be more prevalent in areas of high-crime,
among individuals who have been victims of crime and who fear repeat victimization.
A sizable and consistent body of research fails to support the Death Wish hypotheses (Cullen et al. 1985; Baron & Hartnagel 1996; Hough & Moxon 1985).
For instance, public opinion polling does show a relationship between punitiveness and the fear of crime (Taylor, Schepple & Stinchcombe 1979), but this
seems to be a very modest correlation at best (Roberts & Stalans 1997). Additionally, fear of crime does not seem to have a measurable relationship to views
on capital punishment (Warr 1995) or support for other specific policies like
Three Strikes legislation (Tyler & Boeckmann 1997). Personal victimization
experiences and perceptions of crime salience in ones residential area do not
relate consistently to punitive attitudes (Langworthy & Whitehead 1986; Flanagan, McGarrell & Brown 1985). Finally, at the aggregate level, recorded rates of
crime do not seem to predict public demands for punishment in the way instrumental theories would predict (Wilkins 1991; Mayhew & Kesteren 2002). These
findings lead Tyler & Boeckmann (1997: 252) to conclude that crime-related
concerns are the least important factor in predicting punitive attitudes and suggest that the image of the citizen as supporting punitive public policies because
of fear of crime is inaccurate.
An alternative explanation to the instrumental view is that punitive attitudes serve
an expressive or symbolic function which is of course to beg the question expressive of what? (Garland 2001). There is a rich tradition of theoretical work in
this area (e.g., Durkheim 1933; Mead 1918) which continues today with work such
as David Garlands Culture of Control. The best known explanations for contemporary punitiveness revolve around ontological insecurity or a widespread sense of
anxiety driven by the disembedding processes of modernity as former social
certainties are eroded. (Bottoms 1995, p. 785; see also Vaughan 2002; Garland
2001; Gaubatz 1995). For instance, Bauman (2000) points to the profound anxiety
and insecurity produced by the flexibility of the labor market under the deregulated
202
capitalism favored by neo-liberal states. Indeed, the relationship between economic

insecurity and scapegoating behavior is well known. For instance, in their now classic study, Hovland & Sears (1939) found that the frequency of lynching in the southern US states was negatively correlated with the price of cotton. When farmers suffered the most frustration, they were most likely to redirect their anger on Black men
accused of crimes (these findings were confirmed by Hepworth & Wests recent reexamination of these data).
These accounts provide a fascinating, alternative understanding for what appears to be a rising tide of punitive attitudes among the public. Yet, Voruz (forthcoming: 246) has pointed to the conspicuous absence of a theory of the subject including all of its anxieties, hatreds and aggressivities in these largely
macro-sociological accounts of late modern changes in penal politics (see Hollway & Jefferson 1997 for a similar critique of the fear of crime literature). That
is, whereas the rational actor emerges clearly out of the instrumental perspective, it is unclear what motivational and personality framework is thought to be at
work in the anxiety-driven expressive explanations of punitive attitudes.
One natural area to explore in any effort to understand human irrationalities
and insecurities is the time-tested (if much maligned 3 ) tradition from Freuds
Totem & Taboo 4 to the Frankfurt School of applying psychoanalytic concepts
to cultural analysis. Drawing on the writings of Jung and Freud as well as more
recent psychoanalytic or Freudian-influenced interpretations of punishment and
punitiveness (esp. Garland 1990; 2001), we will argue that unconscious-level,
ego defense mechanisms may help to explain differences in punitiveness among
the public. Like our predecessors in this terrain (most notably Adorno et al.
1950), our approach to psychoanalytic theory, here, will be entirely uncritical5 .
3
Methodologically, the Freudian interpretive framework has been described as something

like a blank check (Smith, 1971, p. 192). After all, even the individual being analysed has
no great access to the unconscious without the analysts help. Freud (1935/1990, pp. 49-50)
writes: As far as the patient is concerned this sense of guilt is dumb; it does not tell him he
is guilty; he does not feel guilty, he feels ill. Therefore, only analysts can diagnose things
like unconcious guilt and they are likely to see evidence of these invisible processes everywhere, even under contradictory circumstances. (If a patient claims to feel guilty, it may
be evidence of guilt; if they claim not to feel guilty, it may also evidence of this underlying
guilt being projected outward).
Published in 1912, Totem & Taboo was the first of Freuds five works with social and cultural themes. This essay will focus primarily on the fourth of these, Civilisation & Its Discontents (1930).
Additionally, we should warn that ours is a selective reading of the massive and highly diverse psychoanalytic canon. It is also an amateur one; we are not trained psychoanalysts, nor
do we consider our scholarly orientations to be psychoanalytic in any general sense. To
compensate for this, our essay over-relies on quotations from original and secondary
203
We are of course well aware of (and indeed agree with in large part) the many
criticisms of Freudian social analysis. These failings (e.g., failure to meet Popperian standards for falsification; lack of conceptual limits; masculine bias; tendency to be just a bit weird and pervy at times) have been repeated so often that
we need not rehash them again here. Of more value, we think, is a fair hearing
for psychoanalytic voices on the question of punishment and society.
After all, for all its problems, psychoanalytic theory has stood up remarkably
well to the tests of time. Indeed, a recent state-of-art review of laboratory-based,
experimental research has found substantial support for many (but not all) of the
processes of defense Freud outlined amounting to a rather impressive positive
testimony to Freuds seminal theorizing (Baumeister, Dale & Sommer 1998:
1083). Indeed, there seems to be something of a psychoanalytic comeback6 occurring in criminology. Interestingly, this is more evident in the realm of punishment theory (e.g., Anderson 2000) than in criminological discussions of delinquent etiology, where psychoanalytic theory once held considerable sway (e.g.,
Aichhorn 1925; Redl & Wineman 1951). Writing in 1990, Garland noted that
Such psychoanalytical literature as there is on [punishment] is often crude and
unpersuasive (1990: 240). Since that time, psychoanalytic concepts have been
usefully employed to shed light on the discussions of the fear of crime (Hollway & Jefferson 1997), the scene of punishment (Valier 2000), public views
on crime and justice (Duncan 1996), corporal punishment (Gruen 1999) and vigilante justice movements (Evans 2003). Psychoanalysis is unlikely to re-ascend
the pedestal of paradigmatic privilege it occupied in the early years of the 20th
Century, but it seems to be recovering from the over-reaction against it by contemporary social science.
On the other hand, psychoanalytic interpretations should not be used as a sort
of ideological weapon. That is, beginning at least with Adorno and colleagues
(1950) research on the authoritarian personality Freudian psychology has been
invoked almost entirely to account for views with which researchers disagree.
Although authoritarianism is held up to psychoanalytic scrutiny, for instance,
liberalism is praised as rational realism (see esp. Martin 2001). This misuse of
psychoanalytic theory has led Tetlock (1994, p. 526) and others to ask whether
such political psychology is really scientific or whether it simply seeks to advance certain moral or political causes by stigmatizing groups with whom we
sources, but we are still in peril of playing a little fast and loose with some incompletely digested Freudian concepts.
The efficacy of Melanie Kleins (1985) psychoanalytic theory, in particular, for such understandings has been argued for and demonstrated on several occasions in contemporary
criminological work (e.g. Hollway and Jefferson, 2000; Valier, 2000).
204
disagree?. To be perfectly balanced, therefore, our analysis should include also

seek to account for highly non-punitive attitudes among the public from a psychoanalytic perspective. Certainly, it could be argued that individuals harboring
such permissive, liberal or soft-hearted views have equally compelling unconscious reasons for wanting to believe that deep down even the worst offenders
are basically good. Sympathy and concern for offenders is no more right or
wrong than punitiveness and seems, from most evidence, to be equally a product of both rational thinking and emotive impulses. This analysis will have to
wait, however, due to space constraints.
Punitiveness as shadow projection

The basic elements of Freuds psychic system (e.g., the ego, the id and the superego) are well known and need not be reviewed here in depth.7 These elements
serve to illustrate the conflict between the individuals search for satisfaction and
the demands of the external world. The ego is rational and externally directed, in
contrast to the id, which is instinctual, primitive, chaotic and unconscious. Although derived from the id, the ego does not share its instinctive pursuit of pleasure (or, more accurately, avoidance of displeasure); Freud regards the ego rather
as a frontage or faade for the id. The ego tends towards synthesis, mediating
between the claims of the id and the objections of the external world. In order to
preserve its adaptability to the external world, the ego must tame the demands of
the id, repressing desires that cannot be satisfied, either temporarily or permanently. The third element in Freuds model of the mind is the super-ego, an
agency that watches over the ego, functioning as a conscience that alerts the ego
when it falls short of the ideal. The ego must then deal with conflict between the
competing demands of the id, the super-ego and the external world. It is here that
the defence mechanisms (e.g., repression, projection, denial and sublimation)
come into play, allowing the ego to cope with the pressures upon it.
In Instincts and Their Vicissitudes (1915: 14), Freud explains how the ego,
acting under the dominance of the pleasure principle, responds to uncomfortable
internal stimuli by separating off a part of its own self, which it projects into the
external world and experiences as hostility. The idea of projection is explored
further in Beyond the Pleasure Principle (1920). According to Freud, the individuals ego defense system receives stimuli or excitations from within as well
as from the external world. However, while it is provided with a protective shield
7
In The Question of Lay Analysis (1926), Freud provides a useful introduction to his theory
by borrowing the discursive technique generally associated with Plato to explain the tenets
of psychoanalytic practice.
205
against stimuli from without, the system has no protection against excitations
from within. In order to deal with internal excitations that produce too much displeasure therefore, such stimuli are treated as though they are acting from the
outside. This allows the system to utilize the protective shield in its defense. The
British Psychoanalyst Melanie Klein describes this as the splitting of the ego
where parts of the self that are feared as bad are split off through projection and
usually identified as belonging to an outside object (or person) (Hollway & Jefferson 2000: 20).
In Jungian terms, this sort of scapegoating is a particular expression of general
problem of shadow-projection or denying the shadow (see esp. Perera 1986).
The shadow is an unconscious part of the personality that the conscious ego rejects or ignores. Ego and shadow, although separate, are thought to be inextricably linked together in much the same way that thought and feeling are. The ego,
however, is in perpetual conflict with the shadow, in what Jung refers to as the
battle for deliverance. (Henderson 1964: 110). According to the Jungian analyst
Joseph Henderson (1964):
Wherever Persona is, Shadow is also. Wherever good is, is evil. We first
know the shadow as the personal unconscious in all that we abhor, deny and
repress: power, greed, cruel and murderous thoughts, unacceptable impulses, morally and ethically wrong actions. This idea of projection raises an
interesting point. It means that the shadow stuff isnt out there at all; it is
really in here; that is inside us. Shadow projections have a fateful attraction to us. It seems that we have discovered where the bad stuff really is:
We have found the beast, the demon, the bad guy. But does Evil really
exist, or is what we see as evil all merely projection of our own shadow
side? Jung would say that there really is such a thing as evil, but that most of
what we see as evil, particularly collectively, is shadow projection.
In Punishment & Modern Society (1990: 240) and more recently in the subtext of
Culture of Control (2001: 253, n. 71), David Garland argues that punitive attitudes against offenders (constructed as they are as alien others) could possibly
be understood in this way. In a play on words on the well-known Freudian idea
that some criminals offend from a sense of guilt that is, they commit crimes
because unconsciously they want to be punished for something Garland develops the notion of punishers from a sense of guilt. This is the idea that some
punitiveness among the public is motivated more by a sense of self-cleansing
than social control.
The most vehement punishments are reserved for those guilty of child
abuse, illegal drug use, or sexual violence precisely those areas in which
mainstream social and cultural norms have undergone greatest change and
206
where middle class ambivalence and guilt are at their most intense (2001:
195-196).
In the subtext, however, Garland offers an even more provocative possibility:

Could it be that the extraordinary public fears and hostilities in respect of
certain crimes against children stem from the residual guilt and ambivalence
that families feel about their choices and the vulnerabilities that they seem
to cause? If so, the pedophile and the drug dealer are screens upon which we
project our guilt as well as our anxieties (2001: 263-264, n. 64).
Rather than recognizing, in line with the narcissism of small differences (Freud
1930/2003: 50) that those we reject closely resemble ourselves, we fall back on
archetypes of Us and Them. Yet, as Jacqueline Rose (1993: 53) argues, The
guilt of the criminal establishes the innocence of the society; but, like all oppositions, it risks a potential identification between its terms (cited in Valier 2000:
388).
Understanding the shadow

Although the metaphor of shadow projection is useful for understanding the
function that punitive attitudes might serve, the question still remains as to what
this shadow is that is being projected. Extant psychoanalytic theory on punitiveness seems to suggest several different possibilities, including but not limited to
the following, each of which will be discussed in turn:
x a sense of inferiority or shame at our own insignificance
x guilt over our own role in the creation of the crime problem
x sublimated jealousy and admiration for the criminals exploits
x sadistic impulses to humiliate others and
x guilt regarding our own sexual desires.
Inferiority and Shame

Adlers much-abused notion of the inferiority complex, suggests that aggression is often the result of concerns that one lacks in some desired quality (e.g.,
masculinity, power). Along these lines, James Gilligan (1996) argues that punishment is symbolic castration (or emasculation), hence a deflection of shame
from one person to another (p. 150). Gilligan answers the question of What
emotional gratification are people seeking when they advocate punishing other
people harshly? (p. 182) by arguing that the motives behind crime and punishment are emotionally and symbolically identical. According to Gilligan, The
greatest fear in each instance is that of being shamed or laughed at (p. 182). Punitiveness, like criminality then, is a means of warding off underlying feelings
207
of personal insignificance or worthlessness (p. 183). The case of Bernard Goetz,

the Subway Vigilante is often explained in these terms:
Beyond practical danger, Goetz was intent on not suffering further humiliation not simply the humiliations that muggers could inflict, but the humiliation of his own fear, of continuing in the world with the common, cowardly wish to believe that such things would not happen to him (Katz 1988:
323; see also Duncan 1996: 79).
The humiliation that many of us feel about our own powerlessness might explain
the widespread public support and sympathy that Goetz received.
Scapegoating and guilt

Additional clues as to the nature of our cultural shadow might be found by examining more closely the societal choice of scapegoats. Offenders who are singled
out for this projection typically fall into two broad stereotypes: the gangster
and the sexual predator. The gangster archetype might be described as being
comprised of an underclass group: young, poor, male, gang-affiliated, drugusing, drug-selling, weapon-carrying minority member (or some combination of
3 of these at least).
As Garland notes, the states denial in relation to this category of offender
seems to stem from its need to defend itself from the knowledge that its economic choices have marginalized and excluded a section of the citizen body:
These middle class fears were, no doubt, overlaid by an element of guilt
and bad conscience. This was, after all, the society that these classes had
chosen By opting for the market, getting the state off their backs, and
freeing up individuals and investments they had undone the delicate web
of solidarity and community Choosing freedom comes at a cost, and all
too often it is the poor and the powerless who are made to pay (2001: 156157).
The middle class is hardly unaware of the promotion of criminality that is inevitable and inherent in processes of ghettoization, prisonization, stigmatization and
the social exclusion of the poor. Such guilty knowledge is not easy to live with,
and therefore might need to be projected through a process of responsiblilizing
offenders. The idea that our offenders are only a stand-in or scapegoat population symptoms of an unequal and unfair society that we have partially created
and fully benefited from may simply be too uncomfortable for most of us to
accept.
This might help to explain why forcing confessions out of wrong-doers has
been a nearly universal characteristic of justice systems worldwide (see Foucault
1988: 42). If we are to punish (or arrest, convict, study, or classify) a person as
208
an offender, the individual needs to be made responsible for the offence. In the
face of a body of social science research that seems to exculpate offending behavior by shifting blame to parents, schools, communities, capitalism and globalization, there is no small comfort in at least having the scapegoat him or herself
taking full responsibility for the crime. Parole board member David Tidmarsh
(1999: 50), for instance, writes:
I believe that Parole Board members are fully aware of the social antecedents of crime but, faced with individuals and the actual harm they have
done, do not find these concepts helpful and go along with the policy of the
prison system, based on persuading offenders to accept responsibility for
their offences.
Psychoanalytic theorists, rather predictably for Freudians, often draw an analogy

between scapegoats and the rejection of feces (see Duncan 1996; Neumann 1973:
128).
What is rejected is initially acceptable to the infant and its caretaker. Excrement is part of what was once nourishment necessary for survival, the inevitable by-product of life and body-ego health, expelled proudly by the
child as an expression of the creative capacity to form material life. Yet it
has traversed the bodys boundary and can be considered alien and dirty
(Perera 1986: 13).
Likewise, the creation of criminal outcasts is an inevitable by-product of neoliberal social policies when combined with a deeply rooted consumerist value
system (see especially Vaughan 2002) and subterranean values emphasizing individualism, machismo, and dominance over others (Matza & Sykes 1961). We
do our best to keep this waste out of view, but we cannot stop producing it (Douglas 1966/2002).
Envy and admiration

The idea that responses to crime evoke a wider range of emotions than those that
are publicly expressed has been explored by several commentators (see Menninger 1968; Katz 1988; Karstedt 2002). The complex and at times conflicting nature of penal emotions is also touched on by Garland (1990). Although he is careful to note that such explanations are inherently controversial and carry little
weight outside of individual case histories (p. 65), Garland outlines an account
of punishment psychology that moves beyond fascination to identification:
the punitive energy shown by some individuals towards criminals can
be interpreted as a guilty and masochistic response to their own tendency to
fantasize an identification with the criminals exploits (1990, p. 65).
209
In her at times remarkable psychoanalytic tract, Romantic Outlaws, Beloved

Prisons, Martha Grace Duncan draws a parallel with known cases of parents who
unconsciously encourage criminality in their children to gratify their own unacknowledged criminal impulses (1996: 66). In an altering gratification of the id
and the superego, the same parents who unwittingly turn their children into
criminals in this manner frequently go on to denounce their children to the authorities (p. 66). Duncan argues that, like such conflicted parents, the public
harbors highly ambivalent feelings toward criminality, involving both a pervasive admiration reflected in our seemingly endless appetite for media products
that claim to delve into the minds of serial killers and rapists and a simultaneous resistance to this admiration (see also Garland 1990: 65 and 239; Menninger 1968). Duncan sees punitiveness as a reaction-formation through which
feelings of attraction are channeled into their opposite loathing or disgust. She
suggests that this process of attraction and repulsion mirrors our unconscious
feelings regarding fecal matter which accounts for the ubiquity of the association
of filth with criminal offenders going back millennia and traversing cultures
and continents in a way that suggests an archetypal symbol (p. 187).
Sadism and Schadenfreude

Freuds thesis in Civilisation & Its Discontents rests on his contention that the
fundamental hostility of human beings towards one another constitutes a constant
threat to civilised society. Rather than loving his neighbour and reserving his
power to defend himself from harm, man is led by this ubiquitous aggressive
drive to seek opportunities to exploit and sadistically attack others:
Hence, their neighbour is not only a potential helper or sexual object, but
also someone who tempts them to take out their aggression on him, to take
possession of his goods, to humiliate him and cause him pain, to torture and
kill him (1930, p. 48).
The Frankfurt Schools Erich Fromm, one of the best-known public analysts,
renowned for marrying Freudian and Marxist concepts in his cultural critiques,
draws on this argument in his explanation for punitive public attitudes. In his two
recently re-discovered articles on the psychology of the criminal justice system
from the 1930s, Fromm seeks to answer the classic question of why society
would cling so tightly to a penal system that clearly is not effective at accomplishing its stated aims. One reason for this, he argues, is that punitiveness is
meant to provide the masses with a form of gratification for their sadistic impulses (see also Alexander & Staub 1928/1956; Mead 1964; Nietzsche 1887/
1989). Fromm (1930/2000) writes:
210
What is commonly termed a sense of justice (Gerechtigkeitsgefhl) or consciousness of justice (Rechtsbewusstsein) in the masses is nothing but an
expression of certain libidinal impulses of a sadistic or aggressive type
(p. 126).
Although Fromm (1970) theorizes that sadism is a natural impulse like Freuds
death instinct, he argues that sadistic impulses could be cultivated and magnified
through social processes. He writes,
Sadism is the great instinctual reservoir, to which one appeals when one
has no other and usually more costly satisfactions to offer the masses
(p. 141).
Moreover, it is understandable that the state might appeal to these impulses because they can be gratified in a manner that is harmless for the state (Fromm
1930/2000: 126). Punitive policies, then, can channel the publics natural aggression against the oppressive and ruling class onto the criminal in a way that
is clearly counter-revolutionary (Fromm 1931/2000: 146).
Friedrich Nietzsche also developed an account of what he perceived as a
greater tendency towards sadistic punishment among common men. For Nietzsche, pleasure in punishment was one manifestation of the sadistic impulses generated by the will to power over others (1887/1989). In punishing wrongdoers
(whether directly or indirectly), the lower classes experience a sadistic power
which is heightened by its rarity:
the enjoyment of violation will be the greater the lower the creditor
stands in the social order In punishing the debtor, the creditor participates in a right of the masters: at last he, too, may experience for once the
exalted sensation of being allowed to despise and mistreat someone as beneath him or at least, if the power and administration of punishment has
already passed to the authorities, to see him despised and mistreated
(1989, p. 501).
But Nietzsche too subscribes to the notion that sadistic aggression transcends
social barriers; the ruling classes are not less prone to taking pleasure in cruelty,
just more skilled at concealing it.
Sex, guilt and ambivalence

The process of shadow projection might be most clearly evident in regards to
attitudes regarding the second general category of scapegoats on the punitive
landscape: namely, sex offenders and pedophiles. Much punitive energy is dedicated to establishing the difference between sex offenders and ordinary people,
even between sex offenders and other violent offenders. In the UK and the US
alike, political rhetoric and the introduction of a range of increasingly punitive
211
sentences construct the sexually violent predator as a real and constant threat to
children (see Simon 2000; Home Office 2002).
However, the instantiation of the sexual predator as late modernitys most
suitable enemy involves quite a considerable degree of denial in particular, of
empirical evidence about sex offenders and offending. What we know about sex
offenders suggests that rather than being a discrete offender group, they bear significant resemblance to other offenders, and indeed the male public at large
(Grubin 1998). A significant amount of empirical evidence testifies to the shared
guilt that lies beneath punitive responses to sex offenders. Surveys note that sexual interest in children exists in non-convicted populations such as male university students as well as among pedophiles (Abel et al. 1987; Briere & Runtz
1989). Estimates of sexual assault, rape by previously-known assailants and rape
within the context of war identify sexual violence as a prevalent rather than an
aberrant behavior (see Simon 2003).
Of course the ultimate denial relates to ambivalence towards the sexuality of
our own children. For Freud, of course, the prohibition of incest is one of a number of repressive restrictions placed on men and women by civilization. However, in late modern civilization, child sexual abuse is a perennial moral concern,
and ambiguous feelings (fuelled, perhaps, by the eroticization of children in art
and the popular media) are projected at a safe distance onto stranger offenders
and pedophiles. Again, this involves the denial of empirical evidence which indicates that some 80% of child sexual abuse is carried out by offenders known to
their victims (Grubin 1998).
This theme is taken up by Jessica Evans in a recent paper about the responses
of people on Portsmouth, Englands Paulsgrove Estate to the presence in the
community of alleged pedophile offenders. In a psychoanalytic exploration of the
vigilante protests on the estate, Evans (2003: 177) asks, What might the [protestors] have shared unconsciously with the pedophile that was so terrible to acknowledge that it must be cast out as belonging to the pedophile alone? In other
words, as Herman Hesse wrote in Demian, If you hate a person, you hate something in him that is part of yourself. What isnt part of ourselves doesnt disturb
us (see also Adorno et al. 1950).
3.6
Summary and Shortcomings
The review above was intended to provide some flesh to the often heard argument that punitive attitudes are just a form of scapegoating. Still, almost none
of the different accounts above have received any empirical attention, and indeed
theoretical accounts based on unconscious defense mechanisms as these are, are
212
inherently difficult to falsify. Also, even though none of these different versions
of the scapegoat hypothesis is explicitly intended to pathologize punitiveness, the
fact that most of these accounts do not offer parallel explanations for nonpunitive attitudes certainly gives that impression. Not only is more research
therefore needed in this case, but also a more balanced research approach not tied
to any particular ideological agenda (Martin, 2001).
Conclusion: Owning our own shadow

In several recent reviews of punitive public attitudes (e.g., Roberts et al. 2003;
Roberts & Hough 2002), the authors have concluded with a section on how best
to work with, counter or respond to punitive attitudes through public education
campaigns, increased citizen involvement in criminal justice activities and the
like. This sort of discussion of direct research implications is usually the point at
which most psychoanalytic discussions of punishment tend to concede to their
critics. Martha Duncan (1996: 186), for instance, admits that trying to draw normative judgments and practical solutions from a psychoanalytic model is like
squeezing blood from a stone:
For whereas law is a system designed to transform gray into black and
white, psychoanalysis is a system designed to transform black and white
into gray. And a system that yields black-and-white results is better suited to
offer unequivocal proposals for action than a discipline that focuses on
complex idiosyncratic meanings.
Indeed, even in the heyday of Freudian psychology, the theory had little impact
on criminal justice outside of the world of offender treatment (see for example
Redl & Wineman 1951). Anderson (2000) writes: Even though much of the
educated public had embraced psychoanalytic concepts [by 1931], the legal system itself had shown little inclination to change from its older punishmentoriented approaches centering around deterrence and rehabilitation after punishment (p. 99). To some degree this is inherent in the nature of the psychoanalytic
enterprise: it is no coincidence that the most enduring legacy of psychoanalysis
has been in literary and arts criticism.
Freuds conclusion to Civilization and Its Discontents certainly does not provide much in the way of consolation, as he himself admits. His final question
raises the possibility that culture, like a repressed individual, may have become
neurotic (1990: 80). The origin of this neurosis might be the conflict between the
desire to punish and the denial of this desire based on an unconscious sense of
guilt. While Freud makes a cautious recommendation that society should be co-
213
erced into (perhaps interminable) therapy, he does not offer much of a strategy
for initiating this process.
In a way, however, in the psychoanalytic framework, the diagnosis itself is the
solution. Smith (1971: 193) writes:
The analyst detects illness; he determines that the illness is due to a sense
of guilt which the patient (individual or society) has hidden from itself; he
announces the means of relief (further analysis); and then proceeds, through
an elaborate process of discovery (prolonged analysis), to a precise determination of what has produced the sense of guilt. Making the sense of guilt
conscious will not wholly solve the problem of illness, but it makes the
problem (the political one) more amenable to solution.
Self-knowledge achieved through therapy, therefore, is not sufficient to guarantee change, but it is necessary (see also Cohen 2001).
A radical reading of the psychoanalytic literature would likely go further,
however. For instance, Jung famously called on society to own our own
shadow (Johnson 1993), arguing that we need to take on a sense of ownership
over the them that the us has created. Jungian psychologist Eric Neumann
(1969: 130) writes:
In contrast to scapegoat psychology, in which the individual eliminates his
own evil by projecting it onto the weaker brethren, [here] the exact opposite
is happening. The individual assumes personal responsibility for part of
the burden of the collective, and he decontaminates this evil by integrating it
into his own inner process of transformation. If the operation is successful,
it leads to an inner liberation of the collective, which in part at least is redeemed from this evil.
Transcending what Cohen (2001) calls the anachronistic ideals of public life,
fraternity, solidarity, universality and common citizenship (p. 290), this might
involve something more akin to Hannah Arendts (1945/1971) notion of organized guilt and universal responsibility.
This truth and reconciliation (or healing the split in Kleinian terms) may require
the heroic and conscious venturing into the otherworld places considered evil, archaic, terrible those very wilderness regions that were originally the scapegoatidentified individuals personal hell (Perera 1986: 102). Owning the shadow means
accepting lifes challenge to live both the good and the bad (Henderson 1964: 112).
This can be a painful thing to accept, of course Perera (1986) calls shadow ownership the sacrifice of the ideal of perfection for one of wholeness but Jungian theory assures us this is a fair trade. Jung (1964) argues:
If we could see our shadow (the dark side of our nature), we should be
immune to any moral and mental infection and insinuation. As matters now
214
stand, we lay ourselves open to every infection, because we are really doing
practically the same thing as they. Only we have the additional disadvantage
that we neither see nor want to understand what we ourselves are doing, under the cover of good manners (p. 73).
References
Abel, Gene G., Becker, Judith V., Mittelman, Mary, Cunningham-Rathner, Jerry, Rouleau, Joanne L. & Murphy, William D. (1987). Self-reported sex crimes of nonincarcerated paraphiliacs. Journal of Interpersonal Violence, 2, 3-35.
Adorno, Theodor W., Else Frenkel-Brunswick, Daniel J. Levinson & R. Nevitt Sanford
(1950). The Authoritarian Personality. New York: Harper.
Aichhorn, August (1925). Wayward Youth. New York: Viking Press.
Alexander, Franz & Hugo Staub (1928/1956). The Criminal, the Judge, and the Public.
Glencoe, Ill: Free Press.
Anderson, Kevin (2000). Erich Fromm and the Frankfurt School Critique of Criminal
Justice. In Kevin Anderson & Richard Quinney (Eds.) Erich Fromm and Critical
Criminology: Beyond the Punitive Society (pp. 83-119). Urbana, IL: University of Illinois Press.
Arendt, Hannah (1945/1971). Organized Guilt and Collective Responsibility. In Roger
W. Smith (Ed.) Guilt: Man and Society (pp. 255-267). Garden City, NY: Anchor
Books.
Baron, Stephen W. & Timothy F. Hartnagel (1996). Lock em up attitudes: toward
punishing juvenile offenders. Canadian Journal of Criminology 38, (2): 191-212.
Bauman, Zigmunt (2000). Social Issues of Law and Order. British Journal of Criminology, 40 205-221.
Baumeister, Roy F., Karen Dale & Kristin L. Sommer (1998). Freudian Defense
Mechanisms and Empirical Findings in Modern Social Psychology: Reaction Formation, Projection, Displacement, Undoing, Isolation, Sublimation, and Denial. Journal of Personality 66:1081-1124.
Bottoms, Anthony E. (1995). The politics and philosophy of sentencing. In C. Clarkson,
Chris & Rod Morgan (Eds.) The politics of sentencing. Oxford: Clarendon Press.
Braithwaite, John (1989). Crime, shame and reintegration. Cambridge, UK: Cambridge
University Press.
Briere, John & Runtz, Marsha (1989). University males sexual interest in children:
predicting potential indices of paedophilia in a non-forensic sample. Child Abuse
and Neglect, 13, 65-78.
Cohen, Stanley (2001). States of Denial. Cambridge: Polity.
Cullen, Francis T., Gregory A. Clark, John B. Cullen & Richard A. Mathers (1985).
Attribution, Salience, and Attitudes toward criminal sanctioning. Criminal Justice
and Behavior 12:305-331.
215
Cullen, Francis T., John B. Cullen & John F. Wozniak (1988). Is rehabilitation dead?
The myth of the punitive public. Journal of Criminal Justice 16 (4).
Douglas, Mary (1966/2002). Purity and Danger: An Analysis of Concepts of Pollution
and Taboo. London: Routledge Classics.
Durkheim, E. (1933). The Division of Labour in Society. NY.
Elias, Norbert (1939/1978). The Civilizing Process. Oxford: Oxford University Press.
Evans, Jessica (2003). Vigilance and Vigilantes: Thinking Psychoanalytically about
Anti-Pedophile Action. Theoretical Criminology, 7, 163-189.
Flanagan, Timothy J., Edmund McGarrell & Alan J. Lizotte (1989). Ideology and
crime control policy positions in a state legislature. Journal of Criminal Justice
17:87-101.
Franz von, M.-L. (1964). The Process of Individuation. In Carl G. Jung (Ed.) Man and
his Symbols. New York: Laurel.
Freud, Sigmund (1930/2002). Civilization and its Discontents. Hammondsworth: Penguin.
(1933/1991). Why War? In Civilization, Society and Religion. Hammondsworth:
Penguin.
(1935/1990). The Ego and the Id. New York: W. W. Norton.
(1950/1989). Totem and Taboo. New York: W. W. Norton.
(1921/1991). Group Psychology and the Analysis of the Ego. In Civilization, Society
and Religion (pp. 95-178). Hammondsworth: Penguin.
Fromm, Erich (1930/2000). The State as Educator: On the Psychology of Criminal
Justice. In Kevin Anderson & Richard Quinney (Eds.) Erich Fromm and Critical
Criminology: Beyond the Punitive Society (pp. 123-128). Urbana, IL: University of
Illinois Press.
(1931/2000). On the Psychology of the Criminal and the Punitive Society. In Kevin
Anderson & Richard Quinney (Eds.) Erich Fromm and Critical Criminology: Beyond
the Punitive Society (pp. 129-156). Urbana, IL: University of Illinois Press.
(1970). The Crisis of Psychoanalysis: Essays on Marx, Freud, and Social Psychology. New York: Holt, Rinehart & Winston.
Garland, David (1990). Punishment and Modern Society: A study in social theory. Chicago, IL: University of Chicago Press.
(2001). The Culture of Control. Chicago: University of Chicago.
Gaubatz, Kathyrn Taylor (1995). Crime In The Public Mind. Ann Arbor, MI: University of Michigan Press.
Gilligan, James (1996). Violence: Our Deadly Epidemic and Its Causes. New York:
Putnam.
Greenberg, David F. (2002). Striking out in democracy. Punishment and Society, 4(2),
237-252.
216
Grubin, Don (1998). Sex Offenders Against Children: Understanding the Risk. Police
Research Series, Paper 99. London: Home Office.
Gruen, Arno (1999). The need to punish: The political consequences of identifying with
the aggressor. Journal of Psychohistory 27(2): 136-154.
Henderson, Joseph L. (1964). Ancient Myths and Modern Man. In Carl G. Jung (Ed.)
Man and his Symbols. New York: Laurel.
Hepworth, Joseph T. & Stephen G. West (1988). Lynchings and the Economy: A TimeSeries Reanalysis of Hovland and Sears (1940). Journal of Personality and Social
Psychology, 55, 239-247.
Hollway, Wendy & Tony Jefferson (1997). The Risk Society in an Age of Anxiety: Situating Fear of Crime. British Journal of Sociology, 48, 255-266.
(2000). Doing Qualitative Research Differently. London: SAGE.
Home Office (2002). Justice For All. Cm 5563. London: HMSO.
Hovland, Carl I. & Robert Sears (1940). Minor Studies of Aggression: Correlation of
Lynchings with Economic Indices. Journal of Psychology, 9, 301-310.
Johnson, Robert A. (1993). Owning Your Own Shadow: Understanding the Dark Side
of the Psyche. San Francisco: Harper.
Jung, Carl G. (1936/1981). The Concept of the Collective Unconscious. In The Archetypes and the Collective Unconscious, 2nd Edition. Princeton, NJ: Princeton University Press.
(1964). Approaching the Unconscious. In Carl G. Jung (Ed.) Man and his Symbols.
New York: Laurel.
Karstedt, Susanne (2002). Emotions and criminal justice. Theoretical Criminology.
Katz, Jack (1988). Seductions of Crime. New York: Basic Books.
Klein, Melanie (1985). Love, Guilt and Reparation and Other Works 1921-1945. London: Hogarth Press.
Laing, R.D. (1960). The Divided Self. London: Tavistock.
Langworthy, Robert H. & John T. Whitehead (1986). Liberalism and fear as explanations of punitiveness. Criminology. 24, 3:575-591
Jon Martin Levi (2001). The authoritarian personality, 50 years later: What lessons are
there for political psychology? Political Psychology, 22: 1-26.
Matza, David & Gresham Sykes (1961). Delinquency and Subterranean Values, American Sociological Review 26, 712-719.
Mayhew, Pat & Van Kesteren, John (2002). Cross-national attitudes to punishment in
Changing Attitudes to Punishment: Public opinion, crime and justice. Julian Roberts
& Michael Hough (Eds.) Willan Publishing.
Mead, George H. T. (1918). The psychology of punitive justice. American Journal of
Sociology 23: 577-602.
217
Menninger, Karl (1968). The Crime of Punishment. New York: Viking Press.
Neumann, Eric (1966). Depth Psychology and a New Ethic. New York: Putnam.
(1973). The Child. New York: Putnam.
Nietzsche, Friedrich (1887/1989). On the Genealogy of Morals. New York: Vintage.
Perera, Sylvia Brinton (1986). The Scapegoat Complex: Toward a Mythology of
Shadow and Guilt. Toronto: Inner City Books.
Redl, Fritz & Wineman, David (1951). Children who hate: The disorganization and
breakdown of behavior controls. New York: The Free Press.
Roberts, Julian & Mike Hough (2002). Changing Attitudes to Punishment: Public opinion, crime and justice. Willan Publishing.
Roberts, Julian V. & Loretta J. Stalans (1997). Public opinion, crime, and criminal justice. Crime and Society Series. Boulder, CO: Westview Press.
Roberts, Julian, Loretta J. Stalans, David Indermaur & Mike Hough (2003). Penal Populism
and Public Opinion: Lessons from Five Countries. Oxford: Oxford University Press.
Rose, Jacqueline (1993). Why War? Psychoanalysis, Politics and the Return to Melanie
Klein. Oxford: Blackwell.
Simon, Jonathan (2000). Megans Law: Crime and democracy in late modern America. Law and Social Inquiry, 25, 1111-1150.
Simon, Leonore (2003). Matching legal policies with known offenders. In B. J. Winick
& J. Q. La Fond (eds) Protecting Society From Sexually Dangerous Offenders.
Washington: American Psychological Association.
Sutherland, Edwin H. & Cressey, Donald R. (1974). Criminology, 9th edition. New
York: Lippincott.
Taylor, D. Garth, Kim Lane Schepple & Arthur L. Stinchcombe (1979). Salience of
crime and support for harsher criminal sanctions. Social Problems, 26, 413-424.
Taylor, Shelley E. (1989). Positive illusions: Creative Self-Deception and the Healthy
Mind. New York: Basic Books.
Tetlock, Philip E. (1994). Political Psychology or Politicized psychology: is the road to
scientific hell paved with good moral intentions? Political Psychology, 15: 509-530.
Tyler, Tom R. & Robert J. Boeckmann (1997). Three strikes and you are out, but why?
The psychology of public support for punishing rule breakers. Law and Society Review 31 (2): 237-265.
Valier, Claire (2000). Looking Daggers: A Psychoanalytic Reading of the Scene of Punishment. Punishment and Society, 2, 379-394.
(2002). Punishment, Border Crossings and the Powers of Horror. Theoretical Criminology, 6, 319-337.
Vaughan, Barry (2002). The punitive consequences of consumer culture. Punishment
and Society, 4, 195-211.
218
Voruz, Veronique (forthcoming). Recent Perspectives on Penal Punitiveness. In Bruce

Arrigo (Ed.) Critical Explorations in Crime, Law and Society. Albany, NY: SUNY
Press.
Warr, Mark (1995). Public opinion on crime and punishment. Public Opinion Quarterly 59: 296-310.
Wilkins, Leslie T. (1991). Punishment, crime and market forces. Aldershot, UK and
Brookfield, VT: Dartmouth Publishing
Zimring, Franklin E., Gordon Hawkins & Sam Kamin (2001). Punishment and democracy: Three strikes and youre out in California. New York: Oxford University Press.
219
Public Opinion and Community Penalties1
SHADD MARUNA AND ANNA KING
In this paper we will address three main questions:

What do we know about public opinion and non-custodial penalties?
What accounts for differences in attitudes between individual members of the
public?
What implications does this research have for efforts to garner public support
for non-custodial penalties?
Throughout, we will draw on the empirical and theoretical literature on public
attitudes toward punishment (esp. Roberts and Hough, 2002, Roberts et al., 2003,
Tyler and Boeckmann, 1997). Moreover, we will draw on the emerging findings
from our own University of Cambridge study of punitive and non-punitive attitudes involving almost 1000 British adults.
First, though, we should briefly address the issue of why this paper exists at all
that is, why an understanding of public opinion should even matter in the study
of non-custodial penalties. Some degree of public acceptance of and confidence
in criminal justice practices is clearly necessary for a well-functioning system.
Justice systems that are not seen as legitimate by the majority of a population
presumably need to resort to brute force and intimidation to enforce the law of
the land. Still, these are extreme cases; one can certainly imagine a wellfunctioning criminal justice system that operates with little or no attention to the
minutiae of public concerns. Indeed, according to Ryan (2003) and other criminal
justice historians, for much of the last century criminal justice policy in England
and Wales was almost entirely in the hands of a small, male metropolitan elite
and public opinion was taken as something that had to be managed and circumvented rather than acted upon.
Originally published in Bottoms, T., Rex, S. & Robinson, G. (Eds.), Alternatives to Prison:
Options for an Insecure Society. Cullompton: Willan. Reprinted with permission.
220
Shadd Maruna and Anna King
Recent decades, however, have seen a substantial shift away from the expertdriven, bureaucratic model of penal policy to a system driven more explicitly by
symbolic and expressive concerns (Garland, 2001). There has been a distinct
emotionalization of public discourse about crime and law (Karstedt, 2002: 301)
and punishment far from being hidden has taken on a newly emotive and
ostentatious character (Pratt, 2000a) with the return of boot camps, chain gangs,
capital punishment and the like. In short, with the decline of bureaucratic rationalism in the criminal justice system, public emotions about crime and justice
may now be translated into action, rather than simply left at the level of talk
in a way that would have been hard to imagine 35 years ago (Pratt, 2000a). This
is nowhere more clear than in the recent Carter Report (Carter, 2003) and the
Halliday Report (Home Office, 2001) which both assign a central role to the
improvement of public confidence in criminal justice sentencing.
Nonetheless, the relationship between recent policy developments and public
wishes has been anything but direct. A considerable amount of populist punitiveness (Bottoms, 1995) defined as allowing the electoral advantage of a
policy to take precedence over its penal effectiveness (Roberts et al., 2003: 5)
seems to take place with only a caricatured understanding of the publics real
views regarding crime and justice (Roberts et al., 2003). Considerable research
(see review in Roberts and Stalans, 1997) suggests that the public is not nearly as
punitive as sentencers, politicians and public officials assume they are. Yet, as
Richard Korn (1971) once lamented, the public is one of criminal justices sacred cows often deferred to but never consulted. For instance, Morgan (2002)
points out that when US President George W. Bush justified the treatment of
prisoners held at Guantanamo Bay, by saying they were being treated no better
than the American public thought they should be treated, he hardly was basing
his assessment on survey data from a representative sample.
As such, systematic reviews of public opinion regarding crime and justice are
an important corrective against this failure to communicate between criminal
justice and the public (Flanagan and Longmire, 1996). As Roberts (1992) eloquently demonstrates in his review, not only is it important to dispel the misperceptions the public holds towards crime, but it is equally urgent to dispel the
misperceptions that criminal justice experts and policy makers have towards the
publics opinion on crime and punishment. Dispelling myths about public opinion might be most crucial in the area of non-custodial sentences as Flanagan
(1996) suggests that perceived public opinion (emphasis added) is the greatest
obstacle to the success of community-based penalties. Indeed, in a speech at the
annual HM Prison Service Conference in 2002, Lord Chief Justice Woolf la-
Public Opinion and Community Penalties
221
mented the regrettable (but undocumented) fact that neither the public nor
sentencers have confidence in the community alternative (cited in Roberts,
2002: 34). Likewise, in her keynote address to members of the National Probation Service, Probation Minister Beverley Hughes, said, Public credibility is
crucial to our success. Only if, together, we can convince communities of your
role and your reliability will you be able to do your important job effectively
(Teague 2001: 218). In short, sentencers are reluctant to utilise community penalties, regardless of their levels of effectiveness, if they assume that the public
would disapprove of these options.
I. Public Opinion and Non-Custodial Penalties: What Do We Know?

Fortunately, there has been a flood of recent research and writing about public
opinion and criminal justice (e.g., Cullen et al., 2000, Roberts and Stalans, 1997).
In fact thanks in no small measure to the sponsorship of the Esmee Fairbairn
Foundation and its Rethinking Crime and Punishment initiative much of this
research focuses specifically on the British public and has a special focus on attitudes towards incarceration and its alternatives (see e.g., Allen, 2002, MORI,
2002, Roberts and Hough, 2002, Wood and Viki, 2004). The present review
owes an enormous debt to these works, and was made simultaneously easier and
more difficult to write because of their recent publication. Certainly, since there
have been numerous state-of-the-art reviews of the research on public opinion
(see esp. those in Roberts and Hough 2002), there is little need to review this research, study-by-study again here. Instead, in the short space provided, we will
try to critically review this research and make some broad generalisations regarding what the research seems to conclusively show, to date, on the topic. As will
be obvious from the subject headings of these sections, the bottom line is that we
do not know a great deal. Importantly, as (Stalans, 2002:20) and others have argued, researchers have barely scratched the surface of public attitudes regarding punishment. Most often we describe what the public says it wants without
providing information about what underlies the preference:
1. The public is ignorant about criminological research (But so what?)

One of the most often repeated findings is that members of the general public
generally do not know much about the workings of the criminal justice system,
crime rates, or the basics of criminology (Roberts, 1992, Vandiver and Giacopassi, 1997). For instance, in 1998, 59% of the UK public thought that crime
rates had increased in the previous two years, and four out of five people continue to substantially overestimate the proportion of crime that is violent (Mattin-
222
son & Mirrlees-Black 1998). Likewise, the majority of survey respondents very
substantially underestimated the proportion of convicted adult male offenders
sent to prison. (Hough & Roberts 1996). Only a minority (16%) of respondents
correctly identified that most known offenders are adults, not juveniles, and
nearly a third thought crime was committed equally by females. Finally, over
two-thirds thought that the young were becoming increasingly involved in crime
between 1995 and 1997, yet according to official statistics, the number of known
juvenile offenders has remained constant or fallen during this period (Mattinson
& Mirrlees-Black 1998).
This widespread and systematic public ignorance about crime and justice
(Hough & Parks 2002) has been documented so many times in criminological
research (Durham 1993, Morgan & Russell 2000, Tarling & Dowds 1997), one
gets the impression it is a source of considerable frustration for those of us who
devote our careers to providing a scientific understanding of crime. Yet, why
should we expect any different? As Indermauer & Hough (2002) point out, the
justice system is far from the only aspect of government about which the public
are largely misinformed. Surveys of public knowledge of welfare provision or
the health service likely show equal confusion.
Moreover, there is no obvious reason why most people should have to be conversant with the details of sentencing procedures, welfare policy, or any other
public function. Still, as Morgan (2002: 220) argues, This is no reason for not
consulting them about changes which, ultimately, are likely, in some degree, to
affect every household.
To conclude that a degree of public ignorance provides judges and politicians with a free hand to make policies that they deem to be fair and efficient, policies that can then be sold to the public, is to misunderstand what
counts as criminal justice (ibid).
Importantly, academics often decry the misunderstanding of the nature of crime

and punishment among the public (Indermaur and Hough 2002), but how settled
is the criminologists understanding of this essential nature? Crime is a social
construction of course, and whilst we in criminology do our best (considering the
circumstances) to measure it with some accuracy, we are all aware of the impossibility of perfect accuracy in this endeavour. Under the circumstances, then, it
seems arbitrary if not elitist to devalue the publics construction of crime (seeing
as they are the ones who commit, suffer and otherwise experience it) in favour of
the Home Offices official construction (Ryan 2003). The Guardians crime
journalist Nick Davies (2003) captured this well in a recent article titled Exploding the Myth of the Falling Crime Rate:
223
In the last five years, just about all of [the criminal justice establishment]
have thrown their hats and helmets in the air to celebrate a steady fall in
crime. The Home Office said it was all down to its crime prevention work.
The police said it was their new intelligence-led approach. The academics
said it was rising consumption, falling inequality, more alarms, fewer adolescent males, a rise in abortionsor a fall in unemployment. But what it if
it never happened? What if all that research is misleading? What if the truth
is that crime didn't fall at all, that it was only the statistics that fell?
Research on public opinion suggests that the public is rather dubious about official crime statistics, whether they are collected by the police or through the British Crime Survey, and to some degree, of course, this scepticism is completely
warranted.
Still, there is no question that the public, in general, do not approach crime in a
scientific manner (and they clearly dont read the journal articles we spend so
much time writing). At the same time, there is considerable public interest in all
aspects of crime and justice and a seemingly insatiable appetite for often highly
distorted tales of true crime, reality police dramas, and the like. Further, individuals hold very strong beliefs regarding issues of justice and punishment and
have no great concern that these beliefs are not founded in criminological science. This seems to be a major annoyance to criminologists, but it is unlikely to
change.
2. Public Punitiveness is a Myth (But Then So is the Whole Notion of

a Public Opinion)
The public is mad as hell about crime and are not going to take it any more. If the
general public had their way, they would string up every paedophile, rapist, burglar, drug dealer and car thief lounging in the luxury holiday camps that claim to
be prisons and hang them from the highest tree. At least that is the basic consensus among politicians and political commentators of both the political Left and
Right according to Cullen and colleagues (2002). Indeed this sense that the public harbours increasingly punitive attitudes extends to numerous criminological
theorists (e.g., Pratt 2000b) as well.
Considerable research on public opinion and criminal justice has sought to
complicate, if not contradict this myth of the punitive public (Cullen et al.
1997). Certainly, the characterisation of public attitudes as indiscriminately punitive is misleading (Sprott 1999, Irwin et al. 1998, Innes 1993). Turner et al.
(1997) argue that the public is better characterised as reasonably moderate
rather than punitive. Stalans (2002) decides the public is selectively punitive and
selectively merciful depending upon the specific conditions (Stalans 2002: 19).
224
Whereas, Cullen, Fisher & Applegate (2000) interpret the publics general view
toward the treatment of criminals as simply mushy. Indeed, public opinion research shows rather conclusively that although the general public does largely
support harsh punishment for serious offenders, we are also very much in favour
of rehabilitation (Cullen et al. 1988, McCorkle 1993). Applegate et al. (2000)
demonstrate that when forgiving attitudes are measured as well as punitive attitudes, forgiving views outnumber the former.
Importantly, this mushiness may simply be a research artefact. Stalans (2002:
25) writes, Public attitudes toward sentencing only appear mushy because
much research has neglected the critical issue of how people form general attitudes, and how prior attitudes and beliefs about criminals affect sentencing preferences for detailed cases. First of all, punitiveness as a construct is poorly
understood. There is little consensus on how to define this theoretical construct in
the literature, making meaningful empirical approximations difficult (Walker et
al. 1988).
Most often, punitiveness seems to be measured by asking respondents what
they think the goals of corrections should be. Individuals who favour retribution
over rehabilitation are said to be punitive. Yet, certainly there are retributivists
who support minimal punishment (v. Hirsch 1993). Likewise, there are numerous
non-retributivist grounds (incapacitation, deterrence) for supporting harsh and
severe punishments. Focusing on the publics goals for punishment assumes
that punitiveness is goal-driven in an instrumental sense. Yet, research on death
penalty attitudes (e.g. Ellsworth & Gross 1994) demonstrates that rationalisations
people give for punitive or non-punitive attitudes carry little actual weight. As
psychologists report, More often than not ... behavior is influenced by unconscious processes; that is we act and then, if questioned, make our excuses (Jacoby et al. 1992).
That is, death penalty supporters say they would still support the death penalty
even if it was proven that it does not deter crime, and death penalty opponents
say they would cling to their opposition even if it was proven that capital punishment did deter crime. Scales asking about the goals of punishment are very
useful for teasing apart the various intellectual rationalisations individuals provide for their beliefs, but they do not gauge levels of emotional intensity regarding preferences in punishment, whatever the supporting or opposing reason may
be, As pre-eminent psychologists working on the death penalty Ellsworth and
Gross (1994) comment, Hardly anyone has asked respondents questions that
give them the opportunity to express their emotions directly, and some have intentionally confined their response alternatives to those that are rational (p. 32).
225
Research by Doble (2002) suggests that, unlike criminologists, members of the

general public do not see any contradiction in valuing both retribution and rehabilitation (see also Chapter 4). Whereas traditional research designs ask survey
respondents to rank order the goals of corrections (e.g., retribution, rehabilitation, deterrence, etc.) and read much into the findings, Dobles survey design
allows respondents to list more than one goal as the most important. He found
almost no variation in the designated importance of the goals thought to be in
competition (with most respondents favouring punishment, reform and restitution
at the same time). He concludes:
If people simultaneously deem more than one goal to be vital, asking them
which goal is most important can be likened to asking someone who is
hungry, thirsty, cold and tired, what is most important: food, water, warmth
or rest. The answer will be essentially meaningless if the persons real goal
is not to have one, but all four of their needs met (p. 151).
There are, of course, numerous other limitations to the standard efforts of measuring public beliefs. Flanagan (1996) points out that while the most popular
method of researching public opinion is polling, this is limited because: 1) attitudes are dynamic while most surveys are cross sectional; 2) surveys often ask
about very specific attitudes that cannot be generalized from 2 ; 3) public opinion
is rarely if ever monolithic even when surveys might make it appear so.
This last point is clearly true in the case of punitiveness. Research on death
penalty attitudes, for instance, suggest that this issue, like abortion, provokes
clear and stark attitudinal differences on surveys with lots of strongly agreetype responses and few undecideds (Ellsworth & Gross 1994). Moreover, there
are some clear and systematic demographic differences in views toward punishment. In the United Kingdom, men, older people, citizens with lower levels of
educational attainment, and readers of tabloid newspapers seem to hold significantly more punitive views (Allen 2002, Hough & Roberts 1996). These patterns
tend to hold true internationally (see Mayhew & Van Kesteren 2002) with
stronger gender effects, but a less consistent pattern regarding age and punitive
attitudes. Self-identified members of racial minorities in North American samples tend to be much less punitive than non-minorities (Applegate et al. 1997).
Finally, there are substantial cross-cultural differences in punitive attitudes across
place and time. Mayhew & Van Kesteren (2002) found that Western European
countries rank last in support for imprisonment and first in support for community service in an international comparison. Western European countries also
2
For instance, while one might assume attitudes towards the death penalty might be generalizable to support for longer sentences, this is not the case (see Kury et al., 2002).
226
rank lowest in the average length of sentence they recommend for a young recidivist burglar. However, between 1989 and 2000 there was an increase in support for imprisonment in England and Wales of 13 percent with the sharpest increase occurring between 1992 and 1996 (6 %). (Roberts 2002).
So, is the public punitive? The question might not be worth asking. In his review of the public opinion literature, Warr (1995) concludes, Public opinion on
crime and punishment encompasses such a wide variety of issues and attitudes
that it is pointless to attempt to describe it with any one adjective or phrase
(p. 296). Simplistic summaries of where the public stands that seek to capture
the publics true opinion as if there were such a thing as the public and it had
a single opinion about anything, seem largely to be an exercise in futility. As
Flanagan (1996) warns, public opinion, itself, is often created and reified in the
process of collecting and reporting the results of public opinion polls (see also
Savelsberg 1994).
3. The public is probably ambivalent toward non-custodial penalties

Where does this leave community penalties in the publics mind? The most
common finding in public opinion research regarding non-custodial penalties is
that the public is largely unfamiliar with this aspect of the criminal justice system
(Hough & Roberts 1998). This is unsurprising as probation officers and community service work are hardly ever featured in the media, at least compared to prisons, policing or the courts. The National Probation Service (2002) recently commissioned a study of the public perceptions of probation in England and Wales
involving 1,000 telephone surveys and 665 street surveys with members of ethnic
minorities. The findings suggest that 43% of the respondents consider themselves
to know only a little about what probation service does; for the minority sample, that number is even less (31%). Only seven per cent of the public say they
know a lot about what the Probation Service does.
Further, from what little the public know, the public is anything but inspired
and excited about the potential benefits of non-custodial alternatives. Roberts
(2002: 34) writes:
Most members of the public (and indeed, not a few judges) remain rather sceptical about the utility of [community penalties] for crimes of intermediate seriousness, on the basis that community penalties are not severe enough. This is true
around the world.
This seems to be particularly true of traditional probation work, involving reporting to a probation office for monthly supervisions. Cullen, Fisher & Applegate (2000) write: Citizens appear wary of regular probation, a sanction that
involves minimal contact with the offender. When members of the public were
227
asked what they thought could help to reduce crime in Britain, the National Probation Service (2002) research found that only 2% spontaneously mention that
the Probation Service, compared to 77% who cite the police and 13% who mention schools. Moreover, recent polling data in the US suggests that public confidence in the concept and efficacy of probation has declined in recent years (Beto
et al. 2000) 3 .
Indeed, some have suggested that public support for probation in the United
States is dangerously low. One of the invited participants at a US conference on
Rethinking Probation stated this matter quite bluntly: Public regard for probation is dangerously low. We have to realise that we dont have broad public
legitimacy (Dickey & Smith 1998: 6). Another participant described the public
mood toward community corrections as a malaise (5). In recent years, the British probation service has undergone its own period of feeling uncomfortable,
threatened, unsure of its role, and not at all confident of its social or political
credibility (Garland 1997: 3), although not to the same extent as described in the
United States. This period of uncertainty, of course, was one of the factors leading to the repositioning of probation and the formation of the unified National
Probation Service with the explicit goal of restoring public acceptance.
However, the idea that the public is strongly opposed to non-custodial sentencing has little support in the research literature. For instance, when asked to rank
the crime reduction potential of various parts of the criminal justice system on a
scale of1 to 10, respondents to the National Probation Service research rated probation no worse than the prison system. When asked in an abstract way, 47% of
an American sample responded that community sentences are evidence of leniency in the criminal justice system. Yet, when asked about individual community corrections programs, support for each is in the 70 to 80 % range (Flanagan
1996). In fact, the majority of studies reviewed find that community penalties are
largely supported by the public so long as they are used for non-violent rather
than violent offenders (Oregon Crime Analysis Center 1991).
II. But Why? What Do We Know About the Origins of These

Opinions?
Although there has been considerable discussion in criminology about whether or
not the punitive public exists, very little research has sought to ask why pu3
This change in perception may be due to actual changes in practice; where probation 20
years ago may have focused on change and rehabilitation, it may now function solely as an
instrument of surveillance.
228
nitive public attitudes exist 4 . In fact, there is a tendency in some criminal justice
research to take the existence of punitive attitudes for granted and assume that
public hostility to offenders is natural 5 or innate. In Punishment and Democracy,
for instance, Zimring, Hawkins & Kamin (2001) argue that public punitiveness
has probably been a constant in US history (what changed was the level of victimization and opportunities for populist political decision-making). Like most
other observers, the authors therefore seem to accept popular punitivism as a
fact of nature and therefore offer no real explanation for it (Greenberg 2002:
246). Yet, comparative and historical research contradicts this image of punitive
attitudes as a constant. Sutherland & Cressey (1978) write:
There has been no constant desire to make all criminals suffer and the system used for inflicting suffering on (criminals) has changed from time to
time. The punitive reaction to lawbreaking has not been present in all societies (in fact it) varies from time to time even within a given society. A theory which precisely explains or accounts for all of these variations has not
been developed.
Additionally, there seem to be important, individual-level differences in punitive

attitudes within a single cultural context. Apparently, the execution of a Death
Row inmate can evoke disgust and tears from one person while triggering celebration and cheers from another 6 .
But what accounts for these differences? In their important article, Three
Strikes and Youre Out, But Why? Tyler & Boeckmann (1997) argue that there
are two basic theoretical frameworks within which public attitudes toward crime
and punishment can be understood: instrumental theories and expressive or symbolic theories (Gabriel & Greve 2003, Girling et al. 2002, Tyler & Boeckmann
1997, Tyler & Weber 1982, Wood & Viki 2004). Theoretical discussions of the
expressive purposes of punishment have dominated the sociological discussion of
punishment from the works of Durkheim and Mead to contemporary work by
David Garland (1990, 2001) and others. Ironically, empirical public opinion re4
4
Some notable exceptions include the work of Gaubatz (1995) and Tyler & Boeckmann (1997).
Indeed, this argument is made explicit in theoretical work in evolutionary psychology (e.g.,
Fehr and Gachter, 2002), in which support for the punishment of wrongdoers is considered
an almost universal human trait, crucial to the evolution of civilisation.
Different crimes generally elicit different reactions (Stalans, 2002, Altemeyer, 1988), but
there is very little research about the processes that underlie the ways in which they do. For
instance, rehabilitation might be chosen as suitable punishment for a drug offense - hard
time s for a violent offense. However, this is more easily explained at the surface - one consequence may seem to logically go with the crime. However, why a person is sympathetic
towards paedophiles, but then explodes at the mention of white-collar criminals (as one subject in our interview research was), is not as easily explained.
229
search has been criticised for focusing too much attention on the instrumental,
capturing only the rational or knowledge-based aspect of opinion and ignoring
the emotional side (Indermaur & Hough 2002: 201).
1. Instrumental explanations
Instrumental theories (Zimring et al. 2001) suggest that punitiveness is motivated
largely out of self-interest. Punitive attitudes are likely to result when individuals
feel a personal threat to themselves or their communities. People are fed up.
Like Charles Bronson in Death Wish, they are normal, good-natured individuals
who get pushed too far by the crime and disorder around them. Hence, punitive
attitudes should be more prevalent in areas of high-crime, among individuals
who have been victims of crime and who fear repeat victimization.
A sizable and consistent body of research fails to support these instrumental
hypotheses (Cullen et al. 1985, Baron & Hartnagel 1996, Hough & Moxon
1985). For instance, public opinion polling does show a relationship between
punitiveness and the fear of crime (Taylor et al., 1979), but this seems to be a
very modest correlation at best (see Roberts and Stalans, 1997) and frequently
not replicated (Cullen, Fisher & Applegate 2000). Additionally, fear of crime
does not seem to have a measurable relationship to views on capital punishment
(Warr 1995) or support for other specific policies like Three Strikes legislation
(Tyler & Boeckmann 1997). While there has been some support for a link between fear of crime and punitiveness (e.g. Sprott & Doob 1997), a major critique
of the research on fear of crime has been its inability to adequately account for
such complex relationships (Ditton et al. 1999, Holloway & Jefferson 1997).
Moreover, personal victimization experiences and perceptions of crime salience in ones residential area do not relate consistently to punitive attitudes
(Langworthy & Whitehead 1986, Quimet & Coyle 1991, Flanagan et al. 1985).
In fact, only 9 percent of the self-reported victims of violent crimes in the 1998
British Crime Survey favoured the incarceration of their offender (Roberts 2002).
This research counters intuitive beliefs that victims are most likely to seek retributive or harsh consequences for offenders. Walker & Hough (1988) suggest
that one reason for this is that individuals who lack personal experience of victimisation may exaggerate the negative consequences of crime while victims are
more pragmatic. Alternatively, it might be that victims underplay the severity of
harm done in order to better cope with the experience. Other research such as
Tufts & Roberts (2002) find that victimization is not at all predictive of attitudes.
Finally, at the aggregate level, recorded rates of crime do not relate to public demands for punishment in a consistent manner (Wilkins 1991, Mayhew & Van
230
Kesteren 2002). Similar conclusions lead Tyler & Boeckmann (1997) to comment that crime-related concerns are the least important factor in predicting
punitive attitudes and suggest that the image of the citizen as supporting punitive public policies because of fear of crime is inaccurate (252).
2. Expressive-emotive explanations
An alternative explanation to the instrumental view is that punitive attitudes serve an
expressive or symbolic function which is of course to beg the question expressive
of what? (Garland 2001). There is a rich tradition of theoretical work in this area
(e.g., Durkheim 1933, Mead 1918) which continues today with work such as David
Garlands Culture of Control. The best known explanations for contemporary punitiveness revolve around ontological insecurity or a widespread sense of anxiety
driven by the disembedding processes of modernity that have resulted from the erosion of former social certainties (Bottoms 1995; Ranulf 1938; Vaughan 2002; Young
2003). For instance, Bauman (2000) points to the profound anxiety and insecurity
produced by the flexibility of the labour market under the deregulated capitalism favoured by neo-liberal states. Indeed, the relationship between economic insecurity and
scapegoating behaviour is well known. For instance, in their now classic study, Hovland & Sears (1940) found that the frequency of lynching in the southern US states
was negatively correlated with the price of cotton. When farmers suffered the most
frustration, they were most likely to redirect their anger on Black men accused of
crimes (these findings were confirmed by Hepworth & Wests (1988) recent reexamination of these data).
Tyler and Boeckman use the label of symbolic theories, to describe a wide
variety of accounts that link punitiveness to concerns for moral cohesion and the
assertion of community values (These theories are reviewed in considerable detail in Garland 1990). Durkheim (1933), of course, famously argued that punishment served a ritualistic reaffirmation of community values. Similarly, for Mead
(1964), punitiveness offers one of the few occasions on which community members can unite around shared interests and an emotional solidarity of aggression
(227). Garland summarizes this position as follows: Taking part in the emotional
defence of societys interests against criminal depredations, the individuals
aggressions against the outsider are aroused and reinforced, as is his or her
identification with the in-group (64). In their own research, Tyler & Boeckmann
find support for such symbolic theories, concluding that two aspects of the social
environment are key to punitiveness: concerns about the breakdown of family
values and anxiety about population diversity and a lack of community cohesion.
They suggest that future research on punitiveness incorporate the findings of re-
231
search on social identity theory (e.g., Turner et al. 1979) regarding the framing of
group boundaries.
The suggestion from this literature is that not only are peoples attitudes towards crime and punishment emotional rather than rational and utilitarian, but
that they are in fact driven by very deep and very personal psychodynamic histories (see Maruna et al. 2004). Indeed, Mead also pointed out that punitiveness
offers a rare outlet for releasing long-restrained and sublimated hostilities. Finally, Garland (1990) argues that criminals may represent a threat because their
behaviours often express desires which others have spent much time and energy
and undergone much internal conflict in order to renounce (239). For this reason, he argues the public may harbour a resentful and hostile reaction out of
proportion to the real danger it represents.
Because of their roots in depth psychology, these symbolic theories are rarely
explored outside of individual case histories based upon reliable clinical evidence (Garland 1990: 65). One exception is Gaubatzs (1995) Crime in the Public Mind. In an inductive analysis of 24 qualitative interviews with Californians,
Gaubatz concludes that punitiveness is a type of botheredness which results
from a macro social displacement effect. Her grounded theory suggests that the
pace of social change over the last four decades in the United States exceeded
some peoples capacity for change. As a result, peoples rejection of certain social practices (i.e., interracial marriage, homosexuality) remained intact, but suddenly became no longer acceptable with changing social mores. She argues that
the emotional rejection of certain social practices that could no longer be expressed directly was displaced onto attitudes regarding punishment and crime
(criminals being among the last groups of individuals that it was still respectable
to openly hate). This research has been widely criticised on methodological
grounds. Warr (1995a) for instance, questions the small sample size, sample selection procedures and the relation between Gaubatzs conclusions and her actual
data. Nonetheless, Gaubatz provides a fascinating attempt at a personological or
psychodynamic inquiry into punitive attitudes.
3. Core beliefs and values

One of the most promising lines of investigation into the psychological formation
of attitudes towards criminal sanctioning to date is work in the area of attribution
theory. Cullen et al. (1985), for instance, looks directly to variation in peoples
attitudes towards punishment, rehabilitation, capital punishment and punishment
of white-collar crime and finds that the way people explain crime helps explain
some of the discrepancies. George Vold (1958: 258) writes: There is an obvious
232
and logical interdependence between what is done about crime and what is assumed to be the reason for or explanation of criminality. Cullen and his colleagues divide attributional beliefs into either classical (dispositional) views
that crime is a choice or else lay positivist (situational) views that see crime as
a product of circumstances. Numerous previous studies (e.g. Cullen et al. 1985,
Grasmick & McGill 1994) have found support for the idea that punitive attitudes
correlate with classical or dispositional attributions, whereas those who hold
more situational attributions tend to be less punitive 7 .
Psychologists are also concerned with another dimension of attributions (stability versus instability) that is often neglected by criminologists, but might be
equally important in determining punitive attitudes. That is, regardless of ones
beliefs about the origins of criminality, do they believe that people can change?
Such a belief in redeemability (or the instability of criminality) may take precedence over attributions in determining punitiveness. Garland (2001) writes,
Whether the offenders character is the result of bad genes or of being reared in
an anti-social culture, the outcome is the same a person who is beyond the pale,
beyond reform, outside the civil community (p. 185). The belief in people who
are permanently and fundamentally bad almost necessitates their segregation
from mainstream society. A belief in redeemability and human malleability,
therefore, might be a more robust predictor of punitiveness, then than the internality/externality (classical/positivist) dimension of attributions.
4. New Research: University of Cambridge Public Opinion Project

(UCPOP)
A project designed to build on the above work, incorporating both in-depth interviewing as well as survey methodology, is currently underway at the University
of Cambridge. The first phase of this three-phase project is a postal survey designed to test a variety of theories regarding the correlates of public attitudes, but
also to identify two samples: one of British citizens with highly punitive belief
scores and a contrasting sample of individuals with strongly non-punitive views.
6
Importantly, the way in which one explains crime may come after, not follow ones preference for punitive policies. It is equally possible that one may believe crime is a choice as a
way to justify their deeper needs to be punitive. As such, attributions about the causes of
crime may act similarly to deterrent beliefs, To the extent that deterrence beliefs are a cognitive justification for an affective value position, those beliefs and the value position they
protect will be only minimally responsive to cognitive persuasion. (Tyler & Weber, 1982:
242) Such beliefs are thought to be somewhat impenetrable by education because their
source is located not in cognitive processes, but in the affective domains that involve social
values developed during the childhood and adolescent socialisation process (Tyler &
Boeckmann 1997: 254).
233
The second phase of the research (currently underway) involves in-depth, exploratory interviews with members of these two groups, where interviewees talk
about their lives: their experiences with being punished, their experiences of punishing others, experiences witnessing punishment, and their general concerns and
anxieties. The goal is to find five to ten themes that characterise the worldviews
or self-identities of each group. The third and final phase of the research will involve bringing large numbers of the initial survey respondents back for a series
of experiments designed to see whether raising the salience of various anxieties
does consistently lead to increases in punitive responses for randomly selected
participants compared to a control group. It is hoped that this triangulation of
data will provide more comprehensive information on the psychology of punitive
and non-punitive attitudes than has previously been collected.
Below, we briefly discuss the methods of the postal survey below and take a look
at some early outputs that relate to public attitudes about community penalties.
UCPOP: the postal survey

In June 2003, 3,600 surveys were sent to randomly chosen households in six
wards selected for their diversity in income and demographics (as well as their
proximity to the investigators). The sampled areas represent both rural and urban
areas in the East and London, ranging from one of the highest ranking wards on
the 2000 indices of deprivation (Great Yarmouth) to the lowest (Stapleford). The
characteristics of the 941 respondents largely reflected the communities from
which they were drawn (see descriptives below).
Table 1: Some sample characteristics (N=941)
56%
51%
58%
41%
72%
7%
are female. 68% have not completed a university degree.

describe themselves as politically conservative.
consider themselves religious or spiritual persons.
describe the household they were raised in as working class.
report their household income as under 40,000.
report being convicted of offences more serious than a
speeding violation.
47% report knowing someone who has been to prison or who has been on
probation.
As with other samples of the British population, the sample is rather neatly divided with about a third of respondents supporting a greater use of community
penalties, another third supporting the more widespread use of imprisonment,
and the remaining third largely undecided in the middle.
234
What does the sample think of alternatives to prison?

47% agree if prison has to be used, it should be used sparingly and only as
a last option.
31% agree probation or a community sentence (rather than prison) is appropriate for a person found guilty of burgary for the second time.
36% would consider volunteering time or donating money to an organisation
that supported alternatives to prison.
36% would consider volunteering for an organisation that supported
toughening the sentencing laws in the UK.
Hypotheses and Measures

For the purposes of this paper, we have run some preliminary analyses to see
what items included in our survey seem to best predict support for community
penalties among our sample. In our analysis, we include the demographic characteristics (age, gender, etc.) thought to be associated with punitive attitudes, and
predict that instrumental factors (living in high crime areas, victimization experiences) will decrease the likelihood that one will support prison alternatives above
and beyond these demographic characteristics. Likewise, we predict that expressive factors (e.g., anxiety about the economy or the state of Britain) will further
decrease support for alternatives, controlling even for instrumental concerns.
Finally, we also include a test for the impact of core beliefs about crime on
ones support for prison alternatives. In our analysis, we test the effects of these
attributions on punitive attitudes while holding other potential correlates of punitiveness constant. In addition, we added our scale on belief in redeemability
into the equation. This tests a second, neglected dimension of attribution theory
(stability-instability) beyond the internal-external dimension usually tested. The
hypothesis here was that individuals who believed deeply in the notion that people can change and that even the worst young offenders have the ability to turn
their lives around, will be the most likely to support alternatives to prison.
Analyses and Findings

Hierarchical regression analysis: To investigate the unique contribution of each
set of explanatory variables in predicting attitudes supportive of community
sanctions, each model was analysed in four separate steps (see Table 2 below).
This technique lets us assess how each additional model might contribute to the
overall variance explained. In the first model, we measure the possible sociodemographic effects on punitiveness. In the second model, measures of victimisation, perceptions of local crime, and fear of crime measures are added to the
235
equation to test instrumental hypotheses. The third model includes measures are
designed to test expressive hypotheses revolving around social and economic
anxieties. Our measures here include: a standard measure of collective efficacy, a
measure of anxiety about youth (including items such as Young people do not
seem to have respect for anything these days) and an item global crime salience
measuring whether the person believed crime was increasing or decreasing
across the United Kingdom. Finally, in model four we assess the impact of core
beliefs on attitudes toward community penalties. Here we would hypothesise,
based on previous literature, that those who attribute criminality to the internal
disposition of an offender (crime is a choice) would be more punitive and
therefore less likely to support community penalties. Additionally, however, we
hypothesise that those persons who see criminality as a largely unstable trait and
believe in offenders abilities to change their behaviours (reflected in high redeemability scores) will be more supportive of community penalties.
Table 2: Regression Models Predicting Pro-Community Sanction Attitudes
Variable
Model 1
Model 2
Model 3
Model 4
1. Controls
Class Background
Gender (male)
University
Income
Race
Age
Unemployment
.125**
-.083*
.282***
-.080
.064
.080*
.060
.123**
-.099**
.251***
-.127**
.080*
.091*
.060
.106**
-.097**
.155***
-.160***
.091*
.106**
.057
.082*
-.100**
.101**
-.112**
.073*
.142***
-.054
.062
.011
-.229***
.037
.121.037
-.120.037
.029
.090.037
-.066
.165***
-.254***
-.054
-.176***
.140***
-.176***
-.024
-.132**
2. Instrumental
Direct victimisation
Crime salience (local)
Fear of crime
3. Expressive
Collective efficacy and trust
Anxiety about youth
Economic pressure
Crime salience (global)
4. Core Beliefs and valutes
Crime is a choice
Belief in redeemability
Adjusted R2
R2change
* p<.05
** p<.01
-.207***
.300***
.102
*** p<.001.
.133
.035***
.290
.160***
.403
.112***
236
In model one, university degree, class background, gender and age emerge as
significant predictors, suggesting that highly educated, older women of higher
social classes are the most likely to support community penalties for offenders.
As in previous research, education seems to have the strongest impact. Those
with a university education are for more likely to support community sanctions
than the less educated, controlling for other demographic factors. Overall however, these variables only explain about ten percent of the variation in our sample, a result that is consistent with the ten percent usually explained by sociodemographics in similar studies.
In the second model, the socio-demographics act as control variables to test
the effects of three main instrumental theory predictors. When socio-demographics are controlled for, fear of crime is the only instrumental variable that is a significant predictor of pro-community sanctions attitudes (.229***). In addition,
the effect of personal income was suppressed by the instrumental variables (from
-.080 to -.127**). The effects of race and gender are also strengthened when instrumental variables are controlled. Overall, though, the addition of instrumental
predictors to the regression only gives the model four percent more predictive
power.
In model three, when expressive variables are added, the r-square increases
sixteen percent 8 . Almost half of the strength of the relationship between fear of
crime and pro-community sanctions attitudes, as well as between education and
pro-community sanctions attitudes, disappears when expressive variables are factored into the equation (from -.229*** down to -.120* and from .261*** to
.155***, respectively).
Finally, the fourth model adds scores of dispositional attributions and belief in
redeemability to the equation. The addition of these variables helps the model
explain eleven percent more of the variation found in individual differences in
pro-community sanctions attitudes in this sample. Several variables decrease
very slightly in strength or significance (class, education, income, local crime
salience, collective efficacy, anxiety over youth and global crime salience), but
overall there are no significant changes to the other variables. Attributions of internality and a belief in redeemability are shown to have significant and unique
effects on pro-community sanctions attitudes (-.207*** and .300*** respectively).
In sum, expressive predictors and core beliefs and values have a strong effect
on pro-community sanctions attitudes, over and above the effect of both sociodemographics and instrumental factors, as predicted. A belief in redeemability
8
Adding variables to an equation will automatically increase variance explained, but there
will be differences in that increase depending on how the model is specified.
237
was the strongest predictor of support for pro-community sanctions attitudes.

Neither victimization nor fear of crime contributed significantly to the model
once expressive predictors and core beliefs and values were accounted for.
Whilst these findings are only preliminary our research project is not even half
done they do suggest some interesting possibilities for developing a better understanding about public support for community penalties. For instance, the research suggests that the publics lay criminological beliefs and understandings
about why people commit crime may play a greater role than actual experiences
with victimisation in determining support for community penalties. In other
words, they may be a role for public education and working with public views.
III. Implications for Working with Public Opinion

Academics often seem to wish that public opinion could be largely kept out of
the policy-making world. Privately, many of us even long for a system more like
the cosy, elite policy-making world described by Ryan (2003) in his history of
criminal justice policy in Britain. Barring such a radical reversal of recent trends,
though, numerous observers have suggested strategies for working with public
opinion. Roberts & colleagues (2003) outline multiple strategies for mediating
and moderating the impact of public opinion on criminal justice policy including
the establishment of institutional layers of protection as buffers between politicians and the judiciary.
Others argue that in order to win public support, community alternatives need
an image overhaul. Beto, Corbett & DiIulio (2000: 1) argue, Although low ratings [in public opinion polls] obviously are related to poor performance, they
also signal a failure on probation's part to convey an image to citizens of a model
of practice that embodies widely held values and serves overriding public safety
concerns. Likewise, Maloney, Bazemore & Hudson (2001) argue that the US
model of probation has gone the way of the Edsel in terms of performance and
reputation, and like the Ford companys infamous failure, probation needs to be
retired. Importantly, they not only advocate the end of traditional US probation
practice (which they say is based on the rather bizarre assumption that surveillance and some guidance can steer the offender straight), but also dispensing
with the brand name of probation in the United States (which they rightly argue is a vague and uninspiring term). They suggest that a more fitting mission
and name for probation should be community justice. Indeed, community justice as an ideal has been the subject of considerable excitement and theoretical
examination over the past two decades.
238
Drawing on Anthony Duffs communicative theory of justice, Sue Rex (2002) has
argued persuasively that community penalties have the potential to communicate a
message to both the public and the offender him or herself. Unfortunately, the potential message to both is often highly diluted and confused because of a lack of a consensus regarding the rationale behind these penalties (Kalmthout 2002). Below, we
review what works and what probably does not work in terms of changing the publics perception of non-custodial penalties. Some of these conclusions are based on the
experiences of campaigning groups such as the organisation Payback that have made
concerted and substantial efforts to change public opinion around issues of community alternatives (see Bowers 2002)9 .
1. Appeals to Cost-Benefits and Effectiveness Will Only Go So Far

Research conducted by the Centre for Social Marketing at the University of
Strathclyde on behalf of the Rethinking Crime and Punishment initiative found
that statistical arguments about the effectiveness of non-custodial sentences ultimately had little impact on public views about community penalties (Stead et al.
2002). Respondents were sensitive to the highly politicised nature of the crime
debate, and regarded any use of statistics as spin (p. 4). Further, arguments in
favour of community alternatives based on the high costs of imprisonment or the
growing numbers of citizens incarcerated seemed largely unpersuasive to focus
group members. In fact, focusing on the high costs of imprisonment simply reinforced the popular view that prisons were full of unnecessary luxuries, and provoked the retort that they should take away the televisions (p. 3).
If attitudes toward criminal punishment are largely driven by emotive rather
than instrumental concerns, as our own research and the wider criminological
literature suggests, then rational appeals to the benefits of various justice options
will have only limited impact on public views.
2. Public Education Will Help, But is No Panacea

One of the most often mentioned strategies for increasing public confidence in
community sentences is to provide more and better (i.e. research-driven) information about crime and justice to the public (Roberts & Stalans 1997, Stalans
2002, Gainey & Payne 2003). The research evidence in favour of this strategy,
however, is somewhat mixed.
On some level, it is irrefutable that information works. That is, in almost
every survey of the public where such comparisons are made, individuals who
are provided with additional information about various sentencing alternatives
9
www.payback.org.uk
239
are less likely to favour these sentences (e.g., by sentencing hypothetical offenders to prison in You be the judge type exercises) than those who are given
no such information (see Roberts 2002). In particular, explaining the variety of
restitution and compensation alternatives to respondents who are unfamiliar with
community-based penalties has the immediate effect of reducing punitive tendencies in survey respondents (Hough & Roberts 1998). Further, respondents
who express punitive views in the abstract, often moderate those views when
presented with more information about the offenders themselves (e.g., learning
that they have suffered abuse in the past, grew up impoverished or are addicted to
a drug) (Doob & Roberts 1988).
On the other hand, much of the research demonstrating the impact of education on attitudes is demonstrates only very short-term effect 10 . For instance, Gainey & Payne (2003) found that a 35-minute presentation of information about
crime and justice can increase support for alternative sanctions, but the duration
of this effect is unknown as only an immediate post-test was done. Additionally,
much of this research is plagued by what can be interpreted as a Hawthorne Effect participants may modify their views on follow-up surveys simply because
it is so obvious that this is what they are supposed to do (see e.g., MirrleesBlack 2002). Finally, the practicality of introducing these educational efforts on a
large scale are doubtful. Much is made about the impact of deliberative polling, whereby attitudes seem to change in light of an educational encounter of
sorts (ranging from a lecture to a two or three-day seminar on issues of crime and
justice). Yet, considerable research (and personal experience on the part of the
authors!) suggests that even an entire academic term spent learning about the ins
and outs of criminology and criminal justice has a negligible impact on students
attitudes toward crime (Giacopassi & Blankenship 1991, Jayewardene et al.
1977). How much education is really needed to change deep-seated attitudes and
how possible would it be to educate the adult population of Britain in this way?
Bowers (2002) argues that attitudes serve four functional purposes: to organize
vast amounts of knowledge, to express values, to help defend ones ego and to
obtain rewards and avoid punishment. Typically, only the knowledge function is
addressed when it comes to strategies for change. Many initiatives or campaigns
talk about making messages easy to remember and recall, which ignores the
possibility that attitudes exist not only to organize information, but also for other
10
Hough and Park (2002) is a rare exception. They found that long-term attitude change is
possible but not very common, and the intervention in question was quite dramatic (a series
of lectures from politicians and experts). Additionally, they found long-term attitude change
was most likely among more educated individuals.
240
reasons. Many attitudes, such as prejudicial ones, are believed to serve purposes
that have little if anything to do with knowledge organization. As such, education
may do little to change them, As anti-rascist campaigners know only too well,
overcoming these rigid stereotypical schemas is exceptionally difficult, and
generally cannot be achieved with information based initiatives. (Bowers 2002,
p. 25) When attitudes do not persist due to informational discrepancies or deficits, they are not easily altered.
Perhaps the most promising findings regarding education, however, is that the
active participation of citizens in the criminal justice process increases satisfaction with the service and decreases punitiveness (see Allen 2002). Research suggests that when citizens are actively engaged in criminal justice decision-making
whether it is through serving on a jury (Matthews et al. 2003), participating in
restorative justice work (Greene & Doble 2000) or even sentencing hypothetical
offenders through academic exercises (Roberts & Stalans 1997) they are less
punitive and more likely to support community alternatives. Apparently, easy
slogans like hang em high or lock em up become less tenable when individuals are assigned the responsibility of actually trying to turn such general notions into practice. Yet, the average citizens only interaction with the criminal
justice system involves little more than reporting a minor crime to the police.
Further, as Morgan (2002: 225) argues, when he or she does report being the victim of crime, the likelihood of the crime being cleared up or, even if cleared
up, their learning about what happened to their offender is low. Unfortunately,
these realities are not likely to change in the foreseeable future.
Schemes to educate and inform the public about the nuances of sentencing, the
facts about crime, and so forth are noble, well-meaning efforts, but unlikely to
have any more than marginal impact on either public understanding of crime issues or punitive, prison-centric attitudes.
3. The Public Wants Affective as well as Effective Justice

Indermaur & Hough (2002: 210) argue persuasively Anyone who wants to improve
public debate about crime needs to be attuned to [the] emotional dimension [of attitude formation]. Frieberg (2000) aptly describes this as the difference between effective and affective justice. The punishment of criminal offenders is a deeply
emotive issue. Yet, saying attitudes are driven by emotion does not make them
wrong. Academics tend to favour the rational and logical over the emotive and
dismiss the latter as irrelevant and misguided (Tetlock 1994; Ryan 2003). The public, on the other hand, has little problem with gut reactions and supporting what
feels right instead of what they are told is logically correct. If one seeks to sway
241
public opinion, the legitimacy of these other bases for opinion need to be understood
and appreciated. That is, punitive policies and practices have won votes in recent
years because they appeal to the emotional needs of late modern voters and tax payers (Karstedt 2002). Rather than bemoaning this lack of sheer technocratic rationality on the parts of citizens, research on punitive attitudes suggests that those in favour of community penalties would do well to make similarly emotive appeals on
behalf of non-custodial penalties.
Unfortunately, we know very little about what emotive themes are likely to support community sentences because we know remarkably little about the social psychology of non-punitive attitudes. Whereas the authoritarian personality has generated a half-century of research across several academic disciplines, research on the
development of liberal, permissive, forgiving, or non-punitive outlooks toward punishment is badly lacking (Martin 2001)11 . Indeed, little is known about the existence
let alone the origins of public compassion, forgiveness or empathy in regards to
criminal offenders. Little in the sociology of punishment (from Durkheim to Garland) could easily account for this sort of empathetic understanding if it exists (and it
surely does12 ) outside of elite circles suffering from liberal guilt. Reflecting the
popular stereotype that a liberal is just a conservative who has never been
mugged, for instance, Garland (2001) writes: The posture of understanding the
offender (is) more readily attained by liberal elites unaffected by crime or else by
professional groups who make their living out if it (p. 78). As a result, we can imagine (indeed have seen clearly) how the publics punitive attitudes could be awakened
and utilised in support of a repressive criminal justice agenda, but we have little idea
how to promote a more tolerant society.
Our own research and that of others suggests that redeemability is a powerful theme for those who support community penalties. Appealing to the public to
support community alternatives because people can change, and demonstrating
this with human interest stories of transformed offenders might have some value.
Applegate & colleagues (2000) write, Our study shows that the compassionate
side of [public attitudes] the belief in forgiveness can also shape how [citizens] think about crime (742).
11
12
Although, the sociological work of Norbert Elias might be a good starting point for such an
analysis.
Anecdotal evidence abounds of course. For instance, the opposition to the tough punishment
involved in English detention centres in the 1950s came not from elite penal reformers but
from the very prison officers expected to implement these regimes: they found it impossible to do so without suffering severe pangs of conscience, according to Pratt (2000a).
242
Some research suggests that the publics hesitations about community corrections can be alleviated by assuring them that the sentence will be intensive (i.e.
involving more than just standard probation services). In particular, describing
the potential use of electronic monitoring and the like seems to increase public
support for non-custodial penalties somewhat (Brown & Elrod 1995; Dowds
1994; Gainey and Payne 2003). These sort of see how tough we are arguments
can backfire however. Essentially, if the public thinks that offenders on community penalties are really so dangerous that they need constant supervision, then
why bother with electronic monitoring when prison would do quite nicely? The
public holds a deeply entrenched view that equates punishment and control
with incarceration, and that accepts alternatives assuitable only in cases where
neither punishment nor control is thought necessary (Smith 1984: 171). No matter how tough the restrictions, community penalties simply cannot compete with
the iron bars, high walls and razor wire of the prison in the battle for being the
toughest 13 .
Far more evidence suggests that the principles of restitution, community service or giving something back appeal strongly to the public (see Gandy 1978;
Mattinson & Mirrlees-Black 1998; Shaw 1982; Wright 1989). The University of
Strathclyde research indicates that arguments about the values and principles underlying non-custodial penalties were far more meaningful to focus group participants than information regarding the effectiveness or cost benefits of these
sentences (Stead, MacFadyen & Hastings 2002). Notions such as paying back,
making good and restorative justice, for instance, were said to resonate
strongly with focus group members (p. 1). Canadian research by the Angus Reid
Group (1997) similarly found the possibility of victim compensation to be a more
persuasive argument in favour of community penalties than arguments revolving
around the high price tag of imprisonment.
Emotive appeals to the unfortunate circumstances and disadvantaged origins
of most criminal offenders seems to carry little weight with the public. Stead,
MacFadyen & Hastings (2002) report that any argument that appears to be sympathetic to the plight of offenders provoked hostile reactions from their focus
groups with British citizens. Yet, there was greater success when appeals were
based on what Bazemore (1999) calls earned redemption, whereby offenders
earn their way back into society through structured opportunities to make
amends, through positive contributions to their communities. Such demonstra13
At least in the publics mind. Research by Petersilia et al (1986) suggests that many prisoners surveyed would rather spend a short time in jail than very long periods of time on intensive probation.
243
tions send a message to the community that the offender is worthy of further
support and investment in their reintegration (Bazemore 1999). As one participant in a Rethinking Probation focus group argued:
Let me put it this way, if the public knew that when you commit some
wrongdoing, you're held accountable in constructive ways and you've got to
earn your way back through these kinds of good works, [the probation
service] wouldn't be in the rut were in right now with the public (Dickey
and Smith 1998: 6).
By symbolically transforming the probationer into a giver rather than a consumer of help, non-custodial penalties might be seen in a more positive light.
Conclusions
Academics are sometimes uncomfortable with the privileging of public opinion
(Ryan 2003), and they are even more uncomfortable with the privileging of emotions and the non-rational. Yet, supporters of community penalties ignore such
things at their own peril. As Garland (1990: 62) argues, reformers tendency to
ignore public punitiveness or dismiss these views as a form of false consciousness is counterproductive. If such sentiments do exist, and give support to current penal practices, then penal reformers will have to address themselves directly to popular feelings if they intend to produce real change.
The public is not nearly as punitive as some politicians seem to think (Roberts
2002), but their public attitudes about justice and punishment are real, not just
logical mistakes based on faulty information that can be corrected once more
information is provided. Research suggests that the presentation of new information, factual or statistical, and even the open discussion of that new information
(as with deliberative polling) can change attitudes, but that change is limited 14 .
Hough (1996) discovered this in his focus group research with members of the
British public. Although his research found that educating members of the public
about the facts about crime and justice could improve overall levels of confidence in sentencing practices (including the use of non-custodial penalties),
Hough concludes:
It would be a large oversimplification to argue that once peoples ignorance about practice has been corrected, opinion and practice fall into line.
Our respondents were, in the main, very punitive toward offenders. Many
of the groups proposed castration by no means frivolously as a way
of dealing with rapists.
14
There may be as yet unknown effects of ongoing factual re-education, but research suggests
that while repetition may be useful for familiarization it is unlikely to be instrumental in
changing attitudes. (Bowers, 2002, p. 32).
244
Understanding the emotive appeal of castration, hanging, and the mass imprisonment of young, minority males (as in the US context, in particular) may
require new research methodologies in public opinion research. Working with the
members of the public who harbour such emotive views will require a new approach to marketing non-custodial penalties. The uncharted territory in the search
for knowledge on public opinion is that of emotions. While understanding the
relationship of emotions to both punitiveness and forgiveness may be more challenging than mapping a purely cognitive schema, it does not follow that it is beyond our grasp 15 . The exploration of this relationship can only expand our knowledge of attitudes towards offenders. With a better, fuller understanding at hand it
is more likely that better strategies will not be far behind.
References
Allen, R. (2002). What Do The Public Really Feel About Non-Custodial Penalties?, Rethinking Crime & Punishment; Esmee Fairbairn Foundation, 5.
Altemeyer, D. (1988). Enemies Of Freedom: Understanding Right-Wing Authoritarianism, Jossey-Bash Publishers, San Francisco, London.
Angus Reid Group (1997). Attitudes To Crime, Angus Reid Group, Ottawa.
Applegate, B.K., Cullen, F.T. & Fisher, B.S. (1997). Public Support For Correctional
Treatment: The Continuing Appeal Of The Retributive Ideal, The Prison Journal, 77,
237-258.
Applegate, B.K., Cullen, F.T., Fisher, B.T. & Vanderven (2000). Forgiveness And Fundamentalism: Reconsidering The Relationship Between Correctional Attitudes And
Religion, Criminology, 38, 3, 719-753.
Baron, S.W. & Hartnagel, T.F. (1996). Lock Em Up: Attitudes Toward Punishing
Juvenile Offenders, Canadian Journal Of Criminology, 191-212.
Bauman, Z. (2000). Social Issues Of Law And Order, British Journal Of Criminology,
40, 205-221.
Beto, D.R., Corbett, R.P. & Dillulio, J.J. (2000). Getting Serious About Probation And
The Crime Problem, Corrections Management Quarterly, 4, 2, 1-8.
Bottoms, A. (1995) The Politics Of Sentencing Reform, In The Philosophy And Politics
Of Punishment And Sentencing, Vol. (Ed, Morgan, C.C.A.R.) Oxford University
Press, Oxford, 17-49.
15
While experiments on the willingness to punish have been conducted (i.e. Milgram) we
would hope that lines of investigation suggested by this type of review might be undertaken
from a more humanistic perspective than a behavioural one. Such research has tended to reduce human experience to stimulus-response sets and in doing so given us a very clear picture, but of only half of what we need to know.
245
Bowers, L. (2002). Campaigning With Attitude: Applying Social Psychology To Criminal Justice Communication, Payback, Pp. 46.
Carter, P. (2003). Managing Offenders, Reducing Crime (Carter Report), Home Office,
Strategy Unit, London, Pp. 52.
Cullen, F. T., Clark, G. A., Cullen, J. B. & Mathers, R. A. (1985). Attribution, Salience,
And Attitudes Toward Criminal Sanctioning, Criminal Justice And Behavior, 12, 3,
305-331.
Cullen, F.T., Cullen, J.B. & Wozniak, J.F. (1988). Is Rehabilitation Dead? The Myth Of
The Punitive Public, Journal Of Criminal Justice, 16, 4, 303-317.
Cullen, F.T., Fisher, B.S. & Applegate, B.K. (2000). Public Opinion About Punishment
And Corrections, In Crime And Justice: A Review Of Research, Vol. 27 (Ed, Tonry,
M.).
Cullen, F.T., Pealer, J.A., Fisher, B.S., Applegtae, B.K. & Santana, S. (2002). Public
Support For Correctional Rehabilitation In America: Change Or Consistency? (Ch
7), In Changing Attitudes To Crime And Punishment ... Vol. (Eds, Roberts, J. And
Hough, M.) Willan, Pp. 128-147.
Davies, N. (2003). Exploding The Myth Of The Falling Crime Rate, In The Guardianlondon.
Dickey, W.J. & Smith, M.E. (1998). Dangerous Opportunity: Five Futures For Community Corrections: The Report From The Focus Group, Department Of Justice, Office
Of Justice Programs, Washington, Dc: U.S.
Ditton, J., Bannister, J., Gilchrist, E. & Farrall, S. (1999). Afraid Or Angry? Recalibrating The 'Fear' Of Crime, International Review Of Victimology, 6, 2, 83-99.
Doble, J. (2002). Attitudes To Punishment In The Us - Punitive And Liberal Opinions
(Ch 8), In Changing Attitudes To Punishment ... Vol. (Eds, Roberts, J. And Hough,
M.) Willan, Pp. 148-162.
Durham, A.M. (1993). Public Opinion Regarding Sentences For Crime: Does It Exist?,
Journal-Of-Criminal-Justice, 21, 1, 1-11.
Durkheim, E. (1933). The Division Of Labour In Society, Ny.
Ellsworth, P.C. & Gross, S.R. (1994). Hardening Of The Attitudes: Americans' Views
On The Death Penalty, Journal Of Social Issues, 50, 2, 19-52.
Fehr, E. & Gachter, S. (2002). Altruistic Punishment In Humans, Nature, 415, 137-140.
Flanagan, T. & Longmire, D.R. (Eds.) (1996). Americans View Crime And Justice: A
National Public Opinion Survey, Sage, Thousand Oaks, Ca.
Flanagan, T.J. (1996). Public Opinion On Crime And Justice: History, Development
And Trends (Ch 1), In Americans View Crime And Justice: A National Public Opinion Survey, Vol. (Eds, Flanagan, T. & Longmire, D. R.) Sage, Thousand Oaks, Ca,
Pp. 219.
Flanagan, T.J., Mcgarrell, E. & Brown, E.J. (1985). Public Perceptions And Criminal
Courts: The Role Of Demographic And Related Attitudinal Data, Journal Of Research In Crime And Delinquency, 22, 1, 66-82.
246
Gabriel, U. & Greve, W. (2003). The Psychology Of Fear Of Crime Conceptual And
Methodological Perspectives, British Journal Of Criminology, 43, 600-614.
Gainey, R.R. & Payne, B.K. (2003). Changing Attitudes Towards House Arrest With
Electronic Monitoring: The Impact Of A Single Presentation?, International Journal
Of Offender Therapy And Comparative Criminology, 47, 2, 196-209.
Garland, D. (1990). Punishment And Modern Society: A Study In Social Theory, University Of Chicago Press, Chicago, Il.
(1997). Governmentability And The Problem Of Crime: Foucault, Criminology,
Sociology, Theoretical Criminology, 1, 2, 173-214.
(2001). The Culture Of Control: Crime And Social Order In Contemporary Society,
Oxford University Press, Oxford.
Gaubatz, K.T. (1995). Crime In The Public Mind, University Of Michigan Press, Ann
Arbor.
Giacopassi, D.J. & Blankenship, M.B. (1991). The Effects Of Criminal Justice Pedagogy On Student Attitudes, American Journal Of Criminal Justice, 16, 97-103.
Girling, E., Loader, I. & Sparks, R. (2002). Public Sensibilities Toward Crime: Anxieties Of Influence (Ch 7), In Crime: Fear Or Fascination?, Vol. (Ed, Boran, A.) Chester Academic Press, Pp. 153-176.
Grasmick, H.G. & Mcgill, A.L. (1994). Religion, Attribution Style, And Punitiveness
Toward Juvenile Offenders, Criminology, 32, 1, 23-47.
Greene, J. & Doble, J. (2000). Attitudes Towards Crime And Punishment In Vermont:
Public Opinion About An Experiment With Restorative Justice, John Doble Research
Associates, Englewood Cliffs, Nj.
Hepworth, J.T. & West, S.G. (1988). Lynchings And The Economy: A Time-Series Reanalysis Of Hovland And Sears (1940), Journal Of Personality And Social Psychology, 55, 2, 239-247.
Holloway & Jefferson, T. (1997). Doing Narrative Research Differently.
Home Office (2001). Making Punishments Work. Report Of A Review Of The Sentencing Framework For England And Wales (July 2001) (The Halliday Report), Home
Office Communication Directorate, London.
Hough, M. & Moxon, D. (1985). Dealing With Offenders: Popular Opinion And The
Views Of Victims, Findings From The British Crime Survey, The Howard Journal,
24, 3, 160-175.
Hough, M. & Parks, A. (2002). How Malleable Are Attitudes To Crime And Punishment? Findings From A British Deliberative Poll (Ch 9), In Changing Attitudes To
Punishment, Vol. Willan, Pp. 163-183.
Hough, M. & Roberts, J. (Eds.) (1996). No. 64 - Attitudes To Crime And Punishment:
Findings From The 1996 British Crime Survey, Hmso, London.
(Eds.) (1998). No. 64 Attitudes To Crime And Punishment: Findings From The
1996 British Crime Survey, Hmso, London.
Hovland, C.I. & Sears, R. (1940). Minor Studies Of Aggression: Correlation Of Lynchings With Economic Indices., Journal Of Psychology, 9, 301-310.
247
Indermaur, D. & Hough, M. (2002). Strategies For Changing Public Attitudes To Punishment, In Changing Attitudes To Punishment: Public Opinion Around The Globe,
Vol. (Eds, Roberts, J. & Hough, M.) Willan Publishing, Cullompton.
Innes, C.A. (1993). Recent Public Opinion In The United States Toward Punishment
And Corrections, Prison Journal, 73, 2, 220-236.
Irwin, J., Austin, J. & Baird, C. (1998). Fanning The Flames Of Fear, Crime And Delinquency, 44, 1, 32-48.
Jacoby, L.L., Lindsay, D. S. & Toth, J.P. (1992). Unconscious Influences Revealed:
Attention, Awareness, And Control., American Psychologist, 47, 6, 802-809.
Jayewardene, C.H.S., Lang, S. & Gainer, C. (1977). Changing Attitudes To The Criminal Justice System Through Formal Education., Crime And Justice, 5, 126-130.
Karstedt, S. (2002). Emotions And Criminal Justice, Theoretical Criminology, 6, 3,
299-318.
Korn, R. (1971). Of Crime, Criminal Justice And Corrections, University Of San Francisco Law Review, 6, 1, 27-75.
Kury, H., Obergfell-Fuchs, J. & Smartt, U. (2002). The Evolution Of Public Attitudes
To Punishment In Western And Eastern Europe (Ch 5), In Changing Attitudes To
Punishment: Public Opinion Crime And Justice, Vol. (Eds, Roberts, J. And Hough,
M.) Willan, Pp. 93-114.
Langworthy, R.H. & Whitehead, J.T. (1986). Liberalism And Fear As Explanations Of
Punitiveness, Criminology, 24, 3, 575-591.
Maloney, D., Bazemore, G. & Hudson, J. (2001). The End Of Probation And The Beginning Of Community Justice, Perspectives.
Maruna, S. (2001). Making Good: How Ex-Convicts Reform And Rebuild Their Lives,
American Psychological Association, Washington, D.C.
Maruna, S., Matravers, A. & King, A. (2004). Disowning Our Shadow: A Psychoanalytic Approach To Understanding Punitive Public Attitudes, Deviant Behavior, 25,
277-299.
Matthews, R., Hancock, L. & Briggs, D. (2003). Jurors' Perceptions, Understanding,
Confidence And Satisfaction In The Jury System: A Study In Six Courts, Home Office: Research, Development And Statistics Directorate, London, Pp. 4.
Mattinson, J. & Mirrlees-Black, C. (1998). No.111 - Attitudes To Crime And Criminal
Justice: Findings From The 1998 Bcs, Hor, Development And Statistics Directorate,
Pp. 1-4.
Mayhew, P. & Van Kesteren, J. (2002). Cross-National Attitudes To Punishment (Ch
4), In Changing Attitudes To Punishment: Public Opinion, Crime And Justice, Vol.
(Eds, Roberts, J. V. & Hough, M.) Willan Publishing, Collumpton, Pp. 63-92.
Mccorkle, R.C. (1993). Research Note: Punish And Rehabilitate? Public Attitudes Toward Six Common Crimes, Crime & Delinquency, 39, 2, 240-252.
Mead, G.H. (1964). On Social Psychology: Selected Papers, University Of Chicago
Press, Chicago.
248
Mead, G. H. T. P. O. P. J., American Journal Of Sociology, Volume 23: 577-602.

(1918). The Psychology Of Punitive Justice, American Journal Of Sociology, 23,
577-602.
Morgan, R. (2002). Privileging Public Attitudes To Sentencing, In Changing Attitudes
To Punishment: Public Opinion, Crime And Justice, Vol. (Eds, Roberts, J. & Hough,
M.) Willan, Pp. 236.
Morgan, R. & Russell, N. (2000). The Judiciary In The Magistrates' Courts, Research,
Development And Statistics Directorate, U.K. Home Office, London.
Mori (2002). Policing The Possession Of Cannabis: Residents' Views On The Lambeth
Experiment, Mori Social Research Institute.
National Probation Service (2002). Perceptions Of The National Probation Service,
Mori, England And Wales, Pp. 1-30.
Oregon Crime Analysis Center (1991). Intermediate Sanctions, Oregon Crime Analysis
Center, Salem, Or.
Petersilia, J., Turner, S. & Peterson, J. (1986). Prison Versus Probation In California:
Implications For Crime And Offender Recidivism, Rand, Santa Monica, Ca.
Pratt, J. (2000a). Civilisation And Punishment, Australian And New Zealand Journal
Of Criminology, 33, 2, 183-201.
(2000b). Emotive And Ostentatious Punishment: Its Decline And Resurgence In
Modern Society, Punishment And Society, 2, 4, 417-439.
Quimet, M. & Coyle, E.J. (1991). Fear Of Crime And Sentencing Punitiveness: Comparing The General Public And Court Practitioners, Canadian Journal Of Criminology, 33, 2, 149-162.
Ranulf, S. (1938/1964). Moral Indignation And Middle Class Psychology. New York:
Schocken.
Rex, S. & Tonry, M. (Eds.) (2002). Reform And Punishment: The Future Of Sentencing,
Willan Publishing, Cullompton.
Roberts, J., Stalans, L., Indermaur, D. and Hough, M. (2003). Penal Populism And Public Opinion: Lessons From Five Countries, Oxford University Press, Oxford; Ny.
Roberts, J.V. (1992). Public Opinion, Crime, And Criminal Justice, In Crime And Justice: A Review Of Research, Vol. (Ed, Tonry, M.) University Of Chicago Press, Chicago, Pp. 99-180.
(2002). Public Opinion And The Nature Of Community Penalties: International Findings (Ch 3), In Changing Attitudes To Punishment: Public Opinion, Crime And Justice, Vol. (Eds, Roberts, J. & Hough, M.) Willan Publishing, Pp. 33-62.
Roberts, J.V. & Hough, M. (2002). Public Attitudes To Punishment: The Context (Ch
1), In Changing Attitudes To Punishment: Public Opinion, Crime And Justice, Vol.
(Eds, Roberts, J. & Hough, M.) Willan Publishing, Pp. 1-14.
Roberts, J.V. & Stalans, L. (Eds.) (1997). Public Opinion, Crime And Criminal Justice,
Westview Press, Boulder, Co.
249
Ryan, M. (2003). Populists And Publics, In Penal Policy And Political Culture In England And Wales: Four Essays On Policy And Process, Vol. Waterside Press, Winchester, Pp. 109-140.
Savelsberg, J.J. (1994). Knowledge, Domination And Criminal Punishment, American
Journal Of Sociology, 99, 4, 911-943.
Sprott, J.B. (1999). Are Members Of The Public Tough On Crime?: The Dimensions Of
Public `Punitiveness', Journal-Of-Criminal-Justice, 27, 5, 467-474.
Sprott, J.B. & Doob, A.N. (1997). Fear, Victimization, And Attitudes To Sentencing,
The Courts, And The Police, Canadian Journal Of Criminology, 39, 3, 275-291.
Stalans, L. (2002). Measuring Attitudes To Sentencing (Ch 2), In Changing Attitudes
To Punishment: Public Opinion, Crime And Justice, Vol. (Eds, Roberts, J. & Hough,
M.) Willan Publishing.
Stead, M., Macfadyen, L. & Hastings, G. (2002). What Do The Public Really Feel
About Non-Custodial Penalties.
Sutherland, E.H. & Cressey, D.R. (1978) Criminology, J. B. Lippincott, Philadelphia.
Tarling, R. & Dowds, L. (1997). Crime And Punishment, In British Social Attitudes: The
14th Report. The End Of Conservative Values?, Vol. (Eds, Jowell, R., Curtice, J., Park,
A., Brook, L., Thomson, K. & Bryson, C.) Ashgate Publishing Ltd, Aldershot.
Taylor, D.G., Schepple, K.L. & Stinchcombe, A.L. (1979). Salience Of Crime And Support For Harsher Criminal Sanctions, Social Problems, 26, 4, 413-424.
Tufts, J. & Roberts, J. (2002). Sentencing Juvenile Offenders: Public Preferences And
Judicial Practice, Criminal Justice Policy Review, 13, 46-64.
Turner, J.C., Brown, R.J. & Tajfel, H. (1979). Social Comparison And Group Interest In
Ingroup Favouritism, European Journal Of Social Psychology, 9, 2, 187-204.
Turner, M.G., Cullen, F.T., Sundt, J.L. & Applegate, B.K. (1997). Public Tolerance For
Community Based Sanctions, The Prison Journal, 77, 1, 6-26.
Tyler, T.R. & Boeckmann, R.J. (1997). Three Strikes And You Are Out, But Why? The
Psychology Of Public Support For Punishing Rule Breakers, Law & Society Review,
31, 2.
Tyler, T.R. & Weber, R. (1982). Support For The Death Penalty; Instrumental Response
To Crime, Or Symbolic Attitude?, Law & Society Review, 17, 1, 21-45.
Vandiver, M. & Giacopassi, D. (1997). One Million And Counting: Students' Estimates
Of The Annual Number Of Homicides In The U.S., Journal Of Criminal Justice Education, 8, 135-143.
Vaughan, B. (2002). The Punitive Consequences Of Consumer Culture, Punishment
And Society, 4, 2, 195-211.
Wood, J. & Viki, G.T. (2004). Public Perceptions Of Crime And Punishment, In Forensic Psychology (Ed, Adler, J. R.), Pp. 16-36.
Vold, G. (1958). Theoretical Criminology, Oxford University Press, New York.
Von Hirsch, A. (1993). Censure And Sanctions, Clarendon Press; Oxford University
Press, Ny, Oxford.
250
Walker, N. & Hough, M. (Eds.) (1988). Public Attitudes To Sentencing: Surveys From
Five Countries, Cambridge, Uk.
Walker, N., Hough, M. & Lewis, H. (1988). Tolerance Of Leniency And Severity In
England And Wales, In Public Attitudes To Sentencing: Surveys From Five Countries, (Eds, Walker, N. & Hough, M.), Pp. 178-202.
Warr, M. (1995). Public Opinion On Crime And Punishment, Public Opinion Quarterly,
59, 296-310.
Wilkins, L.T. (1991). Punishment, Crime & Market Forces, Dartmouth Publishing, Aldershot, Uk; Brookfield, Vt.
Young, J. (2003). Merton With Energy, Katz With Structure: The Sociology Of Vindictiveness And The Criminology Of Transgression. Theoretical Criminology, 7, 389-414.
Zimring, F.E., Hawkins, G. & Kamin, S. (2001). Punishment And Democracy: Three
Strikes And You're Out In California, Oxford University, New York.
251
Political Culture and Incentives to Penal Populism
DAVID A. GREEN
Abstract
Differences in political culture, or in ways of doing politics, help to account for
differences in national appetites for punishment. The distinction between majoritarian and consensus democracy is explored in this chapter by comparing responses to two child-on-child homicides from the 1990s the James Bulger case
in England and the Silje Redergrd case in Norway. In England, politicians had
incentives to politicise the Bulger case, and were pressed to act demonstrably and
swiftly by emotive and condemnatory press coverage. The Labour Party rhetorically enlisted the Bulger case in order to indicate its new and tougher approach to
law and order and to make the party more electable in the face of rising public
concern about crime. Simplistic tabloid rhetoric went largely unchallenged by
political leaders, and the publics press-fuelled fears about crime were legitimated by traditionally left-leaning politicians eager to distance themselves from
elite expertise. These responses contrast with those to the Norwegian case, none
of which appear to be politically motivated. The consensual nature of Norwegian
politics seems to have decreased incentives to politicise the Redergrd homicide
in the ways seen in England. Explanations for differing levels of cultural and political enthusiasm for punishment are not complete until the implications of political culture are fully considered.
A substantially different and expanded version of this chapter will appear in Crime and Justice: A Review of Research, vol. 36, 2007.
252
DAVID A. GREEN
Introduction
This chapter compares the media, political, and public responses to two child-onchild homicides that of two-year-old James Bulger who was killed by two tenyear-old boys outside Liverpool, England in 1993, and that of Silje Marie Redergrd from Trondheim, Norway who was killed by three six-year-old boys in
1994. The cases played out very differently in each country, and explanations for
why they did and more generally, for why politicians in certain countries experience more powerful incentives to engage in penal populism than others
must include a consideration of the mediating roles played by different political
cultures and the practices and institutions that support them. Ways of doing politics in different kinds of democracy generate incentives for particular styles of
political behaviour. I argue that these incentives were powerfully experienced in
English majoritarian political culture in the wake of the Bulger homicide, and
that the press covered and contextualised the case in narrowly constrained ways
stressing how indicative the homicide was of an ailing English society. The Norwegian press was much more restrained in tone in its coverage of the Silje
Redergrd case, and provided a comparatively high level of expert commentary
to provide audiences with context about the rarity of such an event, as well as the
most advised approaches to dealing with the those impacted by the case, including the three boys responsible.
In the first section of the chapter I present the details of the two homicide
cases and the responses linked to them in the press, in politics and in policy.
Next, I consider the role of political culture in particular, the distinction that
Arend Lijphart makes between consensus and majoritarian democracy and
enlist this distinction as a key component in explaining why the response to the
Bulger case was so rhetorically demonstrative and punitive, and why the responses to the Redergrd case were comparatively so muted. Finally, I explore
the implications of political culture as an explanation for differing cultural appetites for punishment and suggest avenues for future comparative research. Political culture deserves more attention than it has yet received in theories explaining
why nations respond to crime events in different ways.
Responses to Two Child-on-Child Homicides

Two-year-old James Bulger was kidnapped from a shopping centre outside Liverpool on 12 February 1993 by two ten-year-old boys later named as Jon
Venables and Robert Thompson. The abduction itself, which occurred after
James briefly wandered away from his mother while she was being served at the
counter of a butcher shop, was captured on the centres security cameras and the
253
images were later widely publicized and used by police to track down the boys.
Those grainy black-and-white images became iconic and horrifying, representing
for many in Britain a momentous turning point when their faith in the innocence
of childhood was violently shaken. The two boys forcibly led James on a walk of
over two miles, battering him along the way, past at least 38 witnesses, none of
whom intervened forcefully enough to prevent what would happen. As dusk fell,
the boys stopped beside a railway line where they beat James to death with bricks
and an iron bar. They left his partially undressed body on the tracks where it was
later severed by a train. Once caught, Thompson and Venables spent nine months
in secure custody awaiting trial in an adult court, receiving no psychological
treatment during this period.
Outside the court where the accused boys first appeared, several arrests were
made after a mob gathered to hurl stones and abuse at the police van transporting
them. The two were found guilty in November 1993 and sentenced to be detained
at Her Majestys pleasure, the mandatory equivalent of a life sentence. The initial
tariff, or minimum sentence required before parole can be considered, was set at
eight years by the trial judge. In December, the Lord Chief Justice raised the tariff to ten years. In July 1994, the Home Secretary at the time, Conservative Party
parliamentarian Michael Howard, intervened, as was his prerogative until the
practice was later outlawed, and, citing public concern, set a new tariff of fifteen years. After years of legal wrangling1, the boys eventually served the original tariff and were freed on life license with anonymity in 2001 amidst a wave of
press and public condemnation.
On 15 October 1994, five-year-old Silje Marie Redergrd was out playing near
a toboggan run in a suburb of Trondheim, Norway with three six-year-old boys,
enjoying the first snow of the year. The boys were not strangers and had played
with Silje before. Though reports of what happened are hazy, the three apparently convinced Silje to undress partially, and then decided to be mean to her.
As she was held down, they proceeded in turns to punch, kick, and stomp on her.
One beat her with a stone. They left her unconscious to freeze to death. The police investigation ended abruptly when the parents of the boys responsible all
1
In the case of R v. Secretary of State for the Home Department ex parte Venables and
Thompson, the English Court of Appeal ([1997] 2 WLR 67) and a majority in the House of
Lords ([1997] 3 All ER 97) quashed the Home Secretarys decision on the grounds that he
was a government minister rather than a judge and should therefore be barred from setting
sentences. The European Court of Human Rights ([1999] T and V v. UK, December 16th)
later ruled that Thompson and Venables did not get a fair trial and it upheld it is unlawful
for government ministers to set tariffs. The practice was outlawed in the Criminal Justice
Act 2003.
254
DAVID A. GREEN
withheld permission for further questioning. Most of what is known about what
happened to Silje came from one of the boys, who in the early stages told police
he had seen older teenagers attack Silje. The three boys were never named, never
punished in any way, and no punitive sentiment can be found in any of the press
coverage of the case.
A comparative analysis of the newspaper coverage about the two cases reveals
some remarkable differences. The first is the sheer quantity of stories generated about
the Bulger case when compared with the Redergrd case. Table 1 shows the number
of stories generated about both cases that appeared in two newspapers in each country,
one tabloid and once broadsheet, in the yearlong period following each homicide. The
Times and Daily Mirror represented the English broadsheet and tabloid newspapers,
respectively. In Norway, the broadsheet Aftenposten and the tabloid Verdens Gang
(VG) were selected. In this chapter, of most interest are the intrajurisdictional responses to the cases, so only a jurisdictions coverage of the domestic case the Bulger case in the English press and the Redergrd case in the Norwegian press are
compared with any depth.
Table 1: Number of articles in one year per paper, per case
Number of articles in one year
per paper, per case
Bulger
Redergrd
The Times
(broadsheet)
256
2
Daily Mirror
(tabloid)
152
3
Aftenposten
(broadsheet)
19
16
VG
(tabloid)
30
56
The Bulger case investigation and trial proceeded under a flood of international
media attention, much of it characterized by sweeping speculations about moral
decay, the loss of childhood innocence, the erosion of traditional values, the immoral and corruptive nature of media entertainment, and the breakdown of the
traditional family. As Table 2 shows, those whose views were expressed in the
press coverage of the two homicides were very different when tabloid and broadsheet newspaper coverage is compared cross-nationally. Perhaps what is most
striking are the first two rows in the table where the percentage of items containing claims made by experts2 and those made by the public3 appear. Almost half
of Aftenpostens Redergrd case items featured the views of experts and almost
Those claims makers considered experts for the sake of this analysis were coded separately
by group and then aggregated. Disaggregated, these experts included psychologists and psychiatrists, medical doctors, social workers, academics, and non-academic researchers.
This group includes vox pop interviews, letters to the editor from laypersons, and comments
made by those not directly involved in the case.
255
40 percent of VGs did so. Only slightly more than a fifth of The Times coverage
of the Bulger case featured expert views and only 10 percent of the Daily Mirror
coverage did. An even bigger contrast is found when views of the public are examined. No public views were expressed in the Norwegian press in either the
tabloid or broadsheet coverage of the Redergrd homicide, while such views
were featured as prominently in The Times coverage of the Bulger case as expert
views were. Public views were expressed in one quarter of the Mirrors coverage
of the Bulger case two-and-a-half times more prominently than expert views.
So in the English coverage of the Bulger case, few experts were consulted to
provide the context required for the reader to assess how indicative the case was
of the social and moral breakdown it was said to represent. Most of the context
offered by the press was instead self-referential pointing not to research findings or practitioner knowledge but to other media coverage to indicate the state
of the nation. Of course, this is a highly suspect barometer. Coverage of crime in
the press often follows the law of opposites (Surette, 1998), presenting highly
distorted and exaggeratory depictions of the real extent of crime.
Table 2: Domestic case items with views by claims-makers (%)
Claimsmaker
groups
Experts
Public
Police
Clergy
Victims
Perpetrators
The Times
(broadsheet)
22
21
19
10
11
7
Daily Mirror
(tabloid)
10
26
27
8
17
12
Aftenposten
(broadsheet)
47
0
53
7
13
20
VG
(tabloid)
39
0
22
7
17
15
Because most of the informational resources that citizens draw upon to formulate
opinions and attitudes are gleaned from the mass media, the types of claims and
claims makers citizens are exposed to via the media help to shape their beliefs,
opinions, and attitudes. We know, for instance, that those who rely upon the tabloid press for their news tend to be more fearful of crime and less informed about
its prevalence (Fletcher and Allen, 2003; Newton, 1997; Schlesinger et al.,
1991). This is not surprising when the tabloids coverage of crime typically excludes dispassionate, fact-based analyses in favor of memorable and emotive
commentary. One can hardly expect otherwise when, as the Bulger case tabloid
coverage indicates, articles rarely featured the informed views of experts to confirm, contradict or qualify with evidence the prevalent claims from the press and
politicians that the Bulger case was a meaningful omen of the times.
256
DAVID A. GREEN
In contrast, the opinions of the types of experts that the English press coverage
either overlooked or willfully ignored were prominently featured from the start in
the Norwegian coverage of Siljes death. This difference in the comparative legitimacy of expert opinion seems to reflect a modernist optimism in Norway that
has been largely eroded in England. Even the Norwegian tabloids drew heavily
upon expert discourses to legitimate the themes they presented, while the English
papers, most actively in the tabloids but also in the broadsheets, tended instead to
relegate opinions by these same expert groups to positions secondary even to
those of the so-called man on the street. The Norwegian press also featured interviews with families of the victims and perpetrators alike, and there was no evident rush to assign blame for what was, by all available accounts, a tragedy. According to one reporter and his editor at the leading tabloid VG, both of whom
were interviewed for this study in Oslo in 2002 (Laure and Haavik, 2002), press
coverage of the case stopped once the mayor of Trondheim appealed to the assembled reporters to leave the community and to give them peace. The three
boys were swiftly reintegrated into the community under the close supervision of
the hjelpeapparat, a team of social workers and psychologists. In partnership
with the local police, the hjelpeapparat appealed for calm, met the needs of those
affected by the case, and invalidated gossip and rumours. The boys were even
offered early places in kindergarten to facilitate the reintegration. The press coverage in Norway suggests that the event required careful consideration and the
application of expert knowledge and experience. If press coverage is any indicator, similar events and issues seem not to warrant the same degree of careful consideration by elites in England. As one tabloid columnist put it, God protect us
from the ologists because their hackneyed perception is dangerous (Daily
Mirror, 1 December 1993). When elite expertise is stripped of credibility, as
suggested by the Bulger case press coverage in England, issues of this kind are
more open to public scrutiny, critique, and commentary.
English politicians from all political parties were apparently compelled to
comment on and to contribute to rising press and public concern about youth
criminality in the immediate wake of the Bulger homicide, especially regarding
the possible effects of violent media on behavior. This tendency continued
throughout the nine months leading to the trial of Thompson and Venables and
into the period following their conviction and sentencing. The Bulger name was
mentioned in the House of Commons in 28 separate debates between March
1993 and June 2001, and in the House of Lords in 21 debates. There was virtually no comparable Norwegian political commentary about the Redergrd homicide, save for brief and comparatively reasoned sets of remarks made by the cul-
257
ture minister confirming that a bill would be brought before the Storting, the
Norwegian parliament, to subject video rentals to the same regulation as films
shown in cinemas (Snstelie, 1994) and by the childrens minister arguing
that without the cooperation of international media firms, the government was
powerless to limit the flow of violent television shows broadcast in Norway
(Solberg, 1994). Siljes name was never invoked in the Storting.
The concern catalyzed by the Bulger case in Britain was also evidenced in the
general tide of condemnatory attitudes toward criminals evident in the rhetoric at
the October 1993 Conservative Party conference and the increasingly punitive
New Labour law and order platform of the time. It shaped the Criminal Justice
and Public Order Act of 1994, which doubled from one year to two the maximum
sentence in a young offenders institution for fourteen to seventeen year-olds and
which called for secure training orders to make it easier to lock up more persistent ten to fourteen year-old juvenile offenders in secure units. The notion of
the child as doli incapax meaning a child between the ages of ten and fourteen
is so young as to be incapable of doing wrong was a legal presumption of innocence in England and Wales which the prosecution was charged to rebut to
obtain a conviction. This presumption was abolished in the Crime and Disorder
Act of 1998 in the wake of the Bulger case after the newly elected Labour government decided that it flies in the face of common sense (Home Office, 1997).
The Bulger case also played a role in the interest shown by some British politicians in American-style boot camps for young offenders, in calls for the return of
corporal and capital punishment4, and in the striking increases in adult prison
admissions immediately following the Bulger murder (see Newburn, forthcoming). The number of young offenders under the age of 21 in custody rose by 30
percent between 1993 and 1996 (Moore, 2000, p. 115). Figure 1 shows how public concern about crime (MORI, 2002) spiked very sharply when James Bulger
was killed, and the prison population, having been in decline, began to rise once
again at a time when crime was actually beginning to fall.
That the name of James Bulger is still so indelibly inscribed on the British
consciousness is due in part to the confluence of a number of political contingencies. One of these is Tony Blair, without whom significantly less might have
been made of the case. Blair was, at the time of the murder, shadow home secretary, and was charged with articulating the Labour Partys law-and-order strategy
in the wake of four consecutive election defeats by Conservative governments. It
also happened at a time when Blair was intent to show that so-called New Labour
4
Most of the Norwegian press stories about the Bulger case focus on these punitive aspects of
the responses from English politicians.
258
DAVID A. GREEN
had begun to break with the penal-welfarist liberalism that some believed had
rendered the party unelectable in the 1980s and early 1990s climate of rising
crime and increasing public concern about it.
Figure 1: Prison population and public concern about crime in England and
Wales: January 1992-November 1994
In the wake of the Bulger murder, Conservative Prime Minister John Major gave
an interview to the Mail on Sunday, a conservative mid-market tabloid, in which
he called for a crusade against crime and argued that society needs to condemn
a little more and understand a little less (Observer, 28 February 1993). Blairs
choice of news outlets used to articulate his own responses reflected another part
of his get-tough, law-and-order strategy. Uncharacteristically then for a Labour
politician, Blair courted and targeted the tabloid press in a way that Labour politicians had not done before, writing pieces for those newspapers that Labour had
previously trivialized (Hargreaves, 2003; Seldon, 2004). Blair wrote pieces in the
tabloids that echoed Majors sentiments and the tabloids claims. He contended
that there is something very wrong and very sick at the heart of our society
(Daily Mirror, 22 February 1993) and the murder was symbolic of a deep malaise in our country (Sun, March 3, 1993) which required a tough, morally resolute
response (Downes and Morgan, 1997, 2002). Blair equated the rarity and horror
259
of the case with lesser offences and stoked growing fears by claiming that what
also shocks us is knowing that minor versions of the same are happening almost
every day (Daily Mirror, 20 February 1993). He claimed authorship of a piece
in the tabloid Sun newspaper in which he legitimated rather than refuted or qualified the exaggerated tabloid claims that juvenile offending had become more prolific and more sinister: We can debate the crime rate statistics until the cows
come home. The Home Office says crime is falling. Others say it isnt. I say
crime, like economic recovery, is something that politicians cant persuade people about one way or another. People know because they experience it. They
dont need to be told. And they know crime is rising (Sun, 3 March 1993).
Though his critics might say that Blairs direct engagement with and legitimization of the hyperbolic concerns of tabloid editors and their readerships was cynical and would render considered and deliberate approaches to sensitive crime
policy questions far more difficult to achieve in future, that he felt compelled to
do so was understandable given the considerable political influence of the tabloid
press in Britain.5
The Bulger murder provided Blair an opportunity to articulate Labours reworked approach to law and order at a time when the security camera images of
the abduction appeared all over the news, and as the country as a whole struggled
to understand how two young boys could have been capable of such a horrible
and vicious act. Blair had begun using his tough on crime, tough on the causes
of crime rhetoric before Bulger was killed, but the case ensured him an attentive
national audience. It is said that Blairs performance during this period ensured
his place as future Labour Party leader (Rentoul, 2001).
Responses in the aftermath of the Bulger and Redergrd cases have continued
to diverge in the years since the murders. With all the public attention and press
opprobrium that followed in the aftermath of the Bulger case, perhaps what is
most surprising is, first, that both killers have been released and, second, that the
lifelong injunction preserving their anonymity has yet not been seriously violated.6 Thompson and Venables were released in 2001 and were provided with
new identities for fear of vigilante reprisals. The parole boards decision to re5
Conservative Party treasurer, Lord McAlpine, credited the tabloids with swinging the 1992
general election toward the Tories, and the Sun in turn printed the headline Its The Sun
Wot Won It (Sun, 12 April 1992). Labour Party leader Neil Kinnock himself considered the
tabloids to have been crucial, a view supported by post-election MORI poll research and a
Labour Party inquiry (Jones, 2001).
The Manchester Evening News was fined on December 2001 for a significant but nondeliberate breach of the injunction after it published details suggesting the whereabouts of
Thompson and Venables (Guardian, 5 December 2001).
260
DAVID A. GREEN
lease them was met with disbelief by many, and once again the case became
front-page news. In the thirteen years since James Bulger was killed, the case has
been mentioned in news stories on topics as varied as other similar and dissimilar
murders, media censorship, parental responsibility, the unjust leniency of rehabilitation, and the erosion of British sovereignty by European Union human
rights law. The English tabloids continue to mine the Bulger case story. For instance, the Sun and Mirror have recently run conflicting stories variously claiming that Devil Dad Robert Thompson fathered a child with a girlfriend who is
ignorant of his past, that he has moved in with his gay lover, and that he is a heroin addict.7
In contrast, the stories since generated about the Silje Redergrd case have
been few. They have focused on how Beathe Redergrd has forgiven her daughters killers, the efforts she has made to comfort the families of other murder victims, her refusal to accept half a million kroner (around Euros 62,000 or $75,000)
blood money in her words (Snstelie, 1995) from a British tabloid in exchange for an exclusive on her story and, in the British press, her support for raising the age of criminal responsibility in England and Wales from 10 to 15, the
age in Norway.
Ways of Doing Politics

That Silje Redergrds name was never mentioned in the Storting and has faded
from collective memory while James Bulgers became a cause clbre begs to be
explained. Why were English journalists and politicians so vociferous in their
condemnation of the perpetrators and so responsive in policy terms while their
Norwegian counterparts remained so reserved? There are many plausible components of this explanation, including the difference in age of the killers, the differing ages of criminal responsibility and the attendant legal implications, particularities of the news media markets, and the reactions of the victims families (see
Green, forthcoming), but I concentrate here on only one differences in political
culture and practice. I focus most energetically on political culture because it has
considerable explanatory power and it requires more theoretical attention than it
has yet received within criminology. Striking differences in how politics is done
in Britain and Norway, particularly the distinction between majoritarian and consensus democracy, aid significantly in accounting both for why the Bulger and
Redergrd homicides produced such different outcomes, and why penal populism
has such appeal in some countries more than others.
7
Other, more reliable accounts suggest that both Thompson and Venables responded very
well to their rehabilitation and that they also adjusted well to their release (Smith, 2006).
261
The agenda-setting approach (McCombs, 1981, 2005; McCombs and Shaw,

1972) to media research is concerned with the ways in which the media tell the
public not necessarily what to think, but what to think about (Cohen, 1963) and
how to think about them. A political-cultural perspective is employed with similar aims. Described in functional terms, political culture defines the range of
acceptable possible alternatives from which groups or individuals may, other circumstances permitting, choose a course of action Its explanatory power is primarily restricted to setting the agenda over which political contests occur
Culture is unlikely to be of much help in explaining why alternative A was chosen over alternative B but it may be of great help in understanding why A and
B were considered, while no thought was given to C, D, or E. (Elkins and
Simeon, 1979, p. 131, 142).
The basic assumptions underpinning cultural concerns establish the repertoire
of policy options, as they reflect policy actors shared understandings of the
problems they face at a specific time and place. Culture in this way sets the
agenda by pre-selecting the range of available policy options from which policymakers choose.
The assumptions that comprise political culture are dispositions that interact
with structural factors, political practices, and individual agency. Elkins and
Simeon go further to circumscribe the explanatory power of political culture by
limiting its employment to comparative analyses. It can only account for differences in behavior when individuals in similar positions in an organization share a
different set of assumptions than either their counterparts in another nation or in
another time. An explanation is cultural if interjurisdictional differences in similar institutions can be accounted for by referring to divergent assumptions.8 The
ways in which tabloid journalists report similar crime events in different jurisdictions, or the way police respond to those events, are two examples. These distinctions help to hone the definition and utility of a concept that has been loosely
employed in the criminological literature.
Political culture is permissive rather than deterministic (Elkins and Simeon,
1979, p. 133). It permits or constrains the volitions of actors but cannot dictate decisions on its own. Political culture determines the range of policy alternatives
8
Elkins and Simeon argue that culture can only explain behavioral differences when structural variables are first eliminated as explanations, and behavior cannot be predicted based
on knowledge of cultural assumptions alone. Therefore, political culture is a secondorder explanation that cannot be used, on its own, to explain political behavior: The cultural assumptions provide the lens through which these more proximate political forces are
assessed; they influence what kind of interpretation will be placed on political forces, but
alone they cannot account for the result (Elkins and Simeon, 1979, p. 140).
262
DAVID A. GREEN
available in a given jurisdiction and operates at the macro level, preceding the
micro-level constraints encountered by policymakers in their day-to-day decision-making. Political-cultural constraints are in this way pre-rational; they restrict the choice of policy options even before any rational choice considerations
arise. If we understand culture to be a set of cognitive constraints (Stokes and
Hewitt, 1976, p. 837), cultural sensibilities help align action in ways that are cognitively congenial with governing norms. The pre-rational selection process is
usually unconscious, and is governed by norms, assumptions, and sensibilities
that, while not readily made apparent by those who subscribe to them, nonetheless shape perceptions and behavior. Cultural assumptions tend to remain unexamined, and it is only by comparing existing filters with those that differ that the
contingency of cultural distinctiveness is revealed.
The study of policymaking, and of the forces that conduce to influence it, is essentially the study of constraint and choice. Policymaking at the micro-level is
where the processes that concern most policy analysts occur. It is there that the
particular ideologies, interests, and information that guide policy choice decisionmaking come into play (Weiss, 1983). The micro-level is where the electoral consequences of particular policy decisions must be considered, where political selfinterest influences action, where the already bounded array of policy options must
be narrowed down and choices made. Therefore, the use of rational choice theories
to explain political decision-making on the micro-level are not incompatible with
theories of political culture because the latter works as an unconscious filter to accumulate culturally appropriate policy alternatives and to reject others that are not.
For instance, the penal detention of young children is incompatible with Norwegian cultural sensibilities. The kinds of policies intended to address youth offending in England and Wales like abolishing doli incapax and making it easier to try
children in adult courts and to lock them up when convicted are excluded from
consideration in Norway. At a pre-rational stage, certain policy options are preselected for consideration at later stages and others never pass the test of cultural
appropriateness and are discarded out of hand. The concept of political culture is
meant to convey this process of pre-rational filtration.
Along with Street (1994), Elkins and Simeon expand the definition of political
culture to include a range of superficially nonpolitical assumptions. Some of the
assumptions they identify as important to the shaping of political culture include:
1. Assumptions about the orderliness of the universe
2. Presumptions about the nature of causality. Is the world random? Are events
foredestined? Inevitable?
3. What are the principal goals of political life?
263
4. Should one try to maximize gains, or to minimize losses? In other words,

what assumptions are made about the relative payoffs of optimistic or pessimistic strategies?
5. Who belongs to one's political community? Is it a vaguely bounded community, or is it one marked by sharp we-they distinctions? Do the boundaries
vary with types of situations, or are they more or less unchanging? To whom
or to what does one owe any obligation?
6. What types of events, actions, or institutions are deemed political (as opposed
to economic, social, etc.), or is a sharp line drawn at all? Is the political a
positively or negatively valued domain?
7. Assumptions about others their trustworthiness, public spiritedness, etc.
and about how one should relate to them. (Elkins and Simeon, 1979, p. 132).
Additional unmentioned but important assumptions might concern the usefulness
of social scientific research and the expertise of its practitioners, the deference
deserved by elites, and the dominant idealized notions of justice, as in the appropriateness of vengeance versus forgiveness.
In responding to a child-on-child homicide, presumptions about causality will
likely affect the ways in which blame and responsibility are apportioned.
Whether or not events have systematic causes or whether they are largely fortuitous or accidental (Elkins and Simeon, 1979, p. 132) will have implications for
the sort of response deemed appropriate. In Britain, blame was apportioned liberally after the Bulger killing, in many directions at once. Many forces were implicated as responsible, suggesting that the crime was preventable. Furedi (1997)
reports that the British Medical Journal no longer allows the word accident to
appear in it volumes, as its editors believe most injuries to be preventable. Like a
preventable injury, that a preventable crime occurred insinuates that someone is
ultimately to blame. In contrast, there was no apparent rush to attribute blame for
the events that led to Siljes death, save for the focus by some on the influence of
violent television.
Assumptions about the trustworthiness of individuals, and about whether people are basically good or evil, are also important as they imply the role for punishment and forgiveness. Norways religious, racial, and cultural homogeneity
might play some role in facilitating inter-party political cooperation and reducing
the kind of suspicion and othering witnessed in Britain when James Bulger was
killed. With fewer cultural and political cleavages, fundamental attitudes and
values are less often perceived to be in conflict. These broadly shared cultural
assumptions will affect the ways policy makers interact and their style of behavior; what forms of behavior are legitimate; what criteria (scientific, religious,
etc.) should be applied to policies; how the policy field is perceived (as consen-
264
DAVID A. GREEN
sual or conflictual, zero- or variable-sum); [and] who is permitted to participate

(Elkins and Simeon, 1979, p. 142).
There is evidence to suggest that different ways of doing politics have consequences for how justice is achieved and the perceived legitimacy of the political
system. The strong we-they distinctions between political parties, as are made
in Britain, appear to create incentives both for inter-party conflict and for politicizing criminal events as tools to undermine the power and credibility of the
party in power. Only very rarely do British politicians cede credit for successes
to rival parties. In contrast, Norwegian political culture is less divided and less
divisive so as to minimize incentives to politicize criminal events to the extent
witnessed in countries like Britain and the United States. This fundamental difference in ways of doing politics distinguishes two kinds of democracy majoritarian and consensus.
Paraphrasing Almond and Verba (1963, pp. 14-15), Lijphart (1998, pp. 99-100)
defines political culture as a pattern of cognitive, affective, and evaluational orientations toward political objects among the members of a group, or, in simpler
words, a groups pattern of political beliefs, feelings, and judgments. He argues
that consensus democracy is in effect a set of structures and institutions, a pattern
of institutional characteristics that particular democracies have (1998, p. 100),
arranged in ways to produce the maximum amount of consensus between disparate
stakeholders.
Lijphart (1998) points out that his mentor, Gabriel Almond, made similar distinctions when classifying democracies in his 1956 article Comparative Political
Systems. Like Lijphart, Almonds structural and cultural classification discovered geographical patterns, representing Anglo-American and Continental
European systems. Lijphart (1998, p. 106) argues: To my mind, this geographical pattern strengthens the case for cultural explanation: I think that there is indeed, as Almond suggested more than forty [now 50] years ago, a major difference between the political cultures of the American and Continental European
worlds, and that this cultural difference manifests itself inter alia in the different
forms of democracy that are practiced in these two worlds. Lijphart has studied
forms of Western democracy on five continents spanning a period of 150 years,
and he now distinguishes between two types majoritarian and consensus democracy9 (Lijphart 1998, 1999). Consensus democracies like Norway and the
9
The democratic approach now known as consensus democracy has undergone a series of
name changes at the hands of Lijphart over the years. In earlier incarnations it has been
known as the politics of accommodation (1968), consociational democracy (1969), and
as power-sharing (1985), each term generating conceptual problems that subsequent relabelings were meant to solve.
265
other Nordic countries tend to have coalition or minority governments, relatively

strong legislatures rather than the strong executives of majoritarian democracies,
multi-party systems, proportional election systems, and corporatist or coordinated interest group systems (instead of free-for-all competitive pluralism among
interest groups) as found in Britain and the United States (Lijphart, 1998,
p. 103). In short, the majoritarian model of democracy is exclusive, competitive,
and adversarial, whereas the consensus model is characterized by inclusiveness,
bargaining, and compromise (Lijphart, 1999, p. 2). Though he recognizes his
distinction to be on a continuum, and is aware of the danger of blurring the
smaller and important distinctions that exist between the wide range of nations
included in his two categories, he does, however, provide a useful dichotomy to
apply to the study of democracies.
What most clearly distinguishes a consensus democracy from its majoritarian
cousin is the way political actors in each choose to interpret the common definition of democracy government by and for the people (Lijphart, 1999, p. 1).
Majoritarian democracies interpret this to mean that the resolution of the conflict
of citizen preferences ought to be done by majority sentiment. The interests of
the majority determine the course of the policies formulated by their representatives in parliaments and legislatures. In contrast, the consensus model interprets
the meaning of the same definition of democracy to mean, as Lijphart (1999,
p. 2, emphasis in original) indicates, as many people as possible. This is the crux
of the consensus model. It does not differ from the majoritarian model in accepting that majority rule is better than minority rule, but it accepts majority rule only
as a minimum requirement: instead of being satisfied with narrow decisionmaking majorities, it seeks to maximize the size of these majorities.
The ways in which this majority is maximized in consensus democracy has
implications for the stability of the political system, the durability of its policies,
and the trust citizens hold in government and in their fellow citizens. Put simply,
in a political system like Norways, in which power is shared and dispersed
among many players, in which inclusiveness is the rule, and in which opportunities for open deliberation are fostered, there are fewer excluded and discontented
players to attack the governments policies. Potential tensions are diffused and
bridges are built between disparate parties that often would be culturally and
structurally prohibited from cooperating in majoritarian democracies. The inclusive power sharing of the consensus model creates fewer incentives for opposition party members to seek power through conflict. As a result, the policies developed appear to suffer less official criticism and last longer. This contributes
generally to a cycle of more effective policymaking that retains greater public
confidence.
266
DAVID A. GREEN
All of this contrasts with the adversarial model of British or American politics
which is characterized by sharp partisan divisions and inter-party derision. These
majoritarian models are premised on a set of assumptions that run counter to
those underpinning consensual power sharing. The object of party politics is not
share but to remove power from the opposition. This creates incentives for politicians to increase their chances of obtaining power by attacking opponents from
other parties and exposing any flaw in current or planned policies. This conflict
model interacts well with the demands of the news media that are reliant to a
large extent upon conflict to attract consumers (Fuchs and Klingemann, 1995). It
is hard to believe the incessant attack upon existing policies combined with the
concomitant promises of reform and improvement that are in turn attacked in
another raft of criticism does not have effects on the publics confidence in
government.10
Lijphart (1999, p. 6) usefully reminds us that a strong political opposition
whose aim is to become government is not, as others assume (Lawson, 1993,
pp. 192-93), the sine qua non of contemporary democracy. The belief that it is
survives because our cultural criteria of what constitutes a good democracy are
influenced by our acculturation within majoritarian democracies. As Zedner
(1995, p. 518) puts it, exposure to other possible ways of seeing and shaping the
world not only excites us out of the torpor of parochialism but demands that we
regard our domestic topography anew. Comparing political cultures exposes our
own hidden assumptions and reveals new ways of conceiving of political systems.
By applying these political cultural frameworks to the two homicides, the two
countries vastly different reactions become more easily reconcilable. In the political context in Britain at the time, incentives to exploit the Bulger homicide for
political gain were considerable. This politicization had knock-on effects. Traditionally left-leaning politicians, like Tony Blair, legitimated simplified tabloid
discourses and discounted the kind of expertise which might have been able to
assuage growing public concern that Britains moral health was in swift decline.
This failure to contest simplistic, doom-laden rhetoric contributed to the wider
perception that crime was rising and getting more serious, that children were becoming more precocious and thus were more culpable than before for their
wrongdoing, and that punishment and moral exorcism were the most appropriate
responses. In contrast, far fewer of the same incentives faced Norwegian politi10
Similarly, it is also possible that the Labour governments current preoccupation with public
confidence may actually serve to undermine it by constantly calling attention to its apparent
absence.
267
cians when Silje was killed. Journalists and editors deferred to elite expertise
when contextualizing the case. In so doing, they legitimated such expertise, and
were swift to dampen concern by placing the homicide in perspective. Sensationalized press coverage was virtually non-existent.
Future Research Into Penal Populism and Democracy

What does this comparison imply for future research? What testable hypotheses
can be offered which might help enrich our understanding of differing penal cultures? This analysis suggests that future research ought to address the importance
of government legitimacy, public confidence, and trust in others. For instance, is
blame apportioned more eagerly in certain types of political cultures than in others? Is there something about the English way of doing politics that incentivizes
politicians to assign blame more enthusiastically? Can political cultures and
structures foster particular attitudes about the nature of humanity? Do particular
types of political culture and their supporting structures work to undermine citizens trust in government and in each other? Do other types foster this trust?
Garland (1996, 2001) builds the case that expressively punitive crime control
and penal policies have been used in the USA and Britain as a means of offsetting the effects of late modernity, including declining legitimacy and weakening
national sovereignty. This thinking is based upon a range of crisis theories advanced in the 1970s claiming crises of legitimation, governability, and democracy: Although the ideological and theoretical bases for these crisis hypotheses
varied greatly, their diagnoses concurred in one essential point: the demands
made by citizens on democratic governments were increasing, and doing so irreversibly, while, at the same time, the capacity of governments to realize their policy objectives was declining due, among other things, to lower economic growth
(Fuchs and Klingemann, 1995, p. 5). Crozier, Huntington, and Watanuki (1975)
argue that the crisis of democracy is caused by two factors: growing citizen demands and a competitive party system which forces parties constantly to outbid
one another in terms of their policy programmes, thus inflating the demands of
citizens (Fuchs and Klingemann, 1995, p. 6). In majoritarian democracies, it
appears that these kinds of power struggles yield a cycle of only pyrrhic victories
where one side strips power from another, only to be left to deal with the gap in
confidence that was rendered in the quest to seize power in the first place. Anything gained by the compulsion to do something in the face of public, press or
opposition party criticism is undercut by the implicit admission in acting that
previous efforts hitherto were inadequate. It also legitimates the critiques that
triggered the remedial action.
268
DAVID A. GREEN
Table 3 compares levels of public confidence in a number of institutions in

Norway and Britain. With public confidence in government institutions as low as
it is in many majoritarian countries, reformist agendas premised upon strong critiques of current policy and practice are likely to garner widespread public support, at least initially. Partisan concerns to win elections and to dominate legislatures and parliaments demand, especially when public confidence is low, that
opposition politicians begin their campaigns to wrest power from the leading
party by first exposing the governments failures. This can have at least five
plausible effects: First, it can further decrease the publics confidence by confirming the belief that the government has indeed got it wrong. Second, it can
legitimize the practice of this kind of partisan politics which seeks one partys
dominance over others at the cost of overall legitimacy for all political institutions. Third, this in turn can perpetuate the practice, and a pattern is established
whereby the perceived effectiveness of the government in power is consistently
undermined by opposition parties that expose government shortcomings, and in
so doing, further undermine public confidence. Fourth, these practices can practically necessitate political action that promises to do something else in the name
of building public confidence. Fifth, as Roberts (2002) has pointed out, such
promises are likely to raise public expectations and to raise further the stakes of
failure. These are all issues of political culture, subject to change over time but
strongly resistant and durable.
Table 3:
Norway
Britain
Public confidence in selected institutions (%)

Press
Police
1982
Legal
System
84
89
Parliament
78
Civil
Servants
58
41
1990
75
43
88
59
44
1981
66
29
86
40
48
1990
54
14
77
44
44
Source: European Values Survey (1981, 1990) cited in (Listhaug and Wiberg, 1995)
To pull these strands together, it appears that the high level of trust that Norwegians have in their political institutions, and in elite expertise in general, is an
important factor conditioning their more deliberate approach to crime (see Table 3). Accounts of late-modern penal landscapes, in majoritarian countries especially, suggest that the current state of affairs is characterized by increasingly
powerful incentives to politicize crime and disorder in ways that render them
more difficult to address deliberately with effective and considered policies.
269
These incentives exist in part because, as assessed and reported by the mass media, the various publics have little confidence in the ability of leaders to deal adequately with crime. Politicians then must make overtures, often over-dramatic
ones, to these publics in order to assuage their concerns and to meet their demands. Setting aside for a moment the distorting effects that the news media tend
to have on issues of crime and punishment including their failure to contextualize the events on which they are so adept at focusing public and political attention the penal-populist incentives that politicians must manage are driven most
fundamentally by a lack of public trust and public confidence. Inter-party political contests and heroic promises to deal more effectively with crime or to
tackle it, in British parlance are aimed at building the publics confidence in
the governments abilities to do what it says it will do.
The relationship between eroding trust and the rise of penal populism trust in
government, trust in the criminal justice system, trust in particular criminal justice practitioners, and trust between and among citizens themselves needs to be
better understood and should be a central concern for researchers. Building on
the theories of Christie (2000) and Garland (2001), Balvig (2004) summarizes
the existential revolution, evident in Denmark as elsewhere, from the rational
and optimistic modern paradigm to the late modern paradigm characterized
by declining faith in the welfare state and increasing levels of individualization
and estrangement. It is plausible then that increasing trust among citizens would
militate against those conditions that incentivize penal populism. Generalized
trust, or the trust people have in strangers, is, in Uslaners (2002, p. 1) words, the
chicken soup of social life and the engine of social capital. Social capital appears, among other things, to be inversely related to crime levels, imprisonment
rates, and to political cynicism. Some suggest social capital can be engendered
through public engagement exercises. If so, several additional research questions
seem prudent. First, are levels of trust and penal populism inversely correlated
across nations? Second, can we increase public trust in the criminal justice system by allowing members of the public to deliberate about issues central to it?
Third, can instruments of deliberative democracy including Deliberative Polls,
Planning Cells, Citizen Juries, and Consensus Conferences (see Ackerman and
Fishkin, 2004) pay dividends in terms of higher levels of public confidence and
trust?
Broader cultural differences than those shaping just political behavior undoubtedly account for some of the disparity in the responses to the Bulger and
Redergrd homicides. It is possible that the reintegrative ways the killers were
approached in Norway in the aftermath of the homicide simply reflected cultural
270
DAVID A. GREEN
norms and assumptions that could not countenance the punishment and exclusion
of children so young, even after the perpetration of so grave an act. However, as
culture and structure are mutually supporting and interactive, it would be unwise
to reject the possibility that structural features also helped to facilitate the more
compassionate and forgiving outcome. It is likely then that structural adjustments
to facilitate more consensual and cooperative ways of doing politics in majoritarian jurisdictions could work to diminish some of the damaging populist inducements that characterize majoritarian democracies. Barkers (2006) recent
work comparing different ways of doing politics within jurisdictions in the
United States shows how political culture and political practices have direct and
pronounced impacts on the viability of imprisonment over time. Cavadino and
Dignan (2006) have similarly explored the penal tendencies of particular types
of democratic regime in an international comparative context. More work in this
vein is required to gain a better understanding of the interaction between political
culture and practices and penal policy.
Conclusion
The explanations for why the Bulger and Redergrd homicides played out in
ways so diametrically opposed are complex, but the aim of this chapter has been
to begin to address an oversight. It draws attention to political-cultural explanations to account for differences in punitive appetites, explanations that few
criminologists have explored. Ways of doing politics complicatedly interact with
the ways that media cover and contextualize crime. These often antagonistic relationships are nonetheless mutually supporting, and they seem to conduce to cultural and political attitudes in citizens that in turn support the status quo. It seems
that distrustful English citizens expect and value a news media that is aggressively antagonistic toward and distrustful of government. This distrust, on emotive crime and punishment issues at least, extends to those who call themselves
experts the woolly-minded ologists vilified in the tabloids and is rhetorically reinforced by opposition party politicians eager to discredit those in government and regain power for themselves. In contrast, the apparently more trustful and deferential Norwegian citizen still retains considerable trust in fellow
citizens, elected leaders, and elite expertise. This trust is culturally fortified by
institutions which reinforce it, both through the sharing of power and the inclusion of dissatisfied dissenters.
The majoritarian-consensus continuum is offered here with an important caveat. It is not intended to categorize democracies monolithically, and there is no
shortage of examples one might employ to contradict the usefulness of the dis-
271
tinction. For example, despite its inclusion in the majoritarian end of the continuum, Canada appears to have withstood the pressures of late modernity without
resorting to increases in imprisonment, in part by talking tough (through toughsounding, expressive legislation, including mandatory minimum sentences and
higher maximum penalties) but acting more considerately (through severely limiting the numbers of offenders subject to them) (see Doob and Webster, 2006).11
In the opposing direction, Downes and van Swaaningen (forthcoming) argue that
the politics of accommodation that characterizes Lijpharts consensus distinction
has recently unraveled in Holland, meaning perhaps that it can no longer be said
to buck the Garlandian trend toward penal dystopias (Zedner, 2002). In addition, a growing body of Scandinavian criminological literature suggests the Nordic countries are also susceptible to the late-modern appeal to law-and-order politics (Balvig, 2004; Estrada, 2004; Kyvsgaard, 2001; Larsson, 1993, 2001; Mathiesen, 1996, 2003; Snare, 1995; Tham, 2001; Victor, 1995; von Hofer, 2004).
These examples illustrate the dangers of making sweeping generalizations
about the incentives for and against penal populism based on Lijpharts distinction. It is presented here cautiously and in full knowledge that just as consensus
democracies may fall victim to bouts of penal populism, so too might some majoritarian countries be better able to weather those conditions that impel others to
react with penal populism. The hypothesis suggesting that consensus political
culture creates fewer incentives for politicians to adopt tough-on-crime postures
does not suggest that such postures are unknown in consensus democracies. Instead, it suggests that countervailing political-cultural values minimize these incentives. The intention here is only to draw attention to those conditions in consensus democracies that seem to work as a windbreak preventing the force of the
winds of late modernity from blowing through at full strength.
Balvig (2004) argues that the paradigm shift away from modernist conceptions
of crime and punishment has proceeded in Denmark regardless of the parties in
power, and the more punitive, late-modern approach replacing it has arisen independently of changes in crime rates, or even the publics fear of crime. If there is
11
The wisdom of this divide between tough discourse and soft policy can be doubted. The
discovery of a disparity between what is said to be done and what is done can only perpetuate incentives for politicians to make more exaggerated claims that further raise public expectations and, without the attendant changes in practice, increase public disappointment.
The disparity would not be such a potential problem were there not so many overtures being
made to the public, with politicians wooing them amid indications of low levels of public
trust. Though some progressive penal reformers might believe doing good by stealth
(Drakeford and Vanstone, 2000, p. 377) is both necessary and appropriate in todays latemodern penal climate, it is also possible to view it as simultaneously contributing to a vicious circle.
272
DAVID A. GREEN
something unstoppable about the late modern march toward the greater individualization, more estrangement, and increasing atomization that seem to be linked
with appetites for higher incarceration rates, it is important that we understand
why its progress has been slowed in the Nordic countries. It seems highly plausible that the particularities of consensus democracies have played a role by limiting the incentives to politicize crime in the ways seen in more divisively partisan
jurisdictions like Britain and the United States. That the late modern march has
been slowed also suggests that institutional aspects of those political systems
might be explored to consider what specifically could be adopted from them to
limit the damages wrought by the more odious aspects of late modernity. If trust
in strangers is in decline because of estrangement, then it is useful to find ways to
build trust between citizens. If confidence in government is flagging, then it
seems wise to find ways to make governments more responsive and more legitimate (see Barabas, 2004; Green, 2006; Jacobsson, 1997).
References
R v. Secretary of State for the Home Department ex parte Venables and Thompson
(1997a). 2 WLR 67.
R v. Secretary of State for the Home Department ex parte Venables and Thompson
(1997b). 3 All ER 97.
T v. UK and V v. UK, ECHR, Decided 16 December. (1999).
Ackerman, B.A. & J.S. Fishkin (2004). Deliberation Day. New Haven: Yale University
Press.
Almond, G.A. & S. Verba (1963). The civic culture: Political attitudes and democracy in
five nations. Princeton, N.J.: Princeton University Press.
Balvig, F. (2004). When law and order returned to Denmark, Journal of Scandinavian
Studies in Criminology and Crime Prevention, 5: 167-187.
Barabas, J. (2004). How deliberation affects policy opinions, American Political Science Review, 98: 687-701.
Barker, V. (2006). The politics of punishing: Building a state governance theory of
American imprisonment variation, Punishment & Society, 8: 5-33.
Cavadino, M. & J. Dignan (2006). Penal policy and political economy, Criminology
& Criminal Justice, 6: 435-456.
Christie, N. (2000). Crime control as industry: Towards GULAGS, western style.
London: Routledge.
Cohen, B.C. (1963). The press and foreign policy. Princeton: Princeton University
Press.
273
Crozier, M., S.P. Huntington & J. Watanuki (1975). The crisis of democracy: Report
on the governability of democracies to the Trilateral Commission. New York: New
York University Press.
Daily Mirror (1993). We must fight chaos of crime, Daily Mirror, February 20, 1993.
Diamond, A. (1993). Nasty, Mr Kirschner, Daily Mirror, December 1, 1993.
Doob, A.N. & C.M. Webster (2006) Countering punitiveness: Understanding stability
in Canadas imprisonment rate, Law & Society Review, 40: 325-368.
Downes, D. & R. Morgan (1997). Dumping the hostages to fortune? The politics of
law and order in post-war Britain, in M. Maguire, R. Morgan & R. Reiner (Eds.)
The Oxford Handbook of Criminology, 2nd ed., pp. 87-134, Oxford: Oxford University Press.
(2002). The skeletons in the cupboard: The politics of law and order at the turn of
the millennium, in M. Maguire, R. Morgan & R. Reiner (Eds.) The Oxford Handbook of Criminology, 3rd ed., Oxford: Oxford University Press.
Downes, D. & R. van Swaaningen (forthcoming). The road to dystopia? Changes in the
penal climate of The Netherlands, in M. Tonry & C. Bijleveld (Eds.) Crime and
justice: A review of research, Chicago: University of Chicago Press.
Drakeford, M. & M. Vanstone (2000). Social Exclusion and the politics of criminal justice:
A tale of two administrations, Howard Journal of Criminal Justice, 39: 369-381.
Dyer, C. (2001). Paper fined for Bulger order breach, Guardian, December 5, 2001.
Elkins, D.J. & R.E.B. Simeon (1979). A cause in search of its effect, or what does political culture explain? Comparative Politics, 11: 127-145.
Estrada, F. (2004). The transformation of the politics of crime in high crime societies,
European Journal of Criminology, 1: 419-443.
Fletcher, G. & J. Allen (2003). Perceptions of and concern about crime in England and
Wales, in J. Simmons & T. Dodd (Eds.) Crime in England and Wales 2002/2003,
pp. 127-153, London: Home Office Communication Development Unit.
Fuchs, D. & H.-D. Klingemann (1995). Citizens and the state: A changing relationship? in H.-D. Klingemann & D. Fuchs (Eds.) Citizens and the state, pp. 1-23, Oxford: Oxford University Press.
Furedi, F. (1997). Culture of fear: Risk-taking and the morality of low expectation.
London: Cassell.
Garland, D. (1996). The limits of the sovereign state: Strategies of crime control in
contemporary society, British Journal of Criminology, 36: 445-471.
(2001). The culture of control: Crime and social order in contemporary society.
Oxford: Clarendon Press.
Green, D.A. (2006). Public opinion versus public judgment about crime: Correcting the
"comedy of errors", British Journal of Criminology, 46: 131-154.
(forthcoming) Comparing penal cultures: Two responses to child-on-child homicide, in M. Tonry (Ed.) Crime, punishment, and politics in comparative perspective,
Vol. 36 in Crime and justice: A review of research, edited by M. Tonry, Chicago:
University of Chicago Press.
274
DAVID A. GREEN
Hargreaves, I. (2003). Journalism: Truth or dare?, Oxford: Oxford University Press.

Home Office (1997). Tackling youth crime A consultation paper. Juvenile Offenders
Unit (JOU).
Jacobsson, K. (1997). Discursive will formation and the question of legitimacy in
European politics, Scandinavian Political Studies, 20: 69-90.
Jones, B. (2001). The mass media and political communication, in B. Jones, D. Kavanagh, M. Moran & P. Norton (Eds.) Politics UK, 4th ed., pp. 177-204, Harlow, Essex: Pearson Educational Limited.
Kyvsgaard, B. (2001). Penal sanctions and the use of imprisonment in Denmark, in M.
Tonry (Ed.) Penal reform in overcrowded times, pp. 115-118, New York: Oxford
University Press.
Larsson, P. (1993). Norwegian penal policy in the 80s, Chroniques, 8: 81-91.
(2001). 'Norway prison use up slightly, community penalties lots, in M. Tonry (Ed.)
Penal reform in overcrowded times, pp. 124-129, New York: Oxford University
Press.
Laure, J.-E. & S.A. Haavik (2002). VG offices, Oslo, Norway.
Lawson, S. (1993). Conceptual issues in the comparative study of regime change and
democratization, Comparative Politics, 25: 183-205.
Lijphart, A. (1968). The politics of accommodation: Pluralism and democracy in the
Netherlands. Berkeley: University of California Press.
(1969). Cosociational democracy, World Politics, 21: 207-225.
(1985). Power-sharing in South Africa. Berkeley: Institute of International Studies,
University of California.
(1998). Consensus and consensus democracy: Cultural, structural, functional, and
rational-choice explanations, Scandinavian Political Studies, 21: 99-108.
(1999). Patterns of democracy: Government forms and performance in thirty-six
countries. London: Yale University Press.
Listhaug, O. & M. Wiberg (1995). Confidence in political and private institutions, in
H.-D. Klingemann & D. Fuchs (Eds.) Citizens and the state, pp. 298-322, Oxford:
Oxford University Press.
Mathiesen, T. (1996). Driving forces behind prison growth: The mass media, Nordisk
Tidsskrift for Kriminalvidenskab, 83: 133-143.
(2003). Contemporary penal policy: A study in moral panics, A paper presented at
the European Committee on Crime Problems 22nd Criminological Research Conference, Strasbourg: October 20, 2003 [online]
www.coe.int/T/E/Legal_affairs/Legal_co-operation/Crime_policy/ Conferences/
PC-CRC(2003)8E-TMathiesen.pdf [Accessed February 9, 2003].
McCombs, M.E. (1981). The agenda-setting approach, in D. D. Nimmo & K. R. Sanders (Eds.) Handbook of political communication, pp. 121-140, London: Sage Publications.
(2005). Setting the agenda: The news media and public opinion. Cambridge: Polity
Press.
275
McCombs, M.E. & D.L. Shaw (1972). The agenda-setting function of the mass media,
Public Opinion Quarterly, 26: 176-187.
Moore, S. (2000). Child incarceration and the new youth justice, in B. Goldson (Ed.)
The new youth justice, pp. 115-128, Lyme Regis: Russell House.
MORI (2002). MORI political monitor: Most plus other important issues 1974-present,
[online] http://www.mori.com/polls/trends/issues.shtml [Accessed 9 July 2003].
Newburn, T. (forthcoming) "Tough on crime": Penal policy in England and Wales,
Crime and Justice: A Review of Research, 36.
Newton, K. (1997). Politics and the news media: mobilisation or videomalaise? in R.
Jowell, J. Curtice, A. Park, L. Brook, K. Thomson & C. Bryson (Eds.) British social
attitudes: The 14th report: The end of Conservative values?, pp. 151-168, Aldershot:
Dartmouth Publishing Company.
Rentoul, J. (2001). Tony Blair: Prime Minister. London: Little, Brown and Company.
Roberts, J.V. (2002). Public opinion and sentencing policy, in S. Rex & M. Tonry
(Eds.) Reform and punishment: The future of sentencing, pp. 18-39, Collumpton,
Devon: Willan Publishing.
Rose, D. (1993). The messy truth about Britains violent youth: An examination of how
last weeks apocalyptic rhetoric has obscured the facts about juvenile offenders, Observer, February 28, 1993.
Schlesinger, P., H. Tumber & G. Murdock (1991). The media politics of crime and
criminal justice, British Journal of Sociology, 42: 397-420.
Seldon, A. (2004). Blair. London: Free Press.
Smith, D.J. (2006). Without a trace, The Sunday Times Magazine, 23 July 2006.
Snare, A. (Ed.) (1995). Beware of punishment: On the utility and futility of criminal
law, Oslo: Pax.
Solberg, H. (1994). Maktesls mot TV-vold (Powerless against TV violence), VG, 19
October 1994.
Snstelie, E.H. (1994). Lov p vei (Law on the way), VG, 18 October 1994.
(1995). 'Ville ikke ha blodpenger (Will not have blood money), VG, 12 March
1995.
Stokes, R. & J.P. Hewitt (1976). Aligning actions, American Sociological Review, 41:
838-849.
Street, J. (1994). Political culture From civic culture to mass culture, British Journal
of Political Science, 24: 95-113.
Sun (1992). Its The Sun Wot Won It, Sun, April 12, 1992.
Surette, R. (1998). Media, crime, and criminal justice: Images and realities. Belmont,
CA: Wadsworth Press.
Tham, H. (2001). Law and order as a leftist project? The case of Sweden, Punishment
& Society, 3: 409-426.
Uslaner, E.M. (2002). The moral foundations of trust. Cambridge: Cambridge University Press.
276
DAVID A. GREEN
VG (1994). Lov p vei (Law on the way), VG, October 18, 1994.
Victor, D. (1995). Politics and the penal system a drama in progress, in A. Snare
(Ed.) Beware of punishment: On the utility and futility of criminal law, pp. 68-88,
Oslo: Pax.
von Hofer, H. (2004). Crime and reactions to crime in Scandinavia, Journal of Scandinavian Studies in Criminology and Crime Prevention, 5: 148-166.
Weiss, C.H. (1983). Ideology, interests, and information: The basis of policy positions,
in D. Callahan & B. Jennings (Eds.) Ethics, the social sciences, and policy analysis,
pp. 213-245, London: Plenum Press.
Zedner, L. (1995). In pursuit of the vernacular: Comparing law and order discourse in
Britain and Germany, Social & Legal Studies, 4: 517-534.
(2002). Dangers and dystopias in penal theories, Oxford Journal of Legal Studies,
22: 341-366.
277
Methodological Problems in Measuring

Attitudes to Punishment (Punitivity)
HELMUT KURY AND JOACHIM OBERGFELL-FUCHS *
1. Introduction
During the last several years it has frequently been pointed out that attitudes to
punishment (punitivity) have become more severe. Punitivity is, however, far
from being a clearly defined concept. When discussing the topic, we may thus
well allude to different aspects, as the concept comprises not only popular attitudes to punishment but also the relevant criminal legislation and the sentencing
practices of the judiciary (see Kury et. al. 2004; Kury 2007). As far as criminal
legislation is concerned, we have observed during the last decades in most Western countries and not only here increased levels of punitivity in the form of
harsher sanctions or the criminalisation of deviant behaviour that was previously
not sanctioned (inside the family, sexual behaviour).
Yoshida (2008, in this volume) shows that, in Japan, the criminal code has undergone several changes during the last few years, virtually all leading to increased levels of punishment. For Germany, Rengier (2002) has demonstrated
that twenty, thirty years ago there existed a definite movement to decriminalise
some sexual offences but that today this trend has been reversed, and that over
the last few years sentences have become increasingly more severe. He concludes: There is little doubt about the way criminal policy has been evolving for
the last few years given the priority granted to security considerations and the
trend to more severe sanctions. This reversal, after a period of liberalisation
which we call the first revolution of sexual criminal legislation can be seen as a
(second) legislative revolution. The actual hard, or at least harder, line fits in with
international tendencies that have originated primarily in the United States, for
instance with their three strikes and you are out legislations and other catch
words such as nothing works and zero tolerance (p. 25). According to Rckert (2006), 42 paragraphs of the criminal code have been modified in Germany
since 1992 to allow for tougher sanctions. The recent drafts for a new legislation
*
We would like to express our special thanks to Evelyn Shea for her thoroughly translation of
the manuscript.
278
Helmut Kury & Joachim Obergfell-Fuchs
for juvenile offenders also put the main emphasis on security. Whereas the
German prison law for adult offenders, adopted in 1977, names as the primary
aim of imprisonment to help inmates lead in future a socially responsible life
without further conflict with the law and lists the protection of society from further offences only in second place ( 2), the recent Lnder drafts for a code for
juvenile offenders have in part reversed these priorities. This reversal is particularly obvious in the Hamburg draft bill that puts, in the same 2, the security
mandate before the goal of education and rehabilitation. Other Bundeslnder
changed the order of paragraphs established by the Strafvollzugsgesetz and removed the sentencing aims from their prominent place in the beginning to a less
conspicuous one in later paragraphs.
Dnkel and Prksen (2007) are sceptical as to the innovative power of the draft
bills: The current draft bills of the Lnder dont do much more than to legalise already existing different regional practises (p. 65). The discussion about the rank and
the positioning of the two principal aims of sentencing is often coloured by ideological view points. The question is not whether to rehabilitate or to protect society. The
Federal Constitutional Court, in its decision of 31 May 2006 concerning the need for a
specific law for juvenile offenders, has made it clear that between the aim of rehabilitating young offenders and the wish to protect the public from further offences,
there exists no basic contradiction. (BVerfG, 2 Bv thatR 1673/04). Several authors
have nevertheless reached the conclusion that, on an international level, recent legislations tend toward increased punitivity. (see Brandenstein 2006; Sack 2006; Beckett
and Sasson 2004; Garland 1990; 2001).
In many countries we notice an increase in the length of sentences for certain
offences, especially in the United States, but also in Germany. Obergfell-Fuchs
(in this volume; see also Kury and Obergfell-Fuchs 2007), having analysed on
the basis of official statistical information the German sanctioning practices of
recent years, was able to show that between 1980 and 2004 the rate of short
prison terms has decreased with a corresponding increase in longer sentences.
Still, long prison terms remain relatively rare in Germany. The number of inmates granted early release after serving two-thirds of their sentence has visibly
decreased during the last twenty years. But there are also signs of lower punitivity, as more drug addicts are transferred to therapeutic institutions. The author
concludes that, despite some opposing developments, German judges generally
tend toward more punitive measures, but that this trend is less obvious than on
the legislative side. This doesnt come as a surprise as judges tend to follow their
established traditions and habits. They will therefore uphold their old sanction
patterns and only gradually adopt tougher measures with the arrival of a new
Methodological Problems in Measuring Attitudes to Punishment (Punitivity) 279
generation of judges and, maybe, because of pressure exerted by the media (see
Streng 2007). In this perspective, we can expect long-term effects from higher
upper limits of the sentence. This can already be seen from observing the longterm changes in the incarceration rates. Between 2000 and 2005 the number of
inmates in 15 European countries has increased, sometimes even dramatically.
The few countries that show a decrease are mainly Eastern European states that
had very high initial rates. This evolution points to an increase in punitvity both
on the legislative and the judicial level, especially in the formerly liberal Western
countries (see Aebi and Stadnic 2007).
It is more difficult to ascertain whether the attitudes to punishment among the
population have also become more severe. The support for capital punishment,
for instance, has known a constant decrease in Germany since the end of the Second World War. Whereas after the war close to three quarters of the population
voted in favour of the death penalty, today support for capital punishment has
shrunk to one third. A representative survey of the Institut fr Demoskopie Allensbach from 1949 showed that, just before the decision of the Parliamentary
Council to anchor the abolition of capital punishment in the constitution, 74 per
cent of West Germans were in favour of maintaining it. The ones in favour were
in particular the young (up to 30 years old: 81 per cent; over 65 years old: 64 per
cent) and those living in Northern Germany (Lower Saxony: 79 per cent;
Schleswig-Holstein: 86 per cent) rather than in the south (Baden: 65 per cent;
Wrttemberg: 62 per cent). In recent years about 30 to 40 per cent of Germans
declare themselves in favour of capital punishment. This percentage is still
slightly higher in East Germany, but the values are converging (for attitudes to
punitivity in East Germany see, for instance, Ludwig and Krupl 2005). In countries like the United States or the former Soviet Union that still maintain capital
punishment or that have abolished it very recently, the percentage of people in
favour are as a rule considerably higher (see Kury 2004, Kury and Kapanadze
2004; Kury et al. 2006).
There also exist considerable methodological problems for the measuring of
attitudes to punishment. Julian Roberts and colleagues have demonstrated in several studies that punitivity values depend on several variables such as the level of
information of the respondents on a case under question or the operationalisation
of the construct (Roberts 1992; Roberts and Hough 2002b; Doob and Roberts
1983). The results obtained may vary considerably depending on how the questions on punitivity are formulated and in which context they are put (see Kury
1995). This methodological difficulty cannot be neglected considering the impact
punitivity values have on criminal policy. High values are often used by politi-
280
cians to push through tougher measures against crime supposedly in answer to

the demands of a population living in insecurity and fear of crime (see Beckett
and Sasson 2004). The frequent debates on the risks of terrorism after September
11, 2001 provide a good illustration. A survey by the ZDF-Politbarometer of
September 14, 2007, interviewing a representative sample of 1.209 adult Germans, shows nevertheless that 57 per cent of the respondents thought that enough
measures were taken in Germany to prevent terrorist attacks. Yet 65 per cent
agreed to an on-line control of their computers as requested by the minister of the
interior, Schuble, and only 31 per cent were against such a measure. Despite a
perceived level of relative security, many citizens are or seem willing to accept considerable restrictions of their freedom and privacy for still more security.
In particular the standardised polling of punitivity used in virtually all victim
surveys, raises considerable methodological problems. In order to contain the
length of the questionnaire, only a restricted number of items, sometimes only
one, are included, such as the question whether the respondents endorse or reject
capital punishment for serious crimes. This question has actually become the
standard item for punitivity research. Standardised polling of complex issues,
such as punitivity, with few or only one item can neither render a complex construct with any degree of precision nor provide adequate results. An added difficulty lies in the fact that capital punishment has been abolished in Western
Europe in most countries many years ago, and that the question has become at
best a hypothetical one. Respondents may thus vehemently clamour for the reintroduction of the death penalty for certain offences without incurring the risk of
this really happening. In this way they can express their repugnance for such
crimes without actually thinking about an appropriate sanction. Even if the results of such polls do tell us something about the attitudes to punishment, at least
about certain of its aspects, it is not clear to what extent. They certainly do not
mirror the attitudes to punishment, even if they are often sold as such.
Standardised surveys also imply the risk of capturing attitudes about which the
respondents have no clear ideas because they have not previously given the matter much thought (non-attitudes). Standard questionnaires that require only the
checking of a given answer option may induce the interviewees to check a mark
on a scale even if they have no or, at most, a very vague opinion. It is even possible that they only become aware of the existence of a problem when reading
the question. In absence of a strong attitude of their own, they will in all likelihood follow the (presumed) majority views which they know from the media. It is probable that respondents will remember in the first place particularly
reprehensive cases reported by the media and that this will not be without repercussions on their answers. Punitivity values obtained by traditional survey methods may thus be less accurate than we could wish for.
The information the average citizen obtains from the media concerns almost
exclusively the most serious forms of crime, certain spectacular cases, distorting
thus the overall picture of crime (see Kerner and Feltes 1980; Beckett and Sasson
2004; Albrecht 1997). This may then lead to the myth violence: crime reporting by the media concentrates almost exclusively on the most serious cases of
violent crimes, the recipients choose, on the basis of their preconceived ideas
about crime, the information that confirms their views, and thus the arguments
for a strong state that has to fight such a development fall politically on fertile ground (see Albrecht 2001a). The real cause of crime and violence in society can thus be avoided as a sufficient solution to the problem seems already at
hand by simply increasing the punishment. This is particularly true for the reporting of the tabloid press and of some private television channels (seeWindzio and
Kleinmann 2006). The media are the willing servants of the spirit of the age that
does not want to hear any more about rehabilitation but is only interest in knowing if an offender is locked up for a long enough period or whether he should
ever come out at all. What happens to this person in prison? Zero interest. What
should be done to prevent his re-offending? Zero interest. The answer is: lock
him up, preferably for ever (Friedrichsen 2007, p. 134).
In the end this amounts to more of the same. If one tries to get to the root
causes of crime, especially of violent crime, research becomes very difficult.
This is not surprising as we are faced with very complex issues. Albrecht is right
when commenting on the state of current research: We cannot conclude from
the fact that the results from a certain number of studies linking social inequality
and social deprivation on the one hand and crime or violent crime on the other
are possibly not significant and/or inconsistent that there is no such relationship,
as long as we have not examined whether the apparent absence of a link or contradictory results have their cause in specific but different conceptualisations of
the independent variables, the dependent variables and/or the specific constructs
in regard to this link (Albrecht 2001b, p. 218). Numerous studies point to societal causes of (violent) crime (see Albrecht et al. 1991), but such causes are usually ignored, especially on the political level. The problem is essentially reduced
to slack prosecution and, above all, to the absence of tough sanctions.
In our own studies we have tried to show the degree of distortion resulting
form measuring the attitudes to punishment with the usual methods by introducing a second, personal interview in which the respondents were asked de-
282
tailed questions about the extent of their punitivity. Using the same method, we
could also show that the standard measurements of fear of crime, fraught with
similar methodological difficulties, lead to vastly inflated values (Farrall et al.
1997; Farrall and Gadd 2004; Kury et al. 2004a; 2004b; 2005). In the following
pages we will briefly describe our research and the main results.
2. Our own research

Our research is divided into two parts. In part one, we determine how far the results of a quantitative survey on punitivity based on a large standardised questionnaire are confirmed by a subsequent qualitative study, in which a selected
sample of respondents is asked in a second, oral interview about their attitudes to
punishment. In part two, we examine significant correlations of punitivity.
2.1
Hypotheses
Based on the findings of previous research (see Roberts and Hough 2002a; Farrall et al. 1997; Kury et al. 2004a; 2004b; 2005) we proceeded on the following
hypotheses:
The measurement of attitudes to punishment through standardised survey
methods leads to a considerable overestimation of the extent of punitivity in a
given population.
The positioning of the death penalty item within the standardised questionnaire affects the measured extent of punitivity.
The different operationalisations of punitivity used in the professional literature do not cover the same construct.
2.2
Measuring Instruments
For the quantitative part of the inquiry a substantial questionnaire was developed
that contained all essential operationalisations of the construct that are used in the
professional literature. The questionnaire consisted of the following parts:
questions concerning demography (age, sex, etc.)
stress from problems in the neighbourhood (incivilities)
integration in the neighbourhood
conservatism
authoritarian style of child raising
ones own mood, contentedness with oneself as a person
proposed sanctions for 26 given offences varying in seriousness
attitude to capital punishment, not only standard item
assessment of the aims of criminal sanctions, respectively of belief in a just

world
questions about the level of crime in the country and in ones neighbourhood
direct and indirect experiences of victimisation
personal views on the causes of crime
cognitive and emotional fear of crime
attitude to the sanction practice of the courts (see Bergmann and Schill 2006).
In order to capture possible primacy/recency as well as priming effects, the

questionnaire was devised in two versions that differ in the following two points:
in version B, contrary to version A, the standard indicator for punitivity (attitude to capital punishment) is put before the detailed questions concerning the
respondents ideas about punishment and the degree of satisfaction with the
system of criminal prosecution
the order of the proposed sanctions for the different offences was inverted in
version B, i.e. the harsher sentences were listed before the more lenient ones.
For the qualitative part of the study we developed a semi-standardised interview,
because this form of data collection made it possible to obtain comparable information for all participants but also permitted a flexible adaptation to the line of
thought of the discussion partner (Wittowski 1994, p. 3). The interview guide
consisted of a general and a specific part, the latter dealing with the topic of
crime and criminal justice in concrete terms. The attitudes to punishment were
directly addressed through specific examples about which the respondents were
asked detailed questions. Additional background information about the offence
and the offender were added to the 26 examples of the standard questionnaire.
The punitivity of the respondents was assessed at the end of the interview both
based on the answers that were given and on the general impression received by
the interviewers.
2.3
The actual inquiry
The inquiry took place during the last three months of 2005 and the first quarter
of 2006 in Freiburg, a town in the south west of Germany with a population of
roughly 200.000. We chose three neighbourhoods that were known from previous studies to portray different social problems:
Weingarten is a neighbourhood with big apartment buildings; it has high fear

of crime and medium crime rates
St. Georgen, a neighbourhood at the outskirts of town with a rather rural structure and low fear of crime but relatively high crime rates
Sthlinger with medium fear values and relatively high crime rates.
284
The target persons were selected according to the random walk procedure, i.e.
the streets in each neighbourhood were first randomly selected. In each selected
street the two interviewers (both female) visited, at different times of day, every
second house. They targeted the person who opened the door if this person was at
least 18 years old and knew enough German to fill in the questionnaire.
The questionnaire was handed out in person and was picked up by the same
interviewer at the time agreed upon a few days later. If the target person was not
at home at the appointed time, the interviewer would make at least two, and up to
seven attempts to contact him or her. In this way 336 completed questionnaires
could be retrieved out of 450 that had been handed out. This corresponds to a
response rate of 74.6 per cent, which is an excellent result for the region and for a
survey involving a relatively long questionnaire. The anonymity of all respondents was guaranteed, on their request also in written form. Their names did appear nowhere on the questionnaire making it thus impossible to trace back the
answers to a particular individual. When the questionnaire was collected, all respondents were asked if they were willing to participate in an additional oral interview and if they would, in this case, give the interviewer their phone number.
161 persons who had participated in the questionnaire survey (47.9 per cent)
gave their phone number for an oral interview.
The questionnaires were immediately analysed in order to determine the punitivity values of the individual respondents. Once these values were ascertained,
we verified which persons had agreed to an additional interview. We found that
the willingness to participate in such an interview was much lower for those that
showed high punitivity than for the others. From each of the two groups polled
by the two interviewers, the ten persons were selected who showed the highest
punitivity values and who had also agreed to an additional interview. These 20
persons were to be questioned again by the same interviewer who had also been
in charge of the questionnaire survey. As one of them did not participate, a total
of 19 interviews were given according to the pre-established guidelines.
2.4
Description of the sample
Of the 336 survey participants, 176 (52.4 per cent) filled out version A of the
questionnaire and 160 (47.6 per cent) version B. 110 men (32.7 per cent) and 221
women (65.8 per cent) were polled. 5 persons (1.5 per cent) did not indicate their
sex. Our sample shows thus a clear overrepresentation of women. As far as the
age distribution is concerned, the youngest group is overrepresented; the other
groups show a relatively even distribution, except for the very small numbers in
the oldest age bracket (see table 1).
Table 1: Age distribution of the total sample

Age
18 30
108
32.6
31 40
48
14.5
41 50
49
14.8
51 60
48
14.5
61 70
49
14.8
71 80
23
6.9
81 90
1.8
331
100
Total
The sample is relatively evenly spread over the three neighbourhoods: Sthlinger: 112 (33.3 per cent), St. Georgen: 109 (32.4 per cent) and Weingarten: 115
(34.2 per cent). As far as the family status is concerned, 136 (40.5 per cent) of
the respondents are married, 112 (33.3 per cent) single, 32 (9.5 per cent) divorced
or separated, 25 (7.4 per cent) widowed and 20 (6 per cent) are living as a couple
without being married. In regard to the educational level, 85 (25.3 per cent) have
obtained the equivalent of an O-levels certificate or a professional certificate
(Mittlere Reife or Fachschulabschluss), 78 (23.2 per cent) the equivalent of Alevels or a higher professional certificate (Abitur/Fachabitur), 75 (22.3 per cent)
secondary modern school qualifications (Hauptschulabschluss), 63 (18.8 per
cent) a university degree, 19 (5.7 per cent) have studies at the university without
obtaining a degree, and 8 (2.4 per cent) do not have secondary modern school
qualifications. 93.4 per cent of the interviewees were Germans, the others of foreign nationality.
3. Results
As for the differences between versions A and B of the questionnaire, there was
no significant position effect for the capital punishment item. Tendentially, fewer
respondents approve of the death penalty for certain crimes, if the item is placed
after the questions about proposed sanctions for particular offences rather than
before. This may indicate that the respondents were more aware of the variety of
sanction alternatives, even for serious crimes, after having been asked repeatedly
to indicate an appropriate sanction for a variety of offences from a catalogue of
nine answer choices, ranging from there is no need for the state to intervene to
capital punishment. They may thus have opted for a more lenient sanction.
286
Table 2: Mean punitivity for case vignettes (ranked according to increasing

punitivity values; 1 = there is no need for the state to intervene
10 = capital punishment)
Item
Termination of unwanted pregnancy
Use of hashish
Use of heroin
Use of public transportation without ticket
Resisting of police
Shoplifting, value 45
Theft, value 250
Drunk driving
Purse snatching with violence
Rape after consented petting
Battery of an adult person, requiring medical
intervention
Rape within marriage
Car theft
Break and entry
Sexual fondling of a 4 year old girl
Beating of a child, requiring medical intervention
Second break and entry, theft of TV
Forcing of a person into a sexual act
corruption, bankruptcy of the firm, loss of 200
jobs
Break and entry, damage of 4,500
Killing of baby after birth because one has
problems
Act by the extreme right, raid on a centre for
asylum seekers
Injuring a person through use of arms
Pushing a car off the road at high speed
Rape
Sex with 4 year old girl that does not object
n
324
332
335
336
328
333
335
333
333
325
332
Min
1
1
1
1
1
1
1
1
1
1
2
Max
9
10
10
10
9
9
9
10
9
10
9
M
2.65
4.42
5.32
5.38
5.51
5.66
6.16
6.90
7.38
7.48
7.63
S
2.34
2.77
2.82
1.92
2.16
1.75
1.59
1.63
1.45
2.18
1.61
331
335
336
329
333
332
332
332
1
1
2
2
2
1
2
1
10
9
9
10
10
10
10
10
7.84
7.90
7.99
8.03
8.06
8.14
8.23
8.27
1.99
1.28
1.47
1.89
1.74
1.41
1.66
1.57
331
334
3
2
10
10
8.31
8.42
1.06
1.82
329
10
8.55
1.18
335
335
333
331
2
1
2
3
10
10
10
10
8.67
8.69
8.83
8.88
0.97
1.13
0.96
0.95
There were significant differences in punitivity values obtained from the 26 case
vignettes if the answer choices for the sanctions were ranked from mild to severe
(version A) or from sever to mild (version B). Version A shows a mean value of
6.47, version B a mean value of 6.72. The difference is statistically highly significant (p < .005). If the respondent is first presented with lenient forms of sanctions that gradually become more severe, leading up to the level of the death penalty, the average sanction proposal will be significantly milder than in the con-
trary case. If, on the other hand, the respondents are first given the death penalty
option and only after milder alternatives, they tend to opt for harsher sentences,
including capital punishment. This confirms earlier findings that the results of
punitivity can be influenced by the wording of the item (see Kury 1995).
Table 2 shows the results for the average sanction proposals for the 26 offences from the written survey, ranked according to increasing punitivity. Both
questionnaire versions were factored in, i.e. the indicated values correspond to a
mean value from both approaches.
As expected, some offences, resp. socially inappropriate behaviour, such as
smoking It is worth noting that the standard deviations decreases the stronger the
severity of the sanction, i.e. the more serious the offence, the more homogeneous
the sanction proposals of the respondents. The mean values indicate considerable
levels of punitivity. For about two-thirds of the 26 cases included in the written
questionnaire the respondents suggest on average a prison term, either to be
served in prison or on probation. An exception to this is abortion; only a minority
still seems to feel the need for a severe sentence, whereas as the majority considers it a private matter of the woman concerned.
Because of possible methodological effects these results have to be qualified.
The most dramatic effect is produced by the choice of the survey method. As
expected, the punitivity values were considerably higher when measured through
standard written polls than through oral interviews; they were actually twice as
high. As mentioned above, 19 cases of highly punitive respondents were given a
second, oral interview. In 11 of the 19 cases, the punitive attitudes estimated by
the interviewers on the basis of specific pre-set criteria were clearly below those
of the written survey. Only in three cases did the two different methods produce
similar results, and in five cases the punitivity values of the oral interview were
higher than those of the questionnaire. The mean value of the 19 respondents that
had participated in both inquiries was 7.34 (scale from 1 to 10) for the questionnaire and 6.16 for the oral interview. These visible differences cannot be explained away by a tendency toward centre values or by a change of attitude between the two inquiries, the latter for two reasons: the time span between the
questionnaire and the interview was short, and during this brief period no particularly spectacular incident happed on the crime front. Thus, the differences have
essentially to be ascribed to the two survey methods. The differentiated oral
method seems to give a more precise evaluation of the attitude to punishment,
whereas the questionnaire includes in part also stereotypes.
288
In a next step, we examined the different aspects of punitivity reflected in the

26 case vignettes and the factors responsible for producing them. The main results will be presented in the following pages.
In order to reduce the amount of data and to construct only a limited number
of punitivity factors for future multivariate comparisons a factor analysis (principal component analysis) was calculated on the basis of the 26 case vignettes included in the questionnaire. The results from a previous correlation analysis indicated that the item interruption of unwanted pregnancy showed hardly any significant correlations with the other variables. It was thus excluded from further
calculations. Further probatory analyses showed that also the item to provoke
the bankruptcy of a firm through corrupt behaviour, causing the loss of 200 jobs
could not be meaningfully integrated into the factor structure and was also excluded. With the remaining 24 items a principal component analysis with varimax rotation has been calculated.
The anti-image covariance matrix showed clearly that the vast majority of the
non-diagonal elements did not significantly deviate from zero, the values of the
diagonal elements in the anti-correlation matrix (MSA criteria) varying between
.71 and .91. The Kaiser-Mayer-Olkin criterion was .85 (meritorious). The requirements for carrying out the analysis are thus given. Based on the Eigenvalue
criterion, a 6-factor solution could be determined which did however not comply
with the criterion of a simple structure. The Scree test did not produce a clear
result but pointed toward a 5-factor solution. The number of factors was then
gradually changed until finally a 5-factor solution proved to have the most discriminative power, while still allowing for an appropriate explained variance.
These 5 factors explain 53.8 per cent of the total variance; the rotated factor loading matrix is represented in table 3.
As can be seen from the factor loading matrix, factor I (F I) explains serious
property offences, including those causing harm to persons. On factor II (F II)
load in particular serious offences against persons, especially those that had recently been given much public exposure, such as sexual offences, infanticide and
aggressive driving (pushing a car off the road) resulting in death. Factor III (F
III) also focuses on offences against the person, but of the kind for which there
exists a lesser consensus as to their gravity. The fact that also drunk driving falls
under this category may be explained by the respondents anticipating the consequences of driving in such a state. Factor IV (F IV) concerns less serious offences, mainly petty crime. Factor V (F V) consists exclusively of drug offences.
The factors were saved as continuous variables for further analyses with the following statistical values (to facilitate the manipulation, the factor values were
multiplied by 10):
F I:
F II:
F III:
F IV:
F V:
M = -0.23;
M = -0.26;
M = 0.13;
M = -0.04;
M = -0.18;
s = 10.38;
s = 10.26;
s = 9.81;
s = 10.11;
s = 9.99;
min -48.34,
min -54.92,
min -39.56,
min -41.89,
min -22.87,
max
max
max
max
max
22.36,
28.70,
20.56,
20.39,
27.60.
In all cases a higher factor values indicates a higher degree of punitivity.

Table 3: Factor loading matrix case vignettes punitivity
Driving under the influence of
alcohol
Driving without permit
Theft 250
Resisting police
Use of hashish
Break and entry
Bodily harm, adults
Shoplifting 45
Use of heroin
Rape
Sexual assault
Lethal pushing with car
Infanticide
Purse snatching
Sexual abuse
Car theft
Bodily harm with weapon
Child abuse
Second burglary
Rape within marriage
Rape after petting
Acts by the radical right
Sexual fondling of child
Break and entry, 5500
FI
.15
F II
-.04
F III
.46
F IV
.24
FV
.23
h2
.35
-.10
.27
.19
.07
.51
.29
.23
.10
.31
.10
.11
.16
.58
.20
.75
.40
.09
.79
.18
.00
.20
-.01
.70
.03
.06
-.05
.12
.24
.20
.17
.10
.61
.57
.67
.58
.17
.64
.20
.31
.31
.01
.16
.18
.40
.56
.20
-.06
.19
.15
-.01
.07
.35
.00
.11
.05
.42
.14
.05
.13
.16
.03
.36
.63
.11
.78
.76
.44
.32
.17
.76
.73
.60
.14
.23
.30
.75
.14
.17
.25
.10
.04
.14
-.10
.19
-.13
.01
.12
-.06
.10
.07
-.03
.03
.06
.02
.25
.83
.22
.26
.09
.83
-.05
-.02
-.04
.25
.25
.14
-.05
-.02
.21
.05
-.02
-.08
-.02
.21
.01
.60
.65
.55
.74
.52
.49
.68
.73
.56
.57
.54
.47
.47
.51
.65
.50
.56
.70
.70
.69
.49
.61
.62
Apart from punitivity, a series of constructs had been operationalised in the questionnaire by using different variables, such as conservatism, authoritarianism
and belief in a just world. In an attempt to reduce the number of variables for
further analyses, the theoretically constructed scales were given an item- and reliability analysis in order to construct statistically homogenous scales with items
of maximum discriminative power.
290
The scale conservatism consisted initially of 17 items. Through recoding it

was possible to have the values vary between -2, i.e. low parameter value in the
sense of the construct, and +2, high parameter value. The category dont know
was given the value 0. In a first step, using a correlation analysis, all items were
excluded which did not significantly correlate with at least half of the others.
This reduced the number to 10 items, of which one more was eliminated because
of its insufficient discriminative power, leaving a scale of altogether 9 items.
Their mean values, standard deviation and discriminative power coefficients are
represented in table 4.
Table 4: Item analysis of the scale conservatism
To respect law and order
To seek security
To hold on to traditions
Banning of preferential treatment of minorities
Banning of begging by the homeless
Recognition of same-sex marriages (-)
To abide by moral laws
Untidy people are repelling
Behaving well is important
Total scale
M
1.51
1.16
0.11
-0.50
-0.90
-0.69
0.39
-0.46
0.03
s
rit
0.77
.40
0.96
.46
1.24
.33
1.30
.30
1.23
.47
1.56
.40
1.10
.46
1.30
.48
1.36
.59
Cronbach = .75
M = 0.66; s = 6.31; Md = 1
Min = -15; Max = +15
The discriminative power coefficients are acceptable for research scales, so does
the Cronbach of.75 as a measure of internal consistency. The resulting total
scale for assessing the construct conservatism has a mean value of 0.66 and a
median of 1.
The construct authoritarianism consisted initially of 8 variables that were
differently coded. In order to achieve standardisation, the categories dont
know of some variables were set on defaults. For preparing the subsequent reliability analysis, one variable was excluded based on the correlation matrix, and
three other ones were eliminated because of their lack of discriminative power on
the basis of the item analysis. Table 5 shows the final version of the reliability
analysis as well as some statistical values of the condensed scale. The discriminative power coefficients and the Cronbach are in an acceptable range; the total
scale has a mean value of 12.39 with a standard deviation of 3.20.
Table 5: Item analysis of the scale authoritarianism

Children have to learn obedience and respect
Slapping is part of a good upbringing
Rules are needed for when child is home alone
Annoyed if child talks back
Comprehension if child has a different
opinion (-)
Total scale
M
2.69
2.17
2.90
2.66
1.96
s
1.01
1.12
1.22
0.80
0.76
rit
.55
.40
.34
.45
.28
Cronbach = .64
M = 12.39; s = 3.20; Md = 12
Min = 5; Max = 21
The construct belief in a just world/ sense of justice had been theoretically operationalised with 8 variables. One of these had to be excluded from the outset,
as it had been only dichotomosly scaled. In the correlation analysis, respectively
the item analysis, three more variables did not prove to be very useful and were
eliminated. Table 6 shows the final composition of the scale with values from -8
to +6, as well as their statistical values. It can be seen at first glance that these
values of this construct are noticeably weaker than for the two previous scales.
Table 6: Item analysis of the scale Belief in a just world/ sense of justice
The world is a just place
People get what they deserve
Parents punish for a good reason
People have to blame themselves for their
misfortune
Total scale
M
-0.60
-0.10
-0.10
-0.85
s
1.12
1.05
1.03
0.96
rit
.30
.42
.32
.24
Cronbach = .53
M = -1.65; s = 2.69; Md = -2
Min = -8; Max = +6
In the questionnaire was also included a self assessment of some character traits.
These 17 traits were given a principal component analysis with Varimax rotation.
The aim was again to reduce the quantity of data and to construct continuous
scales for further multivariate calculations. The prerequisites for the analysis
were given, the majority of non-diagonal elements of the anti-image covariance
matrix did not deviate significantly form zero, the values of the diagonal elements in the anti-correlation matrix (MSA criteria) lay between .54 and .87, and
the Kaiser-Mayer-Olkin criteria was .77. Because of low communality the trait
tense was excluded from the final solution, which left a total of 16 variables
loading on three factors. They explain 48.1 per cent of the total variance. Table 7
gives an overview of the results.
292
Table 7: Factor loading matrix personal traits

Calm
Explosive
Fearful
Cautious
Willing to take risks
Aggressive
Hostile
Energetic
Balanced
Angry
Insecure
At the mercy of others
Threatened
Sovereign
Content
FI
-.36
.69
.30
-.23
-.34
.76
.62
.14
-.39
.72
.18
.37
.52
.04
-.40
F II
.55
.00
-.07
.12
-.48
-.04
-.03
.57
.65
.01
-.30
-.14
-.10
.71
.41
F III
.23
.00
.55
.63
.23
.09
.23
-.13
.05
.14
.69
.56
.52
-.11
-.15
h2
.49
.48
.39
.46
.40
.59
.44
.36
.58
.54
.59
.47
.55
.52
.36
Facor I is characterised by a high load of items that describe an aggressive-hostile

attitude, such as explosive, aggressive, hostile, angry, but also threatened
and dissatisfied du httest ja bei mir bleiben knnen. Dann httest du jetzt alles,
was du brauchst.. Factor II, on the contrary, contains items with a high load that
characterise a sovereign-calm attitude, such as calm, not very ready to take
risks, energetic, balanced, sovereign and content. The third factor expresses in his highly loading items an anxious attitude, such as fearful, cautious,
insecure, at the mercy of people and threatened. The values of the three factors were again saved for further calculations and multiplied by 10 to make them
more manageable. The following are the statistical values of the newly formed
scales, whereby the higher value indicates a stronger presence of the trait.
F I: M = -1.47;
F II: M = 0.11;
F III: M = -0.14;
s = 9.65; min -19.68,

s = 10.09; min -34.79,
s = 9.87; min -22.37,
max 46.22,
max 29.84,
max 30.80.
The questionnaire also included a series of stress-inducing factors, such as typical manifestations of incivilities, but also the lack of cultural and recreational
facilities, especially for young people. These 18 items were again given a principal component analysis with Varimax rotation. The conditions for an analysis
were met to the highest degree, as the anti-image covariance matrix showed that
the vast majority of the non-diagonal elements were not different from zero and
that the values of the MSA criterion, i.e. the diagonal element in the anti-
correlation matrix, varied between .83 and .95. The Kaiser-Mayer-Olkin criterion
was .92.
A two-factor analysis gave an excellent result, explaining 48.5 per cent of the
total variance. Table 8 outlines the factor loading matrix and the communalities.
Table 8: Factor loading matrix stress factors
FI
.66
.49
.47
.69
.67
.69
.23
.71
.51
.63
.40
.04
.76
.75
.55
.18
.49
Bored, idle youth

Rundown, abandoned buildings
Flying merchants, door to door sales
Vandalised phone boots, park benches
Drug addicts
Spray paint and graffitti on houses
Not enough youth homes, centres
Dirt, garbage in street and parks
Undisciplined drivers
Too many foreigners, asylum seekers
Xenophobia, right wing radicalism
Not enough cultural and recreational facilities
Crime
Violent groups
Car wrecks left about
Not enough sports facilities, sports clubs
Illegally parked cars causing obstructions
F II
.30
.34
.20
.22
.34
.01
.74
.08
.11
.08
.53
.80
.22
.27
.31
.76
.32
h2
.53
.36
.26
.53
.57
.48
.60
.52
.27
.40
.44
.64
.62
.64
.39
.61
.34
As we see from table 8, the typical forms of incivilities load on factor I, ranging
from conspicuous groups of persons, spray painted houses to garbage on the
street and traffic problems. Factor II, on the other hand, is composed of items that
describe a lack of opportunities, especially for young people, such as insufficient
cultural, recreational or sports facilities. The fact that the items xenophobia and
right wing radicalism also load on this factor is an indication that the population
associates these phenomena with the absence of sufficient (sensible) offers for
young people. The factor values, multiplied by 10, were saved for each person
for further analyses. The scales had the following statistical values, whereby
higher value indicate again a stronger presence of the characteristic.
F I: M = -0.51;
F II: M = 0.20;
s = 10.07; min -34.28,

s = 9.92; min -28.00,
max 23.42,
max 22.51.
In addition, we included in the questionnaire a series of variables indicating contentedness with ones present life situation. These were again given a principal
294
component analysis with Varimax rotation. The conditions for a factor analysis
were fulfilled, the vast majority of non-diagonal elements showed no significant
deviation from zero and the values of the MSA criteria, i.e. of the diagonal elements in the anti-correlation matrix, varied between .67 and .90. The KaiserMayer-Olkin criterion was .85. The analysis resulted in a clear 2-factor solution
which explained 50.7 per cent of the total variance. Table 9 shows the factor
loading matrix as well as the communalities of the items.
Table 9: Factor loading matrix contentedness with ones life situation
Contentedness with ones health
Contentedness with ones mental state
Contentedness with ones professional future
Contentedness with ones standard of living
Contentedness with ones family
Contentedness with ones abilities and accomplishments
Contentedness with ones looks
Contentedness with ones self confidence and assurance
Contentedness with ones character
Contentedness with ones contacts with friends
Contentedness with the location of ones apartment
FI
.54
.69
.17
.23
.55
.71
.74
.76
.72
.67
.01
F II
.22
.27
.70
.81
.22
.13
.16
.11
.00
-.00
.68
h2
.33
.55
.52
.70
.35
.52
.57
.60
.52
.45
.46
As we can see from table 9, the item that load on factor I are in particular those
that mirror contentment with oneself and ones milieu, whereas factor II is characterised by items that convey contentment with external conditions. Here again,
the two factors were saved as continuous variables for further analyses, and
higher factor values indicate a stronger presence of the characteristic.
F I: M = 0.07;
F II: M = -0.18;
s = 9.86; min -31.79,

s = 10.12; min -37.35,
max 18.65,
max 18.51.
We had also included in the questionnaire variables to assess the fear of crime. 8
items referred to cognitive risk assessment, i.e. questions about the subjective
likelihood of a victimisation for various offences. In addition the inventory contained two questions bearing on a subjective feeling of insecurity, which rather
cover emotional aspects: the fear of being alone in ones apartment at night and
the standard item of fear of crime, i.e. the feeling of security/insecurity to be
out alone at night in ones neighbourhood. These 10 items were given a principal
component analysis with Varimax rotation. The conditions for carrying out a factor analysis were given, the non-diagonal elements of the anti-image covariance
matrix mostly showed no significant deviation from zero, and the values of the
MSA criteria, i.e. of the diagonal elements in the anti-correlation matrix, varied
between .69 and .88. The Kaiser-Mayer-Olkin criterion was .76. The analysis
after the Scree-test resulted in a clear 2-factor solution, explaining 51.2 per cent
of the total variance. Table 10 indicates the factor loading matrix as well as the
communalities of the items.
Table 10: Factor loading matrix fear of crime
Probability of theft of vehicle
Probability of purse being snatched
Probability of break and entry
Probability of being beaten up
Probability of being attacked with a weapon
Probability of being mobbed
Probability of being raped
Probability of being sexually molested
Fear of being home alone at night
Fear of being out alone at night in neighbourhood
FI
.34
.62
.60
.87
.84
.57
.38
.30
.07
.05
F II
.25
.36
.30
.06
.14
.07
.68
.73
.70
.66
h2
.18
.51
.45
.76
.73
.33
.61
.62
.50
.44
We can see from table 10 that only items referring to cognitive risk assessment
load on factor I, quite independent of the type of offence, except for sexual offences. The latter constitute, together with the two fear items (to be home alone,
to be out alone at night), the second factor. As for content, factor I can be described as measuring the cognitive risk assessment. For the interpretation of factor II, the composition of the sample is important, as two-thirds of the respondents are women for whom the risk of a sexual offence has a considerable emotional impact and will not be perceived simply on a cognitive level. We can thus
assume that factor II mirrors the emotional experience of fear. Again the two factors were saved as continuous variables for further analyses. Higher factor values
indicate a stronger presence of the characteristic.
F I: M = -0.26;
F II: M = -0.14;
s = 9.60;
s = 9.83;
min -21.56,
min -19.19,
max 32.75,
max 46.51.
In the following paragraph we will look into the question as to which characteristics determine or correlate with punitivity in its different manifestations expressed by the factors extracted above. We opted for a multiple regression analysis as the most adapted methodological choice and, given the characteristics of
the possible predictors, for a linear regression analysis. As criteria served the five
different aspects of punitivity described above; the following variables were used
296
as predictors: age (coded in ascending order), sex (dummy coded, 1 female),

formal education (coded in ascending order), net income (coded in ascending
order), own victimisation (dummy coded, 1 victim), and scales described above
of conservatism, authoritarianism, sense of justice; personality aggressivehostile, personality sovereign-calm, personality anxious; stress incivilities,
stress lack of offers; contentedness with oneself, contentedness with living
conditions; fear cognitive risk assessment, fear emotional fear, each in ascending order.
Table 11: Results of the linear regression analyses, representation of the beta
weights (the figures in bold indicate a significant T-value)
Age
Sex
Formal education
Income
Victimisation
Conservatism
Authoritarianism
Sense of justice
PERS I aggressivehostile
PERS II sovereigncalm
PERS III anxious
Stress I incivilities
Stress II lack of
offers
CONT I cont. with
oneself
CONT II cont. with
living conditions
FEAR I cogn. risk
assessment
FEAR II emotional
fear
Model R /adj. R2
PUN I
serious
property
crimes
-.18*
-.09
.01
.22**
-.03
.25**
.02
.06
PUN II serious
crimes against the
person receiving
wide publicity
-.14
.09
-.16*
.03
-.04
.12
.05
-.14*
PUN III
offences
against
the person
-.37***
.18
.07
-.06
.09
-.07
.07
-.09
PUN IV
minor
offences
-.10
.07
.02
-.04
.13*
.15
.09
.05
PUN V
drugrelated
offences
.05
.06
-.08
-.06
.01
.16
.18*
.04
.16*
-.14*
-.03
-.19**
-.10
.09
.02
-.03
-.09
.05
.08
-.02
-.09
-.16*
-.15*
.08
-.03
.12
-.22**
.18**
.01
.10
.08
.05
.01
-.02
-.02
-.03
.02
-.17*
-.01
-.05
.04
.27***
-.11
-.03
.11
-.02
-.07
.06
-.08
-.06
-.14
-.08
.16*
.42***/.11
.35*/.06
.50***/.20
.40**/.10
.59***/.30
Significant T-values are indicated in bold: the level of significance is denoted as follows:
*
p < .05; ** p < .01; *** p < .001.
The results of the regression analysis (table 11) show that the punitive attitudes toward serious property crimes (PUN I) are significantly associated with
higher income, conservative values, a rather aggressive-hostile attitude and a
younger age bracket. The model fit is acceptable, the Multiple R is .42, resulting
in an adjusted explained variance of 11 per cent (R2). These correlations are easy
to interpret, as persons with a higher income stand to lose more from serious
property crime. If we add a conservative stance and a rather hostile attitude,
higher punitivity values are plausible. The fact of belonging to a younger age
group takes on the role of a reinforcing factor, i.e. younger people or persons in
their middle years that are fully integrated in their professional activities prove to
be more punitive in this particular constellation.
The model is less successful in explaining punitivity in terms of serious, much
publicised crime against the person (PUN II; sexual abuse of children, rape). In
this case, the multiple R is .35, and the adjusted explained variance only 6 per
cent. Four predictors show at the 5 per cent level significant beta weights, indicating that punitivity is associated with lower educational levels, reduced belief
in a just world, a less aggressive-hostile attitude and a lesser perception of incivilities. These results too can be, at least in part, conclusively interpreted: individuals with lower educational levels may be more susceptible to distorted media
reporting current for this type of (sexual) offences. The experience of injustice,
linked to the concept of helplessness, is also an explanatory factor. It may actually involve a (deliberate) ignoring of signs of incivilities, as this would put an
additional stress on the personality structure. It seems to concern a group of
rather underprivileged people, unhappy with the world, that would like to see
widely publicised crimes punished more severely. Punitivity is taking place on a
rational level, aggressive and hostile impulses being rather rare. Such an interpretation remains, however, rather speculative, as there is no clearly recognisable
pattern and the values of the explained variance are moderate. This has also to do
with the kind of offences considered, i.e. sexual offences and serious traffic violations, as in both cases, the attitude to (severe) punishment among the population is not very consistent and depends on numerous additional variables.
The next variable, punitivity regarding offences against the person, is clearly
marked by a strong age effect. Younger people are significantly more punitive,
male and female. Anxiousness is also a factor that has to be taken into consideration: less anxious persons score higher on punitivity. The model is satisfactory
with a multiple R of .50 and an adjusted explained variance of 20 per cent. Offences against the person, i.e. the use of violence against another person, are of
greater concern for young people who are frequently victims but also perpetra-
298
tors. This again will more likely be the case for the less anxious, willing to get
themselves involved in riskier situations. In particular young women who do not
perceive themselves as timid and display a correspondingly independent behaviour show little tolerance for violent acts and request rather severe sanctions.
PUN III includes also sexual offences (rape within marriage, rape after petting,
child abuse), the attitudes to which are clearly linked to age and sex. Younger
people are more likely to have been familiar with the public debates about child
abuse, rape by a spouse or violence linked to right wing extremists when growing
up, and may thus be particularly sensitive to such occurrences. The more traditional attitude and viewpoint of older people may, on the contrary, have an excusing effect for offences such as rape within marriage or after the consented
exchange of tenderness. Driving under the influence of alcohol also falls in this
category. The younger generation has an easier time internalising these new
norms than older people, many of whom may have been guilty themselves in this
respect.
As far as the attitude to punishment for petty crime is concerned, the results
are more difficult to interpret. We find more highly punitive attitudes among
people who have been personally victimised, who do not have a pronounced aggressive-hostile attitude and are relatively happy with their actual living conditions. Conservative attitudes are just barely non-significant. The model adaptation is narrowly acceptable with a Multiple R of .40 and an adjusted explained
variance of 10 per cent. Apart from a personal victimisation experience which
understandably might produce a certain desire for revenge, it is more difficult to
explain, from a theoretical point of view, why the other, rather positive attitudes
should cause higher punitivity values. The non-significant but relatively high
beta weights for the variables age (-.10), conservatism (-.15), authoritarianism
(.09), personality traits sovereign-calm (-.09) and incivilities (.12) indicate rather
younger, conservative and authoritarian persons, who do not see themselves as
particularly confident and calm and are rather conscious of signs of incivilities in
their neighbourhood.
Clearly the best model (R = .59; adjusted explained variance of 30 per cent) is
the one concerning punitivity for drug related crimes. Here, we have a number of
significant predictors. Persons with authoritarian attitudes who do not consider
themselves anxious show higher punitivity values. So do those who are more
aware of incivilities, are less happy with themselves and show higher emotional
fear. Just barely non-significant is the reinforcing influence of conservative attitudes. It makes sense that the tolerance of drug related offences is rather lower in
persons with an authoritarian outlook, and the same holds true for a greater
awareness of signs of incivilities that are often associated with the drug scene.
The discrepancy between low levels of anxiousness and heightened emotional
fear of crime is also worth pointing out. Whereas the latter is specifically oriented toward the area of crime and thus plausible as far as the result of the regression analysis is concerned (frequently associated with the perception of incivilities), low levels of anxiousness presumably indicate that an assertive, selfconfident attitude leads to a stronger dissociation from drug addicts, the frailty of
whom should be severely punished.
4. Discussion
It is in general not easy to interpret the content of regression analytical results.
The interpretations given above are thus to be understood rather as possible explanations that need to be confirmed by subsequent research. The data show
overall that, on an attitudinal level, punitivity is a very complex concept and that
there are no unified condition factors that influence this trait. The perception of
incivilities, for instance, is an explanatory factor for a punitive attitude to drug
related offences, but with regard to serious, much publicised violent crime it is
rather a moderating factor, and in other constellations it has no significant impact. In similar form, this holds also true for other factors.
The results show clearly, in confirmation of earlier studies, that punitivity is
no unified construct, that there is no punitivity as such, but only varying aspects
thereof that may take on a different significance in different areas of crime. Attitudes to punishment depend not only on the survey method used but also on
demographic factors, character traits, personal attitudes, and on the knowledge
and the degree of information of the respondents about the type of offences concerned. Such complex relations are already known from other fields, especially
from studies on the effects of media reporting in regard to violence (see Kunczik
and Zipfel 2002; 2006, p. 249 ff). This explains why the media transmission of
attitudes to punishment may not have the same effect on all recipients. For some,
the demand for tough measures will fall on fertile ground, others will pay slight
attention to such headlines, and in yet another group it may even lead to a reduction in punitivity.
Conclusions based on oversimplified survey results are thus to be treated with
great caution. It seems hardly justified to exploit, for political purposes, undifferentiated findings, for instance the perceived wish of the population for tougher
sanctions, yet we see it done all the time. Methodical problems, either linked to
the inquiry itself or to the data on which the findings are based, are too often neglected. The findings become thus almost optional, everyone choosing what they
300
need for their line of argument. Given the political significance of the kind of
sanctions the population wishes for, it is important that criminological research
and in particular research in the field of the empirical social sciences contribute
to the precise measurement of these values.
References
Aebi, M., Stadnic, N. (2007). Imprisonment rates in Europe. ESC Criminology in Europe, 6(2), 1-15.
Albrecht, G. (1997). Anomie oder Hysterie oder beides? Die bundesrepublikanische
Gesellschaft und ihre Kriminalittsentwicklung. In: Heitmeyer, W. (Ed.), Was treibt
die Gesellschaft auseinander? Frankfurt/M: Suhrkamp, 506-554.
(2001a). Gewaltkriminaltitt zwischen Mythos und Realitt. In: Albrecht, G., Backes, O., Khnel, W. (Eds.), Gewaltkriminalitt zwischen Mythos und Realitt.
Frankfurt/M.: Suhrkamp, 9-67.
(2001b). Soziale Ungleichheit, Deprivation und Gewaltkriminalitt. In: Albrecht, G.,
Backes, O., Khnel, W. (Eds.), Gewaltkriminalitt zwischen Mythos und Realitt.
Frankfurt/M.: Suhrkamp, 195-235.
Albrecht, G., Howe, C.-W., Woltersdorff, J. (1991). Familienstruktur und Delinquenz.
Ren Knig zur Vollendung des 85. Lebensjahres gewidmet. Soziale Probleme 2,
107-156.
Beckett, K., Sasson, T. (2004). The politics of injustice. Crime and punishment in America. Thousand Oaks: Sage, 2nd ed.
Bergmann, J., Schill, E. (2006). Die Messung von Strafeinstellungen. Ein Versuch der
Kombination einer quantitativen und qualitativen Herangehensweise. Unpublished
thesis: Universitt Freiburg.
Brandenstein, M. (2006). Strafzweckerfllungen als abhngige Variable der Zeit. In:
Obergfell-Fuchs, J., Brandenstein, M. (Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift fr Helmut Kury zum 65. Geburtstag. Frankfurt/M.: Verlag fr Polizeiwissenschaft, 357-394.
Doob, A.N., Roberts, J. (1983). Sentencing: An analysis of the publics view of sentencing. Ottawa: Department of Justice.
Dnkel, F., Prksen, A. (2007). Stand der Gesetzgebung zum Jugendstrafvollzug und
erste Einschtzungen. Neue Kriminalpolitik 2, 55-67.
Farrall, S., Bannister, J., Ditton, J., Gilchrist, E. (1997). Questioning the measurement
of the rear of crime: Findings from a major methodological study. British Journal of
Criminology 37, 657-678.
Farrall, S., Gadd, D. (2004). The frequency of the rear of crime. British Journal of Criminology 44, 127-132.
Friedrichsen, G. (2007). Kritische Anmerkungen zur Justizkritik. Urteile werden heute
nicht nur im Gerichtssaal gesprochen, sondern lange zuvor auch in der ffentlichkeit. Zeitschrift fr Rechtspolitik 4, 133-134.
Garland, D. (1990). Punishment and modern society: A study in social theory. Chicago:
University of Chicago Press.
(2001). The culture of control: Crime and social disorder in contemporary society.
Oxford: Oxford University Press.
Institut fr Demoskopie Gesellschaft zum Studium der ffentlichen Meinung m.b.H.,
Allensbach (1949). Fllt die Todesstrafe? Ablehnung bei politischen Verbrechen. Eine Untersuchung in Westdeutschland. Allensbach: unpublished manuscript.
Kerner, H.-J., Feltes, T. (1980). Medien, Kriminalittsbild und ffentlichkeit. Einsichten und Probleme am Beispiel einer Analyse von Tageszeitungen. In: Kury, H. (Ed.),
Strafvollzug und ffentlichkeit. Freiburg: Rombach, 73-112.
Kunczik, M. & Zipfel, A. (2002). Gewaltttig durch Medien? Aus Politik und Zeitgeschichte, B44/2002, 29-37.
(2006). Gewalt und Medien. Ein Studienbuch (5th ed.). Kln, Weimar: Bhlau.
Kury, H. (1995). Wie restitutiv eingestellt ist die Bevlkerung? Der Einfluss der Frageformulierung auf die Ergebnisse von Opferstudien. Monatsschrift fr Kriminologie
und Strafrechtsreform 78, 84-98.
(2004). Kriminalittsentwicklung, Sanktionen und die Einstellung der Bevlkerung.
Ein Vergleich zwischen Ost und West. In: Kury, H. (Ed.), Strafrecht und Kriminalitt. Entwicklungen in Mittel- und Osteuropa. Bochum: Universittsverlag Dr.
Brockmeyer, 1-30.
(2007). Mehr Sicherheit durch mehr Strafe? Aus Politik und Zeitgeschichte. Beilage
der Wochenzeitung Das Parlament, 4041, 30-37.
Kury, H., Kapanadze, K. (2004). Strafvollzug in Georgien Ergebnisse einer Umfrage
bei Inhaftierten und Angehrigen der Strafverfolgung. In: Kury, H. (Ed.), Strafrecht
und Kriminalitt. Entwicklungen in Mittel- und Osteuropa. Bochum: Universittsverlag Dr. Brockmeyer, 165-188.
Kury, H., Kania, H., Obergfell-Fuchs, J. (2004). Worber sprechen wir, wenn wir ber
Punitivitt sprechen? In: Lautmann, R., Klimke, D., Sack, F. (Eds.), Punitivitt. Kriminologisches Journal, 36, 8., 51-88.
Kury, H., Lichtblau, A., Neumaier, A. (2004a). Was messen wir, wenn wir Kriminalittsfurcht messen? Kriminalistik 58, 457-465.
Kury, H., Lichtblau, A., Neumaier, A., Obergfell-Fuchs, J. (2004b). Zur Validitt der
Erfassung von Kriminalittsfurcht. Soziale Probleme 15, 141-165.
(2005). Kriminalittsfurcht. Zu den Problemen ihrer Erfassung. Schweizerische Zeitschrift fr Kriminologie (SZK) 4, 3-19.
Kury, H., Keller, M., Mitter, U., Rahmati, N. (2006). Verbrechensfurcht, Punitivitt und
Einstellung zur Polizei Ergebnisse einer Umfrage in Aserbaidschan. In: Kury, H.,
Karimov, E. (Eds.), Kriminalitt und Kriminalprvention in Lndern des Umbruchs.
Beitrge einer Internationalen Konferenz in Baku/Aserbaidschan. Bochum: Universittsverlag Dr. Brockmeyer, 415-486.
Kury, H., Obergfell-Fuchs, J. (2007). Zur Punitivitt in Deutschland. In: Kury, H. (Ed.),
Hrtere Strafen weniger Kriminalitt? Zur Verschrfung der Sanktionseinstellungen. Soziale Probleme 17, 119-154.
302
Ludwig, H., Krupl, G. (2005). Viktimisierung, Sanktionen und Strafverfolgung. Jenaer

Kriminalittsbefragung ber ein Jahrzehnt gesellschaftlicher Transformation. Mnchengladbach: Forum Verlag Godesberg.
Regnier, R. (2002). Entwicklungslinien im Sexualstrafrecht. Vortrag auf dem 23. Triberger Symposium am 28. November 2002.
Roberts, J.V. (1992). Public opinion, crime, and criminal justice. In: Tonry, M. (Ed.),
Crime and justice: A review of research. Chicago: University of Chicago Press, Vol.
16, 99-180.
Roberts, J.V., Hough, M. (Eds.)(2002a). Changing attitudes to punishment: Public opinion, crime and justice. Cullompton: Willan.
(2002b). Public attitudes to punishment: the context. In: Roberts, J.V., Hough, M.
(Eds.), Changing attitudes to punishment: Public opinion, crime and justice. Cullompton: Willan, 1-14.
Rckert, S. (2006). Ab in den Knast. Die Zahl der Verbrechen sinkt, doch das Strafrecht
wird systematisch verschrft. Und immer mehr Menschen werden zu immer lngeren
Gefngnisstrafen verurteilt. DIE ZEIT, Nr. 22 vom 24. Mai 2006, 15-18.
Sack, F. (2006). Deutsche Kriminologie: auf eigenen (Sonder)Pfaden? Zur deutschen
Diskussion der kriminalpolitischen Wende. In: Obergfell-Fuchs, J., Brandenstein, M.
(Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift
fr Helmut Kury zum 65. Geburtstag. Frankfurt/M.: Verlag fr Polizeiwissenschaft,
35-71.
Streng, F. (2007). Sanktionseinstellungen bei Jura-Studenten im Wandel. In: Kury, H.
(Ed.), Hrtere Strafen weniger Kriminalitt? Zur Verschrfung der Sanktionseinstellungen. Soziale Probleme 17, 210-231.
Windzio, M., Kleimann, M. (2006). Die kriminelle Gesellschaft als medial Konstruktion? Mediennutzung, Kriminalittswahrnehmung und Einstellung zum Strafen. Soziale Welt 56, 193-215.
Wittkowski, J. (1994). Das Interview in der Psychologie: Interviewtechnik und Codierung von Interviewmaterial. Opladen: Westdeutscher Verlag.
303
Punitivity within the

Criminal Justice System in Germany
JOACHIM OBERGFELL-FUCHS
1. Introduction
The term punitivity does not really exist in English language, but since several
years, the concept has become more and more common in international scientific
literature on sanctioning, punishment, and their psychological and sociological
impacts. Punitivity means the propensity and the will to punish either for deterrence, for incapacitation or even for making the offender suffering (see Kury et al
2003). Therefore, punitivity is always negatively connotated.
In the last years, a discussion on an increasing punitivity in western states has
been emerged among criminologists. Young (2003), for example, has stated a
punitive turn, Beckett and Sasson (2004, p. 4) wrote about the politics of getting tough, and Pratt et al. (2005) named their reader new punitiveness. One
of the most prominent advocates of the thesis of an increasing punitivity is David
Garland (2001). He describes the rapid development of the law and order politics in the United States and in Great Britain during the last three decades. Based
on the specific risks, needs, and uncertainties of the late modern societies, he argues that these new penal strategies are an answer to the social, cultural and societal changes and therefore not limited to Anglo-Saxon countries. Another, more
recent paper by Sack (2006) draws the same conclusions. Sack describes the
changes in criminal politics in other European countries and that most of them
have participated in the punitive turn.
While Garland and Sack mainly focus on sanctioning and the political and cultural climate, punitivity is not only limited to such societal variables. Kury et al.
(2004) presented a conceptual model of punitivity, differentiating between an
individual level (micro-level), e.g. attitudes towards punishment, a macro-level,
e.g. political discourse, and a judicial level, e.g. sanctioning of offenders. The
data showed that there is not necessarily a consistency between these three levels
within one country, even though a general trend towards an increase in punitivity
can be found.
But this is not only a special problem of Western societies, a study by Kury et al
(2002) showed similar trends in Eastern European countries. The breakdown of the
304
Joachim Obergfell-Fuchs
communist system was mostly followed by a more lenient and mild criminal justice
system in these countries, but meanwhile increasing crime rates strengthened the call
for harsher punishment. Poland, for example, since 1991 member of the European
Council and since 2004 member of the European Union, even discusses the restoration of the death penalty (cf. Lissmann 2006), although the protocol no. 6 to the European Convention on Human Rights declares the abolition of the death penalty in
member states of the European Council1 , furthermore, the death penalty is incommensurate with the statutes of the European Union.
During the last years, a quite intense discussion on punitivity, whether increasing or not, has emerged in Germany too. Winfried Hassemer, the vice president
of the German constitutional court (Bundesverfassungsgericht) brought up the
expression the new desire to punish (Hassemer 2000), which gave reason for
many subsequent papers and symposia. (e.g. Lautmann et al. 2004; Kury and
Obergfell-Fuchs 2006; see the overview by Sack 2006). Despite some objections,
especially concerning the demand for harsher punishment within the population
(e.g. Reuband 2004; 2006), most scholars would sign the statement that punitivity has increased in Germany.
2. Aim of the paper

The following analyses will focus on the punitivity within the criminal justice
system. Such judicial punitivity could be expressed in longer prison sentences, in
an inauspicious ratio between mandatory prison sentences and suspended sentences, or a restraining parole practice. Looking at the United States, in the current discussion more or less the prototype of a punitive country, the adult correctional population has more than tripled between 1980 (1,842,100) and 2005
(7,056,000). The supervision rate grew from 1,117 per 100,000 adults in 1980 up
to 3,150 in 2005, an increase which is extraordinary in comparison with other
western countries (cf. Glaze and Bonczar 2006). But looking at the portions of
probation and parole, the data do not reflect such an extreme increase in punitive
reactions. In 1980 60.8% of those under correctional supervision were under probation and another 12.0% were under parole. In 2005, the sheer percentages do
not have changed dramatically, 59.0% were under probation and 11.1% were
under parole (Glaze and Bonczar 2006, see also Pastore and Maguire 2005). This
means, although the number of supervised persons has increased, the proportions
within this number have remained quite stable, a fact that stresses the need for
differentiated analyses.
1
http://conventions.coe.int/Treaty/Commun/ListeTraites.asp?MA=3&CM=7&CL=ENG; November 28, 2006.
Punitivity within the Criminal Justice System in Germany
305
The main hypothesis of this paper is that there has been an increasing punitivity within the German criminal justice system, which will be reflected in longer
prison sentences, an increase in security measures and a more restricted practice
of early prison release. While we have partly analyzed some of these questions in
earlier papers (e.g. Obergfell-Fuchs 2006; Kury and Obergfell-Fuchs 2006; Kury
et al. 2004), a number of new facets will be discussed to give or more holistic
picture to this question.
3. Sentencing in Germany
In earlier analyses (see Kury and Obergfell-Fuchs 2006; Obergfell-Fuchs 2006)
we have reported that sentencing in Germany became harsher during the last 25
years, especially the sentencing of sex offenders. The data of these studies
showed, that the portion of sentence lengths of 5 years and longer according to
rape and sexual assault has about tripled between 1980 and 2004, according to
sexual abuse of children, the portion of this harshest sanctioning category has
even sextupled. Otherwise, one has to keep in mind that such long sentences for
sex offenders are rather an exception than a rule, in 2004 only about 7% of all
sentences for rape and sexual assault belonged to this category.
In order to analyze a general development of sentencing, prison sentences for
all crimes were summed up. Seven sentencing categories were built, starting with
rather mild sentences from 0 to 6 months, 6 to 9 months, 9 to 12 months, and 1 to
2 years. All these sentences are in common that they can be suspended for probation. Such a decision is to the discretion of the judge, mandatory probation sentences do not exist in Germany. The upper end is represented by sentences which
cannot be suspended for probation, starting with the category 2 to 3 years, followed by 3 to 5 years and 5 years up to lifelong, the latter is the most severe sentence in Germany. For each year, all of these seven sentencing categories add up
to 100%. Figure 1 shows this general overview of all sentences (cf. Statistisches
Bundesamt 1981 2006). For making the development during the time course
more visible, each portion got the reference value 1 in 1980. The following years
until 2004 show the relative decrease or increase of the portion of each category
referring to the given starting point in 1980. Values above the virtual reference
line 1 mark a relative increase in this category, values below 1 a decrease.
The figure clearly depicts that the rate of all three short sentence categories
has decreased between 1980 and 2004, while the relative portion of all longer
sentences has increased. This is not only the fact for sentences which cannot be
suspended for probation it also and especially applies to the category 1 to 2
years. It seems that the courts have executed a shift upwards in the direction of
306
harsher sentences. In comparison to the sanctioning of sex crimes mentioned

above (cf. Obergfell-Fuchs 2006), this shift is rather moderate, the steepest increase can be observed in the category 1 to 2 years with the factor 1.6, in the
most harsh category 5 years to lifelong, the factor is even much more moderate
with about 1.2. But equivalent to sexual offences, longer sentences are quite rare,
in 2004 only about 1.5% of all prison sentences were longer than 5 years, while
45.5% were less than 6 months.
While the shown increase in harsher sanctions seems to reflect a rising judicial
punitivity, there is also a development in the opposite direction. In 1980 65.7% of all
prison sentences were suspended for probation, until 2004 this portion has increased
up to 69.5%. These data suggest a tendency among the German criminal justice
courts to avoid imprisonment, especially short-term imprisonment of less than 2
years. This tendency can also be observed for such crimes, which are intensely discussed in the public, e.g. rape and sexual abuse of children. According to rape, the
rate of suspended sentences has increased from 25.8% in 1980 to 39.2% in 2004, for
sexual abuse of children, the change goes from 63.1% in 1980 to 67.0% in 2004. But
one has to keep in mind that such mild sentences up to 2 years are only imposed on
less severe offences.
Figure 1: Relative development of sentence lengths Germany 1980 2004
1,7
1,6
1,5
1,4
1,3
1,2
1,1
1
0,9
0,8
1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002
0-6 months
2-3 years
6-9 months
3-5 years
9-12 months
5 years- lifelong
1-2 years
Source: Statistisches Bundesamt (1981 2006). Rechtspflege. Strafverfolgung.
2004
307
4. Indicators of punitivity
The general increase in harsher sanctions could be due to a rising number of severe
crimes, and indeed, as the police crime statistics show, the numbers of robberies
(1980: 39.2 offences per 100,000 inhabitants; 2004: 72.4 offences/100,000) or severe physical attacks (1980: 106.4 offences/100,000; 2004: 178.0 offences/
100,000) have heavily increased too in this period of time (cf. Bundeskriminalamt
1981 2005). However, in the same time other severe crimes like rape and sexual
assault showed no significant alterations (1980: 11.2 offences/100,000; 2004: 9.9
offences/100,000), but, as mentioned above, particularly here the harshness of punishment has increased, and there is no plausible evidence that for example rape and
sexual assault itself have become more severe.
Another indicator of an increased judicial punitivity could be the number of
inmates in preventive detention, i.e. a detainment for security reasons after the
completion of the prison term, in Germany so-called Sicherungsverwahrung.
This measure is used for offenders who had been convicted for at least one pertinent or dangerous offence before and who are supposed to be dangerous. Therefore, preventive detention is not a penalty, it is a measure to protect the public
from a potentially dangerous and relapsing offender.
Until the end of the 1990s, preventive detention was supposed to be a phasedout model (cf. Kaiser 1997, p. 446; Kinzig 1996) and it was thought about abolishing this often controversial discussed measure (cf. Weber and Reindl 2001).
Figure 2 shows this declining process from the mid-1970s until 1984, followed
by a short revitalization at the end of the 1980s and a new decline until 1999 (cf.
Statistisches Bundesamt 1976 2006). At the end of the 1990s the situation
changed dramatically. Initialized by some severe sex offences (sexual murders)
against children, a huge media campaign started and an enormous public discussion how to deal with sex offenders arose. Only short time later the German parliament passed a new law including some legislations concerning a less restrictive use of the preventive detention. Since then a steady increase in the use of the
measure can be observed (see figure 2). Meanwhile most inmates in preventive
detention are sex offenders, not at least a result of the new legislations (cf. Kinzig
1996; Obergfell-Fuchs 2006, p. 607f.).
In this paper, some further aspects of a possible change in judicial punitivity
will be discussed. According to German law (Par. 57, 1 German Criminal Code,
Strafgesetzbuch StGB), prison terms can be suspended for parole after twothirds of serving the sentence after at least two months of serving a sentence
when the public needs for security are not impaired, and when the prisoner
agrees. Such a decision has to consider the personality of the offender, his past life,
308
Figure 2: Percentage of inmates in preventive detention (Sicherungsverwahrung) Germany 1975-2005

1
0,9
0,8
0,7
0,6
0,5
0,4
0,3
0,2
0,1
0
1975
1978
1981
1984
1987
1990
1993
1996
1999
2002
2005
Source: Statistisches Bundesamt (1976 2006). Rechtspflege. Strafvollzug Demographische

und kriminologische Merkmale der Strafgefangenen zum Stichtag 31.3.
the circumstances of the offense, his behavior during imprisonment, his future living
circumstances, and, not at least, the dangerousness of a possible relapse. Another legislation allows the release of a prisoner even after the halftime of the sentence (Par. 57,
2 German Criminal Code, Strafgesetzbuch StGB). This is possible after serving at
least six months of the prison term, when the offender is incarcerated for the first time
and the prison sentence is not longer than two years, or when the offenders personality, his development in prison and the situation of the committed offense legitimate
special mitigating circumstances. According to German law there are some more possibilities for suspending a prison sentence, these will be explained later.
Figure 3 gives an overview on the general early release practice in Germany
(light labeled curves). The lines do not depict real big changes during the last 20
years. About 70 percent of all prison sentences have to be completed, while another 30 percent were suspended for parole. One has to keep in mind that these
figures do not include the probation practices, here the numbers are much higher,
as explained above. Unfortunately, due to changes in registration, national overall statistics are only available until 2002. Only estimates, for example by the
Zweiter Periodischer Sicherheitsbericht [Second Periodical Security Report]
(Bundesministerium des Inneren and Bundesministerium der Justiz 2006, p. 630)
are available.
309
Figure 3: Percentage of suspended and completed prison terms in Germany

(GER) and in the federal state of Baden-Wuerttemberg (BaWue)
80
70
60
50
40
30
20
10
0
1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004*
GER suspended
GER completed
BaWue suspended
BaWue completed
Sources: Statistisches Bundesamt (1985 2003). Rechtspflege. Strafvollzug Anstalten, Bestand und Bewegung der Gefangenen; Justizministerium Baden-Wrttemberg, Jahresabschluss
Vollzug 2003 2005
* The percentages for Germany for 2004 are estimates by the Zweiter Periodischer Sicherheitsbericht (Bundesministerium des Inneren and Bundesministerium der Justiz 2006, p.
630)
But comparable statistics are still existing for the different German states, e.g. the
state of Baden-Wuerttemberg. The two curves (solid labels) depicted in figure 3
show an increase in suspended prison terms since the late 1990s. In 1984 the
quota of prison sentences suspended for parole was about 34.1 percent, until
1997 it has not really changed (33.8%). Since then, the quota of suspended sentences heavily increased, it was about 48.3% in 2005. This means, the rate for
suspended and completed prison sentences has almost equaled. A similar tendency, although limited by the estimation mentioned above, can be seen for
Germany in total. These data seem to reflect a decrease in punitivity at least in
the state of Baden-Wuerttemberg in contrast to the above described results.
To clear this apparent discrepancy a view at the different types of suspension
is helpful. As mentioned above, German legislation provides two major types of
early release, the two thirds suspension according to Par 57,1 criminal law
(StGB) and the halftime suspension according to Par 57,2 criminal law (StGB).
Figure 4 shows the percentage of suspended sentences after two thirds of the
prison term at least after two months of imprisonment. Both curves, for Ger-
310
many in total as well as for the state of Baden-Wuerttemberg, show a decrease in

this early release practice during the past 20 years. In 1985 20.9% of all released
prisoners in Germany left the facility because of a two thirds suspension, until
2002 this portion has decreased to 15.4% and the estimate by the Bundesministerium des Inneren and Bundesministerium der Justiz (2006, p. 631) for 2004 depicts a further decrease to 13.1%. In the state of Baden-Wuerttemberg, the situation is quite similar, in 1984 20.1% of all releases were two thirds suspensions,
then the rate decreased and reached a plateau of about 17.5% during the mid1990s. After a short peak in 2000 the quota steadily decreased again until 2004 to
the so far deepest point of 14.0%.
Figure 4: Percentage of suspended sentences after two thirds of the prison term
(Par 57,1 StGB) related to all released prisoners in Germany and in
Baden-Wuerttemberg
25
20
15
10
5
0
1984
1986
1988
1990
1992
1994
GER
1996
1998
2000
2002 2004*
BaWue
Sources: Statistisches Bundesamt (1985 2003). Rechtspflege. Strafvollzug Anstalten, Bestand und Bewegung der Gefangenen; Justizministerium Baden-Wrttemberg, Jahresabschluss
Vollzug 2003 2005
* The percentages for Germany for 2004 are estimates by the Zweiter Periodischer Sicherheitsbericht (Bundesministerium des Inneren and Bundesministerium der Justiz 2006, p. 631)
Figure 4 shows the percentage of suspended sentences after two thirds of the
prison term at least after two months of imprisonment. Both curves, for Germany in total as well as for the state of Baden-Wuerttemberg, show a decrease in
this early release practice during the past 20 years. In 1985 20.9% of all released
prisoners in Germany left the facility because of a two thirds suspension, until
2002 this portion has decreased to 15.4% and the estimate by the Bundesministerium des Inneren and Bundesministerium der Justiz (2006, p. 631) for 2004 de-
311
picts a further decrease to 13.1%. In the state of Baden-Wuerttemberg, the situation is quite similar, in 1984 20.1% of all releases were two thirds suspensions,
then the rate decreased and reached a plateau of about 17.5% during the mid1990s. After a short peak in 2000 the quota steadily decreased again until 2004 to
the so far deepest point of 14.0%.
Concerning the second possibility of early prison release, the half time suspension of the prison sentence, the results are quite similar. This type of early release
requires, as mentioned above, special mitigating circumstances and the criteria
given by law are hard to reach, so the numbers are much smaller than under the
condition of two thirds suspension. Figure 5 shows that after a steep increase in
the mid 1980s the use of this regulation slowly went down in Germany until
2004. In 1987, 2.7% of all prison releases were due to halftime suspensions, the
estimation for 2004 according the Zweiter Periodischer Sicherheitsbericht was
only 1.6%. In the state of Baden-Wuerttemberg, the results are again quite similar, in 1987, the highest portion of 3.4% could be observed. Between 1995 and
1998 there was a rapid decline, followed by a sharp increase in 2000. Since then
the portion decreased again and has reached a value of 1.5% in 2005.
Figure 5: Percentage of suspended sentences after halftime of the prison term
(Par 57,2 StGB) related to all released prisoners in Germany and in
Baden-Wuerttemberg
4
3,5
3
2,5
2
1,5
1
0,5
0
1984
1986
1988
1990
1992
1994
GER
1996 1998
BaWue
2000
2002
2004*
Sources: see Figure 4
With young inmates the situation is comparable, according to Par. 88 (and the
former Par. 89) juvenile court law (Jugendgerichtsgesetz JGG), a positive de-
312
velopment of the young inmate, under consideration of security reasons, can allow the suspension of a prison sentence after at least six months of serving. But
as figure 6 demonstrates, from 1984 until the early 1990s (Germany), respectively the mid-1990s (Baden-Wuerttemberg), the portion has clearly decreased.
For Germany the curve started in 1984 with a percentage rate of 6.8%, then it
continuously declined until 1993 (4.0%). In the following years, the portion remained quite stable and the last available data in 2004 depicted a value of again
4.0%. The situation in Baden-Wuerttemberg is not really different, while the portion was 7.5% in 1984 it decreased until 1997 to a value of 3.8%. After some
modulations in the early 2000s it arrived at 3.9% in 2005.
Figure 6: Percentage of early released inmates from youth prison facilities (Par
88, 89 JGG) related to all released prisoners in Germany and in
Baden-Wuerttemberg
8
7
6
5
4
3
2
1
0
1984
1986
1988
1990
1992
1994
GER
1996
1998
2000
2002
2004
BaWue
All three figures concerning adult offenders as well as the management of juvenile offenders demonstrate that the practice of early prison release seems to have
been declined. For adult offenders this occurred since about the end of the 1990s,
at this time the similar development for juvenile offenders widely seems to be
finished. Together with an increase in sentence lengths as shown above (for adult
offenders), the data reflect an increasing punitivity within the criminal justice
system in Germany. But if this would be true, why is the total number of suspended prison terms (see figure 3) increasing or at least stable? There must be
some other factors influencing this relationship.
313
One of these factors is the postponement of the execution of a prison term for
the therapy of drug addicted inmates. In German law, Par. 35 of the narcotics law
(Betaeubungsmittelgesetz BtMG) provides that a maximum 2 years prison sentence for an offender whose offences are committed because of drug addiction
can be postponed for therapy. After completion of the therapy, the time of the
therapeutic intervention can be deducted from the prison term.
The data depicted in figure 7 show a remarkable increase in the use of this
measure for Germany in total as well as for the state of Baden-Wuerttemberg
during the last decades (cf. Kurze 1995). At the beginning of the 1980s about 1
percent of all suspended prison terms were postponed executions for such therapy reasons, at the middle of the first decade of the 2000s this portion has increased for Germany up to about 5.9 percent, in Baden-Wuerttemberg the increase was even somewhat steeper, here the portion reached about 7.4 percent.
Besides the perception that drug addiction is rather a disease than a crime and
that it should be treated in specialized facilities, the data also reflect the increased
number of offenders and finally prisoners with drug problems and related
crimes (cf. Second Periodical Security Report, Bundesministerium des Inneren
and Bundesministerium der Justiz, 2006, pp 299ff).
Figure 7: Percentage of postponed executions of prison sentences for therapy
(Par 35 BtMG) related to all released prisoners in Germany and in
Baden-Wuerttemberg
9
8
7
6
5
4
3
2
1
0
1984
1986
1988
1990
1992
1994
GER
1996
1998
BaWue
2000
2002
2004
314
To some limited extent these changes contribute to the in total increased

number of suspended prison terms shown in figure 3. But although such inmates
are counted as early released prisoners, it is in fact not a real release in freedom.
Such inmates have to begin and complete a therapy, otherwise they will be incarcerated again. Hence, there lasts a lot of pressure on the clients to complete their
therapy successfully (cf. a general overview by Dolde 1995).
The results presented so far are still somewhat puzzling: On the one hand it
could have been shown that the portion of suspended prison terms has been increasing since many years and on the other hand, the rate of suspensions after
two thirds and after halftime of serving the prison sentence decreased. These contradictory results were to some extent outweighed by the increasing numbers of
postponed executions for therapy according to Par. 35 BtMG, but these numbers
are not able to explain the whole situation shown in figure 3. There must be further explanations and one of them will come from an increased number of early
releases for reasons of mercy.
Like all other decisions in the correctional system, the authority to decide on releases for reasons of mercy belongs to each German federal state (with the exception
of the rather seldom act of mercy by the Federal President). Therefore, the application
is extremely heterogeneous. Generally, three different ways of mercy are defined: the
statewide rules of mercy, the Christmas amnesty, and the possibility of halftime release according to substitute penalties. Usually, the statewide rules of mercy do not
define exact conditions for early clemency, mostly factors like the inmates personality or particular circumstances of the offence are taken into account by the state prime
minister or the state minister for justice.
While this kind of act of mercy is relatively rare, the Christmas amnesty is
much more often used in many federal states. As a rule, prisoners whose regular
term ends between end of October, respectively early November, and the first of
January of the upcoming year can be early released. Normally murder, sexual
offenders and severe violent offender are exempted from this amnesty. In 2006,
about 2000 inmates in German prisons were early released as a result of this
Christmas amnesty (Hartmann and Kraekel 2006).
The halftime release according to substitute penalties is a consequence of
overcrowding in prisons (cf. Die Justiz 1998). A substitute penalty means that the
offender was convicted paying a day fine, but has not been able or has not been
willing to pay this fine. Instead he has to serve an equivalent, rather short, prison
term. Because such a procedure is expensive and lead to overcrowding prison
facilities, e.g. the Ministry of Justice in the state of Baden-Wuerttemberg decided
in 1998, renewed in 2005 and 2006, that the enforcement authority might inter-
315
rupt the term after one half of serving for one year. After this year, without any
relapse, the rest of the term can be remitted in an act of mercy. Community services will be credited. Other federal states like Bavaria focus on crediting community services as an act of mercy (Bayerische Gnadenordnung [Bavarian Rules
of Mercy] Par. 31).
Figure 8 shows a steep increase in the use of acts of mercy, especially in the
state of Baden-Wuerttemberg since the late 1990s. In Germany as total, this increase occurred somewhat later and is hardly interpretably because of the estimated figure for 2004. Here, until 2002, there are no big changes in the percentages (1985: 2.4%; 2002: 2.9%), at best, there was a slight increase. But in 2004
the estimated portion reached 9.0%, it has, in fact, tripled. One has to keep in
mind that all German data refer to practices in the particular federal states and it
might be evident, that some of them followed the example of Baden-Wuerttemberg. The first increase in this south-western state was between 1997 and 1999.
In 1997, the portion of released inmates for reasons of mercy was about 4.9%,
two years later, in 1999, it has become 10.3%. This increase is definitely based
on the above mentioned decision of the Federal Ministry of Justice. Until 2002,
the portions depicted in figure 8 remained on a higher level, finally, they decreased to some small extent. Since 2003, a steep increase is observable, in 2005
20.4% of all released inmates in Baden-Wuerttemberg left prison because of reasons of mercy. Therefore, releases as an act of mercy has become a powerful tool
in regulating prison occupancy.
Figure 8: Percentage of releases as an act of mercy related to all released
prisoners in Germany and in Baden-Wuerttemberg
25
20
15
10
5
0
1984
1986
1988
1990
1992
1994
GER
1996
1998
BaWue
2000
2002
2004
316
5. Discussion
The aim of the paper was to discuss the question whether there has been an increase in punitivity within the criminal justice during the last decades. As expected, the data presented are not able to give a simple answer to this question.
Only looking at the overall sentence lengths in Germany, the punitivity hypothesis seems to be confirmed. Since many years, lately since the early 1990s and
therefore it is not an effect of the rising discussion on sex offenders the portion
of longer sentences has clearly increased. But otherwise the portion of suspended
sentences for probation has increased too which casts doubt on the punitivity hypothesis. It seems that todays offenders have a higher chance receiving a longer
prison sentence than 20 years before, but their probability of being incarcerated
has heavily decreased. In the view of the offender, this development is more
profitable, for him the main issue is to stay free and not the risk of a potentially
longer incarceration in the case of a relapse or a maybe elongated probation period. Furthermore it is meanwhile common sense in German penal policy that the
execution of short prison sentences, especially of those less than 6 months, is less
expedient because it causes costs and marginalizes the offender without any reasonable possibility of therapeutic or rehabilitative interventions (cf. DublerGmelin 2000; Janssen 2002; Sohn 2004). Here, at first glance, an increase in punitivity comes along with decrease in incarceration.
But there are some other developments stressing the necessity of a differentiated look on this phenomenon. One of these developments is the change in preventive detention (Sicherungsverwahrung). At first, the portion of such detainees for security reasons seems to have decreased within the last 30 years, but
since some years, a renaissance of this measure can be observed. In fact, it is a
consequence of the enhanced public as well as judicial attention on sex offenders.
But taking into account that the number of such offenses remains quite stable
since many years, the augmentation of preventive detention can be seen as an
increase in punitivity.
The main goal of the paper was to look at possible changes in the German correctional system in order to identify an increase or even a decrease in punitivity.
Because justice and prison are under authority of each federal state, a comparison
between overall German data (as far as available) and the data of the German
federal state of Baden-Wuerttemberg has been drawn. Unlike the development of
sanctioning and the practice of preventive detention, punitivity seems to decrease
within the correctional system. While the portion of completed prison terms
dropped, the portion of parolees increased. But a closer look at this trend showed
that it were not the numbers of suspended prison sentences at half or two thirds
317
of serving the term, neither for adults nor for juveniles. Such a development
would have been a real indicator for a more lenient practice within the correctional system, but on the contrary, this practice of early release has decreased.
The rise in early releases has been more or less due to two other developments:
On the one hand, the postponed executions of prison sentences for therapy of
drug addicts has heavily increased, a fact which can be attributed to the perception that such inmates are more often ill than evil, on the other hand there has
been a steep increase in releases as an act of mercy, mostly in combination with
the execution of substitute penalties. Latter can be regarded as a measure of reducing overcrowding in prison and can be seen in connection with a general tendency of avoiding short term imprisonment.
Generally speaking, a rise in leniency among the penal system is widely restricted to the remittal of a more or less short rest of the penalty or to the avoiding of short term imprisonments. In contrast, the remittal of longer parts of a
prison term, for example one third or even the half has been more and more reduced within the last 20 years. Sometimes, there are some variations between
Germany in total and Baden-Wuerttemberg as a single federal state, but the general trends are widely comparable.
Summing up, although some developments go in contradictory directions, a
general trend towards an increased punitivity in the German penal system is observable. In this respect, Germany seems to take a similar direction like other
western European countries, e.g. The Netherlands (cf. Althoff and Egelkamp
2006, van Swaaningen 2005) or Great Britain (cf. Garland 2001; Horsfield 2006;
summing up Sack 2006), although it is still far away from the British or even
American situation.
318
References
Althoff, M. & Egelkamp, M. (2006). Innere Sicherheit und Kriminalittsentwicklung. Ambivalenzen der niederlndischen Kriminalpolitik. Neue Kriminalpolitik, 18(1), 33-37.
Beckett, K. & Sasson, T. (2004). The politics of injustice (2nd ed.). Thousand Oaks, CA:
Sage.
Bundeskriminalamt (Ed.). (1981 2005). Polizeiliche Kriminalstatistik Bundesrepublik
Deutschland. Wiesbaden: Bundeskriminalamt; www.bka.de.
Bundesministerium des Inneren & Bundesministerium der Justiz (Ed.). (2006). Zweiter
Periodischer Sicherheitsbericht. Berlin: Bundesministerium des Inneren / Bundesministerium der Justiz.
Dubler-Gmelin, H. (2000). Gutachten der Kriminologischen Zentralstelle zur kurzen
Freiheitsstrafe und zur Ersatzfreiheitsstrafe. Bltter fr Strafvollzugskunde. Beilage
zu Der Vollzugsdienst, 3/2000, 5-6.
Die Justiz. (1998). berbelegung der baden-wrttembergischen Justizvollzugsanstalten;
hier: vollstreckungsrechtliche und gnadenrechtliche Manahmen, 47(4), 144-145.
Dolde, G. (1995). Drogengefhrdete und Drogenabhngige im Justizvollzug. In: A. Dessecker & R. Egg (Eds.). Die Strafrechtliche Unterbringung in einer Entziehungsanstalt (pp.
93-103). Wiesbaden: Kriminologische Zentralstelle.
Garland, D. (2001). The culture of control. Oxford: Oxford University Press.
Glaze, L.E. & Bonczar, T.P. (2006). Probation and parole in the United States, 2005. In:
Buerau of Justice Statistics Bulletin. Washington, D.C.: U.S. Department of Justice.
Hassemer, W. (2006). Die neue Lust auf Strafe. Frankfurter Rundschau, 296, 20.12.2000, 16.
Hartmann, R. & Kraekel, M. (2006). Weihnachtsamnestie fr ber 2000 Hftlinge. Die
Welt. http://www.welt.de/data/2006/12/08/1139149.html?prx=1.
Horsfield, P. (2006). Jugendkriminalpolitik in England und Wales: zwischen neuer Bestrafungslust und prventivem Interventionsrecht. Neue Kriminalpolitik, 18(2), 42-45.
Janssen, J. (2002). ber die Wirkung der kurzen Freiheitsstrafe. Zeitschrift fr Strafvollzug und Strafflligenhilfe, 51(1), 16-21.
Kaiser, G. (1997). Kriminologie (10th ed.). Heidelberg: C.F. Mller Verlag.
Kinzig, J. (1996). Zur Gefhrlichkeit von Sicherungsverwahrten. Bewhrungshilfe, 43(1),
31-40.
Kury, H., Ferdinand, T.N., & Obergfell-Fuchs, J. (2003). Does severe punishment mean
less criminality? International Criminal Justice Review, 13, 110-148.
Kury, H., Kania, H., & Obergfell-Fuchs, J. (2004). Worber sprechen wir, wenn wir
ber Punitivitt sprechen? In R. Lautmann, D. Klimke, & F. Sack (Eds.), Punitivitt.
Kriminologisches Journal, 36, 8th supplement, 51-88.
319
Kury, H. & Obergfell-Fuchs, J. (2006). Punitivitt in Deutschland Zur Diskussion um

eine neue Straflust. In: T. Feltes, C. Pfeiffer, & G. Steinhilper (Eds.), Kriminalpolitik und ihre wissenschaftlichen Grundlagen. Festschrift fr Professor Dr. Hans-Dieter
Schwind zum 70. Geburtstag (pp. 1021-1043). Heidelberg: C.F. Mller Verlag.
Kury, H., Obergfell-Fuchs, J., & Smartt, U. (2002). The evolution of public attitudes to
punishment in Western and Eastern Europe. In: J.V. Roberts & M. Hough (Eds.),
Changing attitudes to punishment. Public opinion, crime and justice (pp. 93-114).
Cullompton, Devon: Willan Publishing.
Kurze, M. (1995). Zurckstellung oder Unterbringung drogenabhngiger Straftter? Zur
Anwendung der 35ff. BtMG und 64 StGB anhand ausgewhlter Merkmale der
Betroffenen. In: A. Dessecker & R. Egg (Eds.). Die Strafrechtliche Unterbringung in
einer Entziehungsanstalt (pp. 77-92). Wiesbaden: Kriminologische Zentralstelle.
Lautmann, R., Klimke, D., & Sack, F. (Eds.). (2004). Punitivitt. Kriminologisches Journal, 36, 8th supplement.
Lissmann, C. (2006). Kopf ab? Die Zeit online, 04.08.2006.
Obergfell-Fuchs, J. (2006). Sinn und Unsinn kriminalprventiver Anstze bei Sexualstrafttern. In: J. Obergfell-Fuchs & M. Brandenstein (Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift fr Helmut Kury zum 65. Geburtstag (pp. 599-632). Frankfurt/Main: Verlag fr Polizeiwissenschaft.
Pastore, A.L. & Maguire, K. (Eds.). (2005). Sourcebook of criminal justice statistics
2003, 31st ed. Washington, D.C.: U.S. Department of Justice.
Pratt, J., Brown, D., Brown, M., Hallswort, S., & Morrison, W. (Eds.). (2005). The new
punitiveness. Cullompton, Devon: Willan Publishing.
Reuband, K.-H. (2004). Konstanz und Wandel im Strafbedrfnis der Bundesbrger
1970 bis 2003. In R. Lautmann, D. Klimke, & F. Sack (Eds.), Punitivitt. Kriminologisches Journal, 36, 8th supplement, 89-103.
Reuband, K.-H. (2006). Steigende Punitivitt in der Bevlkerung ein Mythos? Neue
Kriminalpolitik, 18(3), 99-103.
Diskussion der kriminalpolitischen Wende. In J. Obergfell-Fuchs & M. Brandenstein
(Eds.), Nationale und internationale Entwicklungen in der Kriminologie. Festschrift
fr Helmut Kury zum 65. Geburtstag (pp. 35-72). Frankfurt/Main: Verlag fr Polizeiwissenschaft.
Sohn, W. (2004). Zur Entwicklung der Freiheitsstrafen bis zu zwei Jahren. Zeitschrift
fr Strafvollzug und Strafflligenhilfe, 53(5), 264-270.
Statistisches Bundesamt (1976-2006). Rechtspflege. Strafvollzug Demographische
und kriminologische Merkmale der Strafgefangenen zum Stichtag 31.3. Wiesbaden:
Statistisches Bundesamt; www.destatis.de.
(1981-2006). Rechtspflege. Strafverfolgung. Wiesbaden: Statistisches Bundesamt;
www.destatis.de.
320
van Swaaningen, R. (2005). Public safety and the management of fear. Theoretical Criminology, 9, 289-305.
Weber, H.-M. & Reindl, R. (2001). Sicherungsverwahrung: Argumente zur Abschaffung
eines umstrittenen Rechtsinstituts. Neue Kriminalpolitik, 13(1), 16-21.
Young. J. (2003). Merton with energy, Katz with structure: The sociology of vindictiveness and the criminology of transgression. Theoretical Criminology, 7(3), 389-414.
321
Insecurity Feelings and Punitivity:

Relationship in a National Sample of
Adolescents and Young Adults in Spain
ALFONSO SERRANO-MALLO AND HELMUT KURY
1. Introduction
Empirical evidence suggests that levels of punitivity in the population have
steadily increased, at least in many western countries, in the last years (Tonry &
Farrington 2005). A group of developments, with which punitivity seems to be
related, have had an important impact in our societies: citizens seem to be more
worried, fearful and insecure, with a more than probable loss of quality in their
lives. The opening of the borders in Europe at the end of the 1980s and the immigration of foreigners increased feelings of insecurity. At the same time, the
complexity of life conditions seemed to be out of control, politicians were criticized for not solving problems like the rising costs of the health system, money
for the pensions or the unemployment problem. Late modern Society brought
more complexity and, then, stronger feelings of being unable to solve all the
problems. There were more risks which had to be solved (Garland 1997; 2001;
Beck 1986). The Laws and Criminal Justice system seem to be severer, even in
the face of the fact that crime rates have rarely increased, and in some countries
they even decreased (Bondeson, 2005; Garland, 2001; Roberts et al., 2003; Wacquant, 2005). Critics maintain that Governments and politicians are taking advantage of it to promote harsher criminal policies that, they contend, might bring
them more votes. Some authors even suggest that fear and punitivity have been
promoted by politicians at least in certain countries (Tonry, 2004). Topics of
crime, very often on the basis of single terrible cases like child misuse or killing,
were broadly discussed in the media, especially by tabloids regularly with the
background of demanding sharper punishments (Beckett & Sasson 2004). In
many countries, like the USA, crime policy plays an important role in election
campaigns. Politicians inform the voters about their ideas to fight crime, promise
sharper punishment, no excuse policies, etc. in order to increase their chance to
322
Alfonso Serrano-Mallo and Helmut Kury
be elected. So justice policy gets easier: the promise of harsher punishments is

more or less guaranteed in order to find the applause of the majority, echoed by a
big part of the media. In the research arena, efforts to describe, explain and prevent this punitive trend have become a second big tradition in Criminology, for
some more sophisticated and relevant than the old study of the causes of
crime. The ongoing research on punitivity, as well as on fear of crime, shows
more and more theoretical and methodological problems in this field of empirical
Criminology (see the chapters in this book). To solve these problems and to find
out more valid and reliable results is, so far, very important as fear of crime and
punitivity are topics with an immense influence on politics.
So, Kury and colleagues beside others (Warr 1995; Roberts & Hough 2002;
Roberts et al. 2003) insist that punitivity is a complex concept with many dimensions and that its scientific study is pervaded by methodological problems (Kury et
al., 2004; Roberts & Stalans 2000). Punitivity has different meanings. For example,
at the legislative level, punitivity means more behaviors forbidden by the Criminal law, severer or longer penalties at least according to the law, though maybe
not in practice , more difficulties to avoid prison, etc. At the level of the Criminal Courts, punitivity can equal, even without any change in the Law, harsher
sentences, a higher likelihood of prison terms, a shortage of defendant rights, a
higher percentage of served prison time, that is less early releases from prison,
etc. At the police level, an extended use of stops and arrests, extended use of
force, etc. At the personal level, people can favor more punitive measures in any
or all of these fields. Though these developments can follow, hand in hand, the
same trend, this is not necessary the case. For example, some evidence suggests
that politicians and especially judges and prosecutors have an exaggerated image
of citizens punitivity. This might bring more punitive measures at the legislative
and Criminal Justice levels without a parallel development at the level of the citizens. They can be influenced by media reports, especially from the tabloids, asking for stronger laws and sharper punishment, so whipping up a punitive turn.
In the case of Spain, punitivity has reached levels without precedent in some
of the directions just mentioned. The 1995 Spanish Criminal code was more punitive than the previous, which paradoxically dated from 1973, still in the Franco
era. Since 1995, there have been not least that 20 reforms of the Criminal code.
Most of them follow a punitive trend (Serrano-Mallo 2006a; Zugalda Espinar
2006). The incarceration rate, as showed in Table 1.1, has reached unknown levels. In 2006, there were 63,248 people in Spanish prisons and the incarceration
rate for the country was 141.466 for 100,000 inhabitants. The last official data
(16. 01. 2007) for incarcerated people amounts to 64,554.
Insecurity Feelings and Punitivity: A Sample of Adolescents in Spain
323
Table 1.1: Number of people incarcerated & imprisonment rate for Spain (2006
& 2007)
Number of people in
Spanish prisons
Imprisonment rate
(for 100,000 inhabitants)
2006
63,2481
141.466
2007
64,221
Source: Direccin General de Instituciones Penitenciarias & Instituto Nacional de Estadstica.
These developments have been traditionally unknown in a country with a relative

low crime rate. Spain had a moderate crime rate of about 49.5 officially by the
police registered crimes for every 1,000 inhabitants (that is 4950/100.000 inhabitants) in 2005. In the same year Germany had an official crime rate of 7747/
100.000 inhabitants (2006: 7647, since two years there is a downward trend)3.
This situates the country quite below the European mean and among the countries with the lowest crime rates in the continent. Due to the lack of national, repeated victimization and self-report studies, measurement of crime trends rely on
official data. These data strongly suggest that crime rates have remain stable, or
have even slightly decreased in Spain in the last few years. Despite the limitations of official data in the absence of other sources of information, most Spanish
commentators agree on the stable character of crime trends in recent years (Garca Espaa & Prez Jimnez 2004; Serrano-Mallo 2006b).
Despite this, there are indications, as we have mentioned, of an increase in punitivity in Spain at the level of the Law (the same can be shown for Germany, see
e.g. Hassemer 2001; Rckert 2006; Sack 2006; Kury 2006; Kury & Brandenstein
2006; see also other texts in this volume). At the same time, research suggests
that the increase in the imprisonment rate is a consequence of those changes in
the Law, and not so much of changes in the system of Administration of Justice
itself (Cid Molin 2005).
2
3
Mean of all the year 2006. Last published datum on the national population refers to the
total the 01. 01. 2006 (published the 01. 01. 2007).
Mean of the first 7 weeks of 2007.
Ministerio del Interior. Last published data refers to 2005. For Germany: Bundeskriminalamt (Ed.). 2006. Polizeiliche Kriminalstatistik 2005. Wiesbaden.
324
A different issue is to ascertain the punitivity level and whether there have been increases at the population level. The strong evidence of high and increasing punitivity
in many places can not be generalized elsewhere. Though imprisonment rates are
some times interpreted as a punitivity indicator for the individual level, it is a weak
indicator at best (Hinds 2005; Nelken 2005). At the same time, research has found
diverging trends in different countries and, within punitive countries, important differences among regions and subpopulations (Bondeson 2005; M. Brown 2005; Hinds
2005; Meyer & OMalley 2005; Nelken 2005; Tonry 2004; Wacquant 2005), and
even some commentators come close to say that the discourse on the culture of control is an exaggeration (D. Brown 2005; Meyer & OMalley 2005). Unfortunately,
Spain has a long anti-empirical tradition and, again, there is a lack of research in the
area of punitive attitudes among its citizens. Some general surveys conducted by the
Spanish Centro de Investigaciones Sociolgicas have included a question relative to
the support for death penalty. This is the case, for example, of studies number 2212,
2440 and 2574, conducted in 1996, 2001 and 20044. The results for the questions regarding death penalty are shown below in Chart 1.1. Since these studies, among other
methodological disparities, treat different themes, use different questionnaires and
even wordings and have sample specialties (the most obvious of which is that the
population for the 2001 survey are people 15-29 years old, and for the other two 18 or
older), direct comparisons among them are very difficult to justify.
CIS sample design includes several steps, in all of which pains are taken to preserve randomness. The last step, a random routes process, is probably the most problematic one. Needless to say, not every element in the population has the same,
known probability of being interviewed and these surveys merit the same concerns
of most of their kind. With this caveat in mind, most scholars consider that CIS surveys offer a basis for inferences to the Spanish population, especially when introducing weights. CIS surveys use face to face methodology, without the assistance of
computers. As Chart 1.1 shows, in april 1996, only a 14% felt rather appreciative
(ms bien simpata) for death penalty, while a 72% felt rather disapproval (ms
bien rechazo). In december 2001, and with a population of people 15-29 years old,
17.2% said they were in favor of death penalty and 77.3% against. Finally, in
october 2004, 18% were in favor and 72.2% against.
Total N, respectively, 2499, 2471 and 2491.
325
Chart 1.1: For and against death penalty in Spain

Chart 1.1 - For and against death penalty in Spain
For
100%
90%
Against
7
7
Indifrnt.
NK/A
5.5
9.8
77.3
72.2
17.2
18
80%
70%
60%
50%
72
40%
30%
20%
10%
14
0%
Ap. 1996
Dec. 2001 (15-29 yo)
Oct. 2004
Source: Centro de Investigaciones Sociolgicas
Several issues are probably affecting these answers, not to mention measurement
error related problems. For example, qualitative research suggests that some people might just show their political discontent when arguing in favor of death penalty (see for Germany the quantitative and qualitative survey on punitivity from
Kury et al. in this volume). At the same time, Spaniards are, probably, those most
worried about their self-image among Europeans, which might point to social
desirability bias. Whether and to which extent all of this is biasing the estimates
is unknown, but calls for further caution. For this reason, comparative research
including some Spanish regions that reach the same conclusion that Spain is not a
punitive country at all is problematic (Mayhew & van Kesteren 2002; Roberts
2002).
At any rate, the data, extremely limited as it is, does neither point to an obviously high punitivity among the Spanish population nor to an increase in it.
There is a statistically significant difference (p < .05) between the data for 1996
and 2004, but the wording of the questions are not the same and the 1996 survey
includes the response category indifferent. This and other considerations preclude any comparison. That death penalty (abolished in Spain since the 1978
326
Constitution) is not a priority for Spanish people is pointed by an additional survey conducted by CIS in december 2000 (number 2486). It was asked to respondents if they favored changing the Constitution and, in that case, why in an
open way. 33.6% (N = 835) manifested that the Constitution should be changed,
and then only a 1.3% (N = 11) of those said that the reform should be the inclusion of the death penalty5.
2. The rise of punitivity and its explanation

Not only little has been investigated in Spain about the level and trends of punitivity, but explanations are also missing. Spain is not the only place where this
last concern applies. Though the new punitivity trends, well established in parts
of Europe and North America, have attracted huge attention in Criminology,
there is a relative lack of sound theoretical explanations at the individual level, at
least in comparison with the tradition of the causes of crime (for exceptions that
at least offer grounds for building such explanations, Garland 2001; Tonry 2004).
The understanding of phenomena as punitivity represents an aspiration for many
citizens. At the same time, it seems difficult to counterattack these developments
without knowing anything about its causes. Other studies, especially those establishing correlates of punitivity, are clearly necessary and compatible with explanation, but not enough for those goals.
One of us, Kury, has advanced a theoretical concept that tries to explain why
people are more or less punitive. This concept has been developed in the last
years and is built upon the research that Kury and colleagues have conducted in
Germany and other places (Kury 1996; 2004; Kury et al. 2002; Kury et al. 2006).
Especially interesting is his work in the former socialist, east European countries.
Kury conducted different pieces of research in those countries and was able to
analyze in some parts the impact of the quick changes that took place during the
years following the fall of the Berlin Wall and the dismantlement of the Soviet
Union. In a nutshell, the research found that officially registered crime increased
steadily in these countries, though did not reach in most of the countries until
today the levels of the western countries. Punitivity was on the background of the
experience in the former Soviet Union relatively high and as well increasing
among those people, as well as especially other feelings as fear of crime and insecurity. No doubt, theory was needed to make sense of those developments.
Kurys proposal, then, has the advantage of being built upon and contrasted with
5
Total N for the study = 2486. A similar study conducted in december 1996 (number 2201, N
= 2478) found that 46% were in favor of changing the Constitution, and 19% of those gave
as a reason Severer penalties. Death penalty. Unfortunately, then, at least two different
answers have been aggregated by CIS.
327
sound empirical evidence, of being grounded theory, in a field where scholars

as D. Brown (2005) show their discontent with speculative approaches short of
empirical research.
According to Kury (Kury & Ferdinand, 1999), there is a relationship between the
perception by citizens of the crime rates and especially their increase and victimization risk on one hand and punitivity in the other. The problem is that this perception
is highly fictitious and tends to be exaggerated because it is especially formed with
information coming from the media. Medias image of crime is simple and centered
around the most bizarre and serious episodes, which are absolutely unrepresentative
(Beckett & Sasson 2004, pp. 73-102). One of the most noteworthy characteristics
of contemporary crime news has been its tendency to focus on the rarest types of
crime, such as murder and robbery (p. 75). In Germany reports on crime, especially
in the tabloid media, concentrate beside these crimes on sex offences, so stimulating
the impression that there is an increasing problem which cannot be found in the official crime statistics. Sexual misuse of children, one of the most emotionalized
crimes by the media is dropping since at least 30 years (see Bundesministerium des
Innern Bundesministerium der Justiz 2006, 95-105). The media, then, offers a
highly biased image of crime that influences citizens perceptions and, then, their
level of punitivity (D. Brown 2005; Kury & Ferdinand 1999; Leishman & Mason
2003; Roberts & Stalans 2000). This idea is consistent with research that has found
that better information tends to lead to less punitivity. According to the theory, in
second place, social and political conditions greatly affect reactions and attitudes to
crime: the more insecure people feel for whatever reason, the more punitive they as
well tend to be (Kury & Ferdinand 1999). Punitivity and fear of crime might have a
similar background which can be seen in the fast changes in society, the rising problems people are confronted with, like unemployment, financial problems, public
discussions about world wide problems like pollution, increasing temperature, just to
mention a few. There has never been a period in History in which societies were
changing so fast as nowadays. These changes bring new chances for many people,
but at the same time it creates a fears of becoming the looser at the end. Old rules are
not any more valid, new ones are yet to be established, and this can be especially
seen in the countries of the former Soviet Regime.
The theory is consistent with Durkheims tradition of anomie theory: in an
anomic situation, when a society is perturbed, for example by quick changes, the
value system weakens and crime and suicide rates are predicted to rise. At least
in some cases similar to these, increments in punitivity can be as well predicted.
Kurys original formulation of the hypothesis is built at a macro or sociological
level. It is concerned with punitivity rates among communities and countries and
not with individual tendencies. The question of the level of analysis is a very im-
328
portant one in the social sciences (Alexander & Giesen 1987). Though it is clear
that an integration of macro, micro and even meso levels of analysis is a challenge for the field (Tittle 2000), it also seems apparent that certain relations must
exist among them and that, though it is not advisable to spring from one level
to the other, some relationships there must exist (Hannan 1991). For example, the
level of punitivity of a community is the result of the aggregation of the levels of
punitivity of its members. With this caveat in mind, we propose to extend Kurys
original idea to the level of individuals.
At the individual level, research shows that values play a very important role
in punitivity. For example, Boeckmann and Tyler used a sample of north California residents to study their support for the severe Three Strikes Law, as well as
for other punitive issues. They found that social values were the most important
predictor, followed by evaluations of moral cohesion and social conditions
(Boeckmann & Tyler 1997). Though there are some methodological concerns
with this investigation, especially potential bias due to the lack of a true representative sample, it clearly points to the need to take into account values when looking at feelings of insecurity and punitivity.
Having said that, it is important to keep in mind that punitivity, at least at the
macro level, can dramatically increase very quickly, for example after a terrible
case of crime and corresponding presentation and discussion in the media. Recent research all around the western world as well as Kury and colleagues research in the former socialist countries have found evidence for very quick and
dramatic changes in the levels of punitivity. Insecurity feelings theory offers a
mechanism that is able to offer an explanation for such rapid changes. The feeling of insecurity can, besides feelings of anger, change quite quickly, and this has
the potential to explain equal quick changes in punitivity at the individual and
aggregated level. Values, culture and other variables, as currently used in other
approaches and theories, lack this potential.
Our working hypothesis are, then, that (H1) values will be related to punitivity
in a statistically significant way; & that (H2) feelings of insecurity will as well be
related to punitivity even after controlling for values.
3. This study
The present study, then, tries to test a part of Kurys insecurity feelings and punitivity hypothesis at the individual level, adding at the same time a role for values.
In order to do so, we use recent CIS survey Sondeo sobre la juventud espaola,
2005 (Primera oleada) (study number 2596). 1,433 interviews were conducted
(out of the 1,500 designed) in March 2005. Its universe was formed by adoles-
329
cents and young adults, i.e. 15-29 years old, in Spain. As is usual with CIS, the
sample design follows several steps that try to preserve randomness, being the
last one a random route process. Interviews are conducted face to face in the respondents house, without the assistance of computers (de Leeuw & Collins
1997). Quotas for sex and age apply. Real sampling error can not be computed.
Though not strictly representative, general socio-demographic characteristics of
the sample tend to match those of the general population, so that scholars generally assume that there is some base for inference.
This survey merits most of the well know criticisms related with this methodology.
In addition, questionnaires do not follow any specific theory or set of scientific hypothesis, but are built in rather oportunistic, non-systematic ways. This survey was
performed as a request by the Spanish Instituto de la Juventud, so that the interests of
this institution, mostly descriptive, played an important role in the questions and their
wordings, what gives way to validity and reliability concerns. Unfortunately, insufficient information is given as to evaluate and advance the effects of the different methodological decisions and options followed by the survey. Indications of significant
measurement error in the statistics abound. Non coverage, non response, interviewer,
mode, design, etc. effects are here matters of deep concern, but difficult to evaluate
due to the lack of information. We are left, then, with quite crude data. But this is the
first survey about this important topic in Spain. Fear of crime and punitivity can have
a strong influence on politics so it is also from this point of view very important to
evaluate these backgrounds.
On the other hand, researchers more often than not need to rely on secondary
data that, among other concerns, has not been collected to test any specific group
of criminological hypothesis. This is specially the case in countries with a long
anti-empirical tradition, such as Spain, and where extra-scientific factors may
play a role in grant access. Though in many cases secondary data are as unsatisfactory as the one we have here at our disposal, in the near future it might become even more important due to the limitations for conducting surveys. For example, the requirement of active parental permission by several U.S. States
makes surveys among juveniles extremely costly and at risk of serious bias.
With these important limitations in mind, our data includes certain relevant
variables for our purposes.
3.1
Dependent variables: punitivity
CIS 2596 includes the following question: Independently of what you would
personally do, I would like you to tell me if you are in favor or against [...] applying death penalty to persons with very serious crimes (a personas con delitos
330
muy graves)?. It is a closed question. Table 3.1 shows the results for this question. 32% are in favor and 60% against. Note that the question refers to persons
with very serious crimes, not to death penalty in general. This precludes comparisons with other surveys, at the same time explaining the relatively high percentage of respondents in favor of death penalty.
Chart 3.1: Percentage of respondents in the ICVS 2005 preferring a prison sentence in case of repeated burglary (compared with earlier waves of
the survey; see: Dijk 2007: 88, Figure 3.15)
331
From a German point of view the number voting for death penalty in cases of
very serious crimes is not extremely high. In Germany the last years round about
one third of respondents to similar questions where in favor for death penalty.
The results of the last wave of the International Crime and Victim Survey from
2005 also shows that Spain has in comparison with other countries a not high
punitivity rate. Chart 3.1 shows the results for punitivity found in this broad survey done in 18 countries including Germany and Spain. Respondents were given
the case of a recidivist burglar who has stolen a color TV. They were asked about
their preferred punishment for this case. Results show the percentage of respondents preferring a prison sentence (see Dijk et al. 2007: 88, Figure 3.15). Both
countries participated only in the first (1989) and last waves (2005) of the international survey, not in those of 1992, 1996 and 2000.
Chart 3.2: Percentage of public favouring imprisonment in case of a young recidivist burglar according to the results of last wave of ICVS & actual
prisoners rates in 2002/03 (Dijk et al. 2007: 90, Figure 3.17)
332
According to chart 3.1, punitivity in Spain was reduced, between 1989 and
2005, from 27 % voting for a prison sentence for the young burglar to 17 %. In
Germany there was an opposite development: from a relatively low punitivity
level in 1989 13 % voted for imprisonment for the offender to a higher one in
2005 19 % voted for imprisonment. That is, while there was a reduction in puntivity in Spain, there was at the same time an increase in Germany. Meanwhile
obviously more Germans voted for a harsh punishment than Spaniards (19 % vs.
17 %). The difference is small and should be interpreted with caution because of
methodological problems.
These results are insofar interesting as Spain has a higher imprisonment rate
than Germany. The results of the last wave of ICVS show, as calculated by van
Dijk et al. (2007: 90, Figure 3.17), a small correlation between the imprisonment
rate of a country and its punitivity level: respondents living in countries with
higher imprisonment rates tend to be more punitive in the case of the young burglar. This is not true in comparing Germany and Spain (see Chart 3.2).
The above mentioned wording results in the Spanish survey in higher discrimination or heterogeneity in the responses, which represents an important analytical advantage. We assume that people in favor of death penalty in these cases
are much more punitive than those who are not in favor. Does not know and
does not answer cases are excluded from later analysis. Though in some studies punitivity measurement rests in single questions about death penalty, this
strategy seems limited (Roberts et al. 2003).
Table 3.1: Percentage in favor & against death penalty, Spain 2005
(15-29 years old)
In favor
Against
Does not know
Does not answer
32%
60%
6.6%
1.4%
Source: CIS 2596 (p. 1904). N = 1,433.
CIS 2596 includes as well a question about punishing drug use. It is worded as
follows: Talking in general, do you think drug use should be punished (penalizado)?. It is as well a closed question with three response categories. We treat
this variable as ordinal. Table 3.2 shows the results: 44.4% thinks drug use
should be always punished, 32.2% thinks that only hard drug use should be punished and, finally, 20.4% answers that is should never be punished. Again, we
have a high discriminative question. Does not know and does not answer
cases are excluded from later analysis.
333
Table 3.2: Percentage in favor and against punishing (penalizar) drug use,
Spain 2005 (15-29 years old)
Yes, always
Only hard
drugs
No, never
Does not know
Does not
answer
44.4%
32.2%
20.4%
2.9%
0.1%
Source: CIS 2596 (p. 17). N=1,433.
We use the previous variables to construct an index of punitivity. Index construction is always a problematic issue since the natural order of things is altered to a
certain degree. We assume that the most punitive respondents are those in favor
of death penalty. As a result we give special weight to this variable in our index.
For each individual in our sample, the punitivity index is constructed as follows:
Punitivity = (In favor of death penalty = 3) + (Drug use should always
be punished = 2; (Only hard drugs = 1)
Table 3.3 shows the frequency distribution for the index.
Table 3.3: Punitivity Index

Frequency
Percentage
Valid Percentage
Little punitive
200
14
15.6
293
20.4
22.9
334
23.3
26.1
75
5.2
5.9
130
9.1
10.1
Very punitive
249
17.4
19.4
Total (valid)
1,281
89.4
100
Lost cases
152
10.6
1,433
100
Total
Its range is 0 (little punitive) 5 (very punitive), so that the index has 6 categories. The assumption is that those higher in the index are also higher in punitivity.
As analytical strategy, though OLS is some times used with dependent variables
with 5 or more categories, the researcher should consider special methods for
limited dependent variables.
334
3.2
Independent variables: feelings of insecurity
Insecurity and feelings of insecurity are like punitivity complex constructs,

with many dimensions. Special care is required for operationalization and measurement in a valid and reliable way. Use of surveys not designed to measure it is,
then, highly problematic, and any result should be interpreted with caution. With
these important limitations in mind, CIS 2596 offers at least a limited way to
construct one aspect of feelings of insecurity. We assume that one of the dimensions of feelings of insecurity is an economic one. 6 questions included in the
survey about the degree of satisfaction with work, the general degree of satisfaction, the degree of satisfaction with the personal economic situation, the degree
of satisfaction with the goods the respondent has, the general economic situation
of Spain and the personal and familiar economic situation seem to evoke an economic content or other contents that can be influenced by economic feelings of
security. We assume that respondents ranking lower in these variables have as
well a higher feeling of economic insecurity. This is, though a vital one, the
weakest construct of our study.
As analytic strategy in order to explore the dimensions that these questions
might hide and for data reduction, we conducted a principal components analysis.
We selected a random sample that amounted to the 30% of our total sample. We
run an exploratory principal components analysis and applied a Varimax rotation,
which imposes ortogonality. The optimal solution included two factors with eigenvalues over 1 (2.492 and 1.098) and explained an acceptable 59.8% of the
variance. We have named the first component Economic satisfaction. The variables that load the most on it are those referred to different degrees of satisfaction. As we saw, all of them have at least an indirect economic nature. The second component is more clearly an Economic situation one. This preliminary
finding was confirmed with analysis with the whole sample.
We prefer to think about these two principal components as approximations
for a certain dimension of feelings of insecurity that touches on economic issues.
Needless to say, this is a crude measure and next studies should construct better
ways to operationalize and measure the feelings of insecurity.
3.3
Independent variables: values and other controls
We have added the following independent variables to our models.

1. Characteristics of the place of residence. Residence in a high criminality region. This is a dichotomous variable that takes into account whether the respondent lives in a region of relatively high criminality or not. In Spain, these regions
are Madrid and the Mediterranean arch. Size of community. Its 7 categories refer
335
to the population of the community in which respondent resides. They range

from 2,000 or less inhabitants to more than 1,000,000. These first two variables
have a geographic-structural nature, and form our first theoretical model.
2. Socio-demographic variables. Age, Sex, Education of first salary. Refers to the
education of that person that brings the most money to the household. It can either
be the respondent (21.5% of all answers) or another person (75.9%). Unless the
young or young adult lives on his/her own or with his/her partner or friends, the person with the first salary is usually one of the parents, specially the father. Note that
Spanish youth live with their parents until relatively high ages. Married.
3. Values. Prodelictive values. CIS 2596 questions about the degree of acceptability/unacceptability of different criminal and deviant behaviors, including
vandalism, public intoxication, violent encounters with the police, lack of integrity in exams and public examinations, contract of foreign workers under unfair
labor conditions, steal from general stores or supermarkets, be noisy at night, use
of public transportation without paying and smoking in public places. A 10 response categories scale (from absolutely unacceptable to absolutely acceptable) is used. A factor analysis confirms our initial hypothesis that there is here
a unidimensional construct. Political ideology. Respondents are asked to put
themselves in a 1-10 scale, where 1 represents the political left and 10 the political right. Importance of work in life. Importance of money in life. Both last questions use Likert scales. Its response categories are Not important, Little important, Rather important and Very important. Religious. A dichotomous
variable.
4. Education and stress. Education. With 6 categories. Stress or nervousness in the
last three months. Whether the informant suffered that feeling in the last three months.
A dichotomous variable. According to strain theories, this might produce anger and
other negative sentiments. Whether this might affect punitivity is unknown.
These variables capture to certain partial degree Kury's theoretical proposal (at
the individual level) and at the same time control for a few alternative variables
that research has showed to have a potential impact on punitivity. We have tried
to keep a balance between predicted and control independent variables.
Our analytic strategy tests 4 theoretical models and groups of variables, as we
will see. The first includes independent variables from group 1; the second adds
the independent variables from group 2; the third adds the ones in groups 3 and
4; and, finally, the fourth and total model adds our measures of feelings of insecurity. Table 3.4 shows the descriptive statistics for these variables. A first indication of a missing data is that N = 609 (listwise), a problem about with we will
have more to say later.
336
Table 3.4: Descriptive Statistics (Independent Variables)
1433
1433
1433
1396
1431
1333
1072
1422
1428
1389
1422
1425
891
891
609
0
1
15
1
0
0
-1.15522
1
1
1
2.00
0
0
-3.48253
-3.97255
Maximum
1
7
29
2
4
1
5.89083
10
4
4
7.00
1
1
2.63723
2.31160
Mean
.66
3.94
22.36
1.48
1.99
.11
.0047403
4.58
3.50
3.33
3.5637
.68
.23
-.0391729
-.1120287
Std.
Deviation
.475
1.693
4.358
.500
1.308
.317
1.002383
1.808
.603
.654
1.06477
.469
.421
1.002170
1.024541
Skewness
Statistic (std. error)
-.661 (.065)
.237 (.065)
-.108 (.065)
.099 (.065)
-.169 (.065)
2.444 (.065)
1.120 (.067)
.336 (.075)
-1.07 (.065)
-.672 (.065)
.68 (.066)
-.749 (.065)
-.661 (.065)
-.145 (.082)
-.578 (.082)
Statistic (std. error)

-1.566 (.129)
-.883 (.129)
-1.21 (.129)
-1.993 (.129)
-.888 (.131)
3.979 (.129)
1.536 (.134)
.234 (.149)
1.549 (.13)
.352 (.129)
.105 (.131)
-1.442 (.13)
-.354 (.13)
.347 (.164)
.406 (.164)
Alfonso Serrano-Mallo und Helmut Kury
Region Dq
Community Size
Age
Sex
Education First Salary
Married
Prodelictive Values
Political Ideology
Importance of Work
Importance of Money
Education
Religious
Stress
Economic Satisfaction
Economic Situation
Valid N
Minimum
Insecurity Feelings and Punitivity: Results from Spain
337
4. Results
We have tested our hypothesis using the well known and common Ordinary
Least Squares (OLS) Regression method. We have constructed and tested four
models, from the simplest to the full one, that includes our proximate measures
of insecurity feelings. Tables 4.1 and 4.2 show the Model Summary and the OLS
Regression Coefficients. It is worth repeating that the OLS analysis is only using
689 cases.
a
Table 4.1: Modell Summary

Model
1
2
3
4
a.
* Sig.;
F Change < .05;
Adjusted R Square
.009*
.027*
.146**
.152*
Std. Error of the Estimate

1.717
1.702
1.595
1.589
** < .0005.
Model 1 includes our geographic-structural independent variables living in a region with relative high criminality, i.e. in Madrid or the Mediterranean Arch; and
size of the community where the respondent lives. Both the model and the individual predictor living in a high crime region are related to punitivity in a statistically significant fashion. The model offers a very modest adjusted R2, but it is
statistically significant. The effect of living in the Madrid or Mediterranean area
is quite modest: living in those regions increases by less that half a point (0.349)
the degree of punitivity in our 6 points scale. On the other hand, the smaller the
community size, the less punitivity individuals tend to be, but in a very weak and
statistically non-significant (0.1 > p > .05) way. That result, suggests that, as predicted by Kury, objective crime rates might play a role in punitivity, since living
in the Spanish regions with relatively more crime, and controlling for community
size, is related to punitivity. Since the relationship is predicted by the theory to be
mediated by individual factors, we expect that this effect will disappear as we
add other variables in subsequent theoretical models.
Model 2 adds some individual characteristics. Age, education of the person
with the highest salary of the household and being married are related to punitivity in a statistically significant way. Being older is related to a very little less punitivity. In fact, these results predict that one would need 28 years to be one level
less punitive in our 6 categories scale. The higher the education of the first salary
in the household (usually one of the parents), the less punitive the respondent
tends to be. And, finally, being married is related now to higher punitivity.
Maybe the most relevant finding of this regression analysis is that sex has not
been found to be related to punitivity. Some research has found that women are
less punitive than men (Kury et al. 2002; Mayhew & van Kesteren 2002), some-
338
thing not consistent with our results, when controlling for other individual and
socio-demographic variables. Living in a relatively high crime region is still statistically significant. The model shows a modest adjusted R2 of 0.027. These
variables are predicted to lose their relevance once values and feelings of insecurity are added to our theoretical model.
Our extension of the theory places a basic role in values. It is as well important
to remember that it takes a long time for values to change at the individual level,
so that a theory limited to them would have a hard time predicting and explaining
potential quick changes in punitivity (at the individual level). Model 3 introduces
our value measures. All of them, except importance given to work in life, are related to punitivity (p < .05), when controlling for the others, and in the direction
predicted by the theory and by previous research. This is consistent with our first
hypothesis: values are related to punitivity. There is hardly anything surprising
here. Those with higher prodelictive values, in the sense that are more understanding of criminal and deviant acts, tend to be as well less punitive. Those who
are more conservative or more religious tend to be as well more punitive in a statistically significant way. Those with higher education tend to be less punitive.
The important thing to keep in mind is that all of these variables have an independent impact on punitivity even after controlling for other value related measures. Previous research has found similar results. As we will see later, we did not
find clear evidence of severe multicolinearity in our models, though at the same
time this phenomenon is not very usual in social sciences. The strongest relationship is that of education ( = -0.222). But it is as well a tricky one. It is sometimes thought that people with higher education and usually, then, better informed about crime and penalties favored less severe responses to crime and, in a
way, know better (Roberts et al. 2003). But evidence also shows that these person might be more worried about their self-image and, therefore, carefully
choose their answers; and that they are as well more aware of potential second
intentions of survey questions, therefore changing their true opinions. Until
these alternative interpretations are subjected to further research, caution must
prevail in this point. Maybe surprisingly, as respondents give more importance to
money, they tend as well to be more punitive; while the relation with importance
of work is not statistically significant.
As predicted by the theoretical concept, geographic-structural variables and
socio-demographic variables lose their statistical significance, with the exception
of the education of the person that brings the most money to the respondents
household, that is going to stay with us as well in the fourth, full model. It is also
important to note that the adjusted R2 for the model increases to 0.146, therefore
explicating almost a 15% of the variance.
339
The full model adds our approximations to insecurity feelings and includes,
then, all our independent variables. It shows a statistically significant increase in
the adjusted R2 to 0.152, though modest. The individual estimates for the two
principal components of economic insecurity feelings are in the expected direction (the higher in these variables, the less punitive individuals tend to be), but
only one reaches the level of statistical significance (p < .05), the one we have
termed Economic situation. The best one considers the economic situation, the
less punitive one tends to be (b = -0.16). Though the relation is not very intensive, it adds credibility to our second hypothesis: feelings of insecurity are related to punitivity even after controlling for values. In addition to these considerations, little changes respect to Model 3 either regarding the Model itself or the
individual coefficients. Values keep their significance. Education and Political
Ideology show the higher standardized coefficients (-0.197 and 0.159). In addition, and because the well known limitations of the statistical significance approach, we include the 95% coefficient intervals for the full Model (both for the
OLS and for the Ordinal Regression). They confirm the previous considerations,
but show clearly the lack of precision of our analysis. We will return to this issue
later.
Some of the OLS assumptions are violated to a degree that deserves attention.
In these situations, estimators do not remain BLUE, though their effects differ.
Residuals, to begin with, are not normally distributed. This is easy to detect, even
visually, in the case of large samples. At the same time, its effects with large
samples are not in general a matter of concern. As would be expected, age correlates highly with some other variables, though there are not evident signs of severe multicollinearity. We have detected a problem of heteroskedasticity and the
presence of some outliers. This is more problematic. In order to check the impact
of these violations of the assumptions especially of the OLS regression model,
we have run alternative regression models that perform better under these conditions. Particularly, we have run Robust Logit, Robust Probit and HC2 Robust
regressions (data not shown). All of them confirm our initial results regarding the
support for the general model6 and the theoretically relevant individual coefficients7. These results hardly surprise us since the OLS regression method is quite
robust, at least in the case of not dramatic violations of its assumptions. All of
this adds confidence to our results.
The Robust Logit and Robust Probit show a very modest pseudo-R2 of 0.0623 and 0.0624.
Since the Robust models tend to be more conservative, this comes as well as no surprise.
In some of these models, one (and only one) individual variable looses its statistical significance.
.394 (.144)*
-.068 (.04)+
1
b (st. error)
2.314 (.173)***
.394 (.144)*
b (st. error)
-.068(.173)***
(.04)+
2.314
.108
-.067
.108
B
-.067
-.091
-.131
.092
-.131
.092
-.172 (.054)**
.503 (.224)*
2. a, non-significant; +p<.1; *p<.05; **p<.005; ***p<.0005.
-.091
.095
.095
2
b (st. error)
3.142 (.473)***
.344 (.143)*
b (st. error)
a
3.142 (.473)***
-.036
(.017)*
.344 (.143)*
aa
-.036 (.054)**
(.017)*
-.172
a
.503 (.224)*
a, non-significant; +p<.1; *p<.05; **p<.005; ***p<.0005.
1. DV: Punitivity
2.
DV: Punitivity
Economic
situation
1.
(Constant)
High Crime Area
Community Size
(Constant)
Age Crime Area
High
Sex
Community
Size
Age
Education First Salary
Sex
Married
Education
First
Salary
Prodelictive
Values
Married
Political Ideology
Prodelictive Values
Importance of Work
Political Ideology
Importance of Money
Importance of Work
Education
Importance of Money
Religious
Education
Stress
Religious
Economic Satisfaction
Stress
Economic
EconomicSatisfaction
situation
Model
3
3
b (st. error)
1.056 (.632)+
.25 (.135)+
b (st. error)
a
1.056 (.632)+
a
.25 (.135)+
aa
-.109 a(.051)*
aa
-.109
(.051)*
-.205 (.064)**
a
.146 (.036)***
-.205 (.064)**
a
.146 (.036)***
.372 (.102)***
a
-.36 (.07)***
.372 (.102)***
.283
(.140)*
-.36 (.07)***
a
.283 (.140)*
Model
.077
.141
-.222
.141
.077
-.222
-.119
.153
.153
-.083
-.119
-.083
.069
.069
1,2
-.16 (.062)*
b (st. error)
1.082 (.631)+
.25 (.135)+
b (st. error)
a
1.082 (.631)+
a
.25 (.135)+
aa
-.102 a(.052)*
aa
-.102
(.052)*
-.208 (.064)**
a
.152 (.036)***
-.208 (.064)**
a
.152 (.036)***
.357 (.102)**
a
-.319
(.071)***
.357 (.102)**
.288(.071)***
(.140)*
-.319
a
.288 (.140)*
aa
a
-.16 (.062)*
Table 4.2 - Linear Regression Coefficients for Four Models of Punitivity (OLS)
Table 4.2: Linear Regression Coefficients for Four Models of Punitivity (OLS)1/2
-.095
-.095
.078
.135
-.197
.135
.078
-.197
.159
-.121
.159
-.077
-.121
-.077
.069
.069
4
B
-.281
-.039
-.157
2.320
95% C.I. for
b
-.015
.515
-.046
.108
-.157
2.320
-.024
.050
-.015
.515
-.123
.373
-.046
.108
-.024
.050
-.203
-.001
-.123
.373
-.266
.584
-.203
-.001
-.334
-.082
-.266
.584
.082
.222
-.334
-.082
-.296
.136
.082
.222
.156
.558
-.296
.136
-.460
-.179
.156
.558
.014
.562
-.460
-.179
-.433
.148
.014
.562
-.129
.119
-.433
.148
-.129
.119
-.281
-.039
95% C.I. for b
340
Alfonso Serrano-Mallo und Helmut Kury
341
On the other hand, though OLS regression is usually used in the literature with
limited dependent variables with more of 5 categories, there exist some better
methods to deal with this kind of data. In the last years, and in a continuing expansion, better and more robust statistical methods are built up and become accessible to potential users. In particular, Ordinal Regression methods perform
better with limited dependent variables, as in our case. One of the best known
options uses the Logit Link function. We have run such a regression8. The test of
parallel lines is not significant (p = 0.876), so that the null hypothesis that the
location parameters are the same across response categories can not be rejected.
Table 4.3: Estimators for the Full Punitivity Model (Ordinal Regressiona)
Threshold
Location
[Punit4 = 0]
[Punit4 = 1]
[Punit4 = 2]
[Punit4 = 3]
[Punit4 = 4]
Region dq
Size Com
Age
Sex
Edu First
Married
Delq Vals
Ideolog
Imp Work
Imp Money
Educac
Relig
Stress
Ec Satisf
Ec Situat
Estimation
(error)
-1.397 (.917)
.165 (.914)
1.393 (.916)
1.687 (.916)
2.342 (.919)*
a
a
a
a
a
a
-.404 (.085)***
.145 (.045)**
a
.481 (.134)**
-.44 (.084)***
.563 (.175)**
a
a
-.195 (.082)*
C.I. 95%
Wald
2.318
.033
2.313
3.387
6.491
1.256
.148
.050
.012
.790
.729
22.359
10.520
.310
12.797
27.749
10.308
.089
.023
5.599
-3.195
-1.627
-.402
-.110
.540
-.152
-.073
-.057
-.335
-.174
-.265
-.571
.057
-.387
.217
-.604
.219
-.416
-.171
-.356
.401
1.957
3.187
3.483
4.143
.558
.108
.046
.300
.066
.674
-.236
.232
.216
.744
-.277
.906
.306
.147
-.033
Link func.: Logit.

a. a, non significant; *p<.05; **p<.005; ***p<.0005. Df=1.
We have in the same vein run an Ordinal Regression using the also well known Probit Link
function, with very similar results for the model and for the individual coefficients. It shows
a pseudo-R2 of 0.0624 and confidence intervals a little narrower than those of Table 4.3.
342
This gives way for our analytical strategy. The results show that our model is
statistically significant (p < .0005), though with quite modest pseudo-R2s ranching from 0.08 (McFadden) to 0.237 (Cox and Snell) and 0.246 (Nagelkerke). Table 4.3 shows the parameter estimates for the Ordinal Regression using the Logit
Link function.
The estimates for the general model and for the individual estimates confirm
our OLS results. Even controlling for values and other variables, one of our principal components of economic insecurity are significantly related to punitivity.
The relation runs in the direction predicted by the theoretical hypothesis: the
more economic feelings of insecurity, the more punitive. Coefficient intervals are
still broad, indicating the lack of precision of our estimates.
5. Discussion and conclusions

The study of punitivity is an area where methodological problems abound (Kury
1991; 1995a; 1995b; Kury & Ferdinand 1999). Many of these problems have
been encountered in the present study. These methodological concerns are aggravated by the use of secondary data, originally not collected to test this or other
theory or set of scientific hypothesis. On the other hand, most of the empirical
information in Criminology comes from a relatively limited set of countries,
while there is a pervasive lack of data for others, including Spain. Since punitivity might be a general development, it is necessary to check it in as many places
as possible. In fact, research in some European countries suggests that this trend
might not be as universal or as severe as some commentators have come to believe (Bondeson 2005; Meyer & OMalley 2005). This can be shown also on the
background of the results of the different ICVS waves although these surveys
also face a lot of methodological problems. Unfortunately, data are too scant to
evaluate the level punitivity of Spanish citizens, though the limited one, concentrated in attitudes toward death penalty, suggests that it might not be as high as in
other countries. In spite of it, the legislative and Criminal Justice system are more
and more punitive.
Comparative research is also useful to test theories in places different to those
where they were originally formulated. This has been the topic of our study. Kurys
insecurity feelings and punitivity hypothesis offers a sound explanation for punitivity,
is grounded in empirical evidence and has the potential to make sense of changes in
peoples punitivity, something that seems lacking in approaches that emphasize culture or values, for example (see other papers included in the volume; Indermaur &
Hough 2002). Using a national sample of adolescents and young adults from Spain
we find empirical support for the partial and extended version of Kurys theoretical
343
concept, at the individual level. Different measures of values seem to be related to

punitivity, each with an independent contribution. Even controlling for values, one of
our estimates of feelings of insecurity reaches the level of statistical significance (p <
.05), the one we have termed Economic situation. The two variables that had the
strongest impact in this principal component were The general economic situation of
Spain and The personal and familiar economic situation. The better one considers
the economic situation, then, the less punitive one tends to be (b = -0.16). This adds
credibility to our hypothesis and, then, to the theoretical concept.
Ours is a partial test of Kurys hypothesis for several reasons. The most important
is that insecurity feelings is a very complex construct, with probably many dimensions, as is the case of the concept of punitivity. Our measurement of insecurity feelings has only been able to take into account, and in an incomplete way,
those regarding the economic situation. A second reason is that our survey data
did not measure some variables that Kurys original concept predicts to be relevant. Another critical point is the sample. We had only relative young people in
it. Nevertheless, the study was also important because it is a further step in international cooperation and in empirical research in a country like Spain where empirical Criminology is more or less in the beginning.
The Model is quite modest, and it explains little more that a 15% of the variance.
Values seem to play an important role. Especially political ideology shows one of
the highest standardized coefficients and its coefficient intervals are considerably
narrower than its companion variables, meaning a certain degree of precision that is
helpful for inference. Being Spain a relatively divided country from a political and
value point of view, it might be that there were two subpopulations regarding punitivity. This possibility is suggested by some aspects of the data and of the residuals,
but inspection requires detailed study. Only educations impact is bigger in our sample, but its interpretation is not straightforward. Most of the variables from Models 1
and 2 lose their significance once values are introduced. A word of caution regards
living in a region with a relatively high delinquency rate. It shows such broad confident intervals that its effects can not excluded without further investigation. To be
sure, Kurys original position is consistent with an effect of objective crime and victimization rates; though in our present proposal its effect should be absorbed by the
feelings of insecurity. Due to our methodological problems and especially to our
poor estimates of feelings of insecurity, the impact (or lack of it) of this variable can
not be ascertained with our data.
Our study suffers from important limitations it is important to acknowledge.
We do not mind to insist that the survey data we had at our disposal was not designed to test any theoretical concept. As we pointed, few other opportunities to
344
test complicated theories as those regarding punitivity are available in countries

as Spain. Though it included some acceptable measures, others were quite crude.
This is specially the case of our estimates for feelings of insecurity. New studies
should find better ways to operationalize and measure it and explore the different
dimensions of the construct.
Our analytical strategy faces serious missing data problems (Kalton 1981;
Rubin 1987). The regression analysis uses less than half of the initial observations, and correctional action might be considered in future investigations (Allison 2002) since it is not clear that data can be assumed to be lost at random.
As was mentioned, there are indications of serious error sources in the data
(Biemer & Fecso 1995; Biemer & Lyberg 2003; Groves et al. 2004; Kalton et al.
1989; Schober & Conrad 1997). Unfortunately, lack of information makes it very
difficult to evaluate its impact in the statistics employed in this study. Example of
it is found in the wide coefficient intervals, which evidence the imprecision of
our estimations and render any inference problematic. A partial solution is offered by statistical methods that allow to control for error, as is the case of Structural Equation Models (Bollen 1989).
Ours is an effort to expand Kurys insecurity feelings and punitivity theoretical
concept. It is hardly necessary to mention that any explanation will benefit from
continuing empirical and theoretical research and thought. For example, the theory might specify the theoretical link that unites the feelings and the degree of
punitivity (Serrano-Mallo 2006b). Studies as ours suggest an empirical link, a
(positive) correlation. But why is it that people with higher insecurity feelings are
more punitive, or the same people get more punitive as their insecurity feelings
grow? One hypothesis that, though speculative, might be worth examining is the
following. People who feel insecure might have a preference for a strong State. A
strong State might protect them better from crime and external threat, offer them
a job if they lack one or lose the one they have, assure payment when retired and
so on. A strong state can give security not only in crime affairs but also in other
problems as was shown in the former Soviet States. The rate of fear of crime was
in many of the former Soviet States, like the German Democratic Republic, in
comparison to western states very low. Kossowska et al. (in this volume) show
the example of Poland: in 1987, 74 % of the Polish interviewed answered in a
survey that they thought Poland was a safe country. In 1993, after the dramatic
political changes which brought more freedom for the people but problems for
them and the State as well, only 26 % expressed the same opinion. The power of
the State was dramatically reduced giving people the feeling that it can not any
more press them, that they have more rights but that at the same time it cannot
345
help them in the old way. People, by the same token, might identify a strong
State with a punitive one. In this case, the argument goes, people with insecurity
feelings want the State to be punitive because in the popular imagination this
means as well a strong and protective State.
References
Allison, P.D. (2002). Missing data. Thousand Oaks, Ca. [etc.]: Sage.
Alexander, J.C. & B. Giesen (1987). From reduction to linkage: the long view of the
micro-macro link. In The micro-macro link (J.C. Alexander et al. eds.). Berkeley
[etc.]: University of California Press.
Beck, U. (1986). Risikogesellschaft. Auf dem Weg in eine andere Moderne. Frankfurt/M: Suhrkamp (Risk Society: Towards a New Modernity. 1992. London: Sage).
Beckett, K. & T. Sasson (2004). The Politics of Injustice. Crime and Punishment in
America. 2nd Edit. Thosand Oaks et al.: Sage.
Biemer, P.P. & R.S. Fecso (1995). Evaluation and control of measurement error in
business surveys. In Business Survey Methods (B.G. Cox et al. eds). New York:
John Wiley and sons.
Biemer, P.P. & L.E. Lyberg (2003). Introduction to survey quality. Hoboken, N.J.:
John Wiley and sons.
Bollen, K.A. (1989). Structural equations with latent variables. New York [etc.]: John
Wiley and sons.
Bondeson, U.V. (2005). Levels of punitiveness in Scandinavia: description and explanations. In: J. Pratt et al. (eds.), The new punitiveness. Trends, theories, perspectives. Cullompton, Devon: Willan Publishing.
Brown, D. (2005). Continuity, rupture, or just more of the volatile and contradictory?
Glimpses of New South Wales penal practice behind and through the discursive. In
The new punitiveness. Trends, theories, perspectives (J. Pratt et al. eds.). Cullompton, Devon: Willan Publishing.
Brown, M. (2005). Liberal exclusions and the new punitiveness. In: J. Pratt et al.
(eds.), The new punitiveness. Trends, theories, perspectives. Cullompton, Devon:
Willan Publishing.
Bundesministerium des Innern Bundesministerium der Justiz (Eds.) (2006). Zweiter
Periodischer Sicherheitsbericht. Berlin.
Cid Molin, J. (2005). La suspensin de la pena en Espaa: descarcelacin y reincidencia. Revista de Derecho penal y Criminologa, 16.
Dijk, J.van, R. Manchin, J.van Kesteren, S. Nevala & G. Hideg (2007). The burden of
crime in the EU. Research Report: A comparative analysis of the European Crime
and Safety Survey (EU ICS) 2005. Brussels: Gallup.
346
Garca Espaa, E. & F. Prez Jimnez (2004). Evolucin de la delincuencia en Espaa

y Andaluca: anlisis e interpretacin de las estadsticas oficiales. Mlaga: Fundacin El Monte and IAIC.
Garland, D. (1997). Governmentability and the problem of crime: Foucault, criminology, sociology. Theoretical Criminology 1, 173-214.
(2001). The culture of control. Crime & social order in contemporary society. Chicago y Oxford: The University of Chicago Press.
Groves, R.M., F.J. Fowler, M.P. Couper, J.M. Lepkowski, E. Singer & R. Tourangeau
(2004). Survey methodology. Hoboken, N.J.: John Wiley and sons.
Hannan, M.T. (1991). Aggregation and disaggregation in the social sciences, revised
edition. Lexington, Mass. and Toronto: Lexington Books.
Hassemer, W. (2001). Grnde und Grenzen des Strafens. In: N. Courakis (ed.), Die
Strafrechtswissenschaften im 21. Jahrhundert. Festschrift fr Professor Dr. Dionysios
Spinellis, 399-424. Athen.
Hinds, L. (2005). Crime control in Western countries 1970 to 2000. In: J. Pratt et al.
Willan Publishing.
Indermaur, D. & M. Hough (2002). Strategies for changing public attitudes to punishment. In: J. V. Roberts and M. Hough (eds.), Changing attitudes to punishment.
Public opinion, crime and justice. Collumpton: Willan, 198-214.
Kalton, G. (1981). Compensating for missing survey data. Ann Arbor, MI: Institute of
Social Research.
Kalton, G.,D. Kasprzyk & D.B. McMillen (1989). Nonsampling errors in panel surveys. In: D. Kasprzyk et al. (eds.), Panel surveys, New York: John Wiley and sons.
Kury, H. (1991). Victims of crime Results of a representative telephone survey of
5.000 citizens of the former Federal Republic of Germany. In G. Kaiser et al. (eds.),
Victims and Criminal Justice. Victimological research: stocktaking and prospects
Freiburg i.Br.: Max-Planck-Institut fr auslndisches und internationales Strafrecht.
(1995a). Zur Bedeutung von Kriminalittsentwicklung und Viktimisierung fr die
Verbrechensfurcht. In: G. Kaiser and J.M. Jehle (eds.), Kriminologische Opferforschung. Neue Perspektiven und Erkenntnisse, II Verbrechensfurcht und Opferwerdung Individualopfer und Verarbeitung von Opfererfahrungen. Heidelberg: Kriminalistik Verlag.
(1995b). Wie restitutive eingestellt ist die Bevlkerung? Zum Einflu der Frageformulierung auf die Ergebnisse von Opferstudien. MschrKrim, 78.
(1996). Desarrollo de la delincuencia en Europa oriental y occidental. Una comparacin entre diferentes pases (trans. A. Serrano Mallo). Revista de Derecho penal y
Criminologa, 6.
(2004). Kriminalittsentwicklung, Sanktionen und die Einstellung der Bevlkerung.
Ein Vergleich zwischen Ost und West. In H. Kury (ed.), Strafrecht und Kriminalitt.
Entwicklungen in Mittel- und Osteuropa. Bochum: Universittsverlag Dr. N. Brockmeyer.
347
(Ed.) (2006). Hrtere Strafen weniger Kriminalitt? Zur Verschrfung der Sanktionseinstellungen. Soziale Probleme, Heft 2.
Kury, H. & T. Ferdinand (1999). Public opinion and punitivity. International Journal
of Law and Psychiatry, 22.
Kury, H. & M. Brandenstein (2006). Sobre la cuestin de una Nueva Punitividad
Actitudes sancionadoras y poltica sancionadora. In: Guzmn Dlbora, J.L. and Serrano Mallo, A. (Eds.), Dereche penal y criminologa como fundamento de la poltica
criminal. Estudos en homenaje al Profesor Alfonso Serrano Gmez. Madrid: Dykinson, 369-402.
Kury, H., J. Obergfell-Fuchs & U. Smartt (2002). The evolution of public attitudes to
punishment in Western and Eastern Europe. In: J.V. Roberts & M. Hough (eds.),
Changing attitudes to punishment. Public opinion, crime and justice. Cullompton:
Willan Publishing.
Kury, H., H. Kania & J. Obergfell-Fuchs (2004). Worber sprechen wir, wenn wir ber
Punitivitt sprechen? Versuch einer konzeptionellen und empirischen Begriffsbestimmung. Kriminologisches Journal, 36.
Kury, H., M. Keller, U. Mitter, N. Rahmati (2006). Verbrechensfurcht, Punitivitt und
Einstellung zur Polizei Ergebnisse einer Umfrage in Aserbaidschan. In: Kury, H.,
E. Karimov (Eds.), Kriminalitt und Kriminalprvention in Lndern des Umbruchs.
Beitrge einer Internationalen Konferenz in Baku/Aserbaidschan. Bochum: Universittsverlag Dr. N. Brockmeyer, 415-486.
Kury, H., J. Bergmann, J. Obergfell-Fuchs & E. Schill (2007). Zur Abhngigkeit der
Messung von Sanktionseinstellungen vom methodischen Vorgehen. Freiburg, unpublished paper.
De Leeuw, E. & M. Collins (1997). Data collection methods and survey quality: an
overview. In: L. Lyberg et al. (eds.), Survey measurement and process quality. (Hoboken, N.J.: John Wiley and sons.
Leishman, F. & P. Mason (2003). Policing and the media. Facts, fictions, and factions.
Cullompton, Devon and Portland, Oregon: Willan Publishing.
Mayhew, P. & J. van Kesteren (2002). Cross-national attitudes to punishment. In: J.V.
Roberts & M. Hough (eds.), Changing attitudes to punishment. Public opinion, crime
and justice Cullompton: Willan Publishing.
Meyer, J. & P. OMalley (2005). Missing the punitive turn? Canadian criminal justice,
balance, and penal modernism. In: J. Pratt et al. (eds.), The new punitiveness.
Trends, theories, perspectives. Cullompton, Devon: Willan Publishing.
Nelken, D. (2005). When is a society non-punitive? The Italian case. In: J. Pratt et al.
Willan Publishing.
Roberts, J.V. (2002). Public opinion and the nature of community penalties: international findings. In: J.V. Roberts & M. Hough (eds.), Changing attitudes to punishment. Public opinion, crime and justice. Cullompton: Willan Publishing.
348
Roberts, J.V. & L.J. Stalans (2000). Public opinion, crime and Criminal Justice. Boulder, Co.: Westview Press.
Roberts, J.V. & M. Hough (Eds.) (2002). Changing attitudes to Punishment. Public
opinion, crime and justice. Cullompton: Willan.
Roberts, J.V., L.J. Stalans, D. Indermaur & M. Hough (2003). Penal populism and
public opinion. Lessons from five countries. Oxford, New York: Oxford University
Press.
Rubin, D. (1987). Multiple imputation for nonresponse in surveys. New York [etc.]:
Wiley and sons.
Rckert, S. (2006). Ab in den Knast. DIE ZEIT, 24. 6. 2006.
Diskussion der kriminalpolitischen Wende. In: J. Obergfell-Fuchs, M. Brandenstein
(Eds.), Nationale und internationale Entwicklungen in der Kriminologie. FS fr Helmut Kury zum 65. Geburtstag. Verlag fr Polizeiwissenschaft: Frankfurt/M., 31-71.
Schober, M. & F. Conrad (1997). Does conversational interviewing reduce survey measurement error. Public Opinion Quarterly, 61.
Serrano-Mallo, A. (2006a). Introduccin a la Criminologa, 4th. ed. Madrid: Dykinson.
(2006b). Punitivitt und Gesetzgebung: Die Situation in Spanien. In Kriminalitt
und Kriminalprvention in Lndern des Umbruchs (H. Kury and E. Karimov eds.).
Bochum: Universittsverlag Dr. N. Brockmeyer.
Tittle, C.R. (2000). Theoretical developments in Criminology. In Criminal Justice
2000, 1 The nature of crime: Continuity and change. Washington, D.C.: U.S. Department of Justice.
Tonry, M. (2004). Punishment and politics. Evidence and emulation in the making of
English crime control policy. Cullompton, Devon: Willan Publishing.
Tonry, M. & D.P. Farrington (2005). Punishment and crime across space and time.
Crime and Justice, 33 Crime and punishment in Western countries, 1980-1999.
Tyler, T.R. & R.J. Boeckmann (1997). Three strikes and you are out, but why? The
Psychology of public support for punishing rule breakers. Law and Society Review,
31.
Wacquant, L. (2005). The great penal leap backward: incarceration in America from
Nixon to Clinton. In The new punitiveness. Trends, theories, perspectives (J. Pratt et
al. eds.). Cullompton, Devon: Willan Publishing.
Warr, M. (1995). Public opinion on crime and punishment. Public Opinion Quarterly
59, 296-310.
Zugalda Espinar, J.M. (2006). Contrarreforma penal (el annus horribilis de 2003) y el
anteproyecto de reforma del cdigo penal de 2006. In Derecho penal y Criminologa
como fundamento de la Poltica criminal. Estudios en homenaje al Profesor Alfonso
Serrano Gmez (F. Bueno et al. eds.). Madrid: Dykinson.
349
Perceptions, Emotions and Experiences of Crime: Effects

on Attitudes Toward Punishment in a Canadian Sample
TIMOTHY F. HARTNAGEL AND LAURA TEMPLETON 1
Introduction
The Law Commission of Canada (2003) recently pointed out that the latter part
of the 20th century has seen increased demands for harsher punishment for offenders and a greater emphasis on governments adopting a law and order
agenda. The resulting increased reliance upon the criminal law and punishment
saw Canadian rates of imprisonment reach high levels in the 1990s (Hendrick &
Farmer 2002); the incarceration rate reached 134 per 100,000 adults in 2000 and
2001, declining to 129 in 2003 and 2004 (Statistics Canada 2006).
In line with evidence from victim surveys of relatively stable rates of violent
victimization and of most household property crimes between 1993 and 2004
(Gannon 2005), and a generally declining official rate of crimes known to the
police in recent years (with the exception of a slight increase in 2003) (Gannon
2006), a majority (58%) of Canadians think that the level of crime in their
neighbourhood has stayed the same during the last five years, although 30% still
believe it has increased (Gannon & Mihorean 2005). While for many years most
Canadians held the view that crime rates were increasing regardless of the trends
in official crime statistics or victimization surveys, members of the public now
appear to have begun to absorb the reality that crime rates are declining (Roberts
2001). In 2004, 76% of Canadians said they walked alone in their neighbourhood
at night and many (44%) reported feeling very safe, continuing this positive trend
from 1999 (Gannon 2005).
Paper presented to the European Society of Criminology, August, 2006. We acknowledge

the assistance of the staff of the Population Research Laboratory, University of Alberta for
their data collection and preparation. The financial support of the Social Sciences and Humanities Research Council of Canada through a research grant to T.F. Hartnagel is gratefully acknowledged.
350
Timothy F. Hartnagel and Laura Templeton
In spite of these declining crime trends, more accurate public perceptions and
lowered levels of fear, the Canadian public continues to be quite critical of the
criminal courts and prison and parole systems. In 2004, only 27% of the population felt the criminal courts were doing a good job of determining whether or not
the accused is guilty, of helping the victim (20%), and of providing justice
quickly (15%), although public views of the courts have improved somewhat
over time. There were similarly low levels of satisfaction with the prison and parole system, although the police were rated much more positively (Gannon
2005). Parole received the lowest rating among criminal justice sectors with
fewer than 1 in 5 giving it a good rating (Gannon 2005).
Furthermore, it appears that the severity of punishments actually imposed is
quite consistently exceeded by the punishment levels the public desires since
most people respond to polls about sentencing with the same answer: its too lenient (Blumstein & Cohen 1980; Roberts et al. 2003; Roberts & Stalans 1997;
Bureau of Justice Statistics 1977; Zimmerman et al. 1988). For example, in response to survey questions concerning judicial sentencing, 77% of Canadians
appear to believe that sentences are not harsh enough (Doob & Sprott 1997). Almost three-quarters of Albertans responding to a 1997 survey (Alberta Justice
1998) were dissatisfied that sentences did not match the crime, with a similar
proportion thinking that convicted persons spend too short a time in jail. Tougher
sentencing received the greatest endorsement (by 37% of respondents in the same
poll) as the most effective action for reducing the level of crime. From their recent literature review, Cullen et al. (2000) concluded that the public is punitive
toward crime and get tough attitudes are real and not simply a methodological
artifact. But this punitiveness is mushy (Cullen et al. 2000) and is moderated
under certain conditions (Doob 2000).
A number of studies have documented public preferences for criminal justice
penalties and policies. But in comparison to the number of times it has been described, the number of attempts to empirically explain variation in the publics
punitive response to crime is small (Freiberg 2001) and according to Useem et al.
(2003) there are few rigorous studies. One approach has emphasized the role of
fear and concern about crime in explaining punitive attitudes (Langworthy and
Whitehead 1986; Sheley 1985). Support for more severe punishment to lower the
crime rate or chances of victimization has been seen as an instrumental response
to fear and/or concern about crime (Tyler & Weber 1982). But fear of crime can
also be regarded as an emotional reaction attributable to such influences as media
exaggeration of crime news, political exploitation of criminal justice issues,
and/or victimization experiences (Sheley 1985; Ouimet & Coyle 1991; Sprott &
Doob 1997). There is some research demonstrating the role of emotion in policy
Effects on Attitudes Toward Punishment in a Canadian Sample
351
attitudes (Gault & Sabini 2000), although Indermaur & Hough (2002) have argued for more attention to the emotional dimension of attitude formation. However, the possible impact on criminal justice attitudes of emotions other than fear
has rarely been investigated, although anger directed toward criminals may be a
motivating force leading to greater punitiveness or a desire for more severe penalties. The measurement of fear of crime has also been an issue in the punitiveness literature. The items purporting to measure fear have often referred to hypothetical situations that fail to more directly assess the respondents fear.
Concern about crime or the salience of crime and criminal justice issues to the
public should be distinguished from the emotional or affective element of fear. In
addition, we should not assume that these crime and justice topics are of high or
equal importance to all respondents. Furthermore, many members of the public
are not very well informed or knowledgeable concerning crime and justice.
Those who are more knowledgeable about the amount of violent crime or about
sentencing, for example, may be less likely to adopt more punitive attitudes toward offenders. This knowledge may be based in part on some experience with
the criminal justice system.
The present research utilizes data from a national Canadian sample to test an
explanatory model of attitudes toward punishment for crime. The model includes
the predictor variables of: the emotions of fear, worry and anger; knowledge of
crime; the salience of crime issues; prior experience with crime and the criminal
justice system; as well as several social background variables.
Theoretical framework
Emotions
Previous research attempting to explain punitive attitudes has focused on the role of
fear and concern about crime (Langworthy & Whitehead 1986). This view traces
these punitive attitudes to the increased levels of fear of crime and/or concern about
the problem of crime attributable to such influences as media exaggeration of crime
news, political exploitation of criminal justice issues, and/or victimization experiences (Sheley 1985; Ouimet & Coyle 1991; Sprott & Doob 1997). This hypothesis
suggests that support for more severe punishment results from a desire to lower the
crime rate (Tyler & Weber 1982). Thus increased punitiveness may be due to the
utilitarian desire to reduce, control or prevent crime. This perspective suggests that
in their desire to protect themselves and their communities from the threat of crime,
people respond to personal fears when responding to rule breakers. People support
punishing rule breakers because they are afraid that they, their families, or others in
the community will become victims of crime (Tyler & Boeckmann 1997). Fear of
352
victimization may motivate an individual to look to the courts for safety through
increased penalties for offenders (Sprott & Doob 1997).
Garland (2000, 2001), as part of his work on the culture of control within high
crime societies, seems to adopt a version of this perspective. He has recently argued that a change in the experience of crime by the professional middle classes
and the so-called liberal elites resulted in a transformation in their attitudes and
dispositions toward crime control. This group has historically favored a civilized attitude toward crime stressing social circumstances rather than individual responsibility, remedial treatment rather than punishment to signify their
cultural distinction from the more reactionary views of those groups immediately
above and below them in the stratification hierarchy. In the 1950s and 60s the
professional middle classes were generally separated from crime and insecurity
since they typically lived at a distance from criminal events and their daily routines did not often expose them to the threat of crime. More recently, however,
this class has become less supportive of penal-welfarism and more supportive of
punitive responses to crime due, according to Garland, to their new experience of
crime. Crime became a prominent, much more routine fact of life for both the
urban middle classes and suburbanites. The professional middle classes increasing fear of crime was fueled, then, by changes (increases) in the frequency and
distribution of crime, along with other social changes occurring over the same
time period. As a result, substantial sections of the public became more punitive
in their attitudes toward crime, criminals, and criminal justice policies. Members of the public increasingly express their fear, their aggravation at having to
alter their lifestyles and incur expenses, their dissatisfaction with the system that
failed them. The daily tribulations of minor crime and disorder easily slides into
a concern with crime as such, which in turn connotes violent predatory crime.
The trauma of powerlessness in the face of fear prompts the demand for action
(Garland 2000: 368).
The research evidence on the relationship between fear and punitiveness is
mixed. Some research has reported that fear is an important determinant of punitive attitudes (Langworthy & Whitehead 1986; Thomas & Foster 1975; Thomas
& Cage 1976; Rossi et al. 1985; Hough et al. 1988; Conklin 1971). Sprott &
Doob (1997) found that, independent of victimization, gender and age, those Canadians who fear crime are more likely to view adult sentences as too lenient;
while Canadians who were dissatisfied with their overall personal safety from
crime more often supported the use of prison sentences (Tufts 2000). Tufts &
Roberts (2001) conducted multivariate analyses of the 1999 Canadian General
Social Survey data and found that fear was a highly significant predictor of preference for prison.
353
But Taylor et al. (1979) and Stinchcombe et al. (1980) found only a slight effect of fear on support for the death penalty; and other research has concluded
that fear does not lead to greater punitiveness (Baron & Hartnagel 1996; Brillon
1988; Ouimet & Coyle 1991). According to Tyler & Boeckmann (1997), public
punitiveness was not a function of fear of crime or grievances against the courts
but rather reflected symbolic concerns regarding declining moral and social consensus, as well as underlying social values which have a direct effect on the publics punitiveness. They claimed that crime-related concerns are the least important factor in predicting punitive attitudes. Cullen et al.s (2000) review concluded that fear of crime does not appear to drive public calls for harsher punishment while Beckett (1997) has argued that the assumption that anxiety about
crime drives support for punitive policies is problematic since those who are less
afraid typically express the most support for get tough policies; that is, rural
white men as opposed to women and blacks.
Emotions other than fear such as anger and resentment may also affect attitudes toward punishment (Lerner et al. 1998) but have only rarely been investigated. Durkheim long ago argued that the shock of norm violation produces emotions of anger and indignation along with the desire for revenge (Garland 1990).
Some authors have claimed that anger is the emotion most clearly linked to concerns and values about justice and fair treatment (Karstedt 2002), although it is
far from clear how anger about crime is aroused in those who have not been a
victim and who have no personal experience of crime (Farrall 2001). Lerner et al.
(1998) state that anger is the principal emotion associated with justice judgments;
and that it leads people to rely on stereotypes and easily processed rather than
effort-demanding cues. Alternatively, the endemic insecurity inherent in a competitive and mobile society the fear of falling (Ehrenreich 1989) may create
latent anxieties which, according to Scheingold (1991) eventually produce a good
deal of anger and result in a search for scapegoats against whom to discharge this
anger and through whom to maintain the illusion of control. Crime can be a symbolic substitute for feelings about other forms of incivility or unusual behaviour
(Gaubatz 1994). Gaubatzs (1994) analysis of the views of a small sample of
Americans who displayed a consensus on tough sentences found that when they
felt overwhelmed by social problems, their anger would spill over on to criminals, in which case they would fall back to simple solutions such as the effectiveness of severe punishment.
There is some psychological research suggesting that anger-primed participants subsequently make more punitive attributions than neutral-emotion-primed
participants (Averill 1983; Keltner et al. 1993; Quigley & Tedeschi 1996). Lerner
354
et al. (1998) used participants responses to vignettes describing a harm resulting

from negligence by a worker to measure punitiveness and reported than those
exposed to the anger prime made more punitive attributions than did participants
exposed to the neutral-emotion prime. In a series of studies Gault & Sabini
(2000) provided evidence that emotional dispositions and reactions play an important role in shaping certain political attitudes. More specifically, state anger
(i.e., current emotional mood or contextual reaction) predicted support for punitive actions while trait anger (a relatively stable individual characteristic) was not
a significant predictor of policy preferences.
Ditton et al. (1999a) reported that Scotish survey respondents seemed to be
more angry about the threat of criminal victimization for four offence types than
afraid of it. Victims were more likely than non-victims to be angry and to be
afraid, and this held for both genders. There is some additional evidence of anger
as a response to actual victimization (Maguire 1980; Mawby & Walklate 1997;
Ditton et al. 1999b).
Crime Salience
Concern about crime or the perception or belief that crime is a problem should be
distinguished from the emotional reaction of fear (Furstenberg 1974; Thomas &
Foster 1975). People can perceive or believe that crime is a problem or is increasing without necessarily reacting to that perception or belief with a fear response
(Chiricos et al. 2004). Concern about crime as a problem is a cognitive phenomenon how people regard crime while fear is more of an affective response. The cognitive response involves an assessment of risk or expression of
concern (Chiricos et al. 2004). Crime and justice issues may also be more important or salient for some individuals. They may pay more attention to stories or
news of crime and the criminal justice system, think about issues of crime and
justice more often and/or discuss these topics with others to a greater extent.
Therefore, the potential independent effects of the salience of crime as an issue
or problem should be considered.
Several early studies supported the idea that a greater concern about crime resulted in increased punitiveness (Thomas & Foster 1975; Rankin 1979; Stinchcombe et al. 1980). For example, Rankin (1979) concluded that concern about crime, but
not victimization or fear of crime, contributed to support for capital punishment;
while Thomas & Foster (1975) reported that belief that the crime rate is increasing
appeared to be an important determinant of the willingness to employ punishment.
Thomas & Cage (1976) reported that survey respondents who perceived crime to be
an increasing problem, perceived criminal victimization as probable, and thought of
355
punishment as an effective response to crime recommended more severe penalties

for a range of crimes. However, Cullen et al. (1985) using a scale that included
items on the belief crime was increasing in the neighborhood and in the country as a
whole, perceptions of safety alone in the neighborhood during the day and at night,
and the belief that crime was a serious problem in the neighborhood did not find
any significant relationship between crime salience and attitudes toward criminal
sanctioning. However, Chiricos et al. (2004) reported that concern about crime predicted support for punitiveness, as did the perception that crime is violent and the
fear of crime. Other research has suggested that perceptions of crime salience in
ones residential area do not consistently relate to punitive attitudes (Langworthy &
Whitehead 1986; Ouimet & Coyle 1991.)
Knowledge
The public misunderstands many issues related to crime and justice, including
knowledge of trends in crime rates, the prevalence of violent crimes and recidivism rates (Doob & Roberts 1983; Roberts 1992; Roberts & Stalans 1997), as
well as punishment policy (Zimmerman et al. 1988). Most members of the public
have little accurate idea of the actual severity of the courts; perceptions of sentencing leniency seem to be unaffected by the increased severity of the system;
and people tend to think of violent recidivists when responding to polls on the
harshness of sentencing (Roberts and Stalans, 1997). These inaccurate perceptions of what happens to offenders are biased in the direction that leads the public
to see the system as lenient (Roberts & Stalans 2000). The 1996 British Crime
Survey revealed that large majorities of respondents provided estimates of imprisonment rates which were too low for the crimes of rape, mugging and burglary and this was correlated with the belief that sentences were too lenient. As
Garland (2001) has remarked, Public knowledge and opinion about criminal justice are based upon collective representations rather than accurate information ....
Few polls have investigated public opinion of sentencing as a function of public
knowledge (Hough & Roberts 2002), although giving survey respondents more detailed information about the nature of the offender and the offence tends to result in
less punitive sentencing preferences and in public sentencing preferences very close to
what judges actually do under these circumstances (Doob 2000; Roberts et al. 2003;
Roberts & Stalans 1997; Tufts 2000; Doob & Roberts 1983; 1988; Zamble & Kalm
1990). Hough & Park (2002) used the deliberative poll method in an attempt to
change attitudes to punishment, but there was no evidence that changes in public opinion could be attributed to improvements in knowledge. However, an experimental
study testing the impact of different approaches to improving public knowledge of
356
crime found that knowledge gain was a significant factor in predicting perceptions of
the appropriateness of sentencing when other characteristics such as demographic differences, educational qualifications and the influence of initial belief were controlled
in a multivariate analysis (Mirrlees-Black 2002). Hough & Roberts (2002) carried out
a logistic regression of British Crime Survey data to determine which aspects of public misperception were most closely associated with the belief that sentences were too
lenient. The strongest predictor was the perception that national crime rates had increased a lot, followed by the perception that the use of prison was the same or down.
Over-estimating the proportion of violent crime was also a significant predictor.
Hough and Roberts suggest that correcting public misperceptions about the rising
crime rate will probably dampen enthusiasm for harsher sentencing. However, this
assumes that public attitudes toward punishment are primarily cognitive rather than
affective. Hutton (2005), for one, has argued that attitudes to crime and punishment
probably reflect, at least in part, broader anxieties and insecurities and are thus not
directly connected to accurate information about crime and punishment. In any event,
people often have strong views regarding sentencing that are founded upon a very
poor knowledge base (Roberts & Stalans 2000).
Experience with Crime and the Criminal Justice System

Since most people have little contact with the criminal justice system, often limited to traffic enforcement or the reporting of minor victimizations to the police
(Maruna & King 2004), perhaps we should not be surprised at the publics lack
of knowledge about crime and justice. Criminal victimization is certainly one
type of experience with crime and the criminal justice system. But unless the offence is cleared up and the offender charged, many victims experience with the
system will be quite limited. Victimization, even for a violent crime, was not a
significant predictor of attitudes toward the use of imprisonment in the 1999 Canadian General Social Survey (Tufts & Roberts 2001). Personal experience with
victimization does not consistently relate to punitive attitudes (Langworthy &
Whitehead 1986; Ouimet & Coyle 1991). Only 9% of the self-reported victims of
violent crimes in the 1998 British Crime Survey favoured the incarceration of
their offender (Roberts 2002). Van Dijk & Steinmetz (1988) reported that personal experience of crime and perceived risk of crime were not related to support
for more repressive policies, although those who felt more threatened or were
victims did favor extra government spending on crime control. Rich & Sampson
(1990) found that the independent effect of victimization status on punitiveness
in a multivariate analysis was weak and non-significant.
Of course, there are other forms of experience with the criminal justice system
357
than victimization. People may have appeared in court as a witness or defendant

or have family members or friends who have been in this situation. Some may
have volunteered in community police stations, neighborhood watch, or victim
assistance programs. Others may have served on juries while a few may have
participated in prison visiting programs. There is some evidence that when citizens are actively engaged in the criminal justice system through serving on a jury
(Matthews et al. 2003) or participating in restorative justice activities (Greene &
Doble 2000), they are less punitive. However, there is only limited research on
the possible effects of such experiences on attitudes to punishment.
Hypotheses
The present research tests a model predicting punitiveness toward offenders that
is based upon these theoretical perspectives and prior research. Since we do not
regard these perspectives as mutually exclusive, we will examine their relative,
net contribution to the explanation of punitive attitudes toward crime/criminals.
More specifically, the emotions perspective predicts a positive effect of fear of
and anger about crime on punitive attitudes, net of other predictors. However,
given the weak evidence for an effect of fear, we expect anger to be a stronger
predictor of punitive attitudes. The more salient crime is to individuals the more
likely they are to respond punitively. Those who believe or perceived that crime
is a problem should express more punitive attitudes toward offenders. However,
this is more likely for an overall or general perception of crime as a problem than
for crime in ones own neighborhood since fewer people typically perceive the
latter as a problem. In addition, we expect those who think more about crime to
express more punitive attitudes. More accurate knowledge of crime should be
negatively related to punitiveness. Those who have more accurate and therefore
more realistic ideas about crime and criminal justice should express less punitive
attitudes toward criminals. Finally, with respect to experience with crime and the
criminal justice system, the nature or type of such experience needs to be distinguished. The research literature suggests that victimization will have little or no
effect on attitudes toward punishment. However, other types of experience may
have a negative or a positive effect depending upon the type of experience.
In addition to testing these predictions drawn from the four perspectives, we
also include three demographic background variables gender, age, and education often found to be related to punitiveness. Thus males are expected to be
more punitive than females (Hough et al. 1988; Langworthy & Whitehead 1986;
Stinchcombe et al. 1980), although there is some evidence that this may vary
with specific offences and measures of punitiveness (Tufts 2000; Useem et al.
358
2003). Age is predicted to be positively related to punitiveness (Brillon 1988;

Cullen et al. 1985; Hough et al. 1988; Langworthy & Whitehead 1986; Rich &
Sampson 1990), although again there is some evidence that when presented with
hypothetical crime scenarios younger Canadians are more supportive of prison
sentences when compared to older Canadians (Tufts 2000). Education should be
negatively related to punitive attitudes (Brillon 1988; Hough et al. 1988; Walker
et al. 1988b; Rich & Sampson 1990; Hogan et al. 1997; Tufts 2000). We also
control for the population size of the respondents area of residence and their
self-identification as a visible minority.
Data and method

Data for this study were obtained from a survey conducted by the Population Research Laboratory (PRL) at the University of Alberta between March 22nd and June
6th, 2005 using computer-assisted telephone interviewing facilities. The study proposal and survey instrument were approved by the Research Ethics Board of the
Faculties of Arts, Science and Law. A telephone sample was purchased from Survey
Sampling International and was screened for disconnects by the vendor. A random
digit dialing approach was used to ensure that respondents had an equal chance of
being contacted whether or not their household was listed in the telephone directory.
Interviews were conducted between the hours (local time) of 9: 30 am to 2: 30 pm
Monday through Friday, 4: 00 pm to 9: 00 pm Monday through Thursday, Saturday
10: 00 am to 4: 00pm, and Sunday 2: 00 to 8: 00 pm.
Four regions of Canada were selected for sampling. Three of these regions (Pacific, Prairie and Central) were comprised of single provinces (British Columbia,
Alberta and Ontario, respectively). The Atlantic region included three provinces:
New Brunswick, Nova Scotia, and Prince Edward Island. Sample sizes proportional
to the population of each selected region were established; these regional quotas
were further stratified by gender. Respondents were eligible for participation if they
were at least 18 years of age and their quota group was not full. The final sample
contained 1006 completed interviews: 49.9% male (n = 502) and 50% female (n =
504). The average length of the interview was 27.88 minutes (SD=9.7), with an average of 3 attempts to obtain a completed interview. Approximately 24% of records
resulted in a final disposition of no answer, busy tone, answering machine, callback,
or will call lab. The overall response rate for the study was 38.4% calculated by dividing the number of completes (1006) by the number of completes plus refusals
(1410), incompletes (27), and language or communication barrier (174). The response rate was highest in Alberta (54.6%) and lowest in British Columbia (35.3%).
359
Measurement
There has been controversy regarding the measurement of public attitudes toward
punishment (Roberts & Stalans 1997; Roberts et al. 2003). Those arguing that the
public favors a punitive response to crime usually cite national telephone polls
that ask only one or two questions about crime policy; e.g., whether the courts
are harsh enough. If the interest is in a general or global opinion, then broadly
worded poll questions may be sufficient. But as Cullen et al. (2000) note, support
for a policy or punishment may vary under different conditions or when applied
to cases that have certain attributes; and measuring these more specific opinions
requires a series of questions. This seems particularly critical in studies of public
opinion concerning sentencing where various sentencing options or alternatives
seem to be required (Walker et al. 1988a). A general survey question concerning
harshness of sentencing is more likely to tap into stereotypical images of criminals and distort punishment preferences (Roberts & Stalans 1997; Stalans 2002).
We employed a scale of punitiveness developed by Chiricos et al. (2004) which
we modified slightly to fit the Canadian context. It was composed of seven items
referring to ways of dealing with crime in Canada (i.e., make sentences more severe for all crime; use the death penalty for some murderers; make prisoners
work on chain gangs; take away TV and recreation privileges from prisoners; use
more mandatory minimum sentencing laws; lock up more juvenile offenders;
send repeat juvenile offenders to adult courts) to which respondents were asked
to indicate on a ten-point scale their degree of support. Responses to the seven
items were added together and divided by seven. Our version of this scale had an
Alpha reliability of .844. Table 1 displays the means, standard deviations and
range for this and the set of predictor variables. The mean score on the dependent
variable of punitive attitudes was 6.74 (out of a maximum of 10) and a median of
7.07. Interestingly, the modal response was 10 (N = 42). So there is some skew
toward greater punitiveness, although approximately 20% of respondents scored
below 5.
The background variables of age, gender, and education (total years of schooling)
were measured in standard ways. There were approximately equal numbers of
females and males; only 17% of respondents were under 30 years of age and another 17% were 65 or older; and the average number of years of schooling was
slightly more than 15 (12 = high school graduation). Respondents were also
asked if they considered themselves a member of a visible minority (yes or no),
with approximately 19% so identifying themselves; and the approximate size (in
categories) of the population in the location where they presently lived (approximately a third lived in areas of less than 25,000; 69% in areas of less than
300,000; while 23% lived in areas of 700,000 or more).
360
Table 1:
Means, Standard Deviations and Range
Variable
Background
Gender
Age
Education
Visible Minority
Population Size
Knowledge
% Violent Crime
Crime Salience
Crime Problem
Crime Neighborhood
Think Crime
Experience
Victim
Personal Experience
Emotions
Fear
Worry
Anger
Punitive Attitudes
Mean
SD
Minimum
Maximum
1.49
46.89
15.34
1.83
3.52
.500
16.191
3.332
.373
1.745
1 (M)
18
4 years
1 (yes)
1 (<5,000)
2 (F)
87
25 years
2 (no)
6 (700,000>)
33.93
24.577
1%
100%
3.89
2.14
3.25
1.044
1.297
1.268
1 (SD)
1 (SD)
1 (SD)
5 (SA)
5 (SA)
5 (SA)
1.40
1.48
.491
.500
1 (yes)
1 (yes)
2 (no)
2 (no)
1 (not at all)
1 (SD)
1 (not at all)
1
(not at all
supportive)
10 (very much)
5 (SA)
10 (very much)
10
(supportive)
6.68
2.11
8.25
6.74
2.311
1.212
1.904
2.268
The typical measure of fear (how afraid are you or would you be of walking
alone at night in your neighborhood) has been criticized as too hypothetical to tap
the fear respondents experience in the course of their routine activities (Beckett
& Sasson 2000). Fear of crime for this analysis was measured by presenting four
crime scenarios (a house break in/theft and a simple assault, each with a first offender and then with a repeat offender) and asking respondents to rate the degree
(not at all to very much) to which each situation would make them fearful. Their
scores for each of the four scenarios were added together and divided by four for
their total score on fear (alpha = .902). The same four scenarios and procedure
were used for anger asking respondents to rate the degree to which each situation
would make them angry (alpha = .905). The mean scores indicate respondents
reacted with more anger than fear; and the distribution of anger scores is quite
skewed toward the higher values. However, both means are above the mid-point,
indicating a significant amount of both fear and anger expressed in response to
these crime scenarios. In fact, the modal response for both is 10, the maximum
value, with 30% of respondents with this score on anger (10% for fear). Given
the still somewhat hypothetical nature of this fear measure, we decided to include
a second indicator called worry. This was an index composed of two items: I
361
worry about being robbed or assaulted in my own neighborhood at night; and I

worry that a thief will break into my home while I am at home. They were
added together and divided by two to form an index with an alpha value of .836.
Its mean value of 2.11 suggests that the respondents did not worry to a large degree about being victimized by these offences (36% had the lowest score of 1 and
only 15% were at 4 or higher).
Since prior knowledge and experience may shape interpretations and responses to
crime and punishment, respondents were asked approximately what percent of all
crime in Canada do you think is violent crime? Approximately three-quarters overestimated the correct percent (about 10% to 12%), with the mean at almost 34%. So
there is a substantial amount of misinformation concerning the extent of violent
crime in Canada. We included two measures of experience with crime and the
criminal justice system. Respondents were asked if they had ever been a victim of a
crime. Fifty-nine percent indicated they had been a victim. The second measure
asked if they ever had any personal experience with the criminal justice system and
approximately 52% responded positively. The correlation between these two measures is .240 (p <.001) which indicates that they represent somewhat different aspects
of experience with crime and criminal justice.
Finally, we also included three indicators of the salience of crime for our respondents. We combined two items into an index (alpha = .69) of the overall perception
that crime is a problem: The crime problem is getting worse in Canada these days;
and The risk of being robbed or assaulted by teenage gangs has increased in recent
years. The mean for this index is a fairly high 3.89 (median = 4.0 on a 5 point
scale) indicating a substantial amount of concern about crime in this sample. Secondly, we asked respondents if Crime is a serious problem where I live to measure
their degree of concern about crime in their own neighborhood. Here the mean value
was quite a bit lower (2.14), suggesting that respondents generally were much less
concerned about crime in their own neighborhood areas; only 18% agreed or
strongly agreed with this statement The third measure of crime salience involved
responses to the question I think a great deal about crime and criminal justice issues. The mean score was 3.25; and 42% agreed or strongly agreed with this statement, indicative of a significant degree of salience of crime as an issue or problem.
Results
To test the set of hypotheses we regressed the punitiveness scale on the set of fourteen predictor variables. The correlation matrix appears in Table 2 and shows that
our measure of punitiveness is moderately and positively correlated (.539) with the
perception that crime is a problem: those who regard crime as a problem are more
362
punitive. There is also a moderately strong, positive correlation of punitiveness

with the emotion of anger (.393) and perception of the amount of violent crime
(.315). Those respondents who express more anger over the crime scenarios and
those who believe there is more violent crime are more punitive. Those who
think more about crime and justice issues (.297) or who worry about crime (.270)
are also more punitive, while the more educated are less punitive (-.267). Among
the predictor variables, fear and anger are positively related (.573), as are perception of the amount of violent crime and the view that crime is a problem (.349)
and gender and fear (.346), with women more fearful of crime. Those who worry
more about crime are also more fearful (.344), think more about crime and justice
(.357), and think crime is a serious problem in their neighborhood (.423).
The ordinary least squares regression of punitiveness on the set of predictors is presented in Table 3. Overall, approximately 42% of the variation in punitiveness is explained by these predictors. The predictor with the largest net effect on punitiveness is
the overall perception that crime is a problem (B=.415): those who believe crime is
generally more of a problem are also more punitive in their attitudes toward
crime/criminals net of other predictors. Anger about crime has the second largest net
effect (B=.260): angrier respondents are more punitive. Those who worry more about
crime are also more punitive (.129) net of other predictors, but fear of crime is not a
significant predictor of punitiveness. Of the background variables, only education has
a significant net effect (-.105): the more years of schooling the less punitive are respondents. The variance inflation factor scores are all less than 2.0 indicating that collinearity among predictors is not an issue.
Table 2a: Correlations
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
PUNATT
GENDER
AGE
ED
MINORITY
POP
%VIOL
PROBLEM
NEIGHBOUR
THINK
VICTIM
WORRY
FEAR
ANGER
EXPER
1
1
2
.060
1
3
.037
.047
1
4
-.267***
-.091***
-.073*
1
5
-.073*
-.001
.082**
.040
1
6
-.126***
-.081*
-.065*
.208***
-.085**
1
7
.315***
.249***
-.065*
-.224***
-.136***
-.118***
1
8
.539***
.118***
.099**
-.254***
-.090**
-.094**
.394***
1
363
Table 2b
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
PUNATT
GENDER
AGE
ED
MINORITY
POP
%VIOL
PROBLEM
NEIGHBOR
THINK
VICTIM
WORRY
FEAR
ANGER
EXPER
* p < .05
10
11
12
13
14
15
.160***
-.001
-.071*
-.088**
-.117***
.082*
.252***
.267***
1
.297***
.037
-.002
-.135***
-.128***
-.057
.226***
.292***
.277***
1
.056
.069*
.058
-.097**
-.091**
-.052
.058
.048
-.125***
-.029
1
.270***
.163***
.013
-.117***
-.153***
.080*
.255***
.284***
.423***
.357***
.025
1
1
.259***
.346***
.088**
-.049
-.114***
.009
.212***
.206***
.072*
.174***
.109**
.344***
1
1
.393***
.177***
-.002
-.068*
.006
.001
.209***
.252***
.136***
.170***
-.058
.207***
.573***
1
.021
.047
.047
-.015
-.069*
.056
-.023
-.018
-.071*
-.081*
.240***
.050
.109**
-.006
1
** p < .01
Table 3:
*** p < .001
OLS Regression of Punitiveness on Predictors
or
Background
Gender
Age
Education
Visible Minority
Population Size
Knowledge
% Violent Crime
Crime Salience
Crime Problem
Crime Neighborhood
Think Crime
Experience
Victim
Personal Experience
Emotions
Fear
Worry
Anger
R2 =.429
R2ADJ =.419
SE = 1.729
* p < .05
Std.
Error
VIF
-.269
-.003
-.071
.048
-.041
.138
.004
.020
.173
.038
-.059
-.019
-.105
.008
-.031
1.23
1.06
1.15
1.09
1.12
-1.958
-.691
-3.569***
.274
-1.075
.007
.003
.076
1.32
2.417*
.881
-.130
.125
.068
.056
.054
.415
-.075
.070
1.38
1.39
1.24
12.907***
-2.325*
2.296*
.184
.135
.133
.130
.040
.030
1.11
1.10
1.383
1.037
-.023
.242
.309
.037
.063
.042
-.024
.129
.260
1.89
1.50
1.64
-.626
3.854***
7.418***
df
Regression 14
Residual 762
F = 40.923***
** p < .01
SS
1712.661
2277.874
*** p < .001
MS
122.333
2.989
364
We also conducted the Backward Elimination method of multiple regression

in order to check on the stability of the regression coefficients, with all of the
predictor variables entered as in the first equation. The final results were very
similar to the results above. Only gender emerged as an additional significant
predictor (Beta = -.064; p < .05), with females somewhat less punitive in their
attitudes. The values of the standardized regression coefficients for the other statistically significant predictors were very close to the results in Table 3.
Conclusions
These results provide somewhat mixed support for the hypotheses. The emotions
of fear and anger were predicted to have positive effects on punitive attitudes,
with anger predicted to have the stronger effect. Anger about crime did indeed
have a stronger net effect while fear was unrelated to punitiveness. However, our
alternative measure of fear worry about crime did have a positive net effect,
although less than that of anger. Thus the effects of emotions were generally as
hypothesized. Crime salience had mixed effects on punitiveness. As expected,
those who generally perceived crime to be a problem were more punitive, while
perceiving neighborhood crime as a problem had only a small, negative effect on
punitive attitudes. Thinking more about crime also had only a small but positive
effect. While the effect of knowledge about crime was in the expected direction
the more inaccurate were more punitive , the magnitude of the effect was quite
small. Experience with crime as a victim, as expected, did not affect punitive attitudes. Other types of personal experience of crime, at least as measured here,
also had no effect. Generally, then, all of the hypotheses received some support,
although this varied in terms of the magnitude of the effects and, in some cases,
the particular measure.
These results suggest that the emotional aspects of punitiveness should receive
much more research attention. While there is a fairly substantial literature on the
fear of crime, empirical research focusing upon such other emotions as anger and
resentment and their relationships with reactions to crime and punishment is only
beginning to be explored. The present study was limited in its measurement of
anger to only two types of crime and offenders. Additional research should examine a broader range of crime types and offender characteristics to gauge the
effects of anger on punitiveness under a variety of circumstances. Further work
should also attempt to delineate the sources of such anger and other emotional
reactions to crime. For example, such variables as prior experience with crime
and/or a higher salience of crime and justice may affect these emotional responses. Particular attention should be given to the sources of anger in those who
365
have no prior direct experience of crime. The latter may be related to the use of
stereotypes (Lerner et al. 1998) and the role of the mass media in forming attitudes and reactions to crime. Another potentially fruitful line of inquiry would be
to examine the possible effects of economic and other types of insecurity on anger and its role in punitiveness (Gaubatz 1994). Finally, the relative importance
of state versus trait anger and other emotions in predicting punitiveness needs
further investigation.
Measurement of some important concepts remains an issue for further development. Clearly how fear of crime is conceptualized and measured have significant implications. There is by now a sizeable literature on this subject. Given the
conflicting research results concerning the effect of fear of crime on punitive attitudes, further attention to this topic is necessary. In the present case, fear, measured in a parallel fashion to anger, failed to predict punitiveness net of other predictors. However, we noted that this measure is still somewhat hypothetical in
asking respondents to rate how each scenario would make them fearful. On the
presumption that a more direct measure was required, we included an index of
worry about crime. While this index proved to be a significant net predictor of
punitiveness, worry is more akin to distress, concern or agitation than the dread
or trepidation associated with fear. More conceptual and empirical attention to
these distinctions should be a priority.
Our two measures of experience with crime and the criminal justice system
did not have a significant net effect on punitive attitudes. The absence of an effect of prior victimization is in line with previous research. Other forms of experience with crime or criminal justice could have variable effects depending
upon the nature and/or quality of such experience. Since we were unable to distinguish these various types of experience, it is perhaps not surprising that, overall, personal experience had no effect.
The overall perception that crime has increased was the strongest net predictor
of punitive attitudes, followed by anger about crime. Both the cognitive perception that crime is an increasing problem and an emotional reaction of anger about
crime appear to be important variables galvanizing people to demand more punitive responses to crime. This suggests that the combination of these two elements
a strongly held belief that crime has increased along with a greater degree of
anger about crime and/or criminals may be a particularly potent force leading
to punitive attitudes. This interaction of cognitions and emotions, as well as other
potential interaction effects or conditional relationships should be explored in
subsequent research.
366
References
Alberta Justice (1998). Survey Results: Environics Poll. www.gov.ab.ca/justicesum
mit/survey/survey2.htm
Averill, J.R. (1983). Studies on anger and aggression. American Psychologist 38,
1145-1160.
Baron, Stephen W. & Timothy F. Hartnagel (1996). Lock em up: Attitudes toward
punishing juvenile offenders. Canadian Journal of Criminology 38 (2), 191-212.
Beckett, Katherine (1997). Making Crime Pay: Law and Order in Contemporary
American Politics. New York: Oxford University Press.
Beckett, Katherine & Theodore Sasson (2000). The Politics of Injustice. Thousand Oaks,
California: Pine Forge Press.
Blumstein, Alfred & Jacqueline Cohen (1980). Sentencing of convicted offenders.
Law and Society Review 14, 223-261.
Brillon, Yves (1988). Punitiveness, status and ideology in three Canadian provinces,
84-110; in Nigel Walker and Mike Hough (eds.), Public Attitudes to Sentencing.
London: Gower.
Bureau of Justice Statistics (1977). Sourcebook of Criminal Justice Statistics. Washington, D.C.: U.S. Department of Justice.
Chiricos, Ted, Kelly Welch & Marc Gertz (2004). Racial typification of crime and support for punitive measures. Criminology 42 (2), 359-389.
Conklin, John (1971). Dimensions of community response to the crime problem. Social Problems 18, 373-385.
Cullen, Francis T., Gregory A. Clark, John B. Cullen & Richard A. Mathers (1985).
Attribution, salience, and attitudes toward criminal sanctioning. Criminal Justice
and Behavior 12, 305-331.
Cullen, Francis T., Bonnie S. Fisher & Brandon K. Applegate (2000). Public opinion
about punishment and corrections. Crime and Justice: A Review of Research 27, 1-79.
Ditton, Jason, Jon Bannister, Elizabeth Gilchrist & Stephen Farrall (1999a). Afraid or angry? Recalibrating the fear of crime. International Review of Victimology 6, 83-99.
Ditton, Jason, Stephen Farrall, Jon Bannister, Elizabeth Gilchrist & Ken Pease (1999b).
Reactions to victimization: why has anger been ignored? Crime Prevention and Community Safety 1, 37-54.
Doob, Anthony N. (2000). Transforming the punishment environment. Canadian Journal of Criminology 42 (3), 323-340.
Doob, Anthony N. & Julian V. Roberts (1983). An Analysis of the Publics View of
Sentencing. Ottawa: Department of Justice Canada.
Doob, Anthony N. & Jane B. Sprott (1997). Fear, victimization, and attitudes to sentencing, the courts, and the police. Canadian Journal of Criminology 39, 275-291.
Ehrenreich, Barbara (1989). Fear of Falling: The Inner Life of the Middle Class.
New York: Pantheon.
367
Farrall, Stephen (2001). Anger, fear, and victimization: emotional responses without
emotions. Paper presented at joint meetings of Law and Society Association and
Research Committee on the Sociology of Law (ISA), Budapest, July.
Freiberg, Arie (2001). Affective versus effective justice. Punishment and Society 3
(2), 265-78.
Furstenberg, Frank F. (1974). Public reaction to crime in the streets. American Scholar 40 (Fall), 601-610.
Gannon, Maire (2006). Crime statistics in Canada, 2005. Juristat 26 (4), 1-23.
(2005). General Social Survey on Victimization, Cycle 18: An Overview of Findings, 2004. Ottawa: Statistics Canada.
Gannon, Maire & Karen Mihorean (2005). Criminal victimization in Canada, 2004.
Juristat 25 (7), 1-27.
Garland, David (2001). The Culture of Control. Oxford: Oxford University Press.
(2000). The culture of high crime societies. British Journal of Criminology 40:
347-375.
(1990). Punishment and Modern Society. Oxford: Clarendon Press.
Gaubatz, Kathlyn (1994). Crime in the Public Mind. Ann Arbor, MI: University of
Michigan Press.
Gault, Barbara A. & John Sabini (2000). The role of empathy, anger, and gender in
predicting attitudes toward punitive, reparative, and preventative public policies.
Cognition and Emotion 14 (4), 495-520.
Greene, J. & J. Doble (2000). Attitudes towards crime and punishment in Vermont:
Public opinion about an experiment with restorative justice. Englewood Cliffs NJ:
John Doble Research Associates.
Henrick, D. & L. Farmer (2002). Adult correctional services in Canada, 2000-01.
Juristat: 22 (10).
Hogan, Michael J., Ted Chiricos & Marc G. Gertz (1997). Economic insecurity and
punitive attitudes. Paper presented to the American Society of Criminology, San
Diego.
Hough, Mike & Alison Park (2002). How malleable are attitudes to crime and punishment? Findings from a British deliberative poll, 163-183; in Julian V. Roberts &
Mike Hough (eds.), Changing Attitudes to Punishment. Portland, OR: Willan Publishing.
Hough, Mike & Julian V. Roberts (2002). Public knowledge and public opinion of sentencing, 157-176; in Cyrus Tata and Neil Hutton (eds.), Sentencing and Society:
International Perspectives. Aldershot Hamshire: Ashgate Publishing.
Hough, Mike, Helen Lewis & Nigel Walker (1988). Factors associated with punitiveness in England and Wales., 203-217; in Mike Hough & Nigel Walker (eds.),
Public Attitudes to Sentencing. London: Gower.
Hutton, Neil (2005). Beyond populist punitiveness? Punishment and Society 7 (3):
243-258.
368
Indermaur, David & Mike Hough (2002). Strategies for Changing Public Attitudes to
Punishment, 198-214; in Julian V. Roberts & Mike Hough (eds.), Changing Attitudes to Punishment. Portland, Oregon: Willan Publishing.
Karstedt, Susanne (2002). Emotions and criminal justice. Theoretical Criminology 6
(3), 299-317.
Keltner, D., P.C. Ellsworth & K. Edwards (1993). Beyond simple pessimism: effects
of sadness and anger on social perception. Journal of Personality and Social Psychology 64, 740-752.
Langworthy, R.H. & J.T. Whitehead (1986). Liberalism and fear as explanations of
punitiveness. Criminology 24, 575-591.
Law Commission of Canada (2003). What is a Crime? Challenges and Alternatives.
Ottawa: Her Majesty the Queen in Right of Canada.
Lerner, J.S., J.H. Goldberg & P.E. Tetlock (1998). Sober second thought: the effects of
accountability, anger, and authoritarianism on attributions of responsibility. Personality and Social Psychology Bulletin 24 (6), 563-574.
Maguire, M. (1980). The impact of burglary upon victims. British Journal of Criminology 20 (3), 261-275.
Maruna, Shadd & Anna King (2004). Public Opinion and Community Penalties. in
Anthony Bottoms, G. Robinson & Sue Rex (eds.), Non-Custodial Sanctions in an
Insecure Society. Cullompton: Willan Publishing.
Matthews, R., L. Hancock & D. Briggs (2003). Jurors Perceptions, Understanding,
confidence, and satisfaction in the Jury System: A Study in Six Courts. London:
Home Office, Research, Development and Statistics Directorate.
Mawby, R.I. & S. Walklate (1997). The impact of burglary. International Review of
Victimology 4, 267-295.
Mirrlees-Black, Catriona (2002). Improving public knowledge about crime and punishment, 184-197; in Julian V. Roberts & Mike Hough (eds.), Changing Attitudes
to Punishment. Portland OR: Willan Publishing.
Ouimet, Marc & Edward J. Coyle (1991). Fear of crime and sentencing punitiveness.
Canadian Journal of Criminology 33, 149-162.
Quigley, B.M. & J.T. Tedeschi (1996). Mediating effects of blame attributions on feelings of anger. Personality and Social Psychology Bulletin 22, 1280-1288.
Rankin, J.H. (1979). Changing attitudes toward capital punishment. Social Forces 58
(September), 194-211.
Rich, Robert F. & Robert J. Sampson (1990). Public perceptions of criminal justice
policy. Violence and Victims 5, 109-118.
Roberts, Julian V. (2002). Public Opinion and the Nature of Community Penalties:
International Findings, 33-62; in Julian V. Roberts & Mike Hough (eds.), Changing Attitudes To Punishment. Portland, OR: Willan Publishing.
369
(2001). Fear of crime and attitudes to criminal justice in Canada. Ottawa: Public
Works and Government Services Canada, Cat. No. JS42-99/2001.
(1992). Public opinion, crime, and criminal justice. In Michael Tonry (ed.), Crime
and Justice. Chicago: University of Chicago Press.
Roberts, Julian V. & Loretta J. Stalans (1997). Public Opinion, Crime and Criminal
Justice. Bouldher, CO: Westview Press.
2000. Public Opinion, Crime, and Criminal Justice. Boulder, CO: Westview Press.
Roberts, Julian V., Loretta J. Stalans, David Indermaur & Mike Hough (2003). Penal
Populism and Public Opinion. Oxford: Oxford University Press.
Rossi, P.H., J.E. Simpson, & J.L. Miller (1985). Beyond crime seriousness. Journal of
Quantitative Criminology 1, 59-90.
Scheingold, S.A. (1991). The Politics of Street Crime. Philadelphia: Temple University Press.
Sheley, J.F. (1985). Americas Crime Problem. Belmont: Wadsworth.
Sprott, Jane B. & Anthony N. Doob (1997). Fear, victimization, and attitudes to sentencing, the courts, and the police. Canadian Journal of Criminology 39, 275-291.
Stalans, Loretta J. (2002). Measuring attitudes toward sentencing, 15-32; in Julian V.
Roberts & Mike Hough (eds.). Changing Attitudes To Punishment. Portland, Oregon: Willan Publishing.
Statistics Canada (2006). CANSIM tables: http: //www40.statcan.ca/101/cst01/legal31a.htm.
Stinchcombe, Arthur L., Rebecca Adams, Carol A. Heimer, Kim Lane Scheppele, Tom
W. Smith & D. Garth Taylor (1980). 2Crime and Punishment. San Francisco:
Jossey-Bass.
Taylor, D.G., K.L. Scheppele & A.L. Stinchcombe (1979). Salience of crime and support for harsher criminal sanctions. Social Problems 26, 413-424.
Thomas, Charles W. & Robin J. Cage (1976). Correlates of punitive attitudes toward
legal sanctions. International Journal of Criminology and Penology 4 (August), 239255.
Thomas, C.W. & S.C. Foster (1975). A sociological perspective on support for capital
punishment. American Journal of Orthopsychiatry 45, 641-657.
Tufts, Jennifer (2000). Public attitudes toward the criminal justice system. Juristat 20
(12), 1-22.
Tufts, Jennifer & Julian V. Roberts (2001). Sentencing juvenile offenders. Unpublished
manuscript, Department of Criminology, University of Ottawa.
Tyler, Tom R. & Robert J. Boeckmann (1997). Three strikes and you are out, but why?
Law and Society Review 31 (2), 237-266.
Tyler, Tom R. & Renee Weber (1982). Support for the death penalty: instrumental response to crime or symbolic attitude? Law and Society Review 17 (1), 21-45.
Useem, Bert, Raymond V. Liedka & Anne Morrison Piehl (2003). Popular support for
the prison build-up. Punishment and Society 5 (1), 5-32.
370
van Dijk, Jan J.M. & Carl H.D. Steinmetz (1988). Pragmatism, ideology and crime
control. 74-83; in Nigel Walker & Mike Hough (eds.), Public Attitudes to Sentencing. Aldershot: Gower.
Walker, Nigel, Mark Collins & Paul Wilson (1988). How the public sees sentencing,
149-159 in Mike Hough & Nigel Walker (eds.), Public Attitudes to Sentencing.
London: Gower.
Walker, Nigel, Mike Hough & Helen Lewis (1988). Tolerance of leniency and severity
in England and Wales, 178-202; in Nigel Walker & Mike Hough (eds.), Public Attitudes to Sentencing. Aldershot: Gower.
Zamble, Edward & K.L. Kalm (1990). General and specific measures of public attitudes toward sentencing. Canadian Journal of Behavioural Science 22, 327-337.
Zimmerman, Sherwood E., David J. Van Alstyne & Christopher S. Dunn (1988). The
national punishment survey and public policy consequences. Journal of Research in
Crime and Delinquency 25, 120-149.
371
Criminal Policy Based on Fear of Crime:

Case of Poland
ANNA KOSSOWSKA, IRENA RZEPLINSKA,

DAGMARA WOZNIAKOWSKA and WITOLD KLAUS
1. Introduction
One of the characteristic features of modern societies is the universal fear of
crime, which should be understood as consisting both of subjective fear as well
as of objective risk. When speaking about objective risk, putting aside scepticism
regarding credibility of official statistical data, it needs to be taken into account
that crime at present is, however, different from that in the 60s and 70s. It differs
both in number (it has undoubtedly increased) and type. Changes in type of crime
concern mainly increased aggressiveness of criminal acts. Not only does the
number of aggressive crimes increase but also traditional crime against property
includes more elements of aggression than before. This new character of traditional crime increases the fear.
The aforementioned subjective feeling of fear refers to universal tendency to
exaggerate the risk of crime. We often speak of the phenomenon of fear of
crime, i.e. the fact that the intensity of this fear is disproportional to officially
registered crime rate. It appears it is justified to draw this social phenomenon
from the condition of societies resulting from deep social change. Not only in
countries of undergoing political transformation the dimension and pace of social
change cause uncertainty, fear and feeling of threat verbalized often as fear of
crime.
It is a fact that victimisation becomes more and more common 1 . Crime has
moved from the rare, the abnormal, the offence of the marginal and the stranger,
to a commonplace part of the texture of everyday life. 2 .
The increase in crime combined with the feeling of threat and fear of crime are
a fact influencing behaviour of individuals, as well as their attitude to the gov1
See eg. D. Garland and R. Sparks (2000), Criminology, Social Theory and the Challenge of
Our Times, Brit. J. of Criminology 2000, Vol. 40, No 2.
J. Young (1999), The Exclusive Society. Social exclusion, crime and difference in late modernity. Thousand Oaks. p. 30.
372
A. Kossowska, I. Rzeplinska, D. Wozniakowska & W. Klaus
ernment, as well as the tendency to define an enemy. Often a real perpetrator becomes this enemy but even more often an imaginary, potential perpetrator a
stranger.
This universal fear of crime (or imagined crime) of citizens explains public
support for the doctrine of Law and Order in its various forms. Attitude of frustration, aggression and anger against crime, criminals and supporters of so called
liberal criminal policy becomes more common. Consent for formal social control
increases, there appears an appeal to increase so called severity of punitive measures.
One cannot disregard manipulating social fear of crime, applying the problem
of risk to short-term goals of political campaigns, changing the law to increase its
rigorism, not only basing on scientific analyses and expert opinion but basing on
internal convictions, frustrations and political calculations of policy makers.
Christie 3 draws attention to the fact that in modern developed societies crime
control is one of the few areas where politicians can successfully create their image for the purpose of political campaigns. He explicitly says that crime control
is a platform for self-presentation. This probably results from the fact that in
other areas citizens do not expect too much from the government.
According to Garland 4 , a change of societies attitude towards crime results
from specific combination of threat, insecurity and problems of functioning of
social control. Among many results of this new attitude to crime and crime control Garland lists increased punitiveness of the justice system, departure from the
ideas of resocialisation, change in emotional tone of discourse on crime control
(evident in Poland, particularly in current political debates), putting emphasis on
defence of society against threats at the cost of protection of human rights, as
well as privatisation of security. However, most of these points run politics into
trouble with crime and with specific penal populism.
The presented work intends to show the problem of crime risk in Poland and
its consequences for the penal policy, starting with clearly visible politicisation
of this problem.
2. Politicisation of the fight against crime in Poland

Crime and fight against crime is the part of social life which each government
in Poland, before and after political transformation, drew attention to and used
instrumentally. Yet, only after 1989, in democratic Poland, political parties start3
4
N. Christie (2004), Dogodna ilo przestpstw, Warszawa.

D. Garland (2001), Culture of Control; Crime and Social Order in Contemporary Society,
the University of Chicago Press, Chicago.
Criminal policy based on fear of crime: case of Poland
373
ing in parliamentary elections clearly and in distinctive ways used the slogans of
crime risk and the need to toughen penal policy of the courts.5 How much actual
crime and implemented penal policy justified scaring the society with crime is
a separate issue. For political parties it was one of the comfortable and purely
instrumentally used issues in election campaigns.
In 1997 parliament election campaign, election programmes of all political
parties included calls to increase the security 6 , however, they were quite general.
Sojusz Lewicy Demokratycznej 7 promised in case it wins the elections that it will:
improve the law, increase the effectiveness of police, public prosecution, as well
as justice system, particularly through simplification and acceleration of court
procedures. Unia Wolnoci declared the need to streamline police and justice
actions. Right-wing parties postulated the necessity to toughen punishments, restore death penalty and the necessity of harsher imprisonment conditions.
The party which won the elections in 1997 Akcja Wyborcza Solidarno
included in its programme that the basic duty of the state is to ensure safety. It
made more detailed reference in slogans of prevention of organised criminal
structure spread and decided fight with bribery.
At that time, in the 1990s, the crime rate (crimes per 100,000 inhabitants) rose
in comparison with the years 1989-1990, but already in 2000 it became stable.
Just then (from 2000), the Minister of Justice 8 started a campaign to toughen
criminal law and penal policy which he justified by the surge in crime rate. The
society was threatened with crime without justifying the threat with relevant
data on crime.
Politicisation of crime issue and legal reaction to crime in postmodern societies is discussed
in Garland (2001), pp.139 et. seq.
Information on the problem of crime risk in the 1997 election campaign is based on J.
Widacki (2002), Polscy politycy o bezpieczestwie (midzy demokracj a demagogi). in: J.
Czapska i H. Kury (red.), Mit represyjnoci albo o znaczeniu prewencji kryminalnej, wyd.
Zakamycze, s. 95 i dalsze.
Sojusz Lewicy Demokratycznej is a left-wing party; Unia Wolnoci a liberal party (now renamed Partia Demokratyczna); Akcja Wyborcza Solidarno is a coalition of conservative parties (ceased to exist); Platforma Obywatelska a conservative-liberal party;
Samoobrona a farmers populist party; Prawo i Sprawiedliwo a conservative-populist
party; Liga Polskich Rodzin an extreme right-wing party.
Now the President of Poland Lech Kaczyski.
374
In the last parliamentary elections, in 2005, all political parties included slogans, references and evaluations concerning crime, crime threat and crime response penal policy applied by the courts and justice system operation.9 For
example in its political programme Platforma Obywatelska made reference to the
bad system where criminals walk with impunity (because their cases fall under
the statute of limitations), the courts are less severe for perpetrators than for victims, and where access to courts is purely formal due to the long time of proceeding. Another political party Samoobrona listed: inequality before the law,
cases of corruption in courts, applying detention and punishing the poor while
avoiding the same towards the rich and those with political connections.
The most complete assessment of situation concerning crime and proposals of
changes were included in the programme of Prawo i Sprawiedliwo. It announced
that the crime rate in Poland is high, and its systematic growth observed since the
1990s causes the citizens to feel threatened with crime. The culprits were the
liberal penal policy and the bad functioning of police, courts and public prosecution service. The new Polish penal code from 1997 was defined as liberal and
allowing to sentence so low penalties that it argued with basic sense of justice.
According to that programme, the most dangerous criminals are to be effectively
isolated in jails, the longest imprisonment is to be 25 years, life imprisonment
could be sentenced without the right to conditional early release, property obtained through criminal acts was to be confiscated by the state, penalties for reoffenders toughened, defence of necessity broadened and list of circumstances excluding criminal liability in case it is overstepped extended, penalties for offences against health and life were to be toughened. Political critics of the penal
code defined it as liberal, protecting criminals, not citizens from crime, not protecting the society from an increasing crime rate, particularly by abandonment of
general prevention understood as an imperative to scare potential imitators of
punished perpetrator, aiming at resocialisation alone and thus contrary to the idea
of justice 10 .
It is worth to note that criticism of the penal code was completely groundless
from the political and criminal point of view the code is constructed in such a
manner that it allows to administer suitable punishments: from lenient to severe
and very severe.
9
10
Presentation based on the article: T. Bulenda (2006), Polityczno kryminalne implikacje

dotyczce bezpieczestwa publicznego w programach partii politycznych w wyborach parlamentarnych 2005 r., Przegld Wiziennictwa Polskiego, nr. 49, p. 41-64.
Quoted according to the co-author of the 1997 penal code: A. Zoll (2003), Czy trzeba
zmienia kodeks karny? Materials from the conference: Nowelizacja kodeksu karnego,
Wydawnictwo Naukowe Scholar, Warszawa, s. 13.
375
In November 2006, local elections in Poland took place. During the election
campaign, one of the right-wing parties with low ratings (Liga Polskich Rodzin,
represented in Polish parliament and a member of current government coalition)
presented the most severe proposal: introduction of capital punishment for
child murder related with abuse.
The Republic of Poland is a party of the European Convention on Human
Rights and Fundamental Freedoms as well as the Additional Protocol No. 6
whose Article 1 reads as follows: The death penalty shall be abolished. No-one
shall be condemned to such penalty or executed. Introduction of death penalty
in Poland is hence impossible. The authors of the idea know it very well. They
treat the idea purely instrumentally counting to win electoral votes. This is because public opinion polls in Poland show that over 70% of respondents declare
acceptance for death penalty for most serious crimes (the percentage is alike
throughout whole Europe)11 . So far public opinion is that severe, cruel punishments are correct in dealing with criminals. Promoting the idea that death penalty is the capital punishment on which the power of law is based and hence forcing the postulate to reintroduce death penalty in Poland is detrimental for development of moral and legal culture of the society this quote from the statement
of Helsinki Committee conveys the reasons why public opinion in Poland cannot
be a decisive advisor. 12
The example of Poland in the past years shows how strongly political the
problem of crime and approach to criminals became, that is penal policy defined
by legislator and applied by the courts. Between the election campaign in 2005
and the local election campaign in autumn 2006 the majority party in the parliament and the coalition used the slogans of crime risk and the need for sever reaction on it as a peculiar technique of exercising power. Criticism of the current
penal code and penal policy implemented by the courts as being too liberal and
lenient became a permanent element of political discourse, as well as contrary
to facts indicating a threat of serious crime and the need of decisive steps to be
taken by the Polish government (read: adequately severe repressive measures).
This clash of liberal (and just wrong) policy as opposed to the only correct,
repressive one, became the axis of political discussion and an indicator of government actions in contemporary Poland.
11
12
Public opinion poll by Centrum Badania Opinii Spoecznej 2004.

Declaration of the Helsinki Committee in Poland published in: Gazeta Wyborcza from 17.
08. 2006 r.
376
Why? What is the explanation for so sharp, clear, and political use of fear of
crime (threat of crime) and penal reaction to crime by formulating a demand
of high sentences by the courts as the only just solution? Why crime becomes
one of the main election slogans? Why do politicians promote the need for severe
and very severe punishments as the only solution to the problem? Why do politicians think that the society can be convinced only by a tough approach, i.e.
slogans on severe, merciless repression in fight against crime? They forget that
everything that is included in the penal code in the list of penal measures is a repression: imprisonment is deprivation of liberty.
The myth of effectiveness of severe repression as a response to crime is deeply
rooted in the social consciousness hence it is readily and effectively used in
political campaigns as Poland shows over the past few years. 13 The myth is
used regardless of official statistical data and criminological research, regardless
of the actual and perceived crime risk. Creation of the problem of crime by politicians is a very rewarding and accurate instrument in the political game. The
problem is also created by the media as they speak and write of the most drastic
and appalling crimes (yet fail to mention that crimes of such character are merely
a fraction of statistics).
An average citizen knows that crime exists and is informed that severe punishments are the best way to fight it. Respect and chances of winning the elections are awarded to those who advocate tough punishment of criminals. In this
way, the politicians teach the society the only way of reaction to crime: severity and mercilessness. They teach that other punishments not related to imprisonment are ineffective because they are too mild that is inadequate in the
fight against crime. Hence, they show the society that every criminal should
get to jail because this is the only suitable punishment.
But what is the actual crime risk in Poland? At present, it is measured with
two indices: police statistics and victimisation research supported by research on
sense of security in the Polish society.
3. The fear of crime and state of crime in Poland at the beginning

of the 21st century
Ten years ago, the Polish criminologist M. Filar said that fear of crime grows
geometrically in relation to the growth of the very crime rate 14 . Since then, the
sentence has been repeated in various reports, studies and articles. It is worth to
13
14
Cases of other countries, eg. UK and USA are discussed in: D. Garland (2001), pp. 139-141.
M. Filar (1995), Siedem mitw o poytku z wizienia, Gazeta Wyborcza from 17 July 1995 r.
377
have a closer look on how Polish fear of crime looks like when crime stabilises
or even decreases.
The two things Poles fear most are damage to their or their family members
health and, secondly, crime. Unemployment comes next, alcoholism, drug abuse,
AIDS and hooliganism follow shortly 15 . As far as crime risk is concerned, Polish
surveys show which crime people fear most at the place they live. The results
indicate theft and robbery (46%), armed robbery (41%) and reckless driving
(39%). What is listed later are acts which perhaps are not dangerous in themselves but make people uncomfortable: aggression of people under influence of
alcohol or drugs (26%), being provoked by aggressive juveniles (25%), noise and
impolite behaviour of neighbours (10%), vandalism (29%). 16 It seems that Poles,
speaking of fear of crime, take both seriously: crime, and also less dangerous
but unpleasant petty offences.
Table 1:
Dynamics of crime in Poland in 1990-2005 17 according to police

statistics
Year
Total
2005
2004
2003
2002
2001
2000
1999
1998
1997
1996
1995
1994
1993
1992
1991
1990
15
16
17
1.379.962
1.461.217
1.466.643
1.404.229
1.390.089
1.266.910
1.121.545
1.073.042
992.373
897.751
974.941
906.157
852.507
881.076
866.095
883.346
Reported Crimes:
Road
Street
Crimes
Crimes
196.486
1.000.096
177.296
1.085.295
168.827
1.101.387
163.012
1.083.854
138.817
1.107.073
19.894
1.133.162
20.505
1.020.654
34.442
931.037
40.202
838.941
34.863
761.664
35.005
835.641
32.030
780.781
29.578
734.956
35.568
721.082
30.753
721.960
29.141
760.325
Business
Crimes
136.801
152.148
151.596
109.698
103.521
84.260
60.393
78.496
82.296
73.911
83.893
74.956
64.055
108.051
84.590
78.337
Number of
Suspects
Detectability
594.088
578.059
557.224
552.301
533.943
405.275
364.272
396.055
410.844
381.911
423.896
388.855
299.499
307.575
305.031
273.375
58.6
56.2
55.2
54.9
53.8
47.8
45.0
50.5
53.5
54.4
54.2
54.2
53.5
53.1
48.3
40.0
A. Stankiewicz (2006), Polakw strachy wasne, Rzeczpospolita, 25 lipca 2006 r.

A. Biekowska (2006), Czego boj si Polacy, Policja 997, nr 17, sierpie 2006 r.
Statisctics of Police Headquarters http://www.policja.pl/portal.php?serwis=pol&dzial=3&id=1.
378
Being fully aware of all reservations concerning comparison of police statistics or self-reports for subsequent years and judging crime growth on this basis,
one cannot disregard that crime in Poland had grown since the 1980s and later on
grew only to finally reach a flat line and even a slow decrease. What is more, it is
apparent that the street crime which is feared most, is falling as well. Table 1
shows the crime growth in Poland in the years 1990-2005.
Also self-report studies show the tendency, if not of a sharp fall, surely of stabilisation of crime rate. In 2005, as compared to 2000, less persons fell victims of
theft, assault and assault and battery. In 2005, as compared to 2003, it can be
seen a decline in the number of persons who experienced burglaries to houses
and cars (detailed data in table 2).
Table 2:
Percentage of persons declaring an experience of having been a

crime victim in 1993-2003 18
Experience
1993
1995
2000
2003
2005
Theft
5,1
5,4
6,8
5,5
5,8
Assault and battery
1,6
1,7
1,6
1,3
1,2
Burglary
1,2
1,2
2,1
4,0*
3,5
* since 2003 question concerned burglary to a house or a car.

Data from: 1993-1995 Czapiski, 1998; 2000-2005 Diagnoza spoeczna.
Most recent victimisation studies also show a drop in crime in Poland 19 . Along
with a stabilisation or even a slight decline of crime rate, detectability is on the
rise (see table 1), as well as Poles trust in the police. 24 July 1990, when the police forces were formed 20 only 56% of respondents said they trusted in police,
now it is 70% of Poles (see Chart 1).
18
19
20
J. Czapiski, T. Panek (red.) (2006), Diagnoza Spoeczna 2005, Wysza Szkoa Finansw i
Zarzdzania w Warszawie, Warszawa, p. 179.
Unauthorised statement of prof. Andrzeja Siemaszko, Director of Instytut Wymiaru Sprawiedliwoci (Institute of the Judiciary), the author of the first Polish part of international victimisation
studies.
From 1944 r. till 1990 r. Poland had Citizens Militia.
Chart 1:
379
Trust in Police over the past 16 years (values in percent) 21
Recent years also show a rise in Poles security, both in general and in the neighbourhood.
Table 3:
Evaluation of security in the country 22 (data in percent)

Opinions of the respondents
Is Poland, in
your opinion,
a safe country?
III
87
74
22
4
YES
NO
Hard to say
Table 4:
Im not afraid
Im afraid, but not
very much
Im very afraid
Im afraid
Hard to say
22
23
IV
95
19
79
2
IV
96
21
77
2
IV
97
24
75
1
III
98
22
76
2
IV
99
33
64
4
II
00
28
70
2
IV
01
18
81
1
IV
02
27
71
2
IV
03
34
62
4
III
04
33
65
2
III
05
46
49
5
IV
06
43
53
4
Crime risk awareness23
Are you afraid of

becoming a victim of
crime?
21
XI
93
26
67
7
Opinions of the respondents

II
IV
IV
IV
00
01
02
03
31
34
32
39
IV
96
30
IV
97
37
III
98
35
IV
99
40
50
46
47
43
46
47
52
17
67
3
15
61
2
15
62
3
14
57
2
21
67
2
18
65
1
14
66
2
III
04
36
III
05
45
IV
06
43
49
49
42
46
10
59
2
14
63
1
9
51
3
9
55
2
Statistics of Police Headquarters, http://www.policja.pl/portal.php?serwis=pol&dzial=3&id=1.

Statistics of Police Headquarters. http://www.policja.pl/portal/pol/1/1837/.
Statistics of Police Headquarters. http://www.policja.pl/portal/pol/1/1837/.
380
As the figures show, the crime rate in Poland has stabilised, one can even
speak of a slow decrease. Safety increases and does it much more dynamically
than crime declines. If so, can we make a point that fear of crime decreases (or
safety increases) geometrically in proportion to the crime rate decrease? This
seems unjustified. Yet, one can risk the hypothesis that fear of crime is related
not so much to the crime rate itself, but to the crime image presented by the media and politicians. We have already mentioned that Polish politicians often use
fear of crime as a tool to gain or to retain power. If we set the figures above
against the information on crime rate provided by politicians we may see an intriguing thing indeed.
The Last parliamentary elections took place in autumn 2005. 2004 and 2005
were years of a very brutal fight for power and an intensive election campaign.
Prawo i Sprawiedliwo, now at power but in opposition then, put a strong emphasis on the issue of public security, threatening with allegedly increasing crime
rate, exposing brutal crimes, and demanding severe punishments. And, in 2004,
security ratings as well as trust in police fall down. Once Prawo i Sprawiedliwo came to power, they changed their tactic of political game of manipulating
with crime data. Now the party speaks of its own efficiency. The Prime minister
at that time, Kazimierz Marcinkiewicz, announced at his 100th day as the PM:
Zbigniew Ziobro 24 goes to no end of trouble to restore the Judiciary to the state
and its citizens. But at the same time we are working on an amendment of penal
code and the Code of Penal Proceedings. Among others, a simplified procedure
to examine offences and less serious crimes, so-called 24-hour procedure. (...)
We have introduced far reaching changes in the police and public prosecution.
We seek to discipline officers by dismissing those corrupt, who cooperate with
organised crime (...). 25 Soon, the security polls and trust in police were on the
rise.
The phenomenon of this significant increase in public security accompanied
by low crime rate decrease can be explained in another way. According to
Stanley Cohen moral panic occurs in particular in the periods when the society
defines moral values and their limits anew. As the Poles seem to react more than
optimistically to a little crime rate fall, it may mean that Poles ceased to trust unquestioningly politicians and the media, and came to the point where they are
less prone to manipulation. At least in a longer term, as the temporary panic in
2004 has been already spoken about.
24
25
Polish Minister of Justice.

A speech helt by the Prime Minister Kazimierz Marcinkiewicz at that time in the Polish Parliament. http://www.kprm.gov.pl/1433_15811.htm
381
Whatever the reasons of the increasing insecurity of Poles are, it is apparent

that further toughening of laws has no relation to crime increase or feelings of
insecurity, and does not meet social expectations but is merely a political game.
How, in light of the above solutions, does Polish penal policy look like, both in
legislative actions and in judicature operation?
4. Legislative changes in criminal law at the turn of centuries

and their reasons
Inn 1997, Poland got the result of a long work upon a new criminal law, and, as
an effect, a new penal code was passed. The legal acts, drafted by the most eminent Polish specialists in penal law, aimed for greater flexibility of such law by
allowing the judges greater discretion in sentencing. The changes broke with
provisions based on retributivism and punishment, based on the principle of general prevention. It was assumed that imprisonment should not be the queen of
punishments and needs to be applied on the ultima ratio principle. As a result,
the possibilities of sentencing other punishments than imprisonment and other
alternative measures, including mediation, had increased. The principle of humanitarism was exposed as one of the most basic principles in criminal law (Article 3 of the penal code) 26 . Such legislation was based on modern scientific findings in penal law and criminology. However, it is worth to note that in spite of all
those changes the 1997 penal code should be considered more severe and repressive than the 1932 Polish penal code. 27 This shows that authors of the new laws
did not manage to disregard the pressure of public opinion caused by a surge in
detected crimes at that time.
Yet the Codes were passed at a wrong time indeed in the very end of the
term of a left-wing government. Only because of this they were immediately
strongly criticised, particularly by right-wing politicians (what has already been
discussed above). The main charges were excessive liberalism and leniency
towards perpetrators. Critics (among them several scientists) 28 , basing on retributive punishment theories, claimed that the new penal policy will not be a power26
27
28
More on this issue eg. in: M. Melezini (2003), Punitywno wymiaru sprawiedliwoci
karnej w Polsce w XX wieku, Biaystok, pp. 165 et. seq.; M. Filar (2006), Quo vadis iustitia
poenalis poloniae?, in: W. Czapliski (red.): Prawo w XXI wieku. Ksiga pamitkowa 50lecia Instytutu Nauk Prawnych Polskiej Akademii Nauk, Warszawa, p. 200-201.
See T. Szymanowski (1998), W kwestii represyjnoci kodeksu karnego z 1997 r. na tle
kodeksw karnych z 1932 i 1969 r., Przegld Prawa Karnego nr 18.
Main arguments in this discussion are listed in: J. Mozgawa, J. Szumski (2001), Kodeks karny w
ocenie doktryny i publicystyki, in: A. Zoll (red.): Racjonalna reforma prawa karnego, Warszawa,
p. 21 et. seq.
382
ful weapon in fight with surging crime. Consequently, work on an amendment of

the said laws was started immediately.
The new approach to punitive measures was briefly presented in his 10 principles of penal liability 29 by one of the new advocates of the amendment J.
Kochanowski (at present the Ombudsman). They are based on the idea of retributive justice and assume that just punishment has a priority before its utilitarian
purpose, and social purpose before the purpose which the punishment should
have for the convicted (principle 9). They also contained the idea that a condition to acknowledge the dignity of a person is that the subject of this dignity is a
person who has a factual possibility to be responsible, i.e. who is a normal person
and an adequately mature one (principle 6). Such words, contrary to the idea of
human rights based on innate dignity of a human being30 , sound extremely dangerous, particularly in the mouth of the Ombudsman a person who guards the
observance of human rights by public authorities in Poland.
The new approach presents a criminal as a social enemy a person who should
be fought against. Hence, the idea of a fight with crime based on a fight with criminals, not a suitable creation of social policy which does not favour occurrence of
crime as a social phenomenon. Foucault says that since the 19th century the government aims to impose certain outline to the way criminals were perceived so
that they are imagined as those who constantly lurk nearby, are ubiquitous and always dangerous.31 With such argumentation the society is convinced that a criminal is a person which should be punished for the committed act and the very imprisonment is a trifle, which should be accompanied by additional afflictions. The main
purpose of punishment is with such an approach to disturb the criminals committing subsequent crimes whereas the only method to achieve this is to put him in
prison32 .
Such an approach resulted in the need to amend the laws since the laws did not
suit the achievement of goals assumed by politicians. But the amendments, since
the new code was passed, were introduced in numbers never seen before. By
29
30
31
32
Published in 2006 in one of the Polish newspapers. Quote based on: J. Kochanowski (2006),
10 zasad odpowiedzialnoci karnej in: Wspczesny retrybutywizm. Powrt kary
sprawiedliwej, Ius et Lex 2006, nr 1, p. 193.
Expressly contained eg. in the preamble to the International Pact of Citizens and Political
Rights from 1996 or Article 30 of the Polish Constitution.
M. Foucault (1998), Nadzorowa i kara. Narodziny wizienia, Warszawa, p. 281. Along
with this campaign, a new one, opposing the alleged comfort of prisons was introduced. It
aimed at assigning hard labour to prisoners.
Por. A. Gaberle (2006): Rozstajne drogi. Wymiar sprawiedliwoci midzy retrybutywn a
naprawcz reakcj na przestpstwo, in: W. Czapliski: Prawo w XXI wieku , pp. 209 et
seq.
383
September 2004 the code was amended 24 times, including the first amendment
already during its vacatio legis 33 . Some of these should be considered rather as a
far-reaching reform than a mere amendment (eg. in 2003 where almost half of
penal code provisions were amended). Some of the changes were meant to adjust
Polish law to European Union or other international requirements, yet those were
only a small part of changes in general. Most changes were grounded ideologically and introduced more severe punishments34 lifting lower and upper limits
of imprisonment terms, classifying new crimes (eg. drunk driving, possession of
any amount of drugs, extending the subject of bribery), basing penal policy on
imprisonment and limiting judges discretion in choosing a penalty. Most of the
changes contradicted scientific achievements of penal law and criminology, and
most of them were never consulted with academic circles. The changes served
only the political interests and were a reflection of social tendency or a reaction
on single, but brutal or commonly condemned and hence most publicised, crimes 35 .
The changes introduced were justified by politicians in various ways. The
most frequent arguments were: crime rate increase and the need to prevent it, the
increase of social crime risk and popular demand for severe reaction on crime,
too lenient punishments applied by the judges due to new laws and finally expectations of the legal circles (judges in particular) for certain changes in current law
to be introduced.
Interestingly, a strikingly similar criticism towards the judges of excessive
liberalism was used by politicians in the late 1930s. Actually, the judges sentenced imprisonment near the lower bracket 54% of all punishments administered in 1937 were an imprisonment of between 1 week and 6 months. Yet, the
result of such policy was a stop to crime rate surge from the early 1930s, and
even its slow decline (both in detected crimes and in particular categories) 36 .
The issue of crime and emotions evoked by crime in the society has been addressed above. It is worth to note that fear of crime was triggered by politicians
themselves. Frequently, as already mentioned above, they referred to an increase
33
34
35
36
The most important changes are discussed e.g. in M. Filar (2006), pp. 203 et. seq.
Similar tendencies occur in other countries, see eg.: Ch. Pfeiffer, M. Windzio, M. Kleimann
(2005), Media Use and its Impacts on Crime Perception, Sentencing Attitudes and Crime
Policy, European Journal of Criminology, v. 2, July 2005, pp. 275 et seq.
Eg. punishments for paedophilia were significantly toughened as few cases of such crimes
were revealed and widely described in newspapers; similarly, the current discussion on
toughening punishments for juvenile criminals and lowering the age limit of penal liability
is influenced by a case of a suicide caused by classmate bullying.
Cf. M. Melezini (2003), p. 207-211, 305-308, 550.
384
of crime in their speeches and programmes of their political parties presented

during election campaigns. The issue was quickly picked up by the media which
shocked the general audience with examples of most brutal crimes. In effect, the
society believed in claims made by politicians that crime is on a dramatic rise, in
particular the most aggressive and serious crime. This caused an increase in fear
of crime 37 . In turn, the politicians used this increase of fear of crime (although
caused to a considerable extent by themselves) as an argument.
The level of citizens fear of crime has also been discussed above. It needs to
be supported by public opinion polls investigating the idea of punishment and
punishments for particular crimes suggested by respondents 38 . On one hand, respondents most frequently said that the purpose of punishment should be: to
cause other people not to commit crimes (32%) and protection of the society
(24%). The purpose of the penal law reform was to increase the security against
crime (77%) and severe punishments (73%). Only a small percentage of respondents indicated other aspects of penal law such as rehabilitation of the convicted
and facilitation of his return to the society or redressing the damage (from several
to a percent of answers). Presumably it is so because those elements are completely neglected in the public discourse which draws the general attention to the
preventive purpose of the punishment and promotes imprisonment39 .
On the other hand, only every third respondent is inclined to bear the additional tax burden of new prisons construction. Moreover, the respondents, when
asked what punishments should be administered for particular crimes, revealed
rigorous views only towards the most serious crimes (murder, rape, serious robbery). Most of the respondents could not see the need to sentence absolute imprisonment in relation to lesser crimes and indicated the need to administer alternative penal measures. Similarly, the results of international IVCS study from
1996 (respondents were asked to indicate a punishment for theft of a television
set, which is not a very serious crime) show that the Polish society is one of the
37
38
39
On the influence of the media and politicians to the perceived crime rist see eg.: Ch. Pfeiffer
et. al. (2005); R. Reiner (2002): Media made criminality. The representation of crime in the
mass media, w: M. Maguire, R. Morgan, R. Reiner: The Oxford Handbook of Criminology,
Oxford, p. 398 et. seq.; D. Downes, R. Morgan (2002): The skeletons in the cupboard. The
politics of law and order at the turn of the millennium, w: M. Maguire et al.: The Oxford
Handbook of Criminology, Oxford.
A 2002 opinion poll by OBOP, results quoted from: T. Szymanowski (2005), Polityka karna
i penitencjarna w Polsce w okresie przemian prawa karnego, Warszawa, pp. 88-90.
Already in the 18th century Cesare Beccaria drew attention to the fact that severe punishment
does not deter from crime, and research on reoffenders decades ago showed that the length of
imprisonment does not translate into later return to crime (see eg. S. Szelhaus (1969) Modociani
recydywici. Spoeczne czynniki procesu wykolejania, Warszawa, pp. 180, 202 et. seq.).
385
least punitive in Europe as compared with other countries (eg. the UK, Czech
Republic, the Netherlands, Canada or even Sweden). Data from subsequent victimisation studies show also a decreasing level of punitiveness in Poland 40 .
An argument often quoted by politicians that changes in laws or penal procedures are expected was not confirmed in research done in 2005 by Komisja
Kodyfikacyjna Prawa Karnego (Penal Law Codification Commission)41 . A questionnaire sent to all judges and public prosecutors on areas and directions of
changes revealed that as much as a half of legal practitioners do not see any need
of changes. Another half showed diverse needs of changes, yet the suggestion of
general toughening of the penal code was supported by a scarce number of respondents (not exceeding several percent, counting all responses aiming at even
partial toughening of current provisions) 42 . The following part will deal with the
question how new law is applied by the judges.
In spite of so many changes introduced to the penal law, the work on new ones
is in progress (in fact, there remain less and less original penal codes). They are
even more punitive. The work in the Ministry of Justice is not consulted with the
legal and academic circles, new projects are not revealed to the public and they
go directly to the parliament to be passed. Grounds for amendments contain false
information (eg. crime rates increase while statistical and empirical data show
another picture) or are based on observation of everyday life and knowledge
obtained from mass media 43 . Their introduction (even of very controversial solutions) is not preceded by any studies, and their essence often contradicts the
results of contemporary criminological research. But as the originators of those
changes say no one has proved that increased penal repression has not had and
cannot have any positive impact on the results of fight with crime in general44 .
M. Melezinis studies, comparing punitiveness of the justice system with crime
rate in Poland since 1918, decidedly show that there is no positive correlation
between those two variables. Yet, opinions against those changes are ignored 45
40
41
42
43
44
45
A. Siemaszko (2001), Kogo bij, komu kradn. Przestpczo nie rejestrowana w Polsce i
na wiecie, Warszawa, p. 123 et. seq.
This Codification Commission ceased to exist. All of its members were dismissed in 2006
(after Minister of Justice changed), and a new panel was appointed.
W. Wrbel (2005), Zmiany w kodeksie karnym w wietle pogldw sdziw i prokuratorw, Pastwo i Prawo, z. 12, p. 3 et. seq.
Grounds for amendment of the penal code and other acts of law, filed with the parliament 4
April 2006 (parliament document no. 485), p. 3.
Op. cit.
After one of the projects was opinioned unfavorably, the whole Commission was dismissed
by the minister and a new one, consisting of government supporters, appointed.
386
and their authors are called liberals by the politicians (this notion recently took a
pejorative overtone in Poland), defenders of criminals rights and hence the enemies of the society. The Polish specialist in penal law, M. Filar, expressed the
opinion that penal law was persistently fiddled with by unstable governments
operating in social anomalies weak governments whose dream was to become a
powerful one. 46
This is the climate for new projects which assume a reintroduction of hooligan nature of an act which is to toughen administered sentence. This is a return
of legal science to 19th century ideas which had remained in communist era but
were repealed later. A lowering of age limit in penal law, so that 15-year-olds
could be held responsible, is being discussed (politicians now discuss only if this
should pertain to crimes and offences or only crimes). Once again, such regulation turns us back to the 19th century and is close to regulations of Kodeks Karzcy Krlestwa Polskiego (Penal Code of the Polish Kingdom) from 1818.47 This
approach is supported by programmes such as zero tolerance in schools, announced by the minister of education in November 2006 which provides a wide
inventory of punishments for offences against school rules. Moreover, the idea to
reintroduce death penalty seems to be a boomerang idea of Polish political discourse (note that it is supported by current Polish ombudsman).
5. The practice of the judiciary changes in sentencing at the

beginning of the 21st century
One of the crowning arguments supporting the need to introduce changes to penal laws was, according to the critics, liberalisation of the judicature under the
new 1997 penal code. However, the studies comparing judicature from before
1997 and after do not show any significant changes. But one can notice an increase in sentences of imprisonment, including imprisonment with conditional
suspension of execution, and a decrease in sentences of fines and conditional release on probation. Similar conclusion can be drawn while comparing the total
number of sentences. Between 1997 and 2001 one can see an increase in sentences of unconditional imprisonment by almost 3%, yet imprisonment with conditional suspension of execution by 9,5%. This is correlated with a decrease of
46
47
M. Filar (2006), p. 198.

Patrz: J. liwowski (1958), Kodeks karzcy Krlestwa Polskiego (1818 r.). Historia jego
powstania i prba krytycznej oceny, Warszawa, p. 313-316.
387
the number of sentenced fines by 14% 48 . This in no way shows a more liberal
approach of the courts in spite of less severe laws.
It was so in the initial period shortly after the new laws came into force. Let us
now take a look at what results in the penal policy were caused by excessive interest
of politicians and how it influenced penal courts. The year 2000 and the appointment
of the new minister of justice, Lech Kaczyski, seems to be a limit here. His less
than one years term in the office brought significant toughening of penal policy.
Detailed distribution of penalties in selected years between 1993 and 2004 is
shown in table 5. It shows several tendencies in Polish penal policy. On one
hand, there is a steady increase in the number of sentences (it has tripled). The
number of persons sentenced for imprisonment also increases there is a particular surge between 2002 a 2004 of almost 140% (please note that the crime rate
in the same period stabilized or even slightly decreased).
Table 5:
Distribution of penalties sentenced by the courts in 1993, 1996, 2000

and 2004 49
sentenced persons total

fine solely
fine and imprisonment
restriction of freedom
penal measures solely
unconditional imprisonment
imprisonment w. conditional
suspension of execution
1993
1996
2000
2002
2004
171 622
36 920
85 232
6 389
29
36 954
227 731
62 082
92 263
10 612
65
31 240
222 815
33 699
65 655
14 796
75
30 687
365 326
75 698
39 156
86 143
89
35 790
513 410
111 491
111 154
71 887
2 565
48 993
91 259
123 669
143 497
214 485
278 338
The period shows a dramatic increase of imprisonment sentences with conditional suspension of execution by 300%. What also draws attention is the change
in sentences of fines and measures restricting freedom such as community service. Here, the increase is also significant. These changes do not, however, stem
from liberalisation of ge110
neral penal policy of the courts, but classification of new crimes in 2001: drunk
driving article 178a of the penal code (114.618 persons convicted under the
48
49
B. Gruszczyska, M. Marczewski (1999-2000), Nieprawomocne orzeczenia sdw w sprawach

karnych w wietle starego i nowego kodeksu karnego, Archiwum Kryminologii, t. XXV, p. 179
et. seq.; T. Szymanowski (2003), Orzecznictwo sdw karnych w pierwszych latach po reformie
prawa karnego a potrzeby jego nowelizacji, w: Czy trzeba zmienia kodeks karny?, Warszawa,
p. 21 et seq.
Dane cyt. za: J. Jakubowska-Hara, J. Kosonoga, H. Kuczyska, B. Kunicka-Michalska, C. Nowak, J. Skupiski (2006), Alternatywy pozbawienia wolnoci w polityce karnej sdw, w:
W. Czapliski: Prawo w XXI wieku , p. 283.
388
article in 2002 50 ) and possession of drugs (total 6.407 persons convicted for
drug-related crimes in 2002). Particularly in case of the former crime, unconditional imprisonment is rare and in 2002 was administered only to 1% of the convicted.
As it can be seen, among the total number of persons sentenced for imprisonment or restriction of freedom in 2002 as much as are those who drove while
intoxicated. In case of sole fines this percentage is over 55% and remains similar
in the subsequent years (2004 49,7%). The total number of fines sentenced in
2004 excluding crimes under article 178a and tax crimes was 46.039. Hence, it
is much lower than in 1996 under the previous penal code 51 .
The distribution of restrictions of freedom has also changed (table 6). While
analysing the distribution of sentenced penalties, one can see an increase in shortterm imprisonments of up to six months (by more than 14%). Yet, in absolute
figures this increase is much bigger, it reaches about 400% (from 18.000 in 1997
to 72.069 in 2002). Similarly, the absolute number of imprisonments between six
and twelve months increased by 146%. These changes can be attributed to less
frequent sentences of fines and replacing this punishment with imprisonment.
Table 6:
Distribution of imprisonment length in 1997 and 2002 52

1997
2002
2002
percent
of total
percent
of total
change from
1997 in percent
shortest penalties (up to 6 months)
12,9
27,5
+ 14,6
short penalties (6 months - 1 year)
55,3
41,7
- 13,6
medium penalties (1 year - 2 years)
27,5
26,7
- 0,8
longer penalties (2 years - 3 years)
2,4
2,4
0,0
long penalties (3 years - 10 years)
1,8
1,6
- 0,2
very long penalties (more than 10 years)
0,1
0,1
0,0
Length of sentence
At the same time, the number of the most severe punishments of 25 years and life
imprisonment rose as well. In the comparable period the increase was double
50
51
52
Data on sentences for particular crimes in 2002 quoted from: T. Szymanowski (2005), p. 75.
The reason for the departure from fines are numerous, from change in fine calculation introduced by 1997 amendment, which became more complicated and caused judges reluctance
to sentence fines, to pauperisation of the society. Still, the fines are not high, in nearly half
of the cases between 125 and 250 Euros, and in 1/3 under 125 Euros. More on this in: J. Jakubowska-Hara et. al. (2006), p. 290-97.
Quoted from: T. Szymanowski (2005), p. 66. Data for 2002 exclude drunk driving and certain drug-related crimes to allow period comparison.
389
(from 54 in 1997 to 108 in 2002), although the number of committed crimes

changed only very slightly. Such an increase is also documented by analysis of
courts sentence for murder in 1997 and 2002. In case of other serious crimes
(classified robbery or rape) there is no apparent change in sentence distribution 53 .
The tendencies discussed above cause an increase of detained persons, both
imprisoned and arrested. The first surge took place in 2000, and since then the
number has been growing (no new prisons were constructed in this period). The
number of detained persons is shown in table 7. A similar growth of prison population can be observed in other countries, particularly where penal policy was
introduced by politicians to their political programmes, eg. in the United Kingdom 54 .
Table 7:
Number of detained persons in 1993, 1996, 2000, 2003 and 2005 55

1993
1996
2000
2003
2005
2006
total
61 562
55 487
70 544
79 281
82 955
87 901
arrested
14 020
13 595
22 032
18 240
13 416
14 819
convicted
47 030
41 110
48 006
60 773
69 191
72 737
512
782
506
268
348
345
those who
have served
The data in table 7 show one irregular year, 2000, characterised by a significant
increase in the number of detained persons. This was caused, to a considerable
extent, by an increase in arrested suspects (they were as much as 31% of detained
persons in total, while in other periods this number was close to 20%). This was
due to an instruction by Chief Public Prosecutor Minister of Justice 56 to public
prosecution to apply to the courts much more often for arrest of suspects. This
policy brought a considerable growth in number of detained and later, as the
judges usually sentenced unconditional imprisonment to those who had been
previously arrested 57 .
53
54
55
56
57
T. Szymanowski (2005), p. 68.

See eg. A. Ashworth (2002): Sentencing, w: M. Maguire, R. Morgan, R. Reiner: The Oxford
Handbook of Criminology, Oxford, pp. 1076-1077.
Fixed day December 31st, since 2006 September 30; data from: J. Jakubowska-Hara et. al.
(2006), p. 282. Data for 2006 from website of Centralny Zarzd Suby Wiziennej: (Central Penitentiary Service) http://www.sw.gov.pl/images/1161167284.pdf
In Poland, the office of Chief Public Prosecutor is held by minister of justice who has hence,
majc realny wpyw na dziaania prokuratury, moe bowiem wydawa obowizujce
podlegych mu prokuratorw zarzdzenia i wytyczne.
If the court does not sentence a person who was arrested for imprisonment, the accused can
request a compensation for ungrounded detention from the Treasury.
390
Second reason for such a high prison population is the decrease in the number
of sentenced conditional releases (releases on probation). In 1997, 19.201 persons were released on probation (63,4% of all released), while in 2005 the number was 22.201 (45,8% of all released). The number of applications by prison
administration with requests for conditional release has also significantly
dropped, by almost 60%. This causes an increase in the number of persons who
leave prisons only after serving full sentence 58 .
Such circumstances cause a significant overcrowding in prisons, and, as a result, an infringement of the standards by the Council of Europe concerning minimum living space per imprisoned person. It also contributes to aggressive behaviour of the detained (towards co-detained, prison guards, and a number of selfaggression acts) and disturbs the work of the prison employees with individual
cases. All this thwarts the achievements of the Polish prison reform performed in
1990s 59 .
Such conditions in Polish penal policy can also be explained by a punitive approach of the judges. Surveyed at the turn of 1999 and 2000 in an evaluation of
some penal law institutions, they responded that temporary detention should not
be an exceptional measure (almost 1/3) and that conditions for conditional release on probation should be toughened (almost a half of respondents). Moreover, only 42,1% of the respondents decided that imprisonment is an exceptional
punishment which should be applied only to perpetrators of most serious crimes.
They also were of the opinion that prisoners are treated too leniently (over a
half), prison conditions are too comfortable (41,2%). Additionally, the judges (as
much as 88%) consider that crime risk in Poland is high or very high, and many
of them (43,5%) believe that wide application of imprisonment may cause a decrease in crime. With such attitudes, it is hard to be surprised that Polish penal
policy is punitive 60 .
58
59
60
T. Szymanowski (2005), p. 239, 279. Dane za rok 2005 z: Informacji o wykonywaniu kary
pozbawienia wolnoci i tymczasowego aresztowania za rok 2005, tabela 22, ze strony
www.czsw.gov.pl
More on this issue in: P. Moczydowski (2003), Wiziennictwo w okresie transformacji
ustrojowej w Polsce 1998-2003, w: T. Bulenda, R. Musidowski, System penitencjarny i
postpenitencjarny w Polsce, Warszawa, p. 77 et. seq.
Although it needs to be noted that other results were obtained in 2005 by Komisja Kodyfikacyjna Prawa Karnego (Penal Law Codification Commitee) asking on changes in penal law
(see above and Wrbel, 2005). It is difficult to explain such a significant difference taking
into account similar respondent groups and methodology. Change in time may provide an
explanation as Codification Committees survey came 6 years later. During that time penal
law was toughened to a significant extent and the judges accustomed to the new penal provisions. Hence perhaps their reluctance, too, changes later.
391
Yet, it needs to be said that other results were obtained in 2005 by Komisja
Kodyfikacyjna Prawa Karnego (Penal Law Codification Commission) in a survey about the needs of changes in penal law 61 . In this survey, judges were generally against changes in penal laws, including if it is 6 years later. Judges asked in
1999 or 2000 were evaluating a new penal code which was in force for a very
short time, and perhaps they were not accustomed to. What is more, over the next
few years the code was amended more than several times, and in 2005 de facto it
was an entirely different legal act. Judges could also be weary about such frequent changes in penal code as well as the political campaign in the media to
toughen the laws (as then election campaign started).
The data presented show clearly the politicisation of the problem of social
control in relation to fear of crime in Polish society. What draws attention is the
discrepancy between the general picture emerging from statistical data and the
one presented by politicians and the media (with emphasis on the latter), particularly in the moments of intensive propaganda campaigns during elections. The
fact that the crime rate in Poland, even if it does not decrease, definitely stabilizes is of no significance, neither is the crime in Poland, in comparison to other
countries in Europe, average.
Interestingly, regardless of the intensity of the fear campaign addressed to the
society, research shows that the level of fear is decreasing. This, however, does
not influence the programmes of penal law reform by the government. The current Polish government, a strong supporter of the policy of law and order, may
cause the Polish society to turn to thinking that severe repression is the only right
reaction to crime. Already at present there is no understanding for penal policy
based on measures of freedom, and there is a general opinion that such measures
are not punishments.
The result of politicisation of social control in Poland namely the increase in
punitive attitudes seems alarming as it translates to the general condition of the
Polish society.
61
See above and Wrbel (2005).
392
393
Problems Associated with Harsher Sanctioning

Trends in Returning to More Severe Punishment in Japan
TOSHIO YOSHIDA
The Japanese criminal justice system at the turn of the century began to assert
more punitiveness, one could say without exaggeration or hesitation. This tendency started even before the terrorist attacks of nine/eleven 2001 against the
World Trade Center and the Pentagon. This paper will explain and discuss the
tightening up of sanction climate in Japan.
1. Increasing punitiveness at the judicial level

One of the latest, most important law revisions is the revised criminal code of
December 1, 2004. This new code, (The law concerning the partial revision of
the Criminal Code and the other laws), came into force on January 1, 2005, and
has served to tighten up sanctions. It is important, therefore, to present and analyze this law first of all.
1.1
Contents of the revised laws
First of all, we will see the General Part of the revised Criminal Code. This new law
has raised the maximum of the determinate-term of imprisonment with or without forced labor from fifteen years to twenty years ( 12, 13 Criminal Code). In the
case of accumulation of offences or the aggravated punishment for repeated offences, imprisonment with or without forced labor for a determinate-term of thirty
years or less (hitherto twenty years or less) can be rendered ( 14 II Criminal Code).
In the case of a commutation either the death-penalty or of life-imprisonment with
or without forced labor, the penalty has been raised for the determinate-term of from
twenty years to thirty years ( 14 I Criminal Code).
Now we will view the Particular Part of the revised Criminal Code. It is not only
the increase of statutory criminal penalties, but also the criminalization of certain circumstances that concerns us in this paper. As for sexual crimes, the Criminal Code
carries more serious punishment, increasing the statutory penalties. The maximum
394
Toshio Yoshida
penalty for the law against forcible indecency and quasi forcible indecency1 has
been increased from seven years to ten years ( 176, 178 I Criminal Code). The
statutory penalty for rape or quasi-rape has been increased from two years or more to
three years or more ( 177, 178 II Criminal Code)2 . The statutory penalty for rape
resulting in death or bodily injury or quasi-rape resulting in death or bodily injury has been raised from life imprisonment with forced labor or a determinate-term
of imprisonment with forced labor for a period of three years or more to a determinate
term of imprisonment with forced labor for five years or more ( 181 II Criminal
Code). The crime type of plural (gang) rape has been newly prescribed ( 178-2
Criminal Code). The convicted will be sentenced to a determinate-term of imprisonment with forced labor for four years or more. In addition, the crime types of plural
(gang) rape resulting in death or bodily injury or quasi-plural (gang) rape resulting in death or bodily injury have been newly prescribed ( 181 Criminal Code).
The criminal will be sentenced to life imprisonment with forced labor or to a determinate-term sentence of forced labor of six years or more.
As for crimes of homicide, the minimum statutory penalty for murder or
voluntary manslaughter has been raised from the determinate-term sentence of
three years or more to five years or more ( 199 Criminal Code). 3 According to
25 Criminal Code, the execution of imprisonment for a term of from one year
or more to five years or less can be suspended, if the criminal is sentenced with
or without forced labor of three years or less or to the fine of 500,000 Yen or less.
1
178 Penal Code Quasi Forcible Indecency, Quasi Rape: (1) A person who commits an
indecent act upon a male or female by taking advantage of loss of consciousness or inability
to resist, or by causing a loss of consciousness or inability to resist, shall be punished in the
same manner as prescribed for in Article 176 forcible indecency. (2) A person who commits sexual intercourse with a female by taking advantage of a loss of consciousness or inability to resist, or by causing a loss of consciousness or inability to resist, shall be punished
in the same matter as prescribed in the preceding Article rape.
Until the Revision, it was said now and then that the minimum (i.e., not less than two years)
statutory penalty for rape was not stringent enough because the minimum statutory penalty
for robbery was not less than five years. But most criminal lawyers recognized that the latter
was too severe. The statutory penalty for robbery, therefore, should be reduced. Actually,
even the conservative draft of the revision of the Criminal Code which the Legislative
Council of the Ministry of Justice drew up in 1974 reduced the minimum statutory penalty
for robbery from five years to three years. Strangely enough, it was ignored during the discussions in the sub-committee. Besides according to 177 of the German Penal Code, the
minimum statutory penalty for obscenity by compulsion and rape is one year, and two years
even in especially serious cases.
Against the background of the increasing statutory penalties, there was an unlawful detention case where a junior school girl was kidnapped by a twenty-eight year old man on November 13th, 1990 in the harbor city of Niigata on the Sea of Japan. She was kept in a room
in his mothers house for nine years and two months. The victim was found and rescued
only on January 28th, 2000.
Problems Associated with Harsher Sanctioning in Japan
395
If the punishment for murder is, therefore, not reduced according 66 Criminal
Code, the convicted cannot be paroled. In this connection, the minimum penalty
of imprisonment with forced labor for organized murder has been raised from
five years or more to six years or more ( 3 of the Law Concerning the Punishment of Organized Crime and so on).
As for crimes of bodily injury, there have been many revisions as well. The
maximum penalty for bodily injury has been raised from ten years or a fine of
300,000 Yen to fifteen years or a fine of 500,000 yen ( 204 Criminal Code). In view
of the increase of the minimum sentences for rape, murder and voluntary manslaughter, the minimum penalty for bodily injury resulting in death has been raised from
two years or more to three years or more. In this connection, the statutory penalty for
illegal arrest or confinement resulting in death or injury has been practically revised since according to 221 Criminal Code. A person who commits a crime proscribed under the preceding Article illegal arrest or confinement and thereby causes
the death or injury of another shall be dealt with by the punishment prescribed for either the crimes of injury or the preceding Article, whichever is greater. So the maximum term of imprisonment with forced labor for bodily injury or bodily injury resulting in death has been raised respectively from ten years to fifteen years and from fifteen years to twenty years. This is true of both obstructing traffic resulting in death
or bodily injury ( 124 Criminal Code) and abortion without consent resulting in
death or bodily injury ( 216 Criminal Code). In both cases, the maximum term of
imprisonment with forced labor has been raised respectively from ten years to fifteen
years and from fifteen years to twenty years.
In view of the increase of the maximum term of imprisonment for bodily injury, the
penalty with forced labor for dangerous driving resulting in bodily injury was also
increased from ten years to fifteen years ( 208-2 Criminal Code). A person who drives
dangerously and thereby causes the death of another shall be punished more stringently. The minimum statutory penalty has been raised from the determinate-term of
imprisonment for a period of one month to one year. The maximum term of imprisonment with forced labor has been raised from fifteen years to twenty years. The maximum term of the imprisonment with forced labor for especially serious cases of bodily
injury and habitual bodily injury has been raised from ten years to fifteen years ( 12, 1-3 of the Law concerning Punishment of Physical Violence and Others).
In crimes against personal freedom, the maximum statutory penalty for illegal
arrest and confinement has been raised from five years to seven years ( 220
Criminal Code)4 .
4
During the discussions in the sub-committee, it was stressed that the value of life was upgraded still higher than before. With regard to it, the minimum statutory penalty for murder
and voluntary manslaughter should be raised. It was mentioned at the same time that the
396
Toshio Yoshida
These rearrangements are tabulated as follows (table 1):

Table 1:
Increase of statutory penalties through the latest criminal law legislation

Offences
Statutory penalties
before the revision
Imprisonment with forced labor for
from six months to seven years or
less
two years or more
two years or more (hitherto rape or
Quasi-rape applied)
Life imprisonment with forced labor
or imprisonment with forced labor for
three years or more
Statutory penalties after the revision

from six months to ten years or less
(in force, January 2005)
Imprisonment with forced labor for three
years or more (in force, January 2005)
four years or more (in force, January
2005)
Life imprisonment with forced labor or
imprisonment with forced labor for five
years or more (in force, January 2005
)
Life imprisonment with forced labor Life imprisonment with forced labor or
or imprisonment with forced labor for imprisonment with forced labor for six
three years or more (hitherto rape
resulting in death or bodily injury or
quasi-rape resulting in death or bodily
injury applied)
Murder, voluntary manslaugh- Death penalty, life imprisonment with Death penalty, life imprisonment with
ter ( 199 Penal Code)
forced labor or imprisonment for
forced labor or imprisonment for five
three years or more
Organized murder ( 3 of law Death penalty, life imprisonment with Death penalty, life imprisonment with
concerning punishment of organ- forced labor or imprisonment with
forced labor or imprisonment with
ized Crime and regulation of
forced labor for five years or more forced labor for six years or more (in
embezzlement of Excess Profits
force, January 2005)
of August 1999)
Bodily injury ( 204 Penal Code) Imprisonment with forced labor for Imprisonment with forced labor for
ten years or less or fine of 300,000 fifteen years or less or fine of 500,000
Yen or less (in force, January 2005)
Yen or less
Aggravated bodily injury and Imprisonment with forced labor from Imprisonment with forced labor from
habitual bodily injury (1-2, 1- one year to ten years or less
one month to fifteen years or less (in
3 of law concerning Punishment
force, January 2005)
of Physical Violence and Others)
Bodily injury resulting in death Imprisonment with forced labor for Imprisonment with forced labor for three
( 205 Penal Code)
two years or more
Dangerous driving resulting in Imprisonment with forced labor for Imprisonment with forced labor for
bodily injury ( 208-2 Penal
from one month to ten years or less from one month to fifteen years (in
Code. A newly prescribed proviforce, January 2005)
sion in 2001)
Forcible indecency, Quasi
forcible indecency ( 176, 178 I
Penal Code)
Rape, Quasi- rape ( 177, 178
II Penal Code)
Gang rape or Quasi-gang rape
( 178-2 Penal Code) (new provision)
Rape resulting in death or
bodily injury or Quasi-rape
resulting in death or bodily
injury ( 181 II Penal Code)
Gang rape resulting in death or
bodily injury or quasi-rape
resulting in death or bodily
injury ( 181 III Penal Code)
(new provision)
minimum statutory penalty for rape should be raised. The value of life is higher than the
value of sexual self-determination. Therefore, murder or voluntary manslaughter should be
punished more stringently. But it should be reconfirmed here that each statutory penalty
should be compared with one another, not emotionally, but reasonably in the whole sanction
system. On top of that, it should be stressed that murder or voluntary manslaughter can be
committed with various motives. In order that human beings can deal with his crime appropriately, a murderer should be able to be sentenced to imprisonment with suspension of sentence, even if one does not use a reduction of punishment according to extenuating circumstances.
Offences
Statutory penalties
before the revision
Dangerous driving resulting in
death ( 208-2 Penal Code. A newly from one month to fifteen years
prescribed provision in 2001)
Robbery resulting in death, rob- Life imprisonment with forced labor
bery resulting in bodily
or imprisonment for seven years or
injury ( 240 1. Alt. Penal Code) more (suspended sentence not
possible)
Illegal arrest and confinement
( 220 Penal Code)

from three months to five years
Kidnapping or abduction of a
minor ( 224 Penal Code)

from three months to five years or
less
Body trade ( 226-2 Penal Code)

(a newly prescribed provision in
2006 )
Transporting a kidnapped, abducted
person to a foreign country ( 226-3
Penal Code) (a newly prescribed
provision in 2006)
Accidental homicide or acciden- Imprisonment with forced labor or imprisonment without forced labor for
tal infliction of injury while
driving an automobile or oper- from 1 month to five years or fine of
ating a motorbike ( 211 II
1,000,000 Yen or less (hitherto profesPenal Code) (a newly prescribed sional accidental homicide or profesprovision)
sional accidental infliction of injury
applied)
397
Statutory penalties after the revision

Imprisonment with forced labor for one
year or more (in force, January 2005)
Life imprisonment with forced labor or
imprisonment with forced labor for six
years or more (so now via reduction of
punishment according to extenuating
circumstances suspended sentence
possible) (in force, January 2005)
from three months to seven years (in
force, July 2006)
from three months to seven years (in
force, July 2006)
from three months to ten years (in
force, July 2006)
Imprisonment with forced labor for two
years or more (in force, July 2006)
Imprisonment with forced labor or
imprisonment without forced labor for
from one month to seven years or fine
of 1,000,000 Yen or less (in force, June
2007)
As for the field of the criminal procedure, the period of prescription has been
lengthened. In crimes whose statutory penalty contains the death penalty it has
been lengthened from fifteen years to twenty five years. In crimes in which life
imprisonment with or without forced labor is prescribed, it has been lengthened
from ten years to fifteen years. In crimes whose maximum term of imprisonment
with or without forced labor is fifteen years or more, the period of prescription
has been raised from five years to ten years ( 250 of the Code of Criminal Procedure).
1.2
Process of Revision Work
Already in December 2003 before the revised Criminal Code was finished, the
conference of ministers combating crime had drawn up an action plan for realizing society ready for crime. One of its priority programs was to take necessary measures in order the re-establishment internal security, for example, by increasing criminal sanctions.
398
Toshio Yoshida
Another group, the project team of the parties in power (the Liberal Democratic Party and the Clean Party) was named Women and Criminal Law. It demanded that the Minister of Justice increase the statutory penalty for rape, prescribe newly plural (gang) rape and prolong the period of prescription. In fact,
sexual crimes have become increasingly the focus of public and political attention. Against this background, the Minister of Justice submitted some questions
to an advisory body of the Legislative Council of the Ministry of Justice on
February 10, 2004: In view of the state of violent crimes and serious crimes for
the last several years, it is necessary to establish formal and substantial revisions
to the criminal law so completely that it can deal with such crimes appropriately.
So the minister wishes that the Legislative Council give its view on the following
main points: y revision of the maximum determinate-term of imprisonment with
or without forced labor, y revision of the statutory penalty for forcible indecency,
rape and rape resulting in death or bodily injury, y revision of the statutory penalty for murder and voluntary manslaughter, y revision of the statutory penalty
for bodily injury and bodily injury resulting in death, and y revision of the period
of prescription. It follows that the Minister of Justice focused on increasing the
statutory penalty for sexual crimes which tended to provoke ungoverned lust for
punishment. Corresponding, the statutory penalties for other crimes, for example,
murder, voluntary manslaughter, bodily injury and so on were also raised.
A proposal of the report was discussed and drafted in the sub-committee Violent Crimes and other Serious Crimes with thirty-one members. The sub-committee was part of the Division of the Criminal Law of the Legislative Council
of the Minister of Justice. This council eventually submitted a report in response
to the ministers inquiry on September 8, 2004. The Diet passed amendments
without modifications on December 1, 2004.
1.3
Critical reflections on the revision
Generally speaking, revisions to the penal code are for the good, but these revisions
are too harsh and illiberal. It matters from which point of view and in which direction
you should proceed. In this sense, these revised laws have to be criticized fundamentally. In the following, some points at issue are discussed in rough outline.
First, the Criminal Affairs Bureau of the Ministry of Justice (in the following abbreviated as Ministry of Justice) explained that on the basis of criminal
statistics, cruel crimes and other serious crimes have increased in the last several
years. Therefore, the criminal law should be tougher, resulting in criminals being
punished much severely. However, it is difficult to follow this argument. Criminal statistics certainly show that reported criminal cases, registered by police, for
399
example rape cases, have increased considerably in the last several years. While
there have been several short, small-scale studies on the decrease and increase of
criminality, these offer insufficient support for such a seeping generalization. One
should remember that no reliable statement has been made of how criminality is
related to criminal statistics.
Criminal statistics do not mirror the real changes in crime, but only the activities of police. The Ministry of Justice stressed that cases of robbery have increased in the last several years. It could not be, however, excluded that the police has applied robbery to such cases which have hitherto been considered extortion, against the background of strengthened public sensitization. In addition,
robbery is not a typical violent crime, but in fact a property crime. Therefore,
robbery was not an area of concentration for the revision. So one cannot understand why an increase of robbery cases should lead to increasing the statutory
penalty for violent crime in general. Besides, it is strange that no criminological
studies of violent crime based on empirical researches were taken into account
during the discussions of the sub-committee.
Second, the Ministry of Justice asserted the unreliable statement that the longer
the term of imprisonment, the more deterrent it works. In the process of the subcommittees discussions, nobody seemed to have any doubts about this statement although no criminologists would believe it. Today criminological experts agree that the
effect of deterrence depends on much more of the magnitude of the risk of being arrested than of the severity of the statutory penalty. It follows from this point of view
that intensifying investigation or making prosecution more effective is much more
important than increasing the statutory penalty to deter potential criminals. Even if the
statement of the Ministry of Justice were valid, it should be taken into account that the
maximum statutory penalty has been applied to only a few criminal cases up to now in
the practice of criminal courts.
Third, in addition to the alleged increase of the deterrent effect through more
severe punishment, great emphasis was placed on the lust for punishment (or
retribution) of actual and potential crime victims, the feelings of justice and
security and the norms of public consciousness during the discussions in the
sub-committee. The Ministry of Justice explained that the citizens conscious
norm demanded crime prevention through much more severe penalties. According to its explanation, this prevention meant the so-called positive general prevention. This satisfied the publics sense of justice as well. However, this explanation distorts the concept of positive general prevention from its original
meaning and has come to mean instead negative general prevention. In addition,
the very fact itself that you feel unsafe in day-to-day life cannot be given as a
400
Toshio Yoshida
reason for increasing the statutory penalty. First of all, it should be made clear
whether the perceived feeling of uncertainty corresponds to the actual high
level of criminality. Without analytical and reasonable studies about heightened
level of crime, criminal law would probably fall into the criminal law of feelings which modern criminal law has so far managed to avoid.
Fourth, the Ministry of Justice has found that it may increase the statutory
penalty with the explanation that the average lifespan of Japanese people is getting longer and longer. A large-scale association for crime victims, the association for tomorrow, had demanded strengthening criminal sanctions, and that in
view of the prolonging of the average longevity, much more severe penalties
have to be inflicted, otherwise the deterrent effect of punishment would be reduced. It should be countered, however, that behavioral freedom is upgraded
more and more in modern society. On top of that, it should be pointed to that the
longer imprisonment is inflicted, the more difficult is the rehabilitation. Therefore, criminal legislation has to aim at reducing the term of sentence of imprisonment.
Fifth, ones attention should also be called to the fact that release on parole has
not been easily permitted. According to 28 Criminal Code, When a person
sentenced to imprisonment with or without labor evinces signs of substantial reformation, the person may be paroled by a disposition of a government agency
after that person has served one-third of the definite term sentenced or 10 years
in the case of a life imprisonment. Then both the person who is sentenced to
imprisonment for a term of thirty years in the case of accumulation of offences
and the person who is sentenced to life imprisonment can be released on parole
after serving ten years. This is because the maximum determinate-term of imprisonment has been increased from twenty years to thirty years. It means actually that the maximum determinate-term of imprisonment, that is, thirty years is
too severe. According to the law before the revision, the maximum determinateterm of imprisonment was twenty years. A person who was sentenced to imprisonment for twenty years could be released on parole after having served six years
and eight months.
Sixth, it should be pointed out that prisons in Japan are already overcrowded (see
2. Increasing punitiveness at the judicial level). One must expect that these prisons
will have to accommodate more and more convicts held for longer terms of imprisonment. This will provide fewer and fewer chances for prisoner rehabilitation. The
overcrowding will make rehabilitation even less effective. The Ministry of Justice has
declared that the aim of rehabilitation will be maintained along with increasing the
term of punishment. But experts have pointed out that these two aims conflict. The
401
much-more-severe-penalty policy of criminal justice adopted along with the revised

criminal law will distress the aim of rehabilitation and push it farther away still.
Seventh, you should be aware that the maximum determinate-term of imprisonment with or without forced labor has been increased and that life imprisonment with
or without forced labor has been maintained. You may think that in general there is a
qualitative difference between determinate-term imprisonment and life imprisonment.
If so, it follows that the latter can be executed literally life-long without release on
parole, and that it does not make any difference whether the prisoner repents of his
past crimes or is diagnosed as not dangerous. Life-long custody without any chance of
being released on parole takes away from the prisoner sentenced to life imprisonment
any hope he may have. As a result it would be another type of severe punishment in
addition to capital punishment.
Eighth, it should be stressed that legally protected interests, for example, life, integrity of body, freedom of behavior and so on should establish a standard when increasing the minimum statutory penalty for each crime type. At the same time, it is important that the statutory penalty for each crime type should be compared with one another. But this point was not systematically discussed during the discussions of the
sub-committee.
Ninth, you should draw your attention to the fact that partial revision of the Penal
Code has not been restricted to selected crime types. For example, according to 197
I Criminal Code A public officer or arbitrator who accepts, solicits or promises to
accept a bribe in connection with his/her duties shall be punished by imprisonment
with labor for not more than 5 years receipt of bribe; and when the official agrees to
perform an act in response to a request, imprisonment with labor for not more than 7
years shall be imposed receipt of bribe in response to an entreaty. According to
197-3 I Criminal Code, When a public officer commits a crime proscribed under
either of the preceding two Articles and consequently acts illegally or refrains from
acting in the exercise of his or her duty, imprisonment with labor for a definite term of
not less than 1 year shall be imposed aggravated acceptance. Consequentially, it
follows that the maximum determinate-term sentence can reach twenty years in aggravated circumstances. In comparison, the penalty for receipt of a bribe in response
to an entreaty is disproportionately severe. Let us show another example. According to
164, A person who, for the purpose of uttering, counterfeits the Imperial Seal, State
Seal or Imperial Signature, shall be punished by imprisonment with labor for a definite term of not less than 2 years counterfeiting of the imperial seal. The maximum
determinate-term of imprisonment for counterfeiting the imperial seal is twenty years.
But on the contrary, according to 165 Criminal Code, A person who, for the purpose of unauthorized use, counterfeits the seal or signature of a public office or public
402
Toshio Yoshida
officer shall be punished by imprisonment with labor for not less than 3 months, but
not more than 5 years counterfeiting of official seals. You can realize that there is a
big difference between the maximum determinate-term of imprisonment for counterfeiting the imperial seal and that for counterfeiting of official government seals. How
is it possible to explain and find appropriate the big difference in a free, democratic,
constitutional state?
As far as the period of prescription is concerned, the Ministry of Justice has given
as reasons or justifications the following points: the publics lust for punishment, the
Japanese peoples prolonged average longevity and the convicts deserving a more
severe penalty. The Ministry of Justice did not sufficiently explain the reason why the
period of prescription should be lengthened at a stroke by ten years from fifteen years
to twenty-five years. It should be stressed that the sub-committees discussions did not
deliberate sufficiently on the negative side of this prolongation, such as the difficulty
of the suspects producing counterevidence.
Finally, one more thing should be said. The work of revision of the Criminal Code
was discussed mainly in the sub-committee on Violent Crimes and other Serious
Crimes of the Legislative Council of the Ministry of Justice. But the subcommittee met only five times from April 19, 2000 to July 2004. On top of that, it
should be mentioned in particular that indeed ten criminal law professors were part of
the sub-committee, but they rarely spoke during the meetings. Since they hardly opposed the revision of either the Criminal Code or the Criminal Procedural Code, they
seem to have believed in the affirmative effect of deterrence by severe punishment,
retribution and vengeance that stretches as the central thread through the whole Revision-Draft.5 They never seemed to doubt the severe crime policy, such as to sentence
many more criminals to much longer imprisonment and to understand them less. One
might be afraid that the science of criminal law is falling on hard times, rather than
being argued social-scientifically, and founded informatively, rationally, and based on
liberal democratic ideas and solidarity of mind.
5
The members of the sub-committee from universities were: IDA Makoto (Keio-University),
Chair person OHYA Minoru (Doshisha-University. Ex-chairperson of the Japan Society of
Victimology), KAWABATA Hiroshi (Meiji-University), KIMURA Mitsue (TokyotoritsuUniversity), Deputy-Chairperson SHIBAHARA Kuniji (Gakushuin-University. Ex-Chairperson of the Japan Society of Criminal Law), TAGUCHI Morikazu (Waseda-University),
NAGANUMA Noriyoshi (Jochi-University), NAKAMORI Yoshihiko (Kyoto-University),
YAMAGUCHI Atsushi (Tokyo-University) und YAMANAKA Keiichi (Kansai-University).
They did not represent, however, the Japan Society of Criminal Law officially. It is also
conspicuous that neither criminologists nor penologists did belong to the sub-committee. Indeed, when the proceedings were released, each speaker was anonymous, although the
names of the members were all listed. Therefore, one does not know who said what. One
can only presume who said what; one can only presume. Whatever the circumstances were,
the anonymity cannot be justified.
1.4
403
Other amendments
A new provision, 208-2 dangerous driving resulting in death or injury has

been added to the Criminal Code in 2001 against the background of the activities
of the groups helping crime victims. These groups demanded among other things
a more-balanced proportionality between crime act and punishment, which
sounds like a good idea in theory, but it is a question that has never been thoroughly studied. They also demanded more consideration of the feelings of victims and much more effective deterrence, the last of which is here considered
regressive. That measure pleaded for increasing the term of imprisonment for
traffic violators causing another persons death or injury (see table 1 above). In
the meantime, however, they complained that it was very difficult to prove such
elements of the offence as under the influence of alcohol or drugs making it difficult for the person to drive safely ( 208-2 I) or deliberately ignores a red signal light or its equivalent at a speed that will cause serious danger to traffic and
so on. This new provision, therefore, will not be able to be easily applied to crime
cases which are actually committed. So a new provision 211 II accidental homicide or accidental infliction of injury while driving an automobile or operating
a motorbike has been added to the Criminal Code. A person who commits such
a crime shall be punished by imprisonment with or without forced labor for not
less than one month and not more than seven years or by a fine of not more
1,000,000 yen (see table 1 above).
In July 2007, the Law concerning the revision of the Code of Criminal Procedure
for improving the legal status of crime victims was passed. Its introduction dealt with
both adhesion procedures and incidental actions. According to the revised Code of
Criminal Procedure, the victim can ask the defendant questions and demand a sentence at the end of trial independent of the prosecutor. For example, he can demand
the most severe punishment, or even capital punishment. This institution of this new
law institution is questionable. It matters whether and how severely the defendant
should be punished from the point of the reasonable person thinking legally. Both
the judiciary and not a few criminal law scholars are regrettably still of the opinion
that retribution and deterrence have an important role to play. They say that retribution
does not mean the lust for punishment on the part of the victims, but instead on the
part of the public, given that retribution is acknowledged in the criminal law. By establishing the victims right to demand a punishment, this new law will unavoidably
bring one-sided private vengeance into criminal procedure. It will mean that indeed
private vengeance is identified with public retribution. It will make it difficult for
criminal judges to keep their appropriate distance from the feelings of victims. This is
virtually a beginning of irrational criminal law.
1,072,712
1,099,091
1,075,804
1,090,149
986,305
966,530
923,769
876,363
836,777
781,896
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
11
14
134
115
117
82
68
59
48
45
32
34
85,020 (51,446: 60.5%)
85,815 (52,856: 61.6%)
84,900 (52,772: 62.2%)
80,201 (49,250: 61.4%)
75,582 (46,523: 61.6%)
73,184 (45,117: 61.6%)
67,067 (42,039: 62.7%)
63,576 (40,034: 63.0%)
61,886 (38,706: 62.5%)
59,773 (36,980: 61.9%)
Imprisonment with forced labor

(with suspension)
Source: The Whitepaper on Japanese criminality, 2006, p. 50.
D: Death penalty. L: Life imprisonment with forced labor.
Total
Rate of finally decided criminal defendants (1996-2005)
Year
Table 2:
3,904 (3,655: 93.6%)
4,215 (4,001: 94.9%)
4,017 (3,763: 93.7%)
3,510 (3,277: 93.4%)
3,003 (2,805: 93.4%)
2,887 (2,708: 93.8%)
2,613 (2,464: 94.3%)
2,530 (2,251: 95.8%)
2,321 (2,181: 94.0%)
2,446 (2,289: 93.6%)
Imprisonment without
forced labor
(with suspension)
689,972
743,553
784,515
837,144
884,088
906,947
1,016,822
1,006,000
1,030,612
1,005,684
Fine
3,014
2,829
26
2,774
2,752
3,713
3,141
3,514
3,757
4,167
4,708
Minor fine
51
38
77
71
81
81
69
69
64
Penal
detention
404
405
2. Increasing punitiveness at the judicial level

When one considers the punitiveness in the practice of sanction, it is first of all
important, to see the figures for criminal defendants having finished appeal 6 . As
the statistics show (table 2), their number sank dramatically from 1,072,712 in
1996 to 781,896 in 2005, a reduction of 27.1% in Japan. The percentage of
criminals sentenced to imprisonment with forced labor whose sentences were
suspended stayed at almost the same level from 1996 to 2005: 61.9% in 1996 and
60.5% in 2005. It seems that Japanese criminal justice has nothing to do with the
global phenomenon of increasing punitiveness. When one sees, however, the
statistics in detail, one gets a different impression. The proportion of convicts
finished with appeal who were sentenced to pay a fine sank from 1,005,684 in
1996 to 689,972 in 2005 by 31.4%. On the contrary, the proportion of such convicts who were sentenced imprisonment with forced labor or imprisonment without forced labor increased from 62,256 in 1996 to 89,069 in 2005 by 43.1%. Furthermore, the proportion of payable fines among all criminal penalties reduced
markedly. The rate of payable fines was 93.6% in 1996, but it sank to 88.2% in
2005. It means that the proportion of fines reduced in favor of imprisonment with
forced labor and imprisonment without forced labor. When one reckons that
those convicts finished with appeal who were sentenced to imprisonment with
forced labor or imprisonment without forced labor without a suspension of sentence, it turns out that the proportion increased from 22,950 in 1996 to 33,823 by
10,873 (47.4%). Plus, capital punishment candidates have increased. The number
of criminal defendants finished with appeal who were sentenced to life imprisonment with forced labor has increased from 34 in 1996 to 134 in 2005 by 100
(294.1%). All this points to an increasing punitiveness on the part of the Japanese criminal court.
For sentencing, the gravity of the criminal act and the criminal record are of course important considerations. But, the defendants personality should also be considered. For example,
is he a kind person or prone to violence? However, it may be difficult to get such information. All three considerations should play a role in sentencing. Plus, an actual increase in the
gravity of the criminal act is sufficient cause to increase the penalty.
406
2.1
Toshio Yoshida
Murder and voluntary manslaughter
As follows, the term of sentences should be considered according to the particular type of criminal act. It is a convincing indicator of a possible increasing punitiveness of the criminal court independent of the increasing crime rate. As table 3 shows, from 1996 to 2002, except for 1999 and 2001, prisoners sentenced
to imprisonment with forced labor for from seven years to ten years or less assumed the largest share. Since 2003, however, prisoners sentenced to imprisonment with forced labor for from ten years to fifteen years or less have taken the
biggest portion (23.9% in 2003, 22.3% in 2004 and 23.2% in 2005).
In 1996, the portion of sentences of ten years and less amounted to 78.4%. In
the same year, the portion of sentences of from fifteen years to twenty years or
less took 16.5%. In 2005, however, the corresponding portion took respectively
62.9% and 31.5%. Prisoners sentenced to life imprisonment with forced labor
have increased from 14 (2.9%) in 1996 to 38 (5.7%) in 2005. Therefore, one
finds that the proportion of longer imprisonments has increased more than short
imprisonments.
Table 3:
Year
Total
1996
483
1997
497
1998
552
1999
528
2000
558
2001
583
2002
600
2003
635
2004
658
2005
665
Different terms of sentences for newly imprisoned with forced labor

(1996-2005)
-1
year
1
0.2%
-2
years
12
2.5%
7
1.4%
9
1.6%
5
0.9%
8
1.4%
3
5.7%
6
1.0%
7
1.1%
12
1.8%
2
0.3%
-3
years
57
11.8%
45
9.1%
53
9.6%
48
9.1%
49
8.8%
48
8.7%
54
9.0%
67
10.6%
49
7.4%
39
5.9%
-5
years
96
19.9%
118
23.7%
128
23.2%
140
26.5%
131
23.5%
153
27.7%
133
22.2%
147
23.1%
135
20.5%
128
19.2%
-7
years
93
19.3%
80
16.1%
96
17.4%
86
16.3%
87
15.6%
90
16.3%
85
14.2%
83
13.1%
111
16.9%
107
16.1%
-10
Years
131
27.1%
138
27.8%
143
25.9%
118
22.3%
139
24.9%
140
25.3%
147
24.5%
125
19.7%
144
21.9%
142
21.4%
D: Death penalty. L: life imprisonment with forced labor

Source: The Yearbook of Correctional Statistics, 1996-2005.
-15
years
62
12.8%
80
16.1%
85
15.4%
96
18.2%
90
16.1%
105
19.0%
116
19.3%
152
23.9%
147
22.3%
154
23.2%
-20
years
18
3.7%
14
2.8%
24
4.3%
16
3.0%
25
4.5%
25
4.5%
38
6.3%
31
4.9%
31
4.7%
55
8.3%
>20
years
1
0.2%
L
14
2.9%
12
2.4%
10
1.8%
16
3.0%
24
4.3%
15
2.7%
19
3.2%
23
3.6%
30
4.6%
38
5.7%
3
0.6%
4
0.7%
3
0.6%
3
0.5%
2
0.4%
2
0.3%
1
0.2%
Problems Associated with Harsher Sanctioning: Japan
2.2
407
Murder on the occasion of robbery, bodily injury on the occasion

of robbery, robbery resulting in death and robbery resulting in
bodily injury
According to table 4, the number of prisoners sentenced to imprisonment with

forced labor or to the death penalty almost steadily increased (from 314 in 1996
to 827 in 2005). In 1996, prisoners sentenced to imprisonment with forced labor
whose sentences were stayed from three to seven years had the biggest portion
(72.6%). At the extreme, prisoners sentenced to imprisonment with forced labor
whose sentences were from seven to twenty years had only 17.5%. Since that
time, the latter has continually increased. Ten years later, in 2005, the corresponding numbers take respectively 56.6% and 31.8%. It is also noticeable that
not only the number of prisoners sentenced to life imprisonment, but also the
proportion has increased. The number of prisoners sentenced to life imprisonment has increased from 20 in 1996 to 84 in 2005 (320%). This shows that the
criminal courts tend to impose more and longer imprisonments on criminals than
before.
Table 4:
Year
Total
1996
314
1997
336
1998
397
1999
474
2000
607
2001
587
2002
727
2003
810
2004
883
2005
827

(1996-2005)
-6
months
-1
year
1
0.3%
1
0.3%
1
0.3%
-2
years
2
0.6%
2
0.5%
1
0.2%
3
0.5%
4
0.5%
3
0.3%
-3
years
2
0.6%
7
2.1%
8
2.0%
4
0.8%
10
1.6%
5
0.9%
3
0.4%
9
1.1%
3
0.3%
12
1.5%
-5
years
143
45.5%
184
54.8%
163
41.1%
218
46.0%
283
46.6%
265
45.1%
269
37.0%
292
36.0%
278
31.5%
267
32.3%
-7
years
85
27.1%
73
21.7%
105
26.4%
139
29.3%
161
26.5%
138
23.5%
227
31.2%
220
27.2%
271
30.7%
201
24.3%
D: Death penalty. L: life imprisonment with forced labor.

-10
years
39
12.4%
35
10.4%
67
16.9%
66
13.9%
85
14.0%
80
13.6%
124
17.1%
115
14.2%
169
19.1%
168
20.3%
-15
years
16
5.1%
18
5.4%
14
3.5%
19
4.0%
28
4.6%
45
7.7%
46
6.3%
72
8.9%
70
7.9%
80
9.7%
-20
years
1
0.3%
1
0.3%
1
0.2%
7
1.2%
5
0.9%
8
1.1%
17
2.1%
11
1.2%
15
1.8%
20
6.4%
16
4.8%
34
8.6%
25
5.3%
33
5.4%
45
7.7%
50
6.9%
81
10.0%
78
8.8%
84
10.2
6
1.9%
1
0.3%
2
0.5%
2
0.4%
1
0.1%
1
0.1%
1
0.1%
408
2.3
Toshio Yoshida
Rape and rape resulting in death or bodily injury
As table 5 shows, the number of prisoners sentenced to imprisonment with forced

labor has increased from 367 in 1996 to 533 in 2005, 46% more. Prisoners sentenced to imprisonment with forced labor for from three years to five years have
always taken the biggest portion. The proportion was staggering between 37.0%
in 1998 and 44.6% in 2003. It is also remarkable that the number of prisoners
sentenced to imprisonment with forced labor whose terms were from five years
to twenty years increased from 54 (14.6%) in 1996 to 185 (33.8%) in 2005.
Tables 3, 4 and 5 show that persons who committed such a serious crime as
murder, rape have been given still longer terms of imprisonment.
Table 5:
Year
Total
1996
367
1997
432
1998
441
1999
494
2000
502
2001
525
2002
507
2003
583
2004
539
2005
538

(1996-2005)
-6
Months
-1
year
1
0.2%
2
0.5%
1
0.2%
1
0.2%
-2
years
45
12.3%
50
11.6%
61
13.9%
47
9.5%
52
10.4%
42
8.0%
29
5.7%
29
5.0%
22
4.1%
24
4.5%
-3
years
119
32.4%
143
33.1%
172
39.0%
163
33.0%
161
32.1%
166
31.6%
157
31.0%
154
26.4%
123
22.8%
90
16.9%
-5
years
149
40.6%
177
40.1%
163
37.0%
210
42.5%
195
38.9%
214
40.8%
207
40.8%
260
44.6%
220
40.9%
232
43.5%
-7
years
32
8.7%
46
10.6%
30
6.8%
50
10.1%
60
12.0%
62
11.8%
70
13.8%
73
12.5%
80
14.8%
91
17.1%
-10
years
20
5.4%
8
1.9%
10
2.3%
20
4.0%
25
5.0%
27
5.1%
27
5.3%
45
7.7%
54
10.0%
53
9.9%
-15
years
2
0.5%
7
1.6%
2
0.5%
4
0.8%
7
1.4%
10
1.9%
12
2.4%
15
2.6%
27
5.0%
31
5.8%
-20
years
1
0.0%
2
0.4%
2
0.4%
4
0.8%
7
0.1%
12
2.2%
10
1.0%
1
0.2%
1
0.2%
2
0.4%
L: life imprisonment with forced labor.

2.4
Death or bodily injury resulting from professional negligence
Traffic accidents are the most common criminal cases in which the offence,
Death or bodily injury resulting from professional negligence, is applied. According to table 6, the number of prisoners sentenced to imprisonment with
forced labor has increased from 506 in 1996 to 685 in 2005, by 179 (up 35.4%).
409
Those sentenced to imprisonment with forced labor for two years or less have
always taken the biggest portion. But now there are fewer and fewer of them.
Their proportion decreased from 95.5% in 1996 to 70.7% in 2005. Inversely, the
proportion of prisoners sentenced to imprisonment with forced labor for two
years or more increased from 4.5% in 1996 to 29.3% in 2005. It is also striking
that the number of prisoners increased from 524 in 2000 to 716 in 2001, by 192
(up 36.6%). It reached its peak in 2002 and has continually decreased since that
time. This is due to a new provision Dangerous driving resulting in death or
bodily injury, which was prescribed in 2001.
Table 6:

(1996-2005)
Year
Total
1996
506
1997
432
1998
486
1999
533
2000
524
2001
716
2002
772
2003
730
2004
699
2005
685
-3
months
4
0.8%
1
0.2%
1
0.2%
8
0.2%
6
1.1%
10
1.4%
4
0.5%
1
0.1%
2
0.3%
-6
months
43
8.5%
44
10.2%
51
10.5%
57
10.7%
47
9.0%
85
11.9%
62
8.0%
59
8.1%
57
8.2%
53
7.7%
-1
year
215 42.5%
176
40.7%
204
42.0%
197
37.0%
201
38.4
263
36.7%
247
32.0%
229
31.4%
205
29.3%
214
31.2%
-2
years
221
43.7%
188
43.5%
209
43.0%
238
44.7%
241
46.0
283
39.5%
299
38.7%
269
36.8%
252
36.1%
215
31.4%
-3
years
20
4.0%
22
5.1%
19
3.9%
31
5.8%
29
5.5%
50
7.0%
120
15.5%
127
17.4%
144
20.6%
150
21.9%
-5
years
3
0.6%
1
0.2%
2
0.4%
2
0.4%
8
1.5%
21
2.9%
35
4.5%
40
5.5%
37
5.3%
46
6.7%
-7
years
-10
years
2
0.4%
3
0.4%
4
0.5%
4
0.5%
2
0.3%
5
0.1%
1
0.1%
1
0.1%
2
0.3%
1
0.1%
Table 7 shows how long criminal defendants with appeal finished were sentenced
to imprisonment without forced labor. Until 2000, prisoners sentenced to imprisonment without forced labor for one year or less took the biggest portion. Since
2001, the number of prisoners sentenced to imprisonment without forced labor
for a period of one year to three years has the biggest portion. It is also noticeable
that the number of prisoners sentenced to imprisonment without forced labor for
two years or more has increased from 1.3% in 1996 to 15.4% in 2005.
410
Toshio Yoshida
Table 7:
Different terms of sentences for newly imprisoned without forced labor

(1996-2005)
Year
Total
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
154
130
104
153
169
191
223
231
210
240
-3
months
1 (1.0%)
1 (0.5%)
-6
months
1 (0.6%)
2 (1.5%)
3 (2.9%)
2 (1.3%)
6 (3.6%)
5 (2.6%)
4 (1.8%)
9 (3.9%)
1 (0.5%)
8 (3.3%)
-1
year
80 (51.9%)
69 (53.1%)
55 (52.9%)
82 (53.6%)
86 (50.9%)
80 (41.9%)
87 (39.0%)
58 (25.1%)
59 (28.1%)
52 (21.7%)
-2
years
71 (46.1%)
55 (42.3%9
43 (41.3%)
66 (43.1%)
69 (40.8%)
98 (51.3%)
110 (49.3%)
124 (53.7%)
98 (46.7%)
143 (59.6%)
-3
years
2 (1.3%)
3 (2.3%)
2 (1.9%)
3 (2.0%)
7 (4.1%)
5 (2.6%)
19 (8.5%)
88 (38.1%)
48 (22.9%)
30 (12.5%)
-5
years
1 (0.8%)
1 (0.6%)
2 (1.0%)
3 (1.3%)
2 (0.9%)
1 (0.5%)
7 (2.9%)
2.5
Dangerous driving resulting in death or bodily injury
According to table 8, the proportion of prisoners sentenced to imprisonment with

forced labor for two years or more was 46.2% in 2002, 63.4% in 2003 and 59.5%
in 2005.
Table 8:
Different terms of sentences for newly imprisoned (2002-2005)
Year
Total
2002
13
2003
93
2004
86
2005
74
-6
months
-1
year
6
6.5%
1
1.2%
1
1.4%
-2
years
7
53.8%
28
30.1%
21
24.4%
29
39.2%
-3
years
3
2.3%
15
16.1%
17
19.8%
9
12.2%
-5
years
2
15.4%
28
30.1%
21
24.4%
22
29.7%
-7
years
12
12.9%
13
15.1%
11
14.9%
-10
years
1
7.7%
1
1.1%
5
5.8%
2
2.7%
-15
years
3
3.2%
2.6
Larceny
According to table 9, the number of prisoners sentenced to imprisonment with

forced labor has increased from 6,139 in 1996 to 9,798 in 2005 by 3659 (up
59.6%). It can be seen that criminal defendants with appeal finished have been
sentenced to still longer imprisonment with forced labor as well. The proportion
of prisoners sentenced to imprisonment with forced labor for term of two years or
more has increased from 32.4% (1,974) in 1996 to 37.0% (3,232) in 2005. In
2002, the proportion reached a peak (39.4%: 3,232). Therefore, one sees that
411
punitiveness at the judicial level has been increasing with regard to both property crimes and non-property crimes.
Table 9:
Year
Total
1996
6139
1997
6242
1998
6529
1999
7157
2000
7321
2001
7546
2002
8213
2003
8902
2004
9595
2005
9798
Different terms of sentences for newly imprisoned (1996-2005)

-3
months
1
0.0%
2
0.0%
1
0.0%
2
0.0%
1
0.0%
3
0.0%
1
0.0%
-6
months
65
1.1%
72
1.2%
76
1.2%
63
0.9%
79
1.1%
65
0.9%
87
1.1%
83
0.9%
67
0.7%
99
1.0%
-1
year
1481
24.1%
1454
23.3%
1558
23.9%
1662
23.2%
1622
22.2%
1551
20.6%
1607
19.6%
1826
20.5%
1906
19.9%
2183
22.3%
-2
years
2618
42.6%
2611
41.8%
2759
42.3%
3014
42.1%
3086
42.2%
3195
42.3%
3287
40.0%
3614
40.6%
3961
41.3%
3837
39.2%
-3
years
1396
22.7%
1420
22.7%
1432
21.9%
1614
22.6%
1728
23.6%
1767
23.4%
2084
25.4%
2158
24.2%
2288
23.8%
2299
23.5%
-5
years
561
9.1%
658
10.5%
667
10.2%
771
10.8%
762
10.4%
916
12.1%
1069
13.0%
1188
13.3%
1227
12.8%
1183
12.1%
-7
years
15
0.2%
24
0.4%
33
0.5%
27
0.4%
38
0.5%
44
0.6%
68
0.8%
110
1.3%
108
1.1%
115
1.2%
-10
years
2
0.0%
1
0.0%
2
0.0%
3
0.0%
4
0.1%
8
0.1%
6
0.0%
19
0.2%
30
0.3%
25
0.3%
-15
years
-20
years
2
0.0%
1
0.0%
1
0.0%
5
0.0%
1
0.0%
4
0.0%
5
0.0%
2
0.0%
1
0.0%
1
0.0%
3. Increasing Punitiveness on the level of requiring time served

3.1
Release on parole of prisoners sentenced to determinate term

imprisonment with forced labor
Now we will consider the situation of release on parole. As table 10 shows, the
proportion of prisoners sentenced to determinate term imprisonment with forced
labor, and then released on parole after having served more than 80 percents of
their sentence gradually decreased from 64.1% in 1996 to 54.8% in 2003, and
then increased and reached 59.7% in 2005. It is true of all categories of sentences. For example, the ratio of prisoners sentenced to imprisonment with forced
labor for the term of five years or more, and released on parole after having
served eighty percent or more of the sentence has increased from 61.3% in 2002
to 69.3% in 2005. The absolute number of parolees whose time served was more
than eighty percent has increased from 7,754 in 2000 to 9,495 in 2005 by 1,741
(22.5%). This means that fewer prisoners have been released on parole until
eighty percent or more of the sentence has been served.
412
Toshio Yoshida
Table 10: Number of prisoners sentenced to determinate-terms of imprisonment

with forced labor and released on parole after having served more
than 80 percent of term (1996-2005)
Year
1996
Number of parolees and of those

whose term was more than 80 %
served (+ percentage)
11,882
7,612
64.1%
Term of
sentence
-1 year
1,769
1,263
71.4%
Term of
sentence
-2 years
5,246
3,414
65.1%
Term of
sentence
-3 years
3,111
1,907
61.3%
Term of
sentence
-5 years
1,339
775
57.9%
Term of
sentence
>5 years
417
253
60.7%
1997
12,373
7,877
63.7%
1,817
1,270
69.9%
5,211
3,384
64.9%
3,415
2,070
60.6%
1,492
872
58.4%
438
281
64.2%
1998
12,515
7,754
62.0%
1,713
1,226
71.6%
5,191
3,299
63.6%
3,567
2,109
59.1%
1,630
857
52.6%
414
263
63.5%
1999
12,981
7,844
60.8%
1,773
1,299
73.3%
5,191
3,217
62.0%
3,690
2,139
58.0%
1,838
952
51.8%
399
237
59.4%
2000
12,879
7,754
60.2%
1,646
1,198
72.8%
4,960
3,048
61.5%
3,739
2,123
56.8%
2,036
1,062
52.2%
498
323
64.9%
2001
14,006
8,292
59.2%
1,735
1,235
71.2%
5,676
3,382
59.6%
4,025
2,274
56.5%
2,053
1,080
52.6%
517
321
62.1%
2002
14,894
8,247
55.4%
1,728
1,205
69.7%
5,849
3,171
54.2%
4,481
2,433
54.3%
2,309
1,115
48.3%
527
323
61.3%
2003
15,273
8,372
54.8%
1,767
1,273
72.0%
5,968
3,083
51.7%
4,413
2,382
54.0%
2,579
1,286
49.9%
546
348
63.7%
2004
16,227
9,250
57.0%
1,759
1,290
73.3%
6,307
3,365
53.4%
4,787
2,699
56.4%
2,768
1,496
54.0%
606
400
66.0%
2005
15,908
9,495
59.7%
1,603
1,170
73.0%
6,044
3,372
55.8%
4,848
2,869
59.2%
2,795
1,656
59.2%
618
428
69.3%
3.2
Release on parole of prisoners sentenced to life imprisonment with

forced labor
Now we will look at the number of prisoners sentenced to life imprisonment with
forced labor and later released on parole, and their custody term. According to
table 11, since 2003, prisoners sentenced to life imprisonment with forced labor
have not been released until twenty or more have been served.
413
Table 11: Number of prisoners sentenced to life imprisonment with forced labor
and released on parole (1996-2005)
Year
Total
1996
1997
12
1998
-14
years
-15
years
-16
years
-17
years
-18
years
-19
years
4
-20
years
1
>20
years
2
15
10
1999
2000
2001
13
2002
2003
14
14
2004
2005
10
10
7
1
11
The average length of term of those sentenced to life imprisonment with forced
labor and later released on parole was about 20 years 5 months in 1996, about
21 years 6 months in 1997, about 20 years 10 months in 1998, about 21 years 4
months in 1999, about 21 years 2 months in 2000, about 22 years and 8 months
in 2001, about 23 years and 5 months in 2002, about 23 years 4 months in 2003
and about 27 years and 2 months in 2005 (unknown in 2004). That is to say, the
average term of custody increased from 20 years 5 months in 1996 to 27 years
and 2 months in 2005 by 81 months (about 7 years).
3.3
Number of prisoners
As expected more and more prisoners have been accommodated in prisons. Their
imposed imprisonment has been still more completely served. More and more
inmates therefore remain therefore in prisons. Table 12 shows that 36,987 prisoners were in prisons on December 31, 1993 and 67,101 prisoners were in custody
on December 31, 2005. The population of prisons has recently experienced an
explosive increase. From 1986 to 1992, the number of inmates had reduced;
however, from 1996 to 2005 the prison population has doubled. Who could have
thought of such a dramatic increase ten years ago? Table 12 shows also that more
and more inmates have stayed still longer. Nowadays lots of prisoners stay in
overcrowded associate cells. The rate of accommodation lay at 116.0% on December 31. 2005. In order to take precautions against over-crowding not a few
prisons have been constructed or rebuilt.
414
Toshio Yoshida
Table 12: Number of prison inmates sentenced to imprisonment with forced

labor as per December 31, 1986-2005
Year
Total
-1 year
-2 years
-3 years
-5 years
>5 years
1986
45,838
6,605 (14.4%)
15,349 (33.5%)
10,358 (22.6%)
7,257 (15.8%)
6,269 (13.7%)
1987
45,732
6,063 (13.3%)
15,223 (33.3%)
10,491 (22.9%)
7,421 (16.2%)
6,534 (14.3%)
1988
45,476
5,563 (12.2%)
14,814 (32.6%)
10,663 (23.4%)
7,646 (16.8%)
6,790 (14.9%)
1989
42,419
4,814 (11.3%)
13,330 (31.4%)
10,033 (23.7%)
7,403 (17.5%)
6,869 (16.2%)
1990
39,698
4,168 (10.5%)
12,108 (30.5%)
9,542 (24.0%)
7,111 (17.9%)
6,769 (17.1%)
1991
37,600
3,776 (10.0%)
11,658 (31.0%)
9,020 (24.0%)
6,729 (18.1%)
6,417 (17.1%)
1992
37,090
3,681 (9.9%)
11,446 (30.9%)
9,125 (24.6%)
6,548 (17.7%)
6,290 (17.0%)
1993
36,987
3,561 (9.6%)
11,231 (30.4%)
9,475 (25.6%)
6,606 (17.9%)
6,114 (16.5%)
1994
37,304
3,353 (9.0%)
11,121 (29.8%)
9,752 (26.0%)
7,048 (18.9%)
6,030 (16.2%)
1995
38,452
3,161 (8.2%)
11,382 (29.6%)
10,251 (26.7%)
7,474 (19.4%)
6,184 (16.1%)
1996
40,251
3,250 (8.1%)
11,574 (28.8%)
10,917 (27.1%)
8,089 (20.1%)
6,421 (16.0%)
1997
41,573
2,979 (7.2%)
11,723 (28.2%)
11,335 (27.3%)
8,947 (21.5%)
6,589 (15.8%)
1998
43,139
3,102 (7.2%)
11,233 (26.0%)
11,599 (26.9%)
9,758 (22.6%)
7,447 (17.3%)
1999
45,185
3,233 (7.2%)
11,978 (26.5%)
12,124 (26.8%)
10,239 (22.7%)
7,616 (16.9%)
2000
49,651
3,568 (7.2%)
13,373 (26.9%)
13,355 (26.9%)
11,103 (22.4%)
8,252 (16.6%)
2001
53,094
3,427 (6.5%)
13,400 (25.2%)
14,429 (27.2%)
12,360 (23.3%)
9,478 (17.9%)
2002
56,729
3,954 (7.0%)
14,234 (25.1%)
15,106 (26.6%)
13,499 (23.8%)
9,936 (17.5%)
2003
60,565
3,782 (6.2%)
14,215 (23.5%)
16,186 (26.7%)
15,141 (25.0%)
11,241 (18.6%)
2004
63,746
3,604 (5.7%)
14,307 (22.4%)
16,878 (26.5%)
16,262 (25.5%)
12,695 (19.9%)
2005
67,101
3,513 (5.2%)
14,368 (21.4%)
17,478 (26.0%)
17,541 (26.1%)
14,201 (21.2%)
5 years >: inclusive of life imprisonment with forced labor.

Source: The White Paper on Crime 2006, p. 331.
4. Punitiveness in the Juvenile Law

4.1
The Revised Juvenile Law of November 28th 2000
Let us begin with taking a look at the revision of the Juvenile Law. On November
28th 2000, the legislature revised the Juvenile Law partially, but not without significance. The legislature aimed at changing the juvenile crime policy from an
education- and rehabilitation-oriented policy to a penalty-oriented policy. The bill
of the revised Juvenile Law, drafted by the ruling coalition, was supported not
only by the ruling conservative parties (the Liberal Democratic Party, the Clean
Party and the New Conservative Party), but also by the opposition parties (the
Democratic Party, and the Liberal Party). According to its supplement, the revised Juvenile Law must be checked again five years later after it came into effect as was proposed by the opposition parties.
415
The background of this new legislation was the growing worries because of a
series of cruel crimes committed by teenagers in Japan. Let us name only a few
examples:
z
In May 1997 in the harbor city of Kobe, a fourteen-year old boy killed his
friend a twelve-year old boy, decapitated him, cut off his head, put into the
mouth a piece of paper with a written demand, and afterwards placed the head
at the gate of his junior high school. On October 17th, 1997, the young criminal
was admitted to a psychiatric-care reform school.
In May 2000 in the small city of Toyokawa in the prefecture Aichi, a seventeenyear old high school student thrust a kitchen knife into a sixty-five year old
housewife. She died a lingering death. He injured her husband seriously, as
well. He said after committing the crime that he broke into the house and attacked the housewife because the door to the house was by chance open. He
wanted to have the experience of killing and seeing exactly how long one could
live in spite of a bloody criminal act. He stated his belief that although killing
young people was forbidden, killing old people could be excused. He felt he
would benefit from such an experience. On December 26th, 2000, he was admitted to a psychiatric-care reform school.
In the same month in the prefecture of Saga, a seventeen-year old boy highjacked a long-distance bus, threatened to kill a six-year old girl with a knife,
took the bus driver and nine passengers hostage for more than 15 hours and
killed a woman on the bus. The police finally stormed the bus by force in the
prefecture of Hiroshima, arrested the unharmed, young man. On September 25th,
2000, he was admitted to a psychiatric-care reform school.
The number of heinous crimes committed by juveniles which were investigated

by police has continuously increased from that time. The number of murders or
voluntary manslaughters increased from 71 in 1990 to 111 in 1999, an increase of
56%. The number of robberies increased still more from 594 in 1990 to 1,652 in
1999, an increase of 117%. At the same time, you should also take into consideration that the mass media reports these scandalous crimes astonishingly disproportionately, and their reporting cannot be controlled. On top of it all, a part of
the mass media gave the impression that the mild Juvenile Law did not to hinder, if not encouraged, repugnant juvenile crime. So, the mass media and some
groups for helping crime victims demanded more stringent sanctions. That not a
few people have felt vulnerable on account of the miserable situation of the Japanese economy, and that they have excessive worries that they would fall victim to
crime, all this has been picked up by these mass media, the public and crime victims. These circumstances have resulted in a hardening of attitudes of the population towards juvenile delinquents. Plus, one cannot contradict the fact that some
decisive, conservative politicians have made the most of the fear of falling vic-
416
Toshio Yoshida
tim, and that the population strongly believes in the deterrence of robust punishment. Besides, people of rather liberal convictions have been unable to advocate
any other effective measures that might prevent a more stringent sanctions policy
or such a get-tough-policy. These various conditions, the increase of serious
crime, the demand for public security and finally the increase in the penal mentality among the population have led to a distrust of the Juvenile Law system.
The Juvenile Law is applied to juveniles from fourteen years old to nineteen
year old. According to the Juvenile Law before revision, the fourteen-year old
juveniles and fifteen-year old juveniles were not allowed to be prosecuted, therefore could not be punished. The revised Juvenile Law has lowered the minimum
age at which time juveniles can be prosecuted from sixteen years old to fourteen
years old. In this case, the sentence of imprisonment shall be executed in a reform school, the original purpose of which was to educate and reform. Even if
juveniles are sentenced to imprisonment with forced labor, they are not obliged to
perform forced labor. They are only obliged to undergo treatment and education
( 56 III of the Juvenile Law).
In addition, all juveniles of age sixteen and older who commit intentionally offences or offences aggravated by results of a crime, for example, murder, voluntary manslaughter ( 199 Penal Code), robbery resulting in death or bodily injury
( 240 Penal Code), rape on the occasion of robbery, rape on the occasion of robbery resulting in death ( 241 Penal Code), and illegal arrest and confinement
resulting in death or bodily injury ( 221 Penal Code) have to be sent to criminal
court in principle and prosecuted criminally ( 20 II of the Juvenile Law). The
legislature has justified the so called principle of referral to the prosecutor under
the guise of protecting the victim. Imprisonment is supposed to be executed in a
juvenile prison. According to the Juvenile Law before revision, such a stringent
reaction against juvenile offenders is reserved only in exceptional cases.
Also according to the revised Juvenile Law, a determinate-term of imprisonment with or without forced labor from ten years to fifteen years can be sentenced when the crimes committed by juveniles aged seventeen or older at the
time of the criminal act deserve actually life imprisonment with or without forced
labor, but can be sentenced to a determinate-term of imprisonment because of
their age ( 51 of the Juvenile Law). According to the Juvenile Law before revision, a determinate-term of imprisonment with or without forced labor of from
ten years to fifteen years was supposed to be sentenced when the crimes committed by juveniles of this age bracket deserve actually life imprisonment with or
without forced labor.
417
According to the Revised Code, a juvenile who is sentenced to life imprisonment with or without forced labor according to 51 I Juvenile Code cannot be
released on parole until after having served ten years of his sentence according to
28 Criminal Code. However, according to the Juvenile Law before revision, he
could be released on parole after seven years. Furthermore, a juvenile aged seventeen or less should be sentenced to life imprisonment with or without forced
labor instead of receiving the death penalty even if he deserved the death penalty.
These parts of the old code were criticized. The critics said that when a juvenile
commits a crime that deserves the death penalty, and has it reduced to life imprisonment, and is released early on parole, the reaction to the crime is mitigated
twice. Consequently according to this view, neither the interest of proportionality
between the criminal act and the punishment, nor that of the desire for more severe punishment among the population was satisfied.
Prosecutors, following the less stringent, educative juvenile procedure, can
now participate in the case with an agreement of the family court when the exact
circumstances of the crime remain in doubt ( 22-2, 32-4 of the Juvenile Law).
This was not allowed at all according to the Juvenile Law before revision. Plus,
the number of judges can be increased to three in certain cases ( 31-4 of the
Court Organization Law). The maximal term of admission in the juvenile classification home has been extended from four weeks to eight weeks in special cases
where the facts were in dispute ( 17 IV of the Juvenile Law).
One should not overlook, however, some points that have been improved in
view of the rule of law. First, in cases where the prosecutor, following the less
stringent, educative juvenile procedure, takes part ( 22-3 of the Juvenile Law),
the defense is obliged to be present. Second, an appeal can be lodged after the
full execution of educative measures. After the death of the juvenile, however,
neither his parents nor his legal representatives has the right to further appeal
( 27-2 of the Juvenile Law). Third, an appeal can be taken against the decision
to admit to a juvenile classification home and the lengthening of pre-sentence
detention. Fourth, the principle neb is in idem applies to the case where the
prosecutor, following the less stringent, educative juvenile procedure, takes part
and the juvenile is acquitted.
In short, the revised Juvenile Law which came into force on April 1st, 2001,
places great emphasis on punishment: many more penalties and much more
stringent penalties. It is obvious that the Japanese legislator wanted to be seen
doing something about juvenile crime, even the familiar and practiced pattern
of solution. What does this something mean? The criminal policy of the legislature seems to be merely an upgrading of crime victims and the urge for retribu-
418
Toshio Yoshida
tion and/or deterrence. In upgrading crime victims, a danger of retrogression

seems to emerge here. There are few criminological studies that show that the
rate of recidivism and trends toward criminality is connected to the kinds and
severities of sentenced and executed punishment or that the withdrawal of harsh
penalties had led to increased crime.
Nevertheless, the legislature has laid down the practiced crime policy in the
wrong direction. In May 2007, the Law concerning the partial revision of the
Juvenile Law was passed in the Diet. The background of this new legislation
was the sensational crimes which two young men committed.
z
In July 2003, in the city Sasebo in the prefecture of Nagasaki, a thirteen-year

old boy kidnapped a four-year old boy and pushed him off the roof of a parking
building, killing him.
In June 2004, in the city of Nagasaki, a twelve year-old primary school girl
stabbed her girl classmate in the neck in the school, causing her to bleed to
death.
The Juvenile Law before the revision applied to juveniles from fourteen to nineteen years of age. According to which, the family court can impose one of the
following protective measures: probationary supervision, confinement to a
self-support home for juveniles or confinement to a home for dependent children,
or confinement to a juvenile training school. Children under fourteen could be
confined to a self-support home for juveniles or to a home for dependent children. But the court was not allowed to confine them to a juvenile training school.
Now according to the revised Juvenile Law, children of about age twelve or older
can be confined to a juvenile training school, as well. The Ministry of Justice has
explained that it is possible that an eleven-year old child be confined to a juvenile
training school. Likewise, the police are authorized to investigate on a noncompulsory basis, as summoning, asking questions and taking compulsory measures, such as an investigation based on compulsory measures, compulsory search
and seizure.
z
4.2
Comparing the courts practice before to after the revision of the

Juvenile Law
When one compares the courts practice before the revision of the Juvenile Law
with the courts practice afterwards, it turns out that, as one may have expected,
the courts reaction to juvenile crimes has tended towards more and more stringency. In 1996, 163 juvenile criminals were sentenced to such penalties in the
first instance. In 2005, 204 juvenile criminals were likewise sentenced. It means
that the number of punished juvenile criminals has increased by 41 persons
(25.2%) (see table 12 below).
419
Let us see this point a little more exactly. In 1996, the juvenile (family) court decided to transfer the cases of 163 juvenile criminals to the criminal court. These young
people were punished in the first instance of the criminal court: 48 persons (29.4%)
and 114 persons (70.0%) were respectively sentenced to imprisonment without suspension of sentence, or imprisonment with suspension of sentence whether with probationary supervision or without it. By contrast in 2005, the cases of 204 juvenile
criminals, the jurisdiction of whom were refused by the family court, and punished in
the first instance of the criminal court, 76 persons (37.3%) and 121 persons (59.3%)
were respectively sentenced to imprisonment without suspension of sentence or imprisonment with suspension of sentence whether with probationary supervision or
without it. The proportion of cases of imprisonment without suspension of sentence
has increased by 7.9%. The minimum rate of imprisonment without suspension of
sentence was 23.6% in 1997. The maximum rate of imprisonment without suspension
of sentence was 40.9% in 2003. The difference amounts to as much as 16.7% (see
table 13 below).
Table 13: Number of imprisonments with or without suspension of sentence in
the first instance of the criminal court
Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
The number of
juveniles sentenced to penalty
163
161
161
158
151
197
236
198
259
204
Imprisonment without
of sentence
48 (29.4%)
38 (23.6%)
48 (29.8%)
46 (29.1%)
61 (40.4%)
69 (35.0%)
92 (39.0%)
81 (40.9%)
96 (37.1%)
76 (37.3%)
Imprisonment with
of sentence
114 (70.0%)
120 (74.5%)
113 (70.2%)
109 (69.0%)
88 (58.3%)
124 (62.9%)
142 (60.2%)
116 (58.6%)
161 (62.2%)
121 (59.3%)
Fine
1 (0.6%)
3 (1.9%)
0 (0.0%)
3 (1.9%)
2 (1.3%)
4 (2.0%)
2 (0.8%)
1 (0.5%)
2 (0.8%)
7 (0.3%)
Source: The White Paper on Crime in Japan 2006, p. 157 and the White Paper on Crime in Japan 2001, p. 159.
The number of juvenile criminals who were sentenced, without appeal, to imprisonment without forced labor and confined to prison amounted to 34 persons in 1996 and
81 persons in 2005. It means an increase of 238%. When one sees the terms of imprisonment of sentenced and confined juveniles for the last ten years, one recognizes
more clearly the present tendency towards much more stringency. The proportion of
juvenile criminals who would be confined to prison for the term of at least two years
or less than five years has decreased respectively from 20.5%, or 50.0%, in 1996 to
4.9%, or 38.3%, in 2005. In contrast, the proportion of juvenile criminals who were
420
Toshio Yoshida
confined for the term of at least ten years has increased from 29.4% in 1996 to 51.9%
in 2005 (see table 14 below). It follows from the statistics that not only more and more
juveniles are being sentenced to imprisonment without suspension of sentence, but
also for much longer terms of imprisonment. Table 15 shows the situation of the sentenced to imprisonment without forced labor.
Table 14: Terms of imprisonment with forced labor of those sentenced with no
appeal and confined juvenile criminals
Year
1996
Total
34
1997
30
1998
36
1999
36
2000
39
2001
47
2002
77
2003
68
2004
84
2005
81
6
months
1
year
1
2.9%
1
3.3%
2
5.6%
2
5.6%
1
2.6%
2
4.3%
2
years
6
17.6%
4
13.3%
4
11.1%
11
30.6%
4
10.3%
4
8.5%
3
3.9%
4
5.9%
8
9.5%
4
4.9%
3
years
6
17.6%
7
23.3%
4
11.1%
2
5.6%
4
10.3%
9
19.1%
9
11.7%
11
16.2%
10
11.9%
9
11.1%
5
years
11
32.4%
10
33.3%
8
22.2%
7
19.4%
9
23.1%
19
40.4%
28
13.2%
28
41.2%
29
34.5%
22
27.2%
7
years
9
26.5%
5
16.7%
8
22.2%
6
16.7%
12
30.8%
8
17.0%
19
24.7%
14
20.6%
24
28.6%
11
13.6%
10
years
1
2.9%
3
10.0%
9
25.0%
6
16.7%
9
23.1%
5
10.6%
17
22.2%
11
16.2%
12
14.3%
31
38.3%
15
years
Life imprisonment
1
2.8%
2
5.6%
1
1.3%
3
3.7%
1
1.2%
1
1.2%
Source: The Yearbook of Correctional Statistics 1996-2005.
Table 15: Terms of imprisonment without forced labor of the sentenced with no
appeal and confined juvenile criminals
Year
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
Total
1
2
1
3 months
6 months
1 year
2 years
1
1
Source: The Yearbook of Correctional Statistics 1996-2005.
3 years
5 years
1
421
5. Increase of penalty mentality

In order to understand public opinion toward punishment, popular attitudes toward capital punishment should be examined to see that the magnitude of punitiveness is measured by the application the death penalty, which is agreed internationally. As far as the attitudes of the Japanese population toward the death
penalty are concerned, data from opinion polls since 1980 are available to us.
Now we can see the results of opinion polls which the Cabinet Secretariat information center carried out in face-to-face interviews with persons twenty years of
age or older (those eligible to vote). As table 16 shows, in 1980 two-thirds of
those responding asked gave the opinion that the death penalty should not be
abolished under the situation at that time (2,434 persons asked of 3,000 or 81.1%
answered). In 1989, more than six-tenths of interviewees expressed that the
death penalty should not be abolished in any way under the situation at that time
(It is possibly unavoidable to receive the death penalty. (2,293 persons asked of
3,000 or 76.4% answered). 76.8% of all interviewees were of the opinion that the
death penalty should be preserved in the future as well. In 1994, more than
seven-tenths of those asked supported the death penalty (2,113 interviewees of
3,000 or 70.4% answered). In 1999 about eight-tenths of those responding supported the death penalty (3,600 interviewees of 5,000 or 72.0% answered). The
opinion poll of 2004 shows that more than eight-tenths of respondees supported
the death penalty (2,048 interviewees of 3,000 or 68.3% answered). So the data
of 2004 shows the highest support proportion. With all these figures, you could
ascertain the growing punitiveness among the population7 .
As reasons for upholding the death penalty, the following were named by interviewees (multiple naming is possible.). Retribution: A cruel criminal should
pay for his crime with his life. Deterrence: Abolishing the death penalty will
lead to more heinous crimes. The death penalty will prevent fairly well an increase of crimes. Incapacitation of dangerous criminals: A cruel criminal can
commit a crime again if you spare his life. Finally, appeasement of crime victims: If the death penalty were abolished, the feelings of victims and their relatives would not be aggravated. As table 17 shows, it is noteworthy that the appeasement function of the death penalty in society has been increasingly named
as reason for supporting the death penalty, that is, 39.7% in 1989, 40.4% in
7
There is a question, however, concerning support for the death penalty as a standard indicator to grasp the penalty mentality of a population. Does it really provide sufficient and
valid information? Against this background of methodological criticism, one could not overestimate the results. However, such methodologically imperfect opinion research could have
a rather large influence on politics and criminal law legislation.
422
Toshio Yoshida
19994, 48.6% in 1999 and 50.7% in 2004. It has been gaining in importance. It is
probably the upswing of victim action in the last twenty years that has built the
background for it.
Table 16: Proportion of supporters of death penalty (1980-2004)
Year
Proportion of supporters of death penalty
1980
(62.3%)
1989
66.5%
1994
73.8%
1999
79.3%
2004
81.4%
Source: The Yearbook for Public Opinion Research, 1980, 1989, 1994, 1999 und 2004.
Table 17: Reasons for upholding death penalty (1980-2004)

Year
1980
1989
1994
1999
2004
retribution
26.9%
56.0%
51.2%
49.3%
54.7%
deterrence
46.2%
53.1%
48.2%
48.2%
53.3%
incapacitation
23.4%
37.9%
33.9%
45.0%
45.0%
appeasement
39.7%
40.4%
48.6%
50.7%
Source: The Yearbook for Public Opinion Research, 1980, 1989, 1994, 1999 and 2004.
As a whole, you could say, the attitudes of the Japanese population toward the
death penalty are retributive. On top of that, the public is convinced that the
death penalty is suitable for prevention of crime and calming the feelings of victims and of the public at large.
6. Conclusion
As background of the belief in much more and much severer punishment it has
to be pointed out on the one hand that the influence of sensational, dramatized
reports on serious crimes in the print media, especially on crimes committed by
juveniles, sexually motivated criminals, and automobile drivers under influence
of alcohol, together with the misleading information of the public connected with
it. The mass media want ultimately and uniquely to increase sales of their product
and the resulting ratings. The perceived development of criminality and the feeling of unease among the population caused by mass media have disguised the
knowledge of the actual development of criminality. On the other hand, it has
to be pointed out that the feeling of solidarity among the population has been lost
423
in the process of urbanization, and the socialization of citizens oriented on traditional social norms and values has become much weaker, although they can have
given rise to lower crime rates.
These attitudes have caused an additional growing feeling of uncertainty
among the citizens. Criminality has been perceived as threatening, and is connected with still increasing disparity between rich and poor to which the conservative neo-liberal projects in the economy and politics (the increasing globalization and deregulation in economy) has led. This has caused additional anxieties.
Citizens have not felt such safety and security as they did earlier in daily life. It
follows from this that they do not easily tolerate criminals in the mean time.
This development has had an effect on the attitudes of the public towards dealing with criminals. Their basic attitudes which have increasingly become much
more punitive against criminal deviationists, connected with the reports of
mass media and the upswing of the victim movement which has stressed the interests of actual crime victims, have led to the perception that penalties inflicted
by the criminal court are too mild. More severe punishments and more rigid
criminal prosecutions have been correspondingly demanded. Both the call for
rigid criminal prosecution and severe sanctions have been picked up by not only
right-wing conservative politicians who have referred to public opinion as an
alibi for realizing their own lust for punishment, but also by liberal politicians.
Both groups contributed to the revision of the Code of Criminal Law, the Code of
Criminal Procedure and the Juvenile Law.
The Criminal Justice system has been under pressure of the population and
populist-oriented politicians to sentence more often and more severe penalties.
In realty, the Judiciary at every level has been much more punitive for the last ten
years.
Regrettably, one can expect that turning away from repressive authoritarian
criminal law is out of the question, both in theory and practice for the near future.
424
Toshio Yoshida
Selective References
Kury, H., Kania, H. & Obergfell-Fuchs, J. Worber sprechen wir, wenn wir ber Punitivitt sprechen? Versuch einer konzepzionellen und empirischen Begriffsbestimmung,
Kriminologisches Journal, 8. Beiheft 2004, pp. 51-88.
Kury, H., Brandenstein, M. & Yoshida, T. Internationales Ausma und Bedeutung der
Verkehrsdelinquenz, Teil 1 3 (Schluss), The Hokkai-Gakuen Law Journal, Vol. 42,
No. 2, 3 (2006), No. 4 (2007).
Kury, H. Strafvollzug und Resozialisierung, Teil 1-2 (Schluss), The Hokkai-Gakuen
Law Journal, Vol. 34, No. 2 u. No. 3 (1998).
Yoshida, T. Der japanische strafrechtliche Schuldbegriff von gestern, heute und morgen.
Recht, Schuld, Strafe, Strafzumessung und Wiedergutmachung, in: Eser, A. and Yamanaka, K. (edited), Einflsse deutschen Strafrechts auf Polen und Japan, 2001, pp.
225-259.
Die Zukunft der japanischen Strafrechtspflege, in: Moos, R. (edited), Festschrift fr
Udo Jesionek, 2002, pp. 535-553.
Confession, apology, repentance and settlement out-of-court in the Japanese criminal
justice system, in: Weitekamp, E. and Kerner, H.-J. (edited), Restorative Justice in
Context, 2003, pp. 173-196.
Strafrecht, Sanktionen und Einstellungen zu Sanktionen in Japan, in: Kury, H., (edited), Strafrecht und Kriminalitt. Entwicklungen in Mittel- und Osteuropa, 2004, pp.
189-208.
425
Which Sentencing Goals Do Victims of Crime in Austria

Support? Impressions from an Urban Victim Cohort
HELMUT HIRTENLEHNER
1. Introduction
Legitimacy is the belief held by members of a society that the way their political
order works corresponds to their own moral principals and what they personally
believe to be right and wrong (Easton 1965). Order appears legitimate if the majority of the persons subject to it feel that it is right (Habermas 1988). Such a
concept of legitimacy is a pillar of representative democracy. The penal law of a
society also requires legitimacy if it is not to contradict the basic principle of democratic government. When we ask how it is possible to justify a system of
criminal law, or the specific form it takes, then penal theories come to the fore.
Penal theories attempt to give meaning to criminal punishment, and therefore
they deal with the legitimacy of criminal law as a whole. Because criminal punishment was the only reaction of the criminal justice system for such a long time
(and it still is the most typical reaction) the legitimacy of a penal system is very
closely connected to the aims of, and reasons for, punitive measures.
Empirically, legitimacy is represented as acceptance. This makes it possible to
carry out empirical research on individual attitudes to a field of policy and to
evaluate a states actions by the opinions and wishes of its citizens. This is exactly what the following study does. The various theoretical justifications for
criminal punishment are compared with the expectations of a group of people
who have been directly affected by conflicts involving criminal law. The focus of
the study is on persons whose rights have been directly abused through criminal
offences the victims of crime.
In 1985, the classic study by Joanna Shapland and colleagues highlighted the
needs and emotions of victims as they went through the criminal justice system.
The expectations and requirements of the victims were outlined in this study.
However, up to now there has been no Austrian research on the attitudes and
preferences that crime victims have concerning the sentencing of offenders. The
following study, whose findings are now being published, should help to fill this
blatant gap in research.
426
H. Hirtenlehner
The central question in this study is how far the needs and opinions of crime
victims provide support for the different penal theories found in Austria today.
The question is raised whether the different sentencing goals are legitimated by
what crime victims require. We investigate which of the theoretical justifications
of punishment are given support by the expectations of persons whose rights
have been violated. Particular attention has been paid to any mediating effect that
the offence has on the victim. If the victims of property crimes or crimes of violence do have different preferences concerning the aims of punishment, it should
be possible to determine any diverging structure in their preferences, and then the
consequences for legal policy can be worked out.
In this paper, firstly the penal theories current in Austria at the turn of the millennium are introduced and these are then contrasted with the opinions and expectations of crime victims. Using a cohort of victims who reported offences in a
selected Austrian provincial capital during the period July 2004 to June 2005, we
investigate which sentencing goals are favoured by the victims of crime and
which are rejected. To obtain quantifiable results and to rank the penal theories
according to the degree of acceptance, the sentencing goal preferences were collected using a standardised written questionnaire.
2. Penal theories and sentencing goals in Austria

Before the expectations of crime victims can be included in a discussion of legitimacy in criminal law, a short outline of the penal theories current in Austria at
the turn of the millennium is necessary. As will soon be obvious, the justification
for penal measures used in the German speaking domain (Blmmel 2002; Jakobs
2004; Momsen/Rackow 2004; Roxin 2001) Austrian penal law and its use are
strongly influenced by work and theses from Germany is only marginally different from the usual penal principles in the Anglo Saxon world (Ashworth 2002;
Ashworth/Wasik 1998; Padfield 1996; von Hirsch/Ashworth 1998). The greatest
difference is the fact that in Austria and Germany the symbolic confirmation of
the legal order is seen as a distinct aim of sentencing. This sentencing goal,
which was inspired by the work of Emile Durkheim (1907; 1911), focuses on the
norm illustrating, and thus norm stabilising, effects of punishment (Moos 1989;
Mller-Tuckfeld 1998; Zipf 1989). We will return to this later.
Firstly, we must differentiate between absolute and relative penal theories.
Absolute penal theories do not link their justification to social aims, (latin absolutas); they see punishment as indispensable for the sake of justice alone. Relative penal theories aim to prevent crime and protect from crime. Absolute penal
theories are against any preventative aim of punishment. They negate all future
Austria: Impressions from an Urban Victim Cohort
427
oriented purposes of penal law and justify punishment retrospectively as in the

interests of justice. The guilt of the offender is balanced out by the punishment.
A crime requires punishment; this is the retribution for the act committed. Absolute penal theories are therefore retribution theories. Their ideological basis is
found in the philosophy of law of Kant (1797) and Hegel (1821). Retribution
theories are not particularly popular in Austria today, but they have not yet disappeared completely 1 .
There are different varieties of relative penal theories. What they all have in
common is that their goal is to prevent crime and protect from crime and this is
the justification they provide for criminal punishment. They are not retrospective;
they look to the future. They address either individual offenders (individual prevention) or the general public (general prevention).
The main idea of Franz von Liszts concept of individual prevention (von
Liszt 1905), which he presented in his Marburg Programme (1882), was the idea
that the sole purpose of punishment was to keep known offenders from committing further offences. This can be effected in three ways by rehabilitation, by
deterrence or by incapacitation. Rehabilitation means that the offender is adjusted
to society. Treatment and education should ensure that the personality of the offender undergoes change and the moral and legal standards of the community are
internalised. Individual deterrence lies in the fear of punishment. The mere threat
of further punishment should be enough to ensure that the offender abides by the
law in future. Incapacitation ensures that the offender who does not want to rehabilitate stays harmless. This means keeping offenders in custody for unlimited
lengths of time, for as long as they represent a danger to society. This custody
can only be justified through the protection accorded to the general public during
the time the offender is imprisoned. Today, the focus of individual prevention is
on rehabilitation, although the aspects of deterrence and incapacitation are also
still relevant.
General prevention is oriented towards the population as a whole. The aim of
general prevention is to control the behaviour of the general public. According to
the prevailing theories, two varieties of general prevention can be differentiated:
general deterrence and norm illustration. General deterrence aims to prevent the
general public from committing crime by threatening punishment and carrying
out this threat. Fear of punishment should stop people from committing offences.
An alternative concept of general prevention is entitled integrative prevention
(Moos 1989; Mller-Tuckfeld 1998; Zipf 1989). In contrast to general deter1
There is a close connection between the absolute penal theories and the just deserts concept popular in Anglo-Saxon jurisdiction (von Hirsch 1976).
428
H. Hirtenlehner
rence, integrative prevention does not want anyone to conform to the norm
merely through fear of punishment, but from the inner conviction that the norm is
right and valid. The punishment of individual offenders should demonstrate to
the general public the content of the legal norm, and in this way validate the
norm which has been broken. As public confidence in the norms validity has
been shaken by the offence, the continued visible enforcement of this norm
should re-establish confidence in its validity. The real aim of the punishment is to
confirm the sense of right and wrong in the norm-keeping members of the community.
The amalgamation theory is a combination of the penal theories already described (Jescheck/Weigend 1996, p. 75ff.; Meier 2001, p. 33ff.). Analytically, it
would be possible to differentiate between classical amalgamation theories,
which combine retribution and prevention, and true preventive amalgamation
theories, which combine only general and individual prevention. Although the
term amalgamation theory is usually reserved for the integration of absolute and
relative penal theories (Burgstaller 1983; Kaufmann 1982), for the study presented here we will use the term also to mean only prevention-oriented amalgamation theories. Many amalgamation solutions are put to the test by the problem
known as antinomy of sentencing goals, where different penal theories demand
different penalties for the same offence.
It is no longer disputed that the expectations and needs of the victims should
be considered in the theoretical basis of modern-day penal law. The question of
whether satisfying the interests of the victims can be seen as an autonomous justification for punishment, and therefore as an independent penal theory, has not
yet been resolved. Recently there have been increasing numbers of attempts to
complement the above theories with the goal of restoration (Hilf 2006; Jesionek
2005; Kilchling 2002). Restoration means returning the economic and psychosocial situation of the victim to what it was before the offence. One of the main
ideas is to gain access to the personality of the victim, as well as to resolve the
conflict between the offender and the victim and provide compensation for the
results of the crime. The victim must receive recognition as a victim, and receive
confirmation that the offender is responsible for the crime. The punishment of the
offender should strengthen the victims self-confidence and help him to deal with
the offence emotionally. This understanding of restoration combines mediation,
restitution and re-personalisation for the victim2 .
Victim-oriented penal theories like this show many connections to the international restorative justice movement (Johnstone 2003).
429
3. State of research
Little is known about which sentencing goals are preferred by victims in Austria 3 . Up to now there have not been any relevant studies. However, there has
been one study, limited to the area of Vienna, touching the sentencing requirements of victims of criminal acts (Hanak 1982). Based on observations during
the proceedings, 26 % of the victims were considered to be primarily restitution
oriented. A need for explicitly repressive sanctions was attributed to 22 %. The
remaining 52 % showed no particular interest in punishing the offender: 15 %
said expressly that they were not interested in any kind of punishment for the
offender; and 37 % were indifferent (Hanak 1982, p. 28 f). These findings make
it clear that reporting a crime does not necessarily mean that the victim has an
interest in the criminalisation and punishment of the offender.
As there is a lack of findings on the sentencing motives of crime victims in
Austria, we must look at studies in Germany and Switzerland. In a survey of victims in an area previously in the Federal Republic of Germany, Kilchling (1995,
p. 495ff.), showed what attitudes toward imprisonment were held by people who
had been victimised. He showed that the goal of rehabilitation was in first place,
chosen over several others. Almost two thirds of all the victims surveyed thought
that the resocialisation of the offender was the ideal purpose of imprisonment.
Rehabilitation of the offender was the dominant goal for all groups of offences.
A range of studies of the punishment motives of the population showed that
the attitudes towards sentencing goals are not mutually exclusive. Just as in the
amalgamation theory, most people support several sentencing goals at the same
time 4 (Arnold/Korinek 1991; Endres 1992; Schwarzenegger 1992; Teske/Arnold
1991).
There are numerous studies on the acceptance of restorative and restitution
elements of criminal justice (Bals 2006; Kilchling 1995; Sessar 1992; Vo 1991).
In summary the findings of the studies show that diversion and informal settlement strategies are completely compatible with the needs crime victims have and
the reactions they expect.
3
4
This lack of research is true for victimology in Austria in general (Krainz 1991).
The high level of support for each sentencing goal shows that several goals must be supported at the same time. The sum of the acceptance percentages is far above 100 %.
430
H. Hirtenlehner
4. Method
4.1
Sampling and data
The basis of this study on the acceptance of various sentencing goals is a standardised mailed survey with people who have identified themselves as the victims of criminal acts by reporting offences to the police. The standardised mailed
survey method was selected in order to allow as large a group of victims as possible to reply, and so to get a representative picture of the prevailing opinions.
The survey was done by post for two reasons: firstly, it is cheaper, and secondly,
because the interviewer effect which must not be underestimated in such delicate topics can be eliminated in this way (Kury 1992).
The population group used for the survey were persons who had reported a
crime to police stations in Linz5 between 1 July 2004 and 30 June 2005 which
led to an investigation by the criminal investigation department. An electronic
record of the cases dealt with is kept by the criminal investigation department in
Linz. The names and addresses of victims who had reported crimes were taken
from this register.
The list of victims obtained from the police is not representative for all the victims of crime. In fact, there is a considerable distortion concerning the severity of
the offence. Many surveys of victims show that the crimes which are actually
reported are far from representative of the crimes which are not reported6 (Kesteren et al. 2000; van Dijk et al 1991). A further source of distortion can be
found concerning the severity of the crime: criminal investigation departments
prefer to work on serious crimes and on crimes which are relatively easy to solve.
(Kunz 2004, p. 239; Steffen 1976). This means that the list of victims used in the
survey is biased towards more serious crimes 7 .
The group of victims registered in the electronic records of the police was further reduced in the next step. The following groups were excluded for methodo5
Linz is the capital of Upper Austria, has about 200,000 inhabitants, and is known for many
types of industry and the good labour market situation. The unemployment rate was 3.7 %
in June 2005 (Electronic data system of the Upper Austrian Labour Market Service,
http:www.ams.or.at/neu/ooe/1400.htm, downloaded 18 June 2006).
There are generally good reasons for not reporting a crime. The reason most often given,
too little damage, shows that these offences tend to be less serious or intrusive. For information on selective reporting of offences and the resulting distortion of crimes registered
with the authorities see Kesteren et al. (2000), Schwind (2003, p. 388ff.) and van Dijk et al.
(1991).
Personnel from the Upper Austrian Executive confirm that petty crime plays a subordinate
role in police records. Minor offences such as theft with low value, shoplifting, bicycle theft,
or minor damage to property are relatively under-represented in their daily work.
431
logical and content reasons: juridical persons 8 , victims of sex offences 9 , persons
who had been the victims of someone in their family (a close relative), victims of
traffic offences 10 , celebrities, victims of offences where the police had noted not
suitable for the press 11 , and offences concerning underage youth.
At the end of this selection process, the police named 1,646 victims who had
been recorded between 1 July 2004 and 30 June 2005 in the electronic data system of the police criminal investigation department in Linz. A self-administered
questionnaire was sent to these people at the beginning of February 2006. Four
weeks later, a reminder was sent, together with a second copy of the questionnaire. The sender was given as the police authorities in both cases 12 .
A total of 694 persons filled in and returned the questionnaire, giving a return
rate of 42 %. 95 % of the respondents said they had personally been the victim of
a criminal act. 5 % named another person as the real victim (generally a relative).
As the idea behind the questionnaire was to find out the sentencing motives of
persons who had been directly affected by the crime; indirect victimisation was
not relevant. After the indirect victims had been removed from the study, the data
set used for further analyses consisted of 659 persons.
4.2
Measurement
The measurement of the approval or rejection of various penal theories by the

victims was conducted according to the Likert process (Likert 1932). Each theory
was presented in the form of three statements, formulated in everyday terms, and
these were rated by the respondents according to a seven-step scale. The state8
10
11
12
Behind this is the idea that the managers of enterprises which have been subject to damage
through criminal acts (e.g. the manager of a supermarket which has been broken into) have not
really suffered personal injury. Such persons are said not to show as much emotional and intellectual concern as persons who have been the victims of crime personally. This is probably why
officials of victimised enterprises have a greater distance to any possible sentences. According to
Kaiser (1993, p. 314) enterprises have an volatile victim characteristic.
Victims of sexual offences were excluded through police sensitivity. The police authorities
wanted to prevent the victim of a sexual crime from being confronted with the incidents
again when it could be avoided.
These are persons who suffered injuries in traffic accidents where the other person involved
was charged with bodily injury caused by negligence. It can be assumed that someone injured in a traffic accident does not usually see himself as the victim of a crime. This would
lead to traffic victims having different needs concerning criminal law to the majority of
classical crime victims.
According to the police authorities, this is just a very small group. Less than 1 % of offences
are classified as not suitable for the press.
Data protection regulations did not allow the names and addresses of crime victims to be
passed on to a university research institute.
432
H. Hirtenlehner
ments on the different theories were randomly mixed in a battery of statements.

A seven-step scale was provided alongside the items, and this ranged from (6) I
agree completely to (0) I dont agree at all.
Information on the wording of the items used and the psychometric characteristics of the scales constructed can be found in the appendix. The results of factor
and reliability analyses of the scale items (Child 2006; Cronbach 1951; Fishman/Galguera 2003) were as follows: The operationalisation of all the penal
theories worked well with the exception of the measurement of the rehabilitation
goal. Except for the resocialisation aim, all the scales were unidimensional with
factor loadings of >70. Communalities, which are not given here, are over 50 %
throughout13 . The extracted factors explain at least two-thirds (at least 63 %) of
the variance of the original indicators. Measured against the internal consistence
(Cronbachs ) the scales were very homogenous (general deterrence, individual
deterrence, incapacitation) to adequately homogenous (norm illustration, restoration, retribution) 14 . The three indicator measurement of the rehabilitation theory
was not successful. The item concerning re-integration of the offender in society does not conform sufficiently to the items which emphasise a change in the
personality of the offender15 . This means that the respondents made a difference
between re-integrating and resocialising the offender. This breaches the homogeneity assumption. The impaired scalability must be taken into account by removing the re-integration item. The scale must be limited to aspects of education and
reform of offenders, meaning that the underlying sentencing goal specifically
refers to the moment of change in the character of the offender 16 .
Subsequently, a single index was calculated for each of the penal theories. The
individual values of the items on a scale were added together, then the sum was
divided by the number of items. This produced a simple index which showed the
degree of acceptance or rejection of each penal theory. The spectrum ranged
13
14
15
16
This means that the items used as indicators were well reproduced by the penal theories
targeted.
A Cronbachs of .70 was taken as the threshold for acceptance of the assumption of homogeneity.
This is shown in a low communality and a low discriminatory power of the re-integration
item. Only 26 % of the variance of this attribute could be reproduced by the common factor.
Besides, a decline in the factor loadings points to an impaired unidimensionality.
From a statistical point of view, combining the education item with the rehabilitation item
appears justified. The common factor explains 71 % of the initial variance. Both indicators
are substantially linked, with a product-moment-correlation of .42. The somewhat lower internal consistency of .59 should not confuse unduly, as Cronbachs always also depends
on the number of variables involved. With the same average correlation, the value of Cronbachs increases with the number of items involved (Cronbach 1951).
433
from 0 to 6, where (0) was the greatest possible rejection and (6) the maximum
acceptance.
4.3
Sample description
Table 1 shows the structure of the surveyed victim population. The average age
was 44 years. A little more than half were male. Only 6 % were nationals of
states other than Austria 17 . Barely a third was resident outside Linz. These were
mostly people who either worked in Linz or lived in the greater Linz district.
Table 1:
Sample Description (N = 659)
Age
19 29 years
30 39 years
40 49 years
50 59 years
60 years and over
Sex
Male
Female
Nationality
Austrian
Other Nationality
Residence
Linz
Outside Linz
4.4
Percent
24 %
21 %
23 %
14 %
18 %
55 %
45 %
94 %
6%
70 %
30 %
Delict structure
It has already been mentioned that the design of the study causes concern about
how far the sample population is representative for victims of crime in general
with respect to the seriousness of the offence they suffered. This means that
much more notice must be taken of the distribution of offences in the sample and
how this compares with the official figures on crime in Linz. Table 2 gives an
overview of the structure of the offences.
17
According to the last census, which was carried out in 2001, 12 % of the population of Linz
are not Austrian citizens.
434
Table 2:
H. Hirtenlehner
Delict Structure
Victims participating
in the survey
Offences reported
in Linz 2004*
Persons
Percent
Number
Percent
Burglary
290
47 %
4.986
23 %
Theft
183
29 %
6.669
31 %
Bodily harm
83
13 %
1.809
8%
Robbery
26
4%
137
1%
Other offences
40
7%
8.066
37 %
622
100 %
21.667
100 %
Total
* Criminal acts according to the Criminal Code.

Source: special analysis of police crime statistics 2004 for the City of Linz from the Federal
Police Authorities.
A comparison between the official delict distribution in Linz and the offences
reported by the victims in the survey shows that there is an unusually high incidence of burglaries in the sample population. Almost half of the victims had reported a burglary 18 . The number of burglaries in the sample was twice as high as
in the official figures. Additionally, violent offences (bodily harm, robbery) were
also more numerous and involved more robbery than usual. The incidence of
theft is about the same in the sample as in the figures for Linz as a whole, but
seems to involve more serious theft than usual. This assumption can be tested by
comparing the value of goods that the victims in the sample reported stolen with
the threshold of 3,000, which is the value at which serious theft begins19 .
17 % of the theft victims reported the stealing of goods worth more than this
amount. However, the figures given in the official crime statistics for Linz show
that only 2 % of theft cases can be called serious theft. This leads to the conclusion that the cases of theft included in the sample are unusually serious.
The cumulative divergences imply that the victims participating in the survey
are not representative for victims of crime in general regarding the delict structure and the severity of the offence. Further results must be interpreted bearing in
mind that the findings have been obtained using a group of victims who have
been subject to much less minor crime than usually is the case among victims of
criminal offences.
18
19
As enterprises, being juridical persons, had been excluded from the study, these people must
have been victims of break-ins in either dwellings or vehicles.
See 128 of the Austrian Criminal Code.
435
5. Results
5.1
Acceptance of Various Penal Theories
Figure 1 gives an overview of the sentencing goals preferred by the crime victims
participating in the survey. It becomes clear that all penal theories are accepted
by a majority of the victims. However, the fact that there is a gradual difference
in the meaning they give to the theories casts some light on the matter.
Fig. 1:
Sentencing Goal Preferences
5,1
Rehabilitation
General Deterrence
5,0
Individual Deterrence
4,9
4,8
Norm Illustration
4,6
Restoration
4,3
Incapacitation
4,0
Retribution
Error Bars: 95 % Confidence Interval
0 ..Dont agree at all; 3 .. Undecided; 6 .. Agree completely
The sentencing goal which found most agreement was the rehabilitation of offenders. This means that the victims first requirement is for the offender to be
resocialised and re-educated. Close behind this comes the desire for deterrence,
firstly, general deterrence for the rest of the population, and secondly, and chosen
just as often, individual deterrence, referring to the offenders behaviour in future. The gap between rehabilitation and deterrence is only small. Increasing the
sense of right and wrong in citizens (norm illustration) is in fourth place.
436
H. Hirtenlehner
Although they do find acceptance from the majority of victims, and can therefore be considered basically acceptable, absolute penal theories have the lowest
level of legitimacy. Retribution oriented concepts of criminal law have much
lower authorization in comparison to other, more strongly goal-oriented penal
theories. Of all the penal theories current in Austria, retribution theories are given
the least support by the needs and attitudes of the victims of crime. Considering
the above-average incidence of serious offences in the sample population, this
result is extremely interesting, particularly in connection with the leading position of rehabilitation. In a more representative sample, we can expect that retribution would retreat even farther into the background. If retribution is at the lower
end of the needs pyramid for victims of serious crime, then prevention-oriented
penal theories will probably find even more acceptance among the whole group
of crime victims.
The same is more or less true for incapacitation. Neutralising any danger that
could arise from the offender by keeping him in safe custody appears acceptable
for most of the victims, but does not reach anything like the acceptance of rehabilitation, deterrence and norm illustration. However, the safety provided by imprisonment is preferred to a pure strategy of retribution.
Victim-oriented sentencing goals take middle place in the preference ratings.
Even for crime victims, restoration has a lower rating than resocialisation, deterrence or norm illustration. But it has a higher rating than absolute penal theories
or incapacitation. It becomes apparent that the crime victims do not think of
themselves first when they state their preferences for sentencing goals. They are
less interested in looking after their own interests and their own development
than they are in the classic goals for the offenders and society. The reasons they
do not place emphasis on restoration could lie in the way the population has been
socialised by the penal law. For many years, victims were neglected in criminal
proceedings and in the discussions on penal policy. This could have led to people
in Austria associating punishment with convicted law-breakers and their potential
followers. For centuries, Austrian criminal law has replaced interpersonal conflict settlement with a formal process in court between the accused and the public
prosecutor. This could have caused the population to believe that punishment for
crime should be geared towards the offender and the general public. In this way it
could have been possible for the current practice in criminal law to create its own
legitimacy.
One should not be misled by the graduated feedback on the various penal
theories, as the man on the street only has a limited understanding of the fine differences in legal theory. Instead of carefully differentiating between the various
437
sentencing goals, crime victims make rather rough demands on penal law, without any such nuances. Empirically, this is shown in the interdependence of the
sentencing goals. It can already be seen in the correlation between the possible
explanations for the sentences that the sentencing goals are not independent of
each other. A high acceptance value for one penal theory also accompanies a distinct endorsement of a different sentencing goal. Only 3 out of 21 coefficients of
correlation are less than .50. The smallest empirical correlation between two sentencing goals is .41 20 .
If the various sentencing goal scales are exposed to a meta-factor analysis
(Child 2006) the different scales are found to be unidimensional (table 3) 21 . A
reliability analysis (Fishman/Galguera 2003) indicates a high level of internal
consistency (Cronbachs = .91). This only means that, for the man on the street,
the diverse penal theories melt together to form an unspecific demand on penal
law. Crime victims do not differentiate carefully between sentencing goals; what
they expect from criminal punishment is rather vague. The detailed differences
between penal theories in jurisprudence seem very artificial to them.
Table 3:
Meta-factor-analysis of sentencing goal preferences

Unrotated
Factor Loadings
Communalities
Rehabilitation
.67
.45
General deterrence
.88
.77
Individual deterrence
.91
.82
Norm illustration
.86
.74
Restoration
.78
.61
Incapacitation
.83
.68
Retribution
.82
.66
5.2
Delict Experiences and Sentencing Goal Preferences
The next step is to find out whether the expectations crime victims have on the
goals of criminal sentences have been formed by their individual experience of
crime. This raises the question of whether there is a variation in the level of acceptance of various penal theories according to the kind of offence. With this
comes the question of whether certain penal theories systematically gain or lose
importance according to the increasing severity of the offence.
20
21
All correlations are significant with p .001.

The factor analysis was calculated as principal component analysis. All factors with eigenvalues > 1should be extracted.
438
H. Hirtenlehner
A range of variance analyses (Turner/Thayer 2001) was carried out to see if

there could be a connection between the preferences for various penal goals and
the offence committed. Three groups of offences were taken from the reported
crimes 22 : theft, burglary and violent offences 23 . The results indicate clearly that
the sentencing goal preferences are independent of the offence (table 4). No significant differences of p .05 could be detected. Victims of theft, burglary or
violence have the same expectations concerning the aims of criminal punishment 24 .
Theft can be ranked as a less serious offence than burglary or violence, so the
independence of the sentencing motives of the offence group can be taken as the
first evidence that the seriousness of the crime has no influence on the sentencing
preference. The findings of offence-specific correlation analyses show similar
results. For each of the three offence groups given above the product-moment
correlation coefficients of the sentencing goal preferences with the seriousness of
the offence were calculated. No significant correlation of p .05 was observed.
Victims of theft did not change their sentencing preferences according to the
value of what they lost. Victims of burglary were not influenced in their sentencing preferences by the value of their damages (the combined value of the goods
lost and the actual damage caused by the break-in). Persons affected by violence
did not link their sentencing preferences to the seriousness of the injuries 25 they
suffered.
Table 4:
Sentencing Goal Preferences by Delict
Rehabilitation
General deterrence
Individual deterrence
Norm illustration
Restoration
Incapacitation
Retribution
22
23
24
25
Theft
Burglary
Violent
offence
5.1
5.1
5.0
4.9
4.7
4.3
4.1
5.1
4.9
4.9
4.7
4.5
4.2
3.9
5.0
5.0
4.9
4.8
4.7
4.5
4.1
F
.08
1.30
.31
.92
.77
.99
.97
d. f.
2
2
2
2
2
2
2
.924
.275
.737
.400
.465
.373
.381
If a person reported more than one victimisation, the last offence was counted.
Because there were few cases, bodily harm and robbery were combined as the group of violent offences.
This result remains stable if the sociodemographic characteristics of the respondents (age,
sex, nationality and domicile) are taken into account as a control variable.
Victims of violent offences had to state whether they had been seriously injured, quite seriously injured, slightly injured or not injured at all during the crime they had reported.
439
All in all, we can conclude from the above findings that crime victims attitudes towards sentencing goals are in no way determined by the form or severity
of victimisation 26 . The victims of different crimes all have approximately the
same expectations concerning the goals of criminal punishment. This means that
the assumption that there is a different legitimacy of current penal theories depending on the kind of victimisation is incorrect, as is the thesis that a preference
for a certain penal theory is dependent on the severity of the offence.
6. Conclusion
Within the scope of this report, a study was presented on the legitimacy of penal
theories current in Austria at the turn of the millennium according to the expectations and needs that victims of crimes have. The findings of this research show
that a prevention-oriented amalgamation theory is preferred by persons who have
suffered damages through criminal acts and who have reported these crimes in an
urban area in Austria. People cannot see much difference between the various
sentencing goals. The elaborate differentiation given by jurisprudence to the different goals can only partly be understood by crime victims. All in all, the sentencing goals rehabilitation, general and individual deterrence and norm illustration were preferred over the goals of restoration, incapacitation and retribution.
Referring to the question of legitimacy posed in the introduction, the outcome is
that the preventive amalgamation theory finds the greatest level of agreement
among victims.
Restoration and restitution aspects lag behind the usual goals of prevention,
which could perhaps be accounted for by the fact that, socially and legally, there
is no tradition of considering victims interests in Austria. Conflicts in society
have been dealt with by penal law for hundreds of years, and during all that time,
the interests of victims have been ignored, which may have led to the legitimacy
of this neglect.
Retrospective retribution theories are still acceptable to the majority of the victims, but take last place on the scale of preferences. Retribution is the least respectable argument for the justification of criminal punishment. Retribution is
not basically rejected, but, in comparison with other, future-oriented goals of
criminal punishment, it seems to be the last choice. This low ranking of any idea
of retribution which is not linked to preventive aims is impressive, particularly
considering that the findings come from a sample population where there is more
incidence of serious offences than among victims of crime in general. We can
26
We have to remind the reader here that victims of sexual offences were not included in the
study.
440
H. Hirtenlehner
presume that the acceptance of prevention-oriented penal theories would be even

higher in a more representative sample of victims, where minor offences have
occurred more frequently.
The prevention-oriented amalgamation theory had a higher ranking in all the
groups of offences studied. Victims of theft, burglary and violence preferred the
same sentencing goals. In all three groups of offences, there was a demand for
prevention-oriented sentencing, restoration goals were slightly less highly ranked
and sentencing goals not linked to social objectives received a conspicuously
lower ranking. As the intensity of the offence (amount of damage and severity of
injury) was not relevant to the sentencing motives, the assumption that the sentencing goal preferences of crime victims are linked to the form and seriousness
of the victimisation must be rejected.
441
Appendix
Rehabilitation
Unrotated
Factor Loadings
Corrected
Item-Total
Correlations
Unrotated
Factor
Loadings
Corrected
Item-Total
Correlations
.76
.32
.84
.42
.51
.20
----(not
included)
----(not
included)
.83
.43
.84
.42
The sentence should have an educational effect on the offender.

An important goal for criminal law is
to re-integrate the offender in society.
The aim of a sentence should be to
rehabilitate the offender.
Eigenvalue:
1.53
Variance Explained: 51 %
Cronbachs D:
.49
General Deterrence
People should be afraid of punishment so they do
not commit offences.
It is necessary to punish an offender so that others
know what will happen to them if they commit such
an offence.
An important task of punishment is to deter others
from committing similar offences.
Eigenvalue: 2.31
The offender should be so shocked by the punishment that he thinks over his behaviour in future.
Punishment should be like a warning shot to keep
the offender from committing further crimes.
The punishment should show the offender what will
happen if he offends again.
1.42
71 %
.59
Unrotated
Factor Loadings
Corrected Item-Total
Correlations
.86
.70
.88
.72
.89
.75
Cronbachs D: .85
Individual Deterrence
Eigenvalue: 2.14
Eigenvalue:
Variance Explained:
Cronbachs D:
Unrotated
Factor Loadings
Correlations
.84
.63
.83
.63
.86
.67
Cronbachs D: .80
442
H. Hirtenlehner
Norm Illustration
Effective penal law is essential for the awareness of
justice in the population.
The punishment of offenders should aim to convince the population that the laws are reasonable.
Offenders should be punished to increase the sense
of right and wrong of the general public.
Eigenvalue: 2.00
Criminal law proceedings should deal with compensation for the victim, too.
The punishment given an offender should make it
clear to the victims that the offence was not their
fault.
The offenders punishment should raise the self
confidence and self esteem of the victim.
The longer an offender is imprisoned the better. At

least he cant be a danger to any other innocent
person.
Imprisonment is justifiable because the general
public is protected from the offender for a time.
Prison sentences should be given to keep offenders
away from society.
It is necessary to punish offenders to compensate

for the wrong they have done.
An eye for an eye, a tooth for a tooth should apply
to offenders too
Punishment should be retribution for the injustice
committed.
.45
.85
.63
.88
.69
Unrotated
Factor Loadings
Correlations
.73
.46
.83
.58
.81
.56
Unrotated
Factor Loadings
Correlations
.86
.69
.89
.74
.88
.72
Cronbachs D: .85
Retribution
Eigenvalue: 1.97
.71
Cronbachs D: .70
Incapacitation
Eigenvalue: 2.30
Correlations
Cronbachs D: .74
Restoration
Eigenvalue: 1.89
Unrotated
Factor Loadings
Unrotated
Factor Loadings
Correlations
.83
.57
.76
.50
.84
.59
Cronbachs D: .72
443
References
Arnold, H. & Korinek, L. (1991). Victimization, Attitudes towards Crime and Related
Issues, in: Kaiser, G., Kury, H., Albrecht, H.-J. (eds): Victims and Criminal Justice.
Volume 2. Freiburg i.Br.: Max Planck Institut. pp. 99 121.
Ashworth, A. (2002). Sentencing, in: Maguire, M./Morgan, R./Reiner, R. (eds.): The
Oxford Handbook of Criminology. Oxford: Oxford University Press. pp. 1076-1112.
Ashworth, A. & Wasik, M. (1998). Fundamentals of Sentencing Theory. Oxford: Oxford
University Press.
Bals, N. (2006). Tter-Opfer-Ausgleich Cui bono? Monatsschrift fr Kriminologie
und Strafrechtsreform, Vol. 89, pp. 131-145.
Blmmel, R. (2002). Der Opferaspekt bei der strafrechtlichen Vergangenheitsbewltigung. Berlin: Duncker & Humblot.
Burgstaller, M. (1983). Sinn und Zweck der staatlichen Strafe, in: Porstner, K. (ed.):
Strafrecht. Vergeltung oder Vershnung. Wien: Geyer Edition. pp. 53-55.
Child, D. (2006). Essentials of Factor Analysis. London: Continuum.
Cronbach, L. (1951). Coefficient alpha and the Internal Structure of Tests. Psychometrika, Vol.16, pp. 297-334.
Durkheim, E. (1907). Les rgles de la mthode sociologique. Paris: Felix Alcan.
(1911). De la division du travail social. Paris: Felix Alcan.
Easton, D. (1965). A Systems Analysis of Political Life. New York and London: Wiley.
Endres, J. (1992). Einstellungen zu Straf- und Sanktionszwecken und ihre Messung.
Monatsschrift fr Kriminologie und Strafrechtsreform, Vol. 75, pp. 309-320.
Fishman, J. & Galguera, T. (2003). Introduction to Test Construction in the Social and
Behavioral Sciences. Lanham: Rowman & Littlefield Publishers.
Habermas, J. (1988). Legitimation Crisis. Cambridge: Polity Press.
Hanak, G. (1982). Diversion und Konfliktregelung. berlegungen zu einer alternativen
Kriminalpolitik bzw. einer Alternative zur Kriminalpolitik. Kriminalsoziologische
Bibliographie, Vol. 9, pp. 1-39.
Hegel, G.W.F. (1821). Grundlinien der Philosophie des Rechts. Berlin: Nicolai.
Hilf, M. (2006). Der Strafrechtszweck der Restoration, in: Jesionek, U./Hilf, M. (eds.):
Die Begleitung des Verbrechensopfers durch den Strafprozess. Innsbruck: Studienverlag. pp. 13-22.
Jakobs, G. (2004). Staatliche Strafe: Bedeutung und Zweck. Paderborn: Schningh.
Jescheck, H.-H. & Weigend, T. (1996). Lehrbuch des Strafrechts. Allgemeiner Teil.
Berlin: Duncker & Humblot.
Jesionek, U. (2005). Das Verbrechensopfer als Prozesspartei, in: Bundesministerium fr
Justiz (eds.): 33. Ottensteiner Fortbildungsseminar aus Strafrecht und Kriminologie.
Wien und Graz: Neuer Wissenschaftlicher Verlag. pp. 41-70.
Johnstone, G. (2003). A Restorative Justice Reader. Texts, Sources, Context. Devon:
Willan Publishing.
444
H. Hirtenlehner
Kaiser, G. (1993). Kriminologie. Heidelberg: Mller.

Kant, I. (1797). Die Metaphysik der Sitten. Knigsberg: Friedrich Nicolovius.
Kaufmann, A. (1982). Strafrechtsdogmatik zwischen Sein und Wert. Kln: Carl Heymanns Verlag.
Kesteren, J. v., Mayhew, P. & Nieuwbeerta, P. (2000). Criminal Victimisation in Seventeen Industrialised Countries. Key Findings from the 2000 International Crime Victim Survey. Den Haag: DOHC.
Kilchling, M. (1995). Opferinteressen und Strafverfolgung. Freiburg i.Br.: Max-PlanckInstitut.
(2002). Opferschutz und Strafanspruch des Staates ein Widerspruch? Neue Zeitschrift fr Strafrecht, Vol. 22, pp. 57-63.
Krainz, K. (1991). Victimology in Austria, in: Kaiser, G./Kury, H./Albrecht, H.J. (eds.):
Victims and Criminal Justice. Volume 1. Freiburg i.Br.: Max-Planck-Institut. pp. 7194.
Kunz, K.L. (2004). Kriminologie. Bern: Haupt.
Kury, H. (1992). Der Einfluss der Art der Datenerhebung auf die Ergebnisse von Umfragen erlutert am Beispiel einer Opferstudie, in: Kaiser, G. & Kury, H. (eds.),
Kriminologische Forschung in den 90er Jahren. Beitrge aus dem Max-PlanckInstitut fr auslndisches und internationales Strafrecht. Freiburg i.Br.: Max-PlanckInstitut. pp. 321-410.
Kury, H., Drmann, U., Richter, H. & Wrger, M. (1992). Opfererfahrungen und Meinungen zur Inneren Sicherheit in Deutschland. Wiesbaden: Bundeskriminalamt.
Likert, R. (1932). A Technique for the Measurement of Attitudes. Archives of Psychology, Vol. 140, pp. 1-55.
Meier, B.D. (2001). Strafrechtliche Sanktionen. Berlin: Springer.
Momsen, C. & Rackow, P. (2004). Die Straftheorien. Juristische Arbeitsbltter, Vol. 36,
pp. 336-340.
Moos, R. (1989). Positive Generalprvention und Vergeltung, in: Melnizky, W. & Mller, O.F. (eds.): Strafrecht, Strafprozessrecht und Kriminologie. Festschrift fr Franz
Pallin zum 80. Geburtstag. Wien: Manz. pp. 283-318.
Mller-Tuckfeld, J. (1998). Integrationsprvention. Studien zu einer Theorie der gesellschaftlichen Funktion des Strafrechts. Frankfurt: Peter Lang.
Padfield, N. (1996). Sentencing: Theory, Law and Practice. London: Butterworths.
Roxin, C. (2001). Wandlungen der Strafzwecklehre, in: Britz, G. & Mller-Dietz, H.
(eds.): Grundfragen staatlichen Strafens. Festschrift fr Heinz Mller-Dietz zum 70.
Geburtstag. Mnchen: Beck. pp. 701-715.
Schwarzenegger, Ch. (1992). Die Einstellung der Bevlkerung zur Kriminalitt und
Verbrechenskontrolle. Freiburg i.Br.: Max-Planck-Institut.
Schwind, H.D. (2003). Kriminologie. Eine praxisorientierte Einfhrung mit Beispielen.
Heidelberg: Kriminalistik Verlag.
Sessar, K. (1992). Wiedergutmachen oder strafen. Einstellungen in der Bevlkerung
und der Justiz. Pfaffenweiler: Centaurus.
445
Shapland, J., Willmore, J. & Duff, P. (1975). Victims in the Criminal Justice System.
Aldershot: Gower.
Steffen W. (1976). Analyse polizeilicher Ermittlungsttigkeit aus der Sicht des spteren
Strafverfahren. Wiesbaden: Bundeskriminalamt.
Teske, R. & Arnold, H. (1991). A Comparative Victimization Study in the United States
and the Federal Republic of Germany: A Description of the Principal Results, in:
Kaiser, G., Kury, H. & Albrecht, H.J. (eds): Victims and Criminal Justice. Volume 2.
Freiburg i.Br.: Max-Planck-Institut. pp. 3-44.
Turner, J.R. & Thayer, J. (2001). Introduction to Analysis of Variance: Design, Analysis, and Interpretation. London: Sage.
Van Dijk, J.M., Mayhew, P. & Killias, M. (1991). Experiences of Crime across the
World. Key Findings from the 1989 International Crime Survey, Deventer: Kluwer.
Von Hirsch, A. (1976). Doing Justice. New York: Hill and Wang.
Von Hirsch, A. & Ashworth, A. (1998). Principled Sentencing: Readings in Theory and
Policy. Oxford: Hart Publishing.
Von Liszt, F. (1905). Strafrechtliche Aufstze und Vortrge. Band 1. Berlin: Guttentag.
Vo, M. (1991). Victims Expectations, Diversion, and Informal Settlement: Results of a
Victim Survey, in: Kaiser, G., Kury, H. & Albrecht, H.J. (eds.): Victims and Criminal Justice. Volume 3/1. Freiburg i.Br.: Max-Planck-Institut. pp. 67-93.
Zipf, H. (1989). Die Integrationsprvention (positive Generalprvention), in: Melnizky,
W. & Mller, O.F. (eds.): Strafrecht, Strafprozessrecht und Kriminologie. Festschrift
fr Franz Pallin zum 80. Geburtstag. Wien: Manz. pp. 479-490.
446
H. Hirtenlehner
447
Different Perspectives, Different Gaps

Does the General Public Demand a More Responsive Judge?
HENK ELFFERS and JAN W. DE KEIJSER
Abstract
Public dissatisfaction with the performance of the criminal justice system is often
connected to the degree to which judges are responsive to public opinion, especially concerning the severity of punishment. This paper examines to what extent
judges are and should be responsive to public opinion by discussing the results of
a survey amongst the general public in concordance with identical questions being posed to a sample of Dutch criminal judges. We distinguish between a real
gap between judges opinion and the publics opinion, a gap as perceived by
judges and a gap as perceived by the public. Findings show that the gap as perceived by judges is the largest gap of all. Rather than clamouring for more responsive judges, the public would rather have an independent and impartial judge
who isolates himself from public opinion when considering a criminal case.
Judges completely misperceive this state of affairs.
448
Henk Elffers and Jan W. de Keijser
Introduction
It is no secret that the general public is not always happy with the functioning of
the criminal justice system. Judges meet criticism first and foremost in matters of
severity of punishment, but at least in systems such as the continental where the
judge has to decide on guilt as well also decisions with respect to proof and
criminal liability regularly meet lack of understanding. Often, this state of affairs
is depicted as implying that a public wish exists that judges should be more responsive, i.e. that they should listen more to public opinion about a case at hand.
This article studies whether this identification of, on the one hand, public disconcert with, on the other hand, a wish for increased responsiveness is correct.
The present paper addresses the issue by clarifying the concept of responsiveness first. After discussing the literature, we introduce the concept of a gap between, on the one hand, what judges feel as an appropriate degree to which they
ought to be responsive, and, on the other hand, what the public expects from
judges in terms of responsiveness. For better understanding of the issues at hand,
we distinguish between a real gap, a gap as perceived by judges and a gap as
perceived by the public. We establish empirically the size of these various gaps
by means of two survey studies, one among 27 Dutch criminal judges, and one
among 529 members of the Dutch general public, embedding the findings in
various background characteristics of respondents.
The concept of responsiveness

The independence of the judge within the boundaries of law and jurisprudence
has always been the corner stone of the rule of law. However, in the second half
of the 20th century, a further desirable characteristic of the system and its judges
has been advocated: responsiveness to societal developments and public opinion.
The term, first having been introduced in the phrase responsive law, by Nonet
and Selznick (1978), was coined as useful for analysing how legislation and interpretation of law can or should react to changing societal circumstances and to
public opinion on such issues. A responsive institution, in their words, perceives
social pressures as sources of knowledge and opportunities for self-correction
(p. 77). Initially, the term was devised as pertaining to law or legislation, but it
can easily be adapted to be applicable to the judge or the court as instruments of
the law. A responsive judge is then, a judge who is aware of public opinion about
a case at hand, or about law in general, and who uses it as a source of knowledge
and information for making decisions.
Responsiveness of the judge can be seen as supporting legitimacy (Hough &
Roberts 1999): the public simply does not view legitimacy of the judicial system
Does the general public demand a more responsive judge?
449
as self-evident, it should earn its keep, and a demonstration of being responsive is

believed to be a powerful argument. Already the influential Dutch philosopher of
law Scholten (1931) argued that judges can only function when they are well
aware of the limits of their discretionary power. Society implicitly delineates the
boundaries of the judges decision space, Scholten stated:
The judge is an organ of society. His judgment is not an individual moral
judgment, but a power vested verdict, binding a community. Hence, he
should know the opinion of those who are subordinated to his power in law.
There is a limit to the degree in which he can act against those opinions: he
has to be aware of those limits, but within these he is a free agent (p. 128).
While Nonet and Selznicks treatise was in fact debating with a rigid bouche de
la loi-type theory of law and opening up the judges mind to the opportunity as
opposed to the necessity of learning from public views and social pressure,
later writers have gone a step further. Some of them (e.g., Huls, Mevis & Visscher 2003) are not satisfied with an open-minded judge as such; they indeed require him to actually follow public pressure. Though it is not quite clear how
far those writers will go in weighing the influence of public opinion in comparison with other arguments, such as the law or the conviction of the judge, they
explicitly demand a leaning towards public pressure.
A conflict between the values of responsiveness and independence and integrity
is looming here (cf. Nonet & Selznick 1978: 77). If responsiveness would lead to
abandoning (formal) independence and unbridled discretion in following popular
notions that are here today but gone tomorrow, it would be counterproductive as
legitimising mechanism. The responsive judge is thus confronted with fundamentally competing values: he should take account of new forces and pressures in his
environment whilst retaining a grasp on what is essential to his integrity.
Obviously legitimacy as a result of responsiveness may be contrasted, or, rather,
be compared to legitimacy from the perspective of procedural justice. In the latter
case legitimacy is primarily connected to the inherent qualities of legal procedure
and interaction between process participants. Legitimacy from the perspective of
procedural justice is predominantly determined by peoples perceptions of fairness
of legal procedures, and process participants evaluation of their treatment by legal
authorities (cf. Lind & Tyler 1988; Thibaut & Walker 1975; Tyler 1988), more so
than by the mere outcome of a legal procedure or even by the degree to which outcomes echo societal developments or public opinion. The primary focus of the present study, however, is on responsiveness as legitimising mechanism, on judges and
the publics perceptions of judges responsiveness to society.
450
As key element in a definition of responsiveness, we identify the fact that the

judge should be aware and prepared to take stock of public opinion in society,
and ready to account for his decisions, even when these are contrary to the opinion in society: Responsive means to be able to explain legal decisions and procedures to a larger public and being able to show that public opinion has been
accounted for or, at least, has been taken into account. Whether the judge is successful in convincing the public should not be an element of this definition, in
our opinion.
Modern judges appear to value responsiveness as a legitimising mechanism. It
is not uncommon for judges to mention public opinion in explaining a sentencing
decision, particularly in cases that enjoy wide media attention. Indermaur (1990:
50) found that 14 out of a sample of 17 Australian judges and magistrates interviewed believed that public attitude should be considered by the sentencer in all,
most or some cases. De Keijser, Van de Bunt & Elffers (2004) observed among
Dutch judges a general adherence to the idea that they do well in taking notice of
public opinion. In many Western jurisdictions, at present, public opinion is thus
said to play an important (indirect) role in sentencing policy and practice (Hough
& Roberts 1999; Roberts, Stalans, Indermaur & Hough 2003; Hutton 2005).
Public disconcert about the criminal justice system

The whole responsiveness as a means for legitimacy discussion is based on the
idea that judges and the public will generally have rather distinct views on the
desirability of various judicial decisions, that there indeed exists a gap between
them. But is that a fact? We aim to scrutinize that commonly held idea in this
paper, and investigate size and nature of the gap between judges and the public,
when it comes to responsiveness in criminal law. Both in the Netherlands and
abroad, authors have diagnosed that the public is not very satisfied with the way
the criminal law system is facing crime. In the Netherlands, De Roos (2000) diagnosed the publics view on the judiciary as a deeply rooted unease. Roberts
and colleagues (2003) observed comparable phenomena in Australia, Canada, the
UK, the USA and New Zealand. Distrust seems to be concentrating on public
distress about the level of punishment (i.e. lack of severity) in criminal cases.
The Dutch Social and Cultural Planning Bureau (Sociaal en Cultureel Planbureau
2002) as well as Van Koppen (2003) showed that an overwhelming majority of
the Dutch public endorses the statement that criminal punishment tends to be
much too lenient; Parmentier et al. (2005) showed comparable results for Belgium. So, it is well documented that the general public is not very much satisfied
with what the criminal judges deliver, and especially so on the severity of punishment meted out to convicted criminals.
451
However, this well-documented state of affairs does not necessarily imply that
the general public demands that the judges should lean more to the opinions voiced
in society in general. Such a conclusion may be too rash. The existence of an unequivocal public opinion may be challenged. Furthermore, it is unclear whether the
public demands that judges should bend to public opinion.
The public opinion on criminal justice?

First the idea about the existence of a well delineated public opinion can be challenged. Indeed, Robert & Faugeron (1978) and Rood-Pijpers (1988) have demonstrated that marked differences exist between opinions of members of the general
public on crime, judicial processes, and punishment. These authors distinguished a
number of homogeneous classes within the public that share within a class attitudes
and opinions about criminal justice (called representations by these authors), while
between classes there exist sturdy differences. So, the idea of the public opinion is
rather coarse.
Moreover, it has been challenged that public opinion is unequivocally requiring
harsher treatment. In recent years research has accumulated showing how outcomes
of penal attitude measurements are to a large extent determined by questioning technique and context provided (cf. Durham III 1993; Green 2006; Hough & Park 2002;
Tonry 2004; Hutton 2005; Roberts, Stalans, Indermaur, & Hough 2003; Stalans
2002; Walker & Hough 1988). From research on public opinion regarding punishment severity we know that there exists a huge difference between judgements in
general and in specific cases. With regard to global questions about sentencing, people tend to respond with very serious crimes such as murder or rape in mind (cf.
Roberts et al. 2003). However, when further questioned about punishment in concrete and less serious criminal incidents, in which they get detailed information
about the circumstances of the crime, the perpetrator and the victim, they may end
up much closer to what criminal judges tend to mete out as punishment (Ashworth
& Hough 1996; Kuhn 2002; Seidman-Diamond 1990). This state of affairs again
shows that the phrase public opinion is overly simplistic: opinions may change
when more or less information is known to the public. Though a general longing for
harsh punishment exists, this by no means needs to imply that in concrete instances
the public is demanding harsh(er) punishment.
Does the public demand more responsiveness?

Our second concern is whether the public indeed desires judges to follow public
opinion. Judges do more than punish the guilty: in a continental system they are
there as well to establish guilt, safeguard the accused against wrongful conviction
or unjustified prosecution as well as against unlawful methods in the hands of the
452
prosecution. There seems to be a tendency to assume that the public looks at a

criminal court case as one in which a surely guilty person has to be punished. But
is that not a gross exaggeration? It may be true that the ordinary laymen is not
well versed in the niceties of penal law, but the sheer realisation that a criminal
case is above all on establishing guilt, is certainly not out of reach for most people. They indeed are well helped by popular detective fiction, in which the innocently accused hero or heroine is the character most frequently popping up. So, it
seems to us a gross simplification to portray the public as having just one demand to judges: harsher sentences, please. Indeed, Walker et al. (1988) as well
as Indermaur (1990) identified substantial support among the public for a judiciary not leaning to public opinion, but maintaining an independent stand.
Researching the gap

Summing up, we feel a need to investigate the extent to which there exists a gap between judges and the public regarding judges responsiveness. What does the public
demand from judges? Is it different from what the judge feels apt to deliver?
Figure 1: Gap from three points of view
Gap from three points of view
en
se lic
as pub
p
ga the
by
en
se s
as dge
p
ga y ju
b
real gap
Figure 1
We will describe what public and judges think in this respect, and then interpret
the findings normatively. The basis of our research is comparing the results of a
453
survey among the general public, with a parallel questionnaire among criminal
judges.
The concept of a gap between public and judges deserves a closer look. We try
to get insight in perceptions of judges on their responsiveness and in the perceptions of the public on judges responsiveness. However, we should realise that
the groups on either side of the gap have only a dim view of each other and do
not have direct access to what the other side actually thinks or perceives. Figure 1
depicts the rather complex situation.
The difference between, on the one hand, what judges actually think of their
own professional groups responsiveness, and on the other hand, what the public
really thinks about judges responsiveness forms the real gap. We should be
aware, however, that both groups do not have direct access to the other sides
views. Judges have a perception of what the public thinks about them that may or
may not be accurate, while also the public may or may not have a good idea of
how judges see themselves. In fact, the groups at either side of the gap base their
reactions and ideas on a perceived gap, which need not coincide with the real
gap. So, apart from the objective real gap, we distinguish two perceived gaps: a
gap as seen by the judges, and a gap as seen by the public. We stress that how
judges cope with the existence of a gap, and how they feel they should react on
it, is in fact based on the gap as seen by them, not on the real gap, as this last one
is unknown to them. Likewise, the publics notion on whether a gap between
their opinion and the judges has to be qualified as problematic or not, is based on
the gap as seen by the public, not on the real gap that is not known to the public.
It is exactly for this reason that we think it worthwhile to chart out the various
gaps, and study their sizes.
Table 1:
Perspectives on the gap
Descriptive
How responsive are judges?
real gap
gap as seen by judges
gap as seen by the public
Normative
How responsive should judges be?
real gap
gap as seen by judges
gap as seen by the public
We may further distinguish between what people (and judges) think about what
is the case (description) and what they deontologically think should be the case
(norm). All three gaps, i.e. real gap, perceived from the one side, perceived from
the other side, may be examined from both a descriptive and a normative point of
view. The six resulting types of gap are schematically presented in Table 1.
454
Research design
The present research relates an empirical study among judges to a survey among
the general public. The first is an offshoot of a qualitative group interview study
among Dutch criminal judges (De Keijser et al. 2004). The second study is a survey among a representative sample from the Dutch population. In both studies a
block of identical questions was incorporated in order to be able to contrast
judges and publics own opinion as well as mutual perceptions on exactly the
same issues. As discussed above, for both groups of respondents we wanted to
discern descriptive perceptions on responsiveness issues as well as normative
opinions. The following set of statements was constructed. The statements designated by D are descriptive statements, while statements designated by N are normative. 1
D1
Judges do not try hard enough to explain their decisions to the common man
D2
Judges decide too often in a way unacceptable to the ordinary citizen
D3
The Dutch judge lives in an ivory tower
D4* Dutch judges are out of touch with society

N1
In order to safeguard their independence, judges ought to isolate themselves from

public opinion
N2
Judges should not pay attention to the publics opinion on a case but on characteristics of the case itself
N3
Judges cannot prevent that some of their decisions will not be understood by the
general public
N4* When the public is outraged about a particular offence, the judge should not sentence that offence harsher
N5* It is more important that a verdict is legally flawless than that it is accepted in society.
Concerning these statements, we requested from our sample of judges and from
our sample of the general public:
1. to indicate to what extent one agreed or disagreed, using a seven-point scale
running from completely disagree to completely agree,
2. to indicate how one expects the average member of the other group will respond to the statement. Specifically, we asked the general public:
How do you suppose Dutch criminal judges would respond to this statement?
And we asked judges:
How do you suppose the average member of the Dutch public would respond to
this statement?
1
Statements designated by * have been mirrored from the original formulation.
455
In discussing the results below we will refer to the questions regarding ones
own opinion as introspective questions, whereas the questions tapping into perception of what the others would answer will be referred to as projective questions.
General public questionnaire

The general public was approached in late 2003 through the existing CAPI-athome-panel of the Netherlands Institute for Public Opinion and Marketing Research NIPO. This standing panel consists of a representative cross section of the
Dutch public over 18 years old, and is fortnightly approached by NIPO with
various surveys. Questionnaires are presented to and can be completed by the
respondent on his home computer; in a typical wave over 90% of the panel
members approached respond to any given questionnaire. The panel attrition over
years is low. A random sample from that panel has been drawn (N=529). 2
The questionnaire consisted of 5 parts:
(a) responsiveness statements: introspective, i.e., own opinion
(b) responsiveness statements (cf. above): projective, i.e., perception of judges
opinion
(c) evaluation marks given by respondent to police, public prosecutors, judges
(d) a listing of typical characteristics that a good judge should have
(e) general statements about the performance of the criminal justice system
Background demographics are routinely available from the panel data base.
Judges questionnaire
Data from judges were gathered in Spring 2003 as a side-product from a group
interview study among 27 criminal judges from 5 primary courts and all 5 courts
of appeal (reported in another chapter). Strictly speaking, the sample of judges is
not a random sample from all judges serving in primary or appeal courts, but a
selection (stratified over the 5 appeal court jurisdictions) of the first 30 of them
that stepped forward as being interested in the topic. As we have no indication
that being interested in the subject will be associated with the content of views
2
Actually, the sample had been drawn as a stratified sample, stratifying according to whether
the respondent had ever in his life visited a court room a characteristic known from a previous study on the panel by Malsch (2005), and kindly made available to us by her. As
analyses, however, did show that with respect to background variables no important differences did occur between strata, and moreover on the responsiveness variables no marked
differences occurred either, we present, for convenience sake, all analyses as if the sample
had been drawn as a simple random sample (n=529) from the population of all Dutch over
18. A total of 146 respondents had previously visited a court room, 383 had not.
456
held on the subject, we analyse the data as if they form a random sample of
judges.
The questionnaire consisted of two sections:
(a) responsiveness statements: introspective, i.e., own opinion
(b) responsiveness statements (cf. above): projective, i.e., perception of Dutch
public opinion.
Analysis
All statements were scored on scales ranging from 1 completely disagree to 7
completely agree. Our analyses are focused on mean scores. On average, the
items standard deviations for the publics sample are all in the order of 1.75
(maximum SD is 2.04). Standard deviations are slightly lower for projective
items. Both for introspective and for projective items the standard deviations in
the judges sample are all slightly lower. Taking into account Cohens well
known criterion for calling differences between means of moderate size when
they are surpassing times the standard deviation (Cohen 1977), we propose to
consider differences sizeable as soon as they surpass a full point of the scale.
Given that a 4 on a seven point scale represents a neutral position (neither
agree, nor disagree), a mean score under 3 is considered to indicate rejection,
between 3 and 5 is indecisive and a score over 5 is considered endorsement.
When we compare means, we will interpret them as different only if they both
are statistically significant at significance level 3 = 0.01, as well as surpassing a
size of 1 point absolutely. Table 2 sums up all differences investigated. The figures that are presented below are all based on the data incorporated in Table 2.
The descriptive gap

Let us first look into the descriptive gap, i.e. the degree to which judges and public agree or disagree about the current level of responsiveness of judges. Figure 2
shows the mean scores of the four descriptive items in the introspective mode.
As we compare quite a number of dependent means, we like to take a rather strict significance level , to safeguard ourselves to some degree against chance capitalization.
457
The real descriptive gap

Notice, first of all, that in Figure 2 high scores represent a non-responsive judge.
The figure shows that on three of the four items judges do rate themselves as
highly responsive: they see their own professional group as in touch with society,
not at all in an ivory tower, and not at all taking unacceptable decisions. They do,
however, plead guilty with respect to their unpalatable way of issuing verdicts;
they are well aware that these are not clear to the commoner. On the other hand,
we see that the public has a rather moderate view on the judge: yes, he is a bit
isolated and not really responsive, but certainly not extremely so. The not in
touch item is even scored below the neutral point. Only on the same issue of
explaining decisions to the public, the public heartily agrees that judges are not
icons of responsiveness.
Figure 2: Descriptive responsiveness gap the real gap
Descriptive responsiveness gap
the real gap
agree

5
public opinion
(n=526)

disagree
judges' opinion
(n=27)
1
do not
explain
enough
too often
unacceptable
verdicts
in ivory
tower
unaware
of public
opinion
In summary, comparing the two points of view that define the real descriptive
gap, we observe a marked difference on three out of four items (difference larger
than 1 point and statistically significant, as indicated earlier), while judges and
the public completely agree on the explanation deficit. Thus, descriptively speaking, there is indeed a real gap between judges and the public.
458
Table 2:
Means and significances

Publics Score
(n=526)
Judges Score
(n=27)
Significant differences
(p < .01)
(0)
1
introspective
2
projective
3
introspective
4
projective
1-2
paired
sample
1-3
two
samples
1- 4
two
samples
2-3
two
samples
3-4
paired
sample
Judges do not try hard enough to explain their decisions to

the common man
5.67
3.92
5.56
6.11
Judges decide too often in a way unacceptable to the

ordinary citizen
4.81
2.96
2.30
4.59
The Dutch judge lives in an ivory tower
4.37
2.98
2.22
4.81
Dutch judges are out of touch with society
3.72
2.31
2.44
4.59
In order to safeguard their independence, judges ought to

isolate themselves from public opinion
5.35
5.54
2.41
2.67
Judges should not pay attention to the publics opinion on a

case but on characteristics of the case itself
5.32
5.76
3.26
2.07
Judges cannot prevent that some of their decisions will not

be understood by the general public
5.12
5.79
5.78
3.22
4.58
5.50
5.00
1.93
4.32
5.47
4.00
2.56
Overview of means and significance of differences

DESCRIPTIVE ITEMS
* When the public is outraged about a particular offence,

the judge should not sentence that offence harsher
* It is more important that a verdict is legally flawless than
that it is accepted in society
(0) Student t-tests for equality of means;
* significant; not significant.
H. Elffers and J. W. de Keijser
NORMATIVE ITEMS
459
The descriptive gap according to judges

In Figure 3 the projective scoring by judges is added to the data from Figure 2.
The line added in the figure thus represents how judges expect the public to have
scored on the descriptive items. Figure 3 shows a rather nice coincidence of
judges projective ideas and the publics real (introspective) idea. On three items
no (statistically) significant difference could be established at all, while on the
not in touch item the difference is significant (at 0.01 level), but not very substantial with a size of 0.82, which falls short of our criterion as established above.
Indeed, we may say that judges have a pretty good image of the publics ideas;
hence the descriptive gap according to judges is concordant with the real gap.
Figure 3: Descriptive responsiveness gap, real gap & gap as seen by judges
real gap & gap as seen by judges
agree

public opinion
(n=526)
judges' opinion
(n=27)

disagree

public acc. to judges

(n=27)
1
do not
explain
enough
too often
unacceptable
verdicts
in ivory
tower
unaware
of public
opinion
Conversely, in Figure 4, the publics projective scoring of the descriptive responsiveness items is added to the data of Figure 2. The line added in Figure 4 now
represents how the public expects judges to have scored on these descriptive
items. The figure shows that the public is also rather well aware of judges opinion. For three out of four items we again observe no (statistically significant) difference: the gap according to the public is much like the real gap. Only on the
explanation-item, the public misperceives the judges position. The public has no
idea that judges are so self-critical in this respect.
460
The descriptive gap according to the public

Figure 4: Descriptive responsiveness gap, real gap & gap as seen by the public
real gap & gap as seen by the public
agree

5
disagree
public opinion
(n=526)

judges acc. to public

(n=526)
judges' opinion
(n=27)
1
do not
explain
enough
too often
unacceptable
verdicts
in ivory
tower
unaware
of public
opinion
Conclusion about the descriptive gap

In summary, a descriptive gap concerning responsiveness does exist. Moreover
the parties on either side of this gap are well aware of each others position, except on the issue of insufficient explanation of verdicts to the common man:
judges know they are not delivering in this respect, while the public would be
surprised to hear that judges are so self-critical.
The normative gap

Turning now to the normative items, on what the judge should do in terms of responsive behaviour, we start with the real gap (based on introspective scores),
between how judges see their own professional group and how the public rates
judges in this respect.
The real normative gap

Figure 5 shows the findings concerning the introspective scores of the public and
of judges on the normative items. Notice that again a high score indicates an isolated judge, a low score a responsive one.
461
Figure 5: Normative responsiveness gap

Normative responsiveness gap
real gap
agree

disagree

public opinion
(n=526)
opinion
judges'
(n=27)
1
ought to
isolate
themselves
ought to
mind case,
not public
misunderstanding
inevitable
punishment
lawfulness
irrespective
above
of indignation acceptation
Looking at the publics (introspective) scores on the normative items in Figure 5,

we observe that the Dutch public does, in fact, not demand a very responsive
judge: all five items get mean scores above 4, and indeed three of the five items
are indicative for a heartily endorsement of a form of isolation: the public likes to
see that judges do insulate themselves from public influence, concentrate on the
characteristics of the case itself, even at the cost of some misunderstanding of
their verdicts. Though not as markedly as these three items, the other two are also
scored on average above the neutral scale-level: the public does not at all clamour
for judges meting out levels of punishment responsive to public indignation, or for
judges valuing public acceptation more than they do juridical craftsmanship.
On the last three items in the figure judges largely agree with the public: no punishment guided by public indignation, even at expense of some misunderstanding,
though they value craftsmanship just as much as acceptance. However, judges have
the opinion that in order to reach those desiderata, they need not insulate themselves
or neglect public opinion, quite in line with what the same judges had discussed in a
round-table-session on the topic (De Keijser et al. 2004).
462
Comparing judges and the public, the isolation and case, not public items
exhibit a considerable difference (larger than 1, significant), while the other items
do not. In terms of a gap, this means that on these two items there is a gap in the
unexpected direction: judges score normatively more responsive than the public
does. For the other three items, no real normative gap exists. We see that in reality judges and public agree on the desirability of a not-too-responsive judiciary,
but judges are quite more optimistic that that can be reached without a necessity
for an all-too-splendid isolation.
Figure 6: The normative gap according to judges
real gap & gap as seen by judges
agree

disagree

1
ought to
isolate
themselves
ought to
mind case,
not public
opinion
public
(n=526)
misunderstanding
inevitable
punishment
lawfulness
irrespective
above
opinion
judges'
(n=27)
to judges
public acc.
(n=27)
We now turn to the normative gap as seen by judges, i.e. the difference between
how judges rate themselves (introspective scores) and how they perceive that the
public would rate them (projective scores). For that purpose, in Figure 6, we add
to the data of Figure 5 judges projective scores on what they believe the public
would score. Figure 6 shows that judges perceive a considerable normative gap.
Incorrectly so judges believe that the public is simply craving for responsiveness.
Almost all projective items have a mean score under 3. Especially regarding public influence on sentences and the need not to insulate themselves, judges fear
463
that the public will have a rather extreme demand for responsiveness. On all five
issues there is a considerable (large and significant) difference with what the
public actually demands. Judges totally misperceive the public demand for responsiveness.
Figure 7:
The normative gap according to the public

real gap & gap as seen by the public
agree

disagree

1
ought to
isolate
themselves
ought to
mind case,
not public
opinion
public
(n=526)
misunderstanding
inevitable
punishment
lawfulness
irrespective
above
to public
judges acc.
(n=526)
opinion
judges'
(n=27)
Conversely, we look into the publics perception of what judges would score by
adding the publics projective scores to the basic data from Figure 5. In Figure 7
it can be observed that the publics projective scores are slightly more towards
the isolated position of judges than the publics introspective scores. The difference and hence the gap as perceived by the public is small. However, this is to a
large extent a misperception: peoples projection of what judges score is biased
towards insulation: on three out of five items (juridical flawlessness over acceptation, need for isolation and concentrate on case instead of on public opinion) judges actually score much more responsive than the public expects (i.e.
perceives) them to be.
464
Conclusion about the normative gap

Neither the judge nor the public desires a very responsive judiciary. Both value
the independent position of the judge who should not let his judgment be influenced by public indignation, but should instead anchor himself in juridical
craftsmanship. The public is prepared to accept a certain amount of judicial insulation. Judges, however, do not even believe that that would be necessary. Judges
are completely unaware that the public does not at all demand a highly responsive judge. Judges thus perceive a huge gap here, which in fact does not exist.
The public, on the other hand, overestimates the degree to which judges would
like to insulate themselves. So, in summary, the presumed and feared normative
gap on responsiveness exists in two ways: i) it is in part a real gap but in the opposite direction than expected, and ii) as far as a perceived gap exists, it is in the
mind of judges, not in the mind of the public.
Relation with background characteristics

Our analyses have been on the level of aggregate opinions. Underlying the central tendencies reported there is a considerable disagreement between people
(standard deviations of about 1.75). This implies that a minority group of a considerable size within the Dutch public does not see judges as responsive enough
and expects judges to respond more clearly to public dissatisfaction. Indeed, in
the last resort the legitimacy of the judge as a necessary part of democratic society is threatened by the existence of a minority that routinely cannot endorse
judges decisions. In this respect, we should be aware of the fact that our demonstration of a majority acceptance of a not-too-responsive judge is by no means a
reason to decide that no problems exist at all.
We analysed the association between background characteristics and the publics descriptive and normative (introspective) rating of judges, by looking into
the average score for the four descriptive items, respectively the average score
for the five normative items, using t-tests for high and low scoring respondents
on various background variables. Grossly speaking, almost no such association
exists. Higher educated people do somewhat stronger prefer an independent, isolated judge, and in fact rate him descriptively lower on this topic. Men do rate
judges descriptively slightly more responsive than women do; those engaged in a
paid job do rate the judge more responsive than jobless people. Religious denomination is not important, neither is degree of urbanisation. Rather remarkably, having personal experience in court does not matter either. So, though we
have observed a rather high variation in the public opinion on responsive judges,
465
we cannot explain these differences in terms of fairly standard socio-demographic variables, nor in terms of previous experience with the law.
Responsive enough, but too lenient: overall satisfactory

Having seen that the public is not at all hoping for an overly responsive criminal
judge, we can wonder whether this state of affairs indicates that the public is satisfied with the criminal law system as such. That, however, is not the case. Just
as in the literature reviewed above, people in our sample endorse the statement
that Generally speaking, criminal offences are punished too lenient (83%
agrees), moreover a clear majority (82%) endorses that The criminal justice system is slow and inefficient and that Too many crimes remain unpunished as a
consequence of errors made by the prosecution or by judges (70%). Thus, although people appear to be dissatisfied with the criminal law system in toto, they
do not seek, in majority, an improvement through a judge who would listen more
to the public.
This seemingly contradictory state of affairs may be clarified by considering
the publics awareness of conflicting values embodied in the criminal justice system. When asked to ponder upon the relative severity of the two basic errors that
can be made in judicial decision making (i.e. false positive decisions and false
negative decisions), 81 percent of our sample agrees that It is better to let a
guilty person go unpunished than to condemn an innocent one. However, after
increasing the stakes, only half of our sample agrees that It is better to let ten
guilty persons go unpunished then to condemn an innocent one. Albeit with
some caution, we infer from these findings that people are aware of a fundamental trade-off: the stricter the focus on punishing the guilty the bigger the chance
of sending innocent persons to prison. However, our findings reported above indicate that the general public seems quite happy to leave this process of weighing
and deciding on such difficult issues in concrete cases to an independent and
well-trained professional judge.
When evaluating the criminal judge other values than responsiveness and punitiveness are also considered by the public, and may turn out to be more important. This is supported by responses to a straightforward question that we posed
to our sample. We asked respondents to rank order 10 typical traits that a criminal judge may be supposed to have. Table 3 shows that the traits Just, Impartial and Independent are top-ranked. Majorities ranking from 71 to 92 percent
include these traits amongst the five most important ones. Remarkably the trait
Austere appears at the lower end, being included in the top-5 by 22 percent of
466
our sample, and mentioned as the most important trait by only 2 percent. Being
Aware of public opinion receives roughly the same verdict from the public,
being included in the top-5 by one third and mentioned as the most important by
a mere 3 percent.
Table 3:
Rank ordering of important traits that a criminal judge should have:

among five most important traits or ranked as the most important
trait (percentages)
Trait of judge
Just
Impartial
Independent
Critical
Good listener
Law specialist
Able to see things from point of view of victims
Aware of public opinion
Austere
Able to see things from point of view of perpetrators
ranked among
the five most
important
92
82
71
54
52
50
40
33
22
4
ranked as
the most
important
42
22
14
2
4
3
7
3
2
0
Earlier, in defining the concept of responsiveness, we contrasted responsiveness

with procedural justice as legitimising mechanism. As far as legitimacy of the
judge is concerned, our findings now suggest that the traits that are considered
most important by the public are closer to procedural justice than they are to responsive justice. Despite a certain degree of dissatisfaction with the criminal justice system in general, the public endorses a just, impartial and independently
operating criminal judge much more than a responsively austere one. It is no
wonder then that, while people voice dissatisfaction with the severity of judges
decisions when it comes to punishment, they evaluate the performance of criminal judges as overall satisfactorily (86%), which remains a better overall evaluation than either the prosecution service or the police get (overall satisfactory, respectively 65% and 70%).
In summary, as the functioning and effectiveness of the criminal court system
is a remote and opaque business to the general public, people tend to wish in an
abstract way for a more effective system, but without being prepared to sacrifice
fundamental principles such as fair trial and protection of the rights of the innocent. In that sense, the public is conversant with the fundamental dilemma of the
responsive judge. The contradiction as observed above is in a way a luxury: the
public, as outsiders, can allow itself the luxury of this contradiction, not facing
the necessity to make decisions in individual cases.
467
Discussion
Judges and the public do disagree somewhat on whether judges are responsive or
not: judges are more optimistic in this respect than the public, and both are reasonably aware of the others position. That is as might have been expected:
members of a professional group are rather more satisfied about their own ability
to communicate than outsiders.
Looking at the normative aspect, we consider the findings surprising: in many
ways no gap exists, and with respect to some points, the public is really asking
for a less responsive judge than judges themselves do. This is unexpected. People
perceive judges as rather unresponsive and slightly isolated, but do evaluate that
as correct and inherent to the position of a judge. People see a certain isolation
and are happy with it. Victims of the gap problem are not so much the public, but
rather the judges: they believe that the public wants them to be much more responsive, but they have a completely wrong perception: that is not true.
How can we explain the existence of the judges misperception of public preferences? We have to realise that the picture sketched above is a picture about the
majority of the public. However, in a public in majority endorsing an insulated
position of the judge, there is of course also a minority voting for the opposite,
i.e. some people would like to have (extremely) responsive judges. So, judges are
rightly afraid of the existence of such people, but they overestimate considerably
their prevalence. It is here that the role of the mass media is interesting. After a
verdict, mass media typically cite people who are dissatisfied with the outcome
of the court proceedings, but the conclusion (either implied by the media or at
least perceived by the judge) that mentioning of dissatisfaction proves the existence of massive dissatisfaction, is simply wrong. In this respect, judges are
themselves the victims of the media. The public is not taking mass media critical
coverage of court proceedings lock, stock and barrel as indicating a general malfunctioning of judges. It is awkward indeed that Roberts et al. (2003: 85) showed
that judges endorsing the need for responsiveness declared to obtain knowledge
about reigning public opinion from newspapers exclusively. Judges indeed appear to have a very negative outlook on the media (e.g. Beyens 2000). Perhaps
we can say that the public is more resistant to media amplification than the
judges are. We mention that this is quite in line with Ditton et al. (2004), who
observed that in shaping their attitudes, peoples own perception and interpretation are dominant over media images. Anyway, a closer look into the actual role
of the media between judge and public seems a proper topic for further research.
What we have shown for now, is that the Dutch public does not demand a much
more responsive criminal judge. Simultaneously, we should not close our eyes
468
for the fact that the public does voice quite some criticism on the criminal law
system as a whole. The public sees no need for a more responsive judge, but it
does want a better criminal law system.
References
Ashworth, A. & Hough, M. (1996). Sentencing and the climate of opinion. Criminal
Law Review, 776-787.
Beyens, K. (2000). Straffen als sociale praktijk: Een penologisch onderzoek naar
straftoemeting. Brussels, VUBpress.
Cohen, J. (1977). Statistical power analysis for the behavioral sciences (3rd ed.). London, Academic Press.
De Keijser, J.W., Van de Bunt, H.G. & Elffers, H. (2004). Strafrechters over
maatschappelijke druk, responsiviteit en de kloof tussen rechter en samenleving. In J.
W. De Keijser & H. Elffers (Eds.), Het maatschappelijk oordeel van de strafrechter:
De wisselwerking tussen rechter en samenleving, 21-51. The Hague, Boom
Juridische Uitgevers.
De Roos, T.A. (2000). Het grote onbehagen: Emotie en onbegrip over de rol van het
strafrecht. Amsterdam, Balans.
Ditton, J., Chadee, D., Farrall, S. & Gilchrist, E. (2004). From imitation to intimidation: A note on the curious and changing relationship between the media, crime and
fear of crime. British Journal of Criminology, 44, 595-610.
Durham III, A.M. (1993). Public opinion regarding sentences for crome: Does it exist?
Journal of Criminal Justice, 21, 1-11.
Green, D.A. (2006). Public opinion vresus public judgment about crime: Correcting the
'comedy of errors'. British Journal of Criminology, 46, 131-154.
Hough, M. & Park, A. (2002). How malleable are attitudes to crime and punihsment?
Findings from a British deliberative poll. In J. V. Roberts & M. Hough (Eds.),
Changing attitudes to punishment: Public opinion, crime and justice (pp. 163-183).
Cullompton, Devon: Willan.
Hough, M. & Roberts, J.V. (1999). Sentencing trends in Britain. Punishment and Society, 1 (1), 11-26.
Huls, N.J.H., Mevis, P.A.M. & Visscher, N. (2003). De kloof tussen rechtspraak en
samenleving: Hoe klantvriendelijk is de rechter?, Justitile Verkenningen: Rechter en
samenleving, 29(1), 11-31.
Hutton, N. (2005). Beyond populist punitiveness? Punishment & Society, 7(3), 243-258.
Indermaur, D. (1990). Attitudes to crime and sentencing: A comparison of court practice and the perceptions of a sample of the public and judges. W.A. College of Advanced Education.
469
Kuhn, A. (2002). Public and judicial attitudes to punishment in Switzerland. In J. V.

Roberts & M. Hough (Eds.), Changing attitudes to punishment: Public opinion,
crime and justice, 115-127. Cullompton, Devon, Willan.
Lind, E.A. & Tyler, T.R. (1988). The social psychology of procedural justice. New York,
Plenum Press.
Malsch, M. (2005). Het grote publiek en de strafrechtspraak: De belangstelling onder de
bevolking. In M. Malsch & J. F. Nijboer (Eds.), De zichtbaarheid van het recht:
Openbaarheid in de strafrechtspleging (pp. 83-97). Deventer: Kluwer.
Nonet, P. & Selznick, P. (1978). Law and society in transition: toward responsive law.
New York, Harper & Row.
Parmentier, S., Vervaeke, G., Doutrelepont, R. & Kellens, G. (Eds.). (2005). Public opinion and the administration of justice. Brussels, Politeia.
Robert, P. & G. Faugeron, C. (1978). La justice et son public: Les reprsentations sociales du systme pnal. Geneva, Masson.
public opinion: Lessons from five countries. New York, Oxford University Press.
Rood-Pijpers, E. (1988). Mensen over misdaad en straf. Arnhem, Gouda Quint.
Scholten, P. (1931). Mr. C. Assers handleiding tot de beoefening van het Nederlands
burgerlijk recht: Algemeen deel (3rd ed.). Zwolle, Tjeenk Willink.
Seidman-Diamond, S. (1990). Revising images of public punitiveness: Sentencing by
lay and professional English magistrates. Law and Social Inquiry, 15, 191-221.
Sociaal en Cultureel Planbureau (2002). Sociaal en cultureel rapport 2002: De kwaliteit
van de quartaire sector. The Hague, Sociaal en Cultureel Planbureau.
Stalans, L.J. (2002). Measuring attitudes to sentencing. In J. V. Roberts & M. Hough
(Eds.), Changing attitudes to punishment: Public opinion, crime and justice (pp. 1532). Cullompton, Devon: Willan.
Thibaut, J. & Walker, L. (1975). Procedural justice: A psychological analysis. Hillsdale,
Erlbaum.
Tonry, M. (2004). Thinking about crime: Sense and sensibility in American penal culture. Oxford, Oxford University Press.
Tyler, T.R. (1988). What is procedural justice? Criteria used by citizens to assess the
fairness of legal procedures. Law & Society Review, 22(1), 103-135.
Van Koppen, P.J. (2003). Verankering van rechtspraak: Over de wisselwerking tussen
burger, politie, justitie en rechter. Deventer, Kluwer (inaugural address Vrije
Universiteit).
Walker, N. & Hough, M. (Eds.). (1988). Public Attitudes to Sentencing: Surveys from
Five Countries. Aldershot: Gower.
Walker, N., Hough, M. & Lewis, H. (1988). Tolerance of leniency and severity in England and Wales. In N. Walker & M. Hough (Eds.), Public Attitudes to Sentencing:
Surveys from Five Countries, 178-202. Aldershot, Gower.
470
471
Responsive But Misunderstood:

Dutch Judges on Their Relation to Society
JAN W. DE KEIJSER, HENK ELFFERS AND HENK G. VAN DE BUNT
Abstract
How do Dutch criminal judges deal with public opinion and public pressure? On
the basis of five focus group interviews with 28 judges from various Dutch
criminal courts, it is shown that judges feel a duty to be aware of public opinion
and public pressure in criminal cases. They feel well able to take stock of such
opinion, selecting professionally only what is relevant and useful. They do not at
all feel threatened in their independence by such pressure. Actually judges believe that public rumors for harsher punishment is due to lack of information on
the side of the public: judges perceive the discrepancies between the public opinion and their own judgments as a cognitive, not as a normative gap. They perceive mass media not helpful in bridging that gap, on the contrary, they feel rather helpless seeing that, in their opinion, media actively misinform the public on
what goes on in court. The validity of judges perceptions is discussed.
472
Jan W. de Keijser, Henk Elffers and Henk G. van de Bunt
1. Introduction 1/2
Until late in the 20th century judges in most Western European countries were not
much bothered or affected by the pressures of public opinion. In fact public
opinion on criminal justice in general and on specific cases, as articulated daily
by modern mass media, is a relatively new phenomenon. Earlier the occupation
of judge could and was largely performed in isolation from public opinion. The
legitimacy of independent legal judgment remained self-evident; authority was
based on the assumption that only judges knew the just and correct application of
the law in each and every case. This is especially so in the case of the Dutch judiciary, which consists of professionals who are appointment for life by the government. Lay people have no role in the Dutch system: there are no juries, nor
has it any form lay judges, such as e.g. the German Schffen system. In a comparative study, the Dutch penal law researcher Tak concludes that in the Netherlands the standards of criminal proof and of the grounds for why a case is considered as proven is below a reasonable minimum, and the same holds, he states, for
the reasoned motivation of the level of punishment meted out (Tak 2003, p, 176).
Moreover, according to Dutch law, it is largely to the judges discretion what
level of punishment is seen as suitable in any concrete case. The minimum punishment foreseen in the law is equal for all crimes (and effectively equal to no
punishment at all), while for every given offence a specific maximum is formulated; this specific sentencing ceiling is always a high one. The judge has ample
choice within these limits of the law. E.g. burglary may be punished by imprisonment during any length between 1 day and 6 years. So, the primacy of punishment choice lies with the judge, not with the lawmaker. The consequence
hereof is that judges have great discretion in sentencing severity. Simultaneously,
they have no obligation whatsoever to give a public account for their choices,
and laymen influence is not present at all. Referring to Montesquieu, the Dutch
system is defended with a plea for a division of powers. Exactly because of his
distance to policy and public, the judge may retain his independence. However,
public and political concern regarding criminal justice has recently increased,
coinciding with rather sharp criticism on the practice of criminal law and the severity of punishment. Many citizens seem to have lost their trust in rehabilitation
as a punishment goal, and demand that punishment is applied for reasons of retribution or incapacitation. The question then arises whether and how the Dutch
1
This research project was financially supported by the Dutch Council for the Administration
of Justice (Raad voor de rechtspraak).
We thank Michael Tonry for his comments and suggestions on an earlier version of this
paper.
Responsive but misunderstood: Dutch judges on their relation to society
473
criminal judge can respond to these changes in society, if he is placed at such

large distance from what is going on in society?
Nonet and Selznick (1978) discussed a development in Western societies
where legitimacy is increasingly served by responsiveness to societal developments and public opinion, rather than by mere correct application of legal rules,
instrumental effectiveness or procedural fairness. A responsive institution, they
argue, perceives social pressures as sources of knowledge and opportunities for
self-correction (p. 77). Responsive, in our own words, means to be able to explain legal decisions and procedures to a larger public and being able to show
that public opinion has been accounted for or, at least, has been taken account of.
The primary focus of the present study is on Dutch judges own perceptions of
their responsiveness to society and the level to which they desire to be so. Do
they perceive pressure from society? Do they perceive a large difference between
their own and the publics and politicians opinion on levels of punishment? Are
they prepared to yield to societal expectations, and if so, to what degree?
Modern judges indeed appear to value responsiveness as legitimizing mechanism. Nowadays, it is not uncommon for judges to give public opinion an explicit
role in explaining a sentencing decision, particularly in cases that enjoy wide
media attention. At present, public opinion is said to play an important indirect
role in sentencing policy and practice (Hough & Roberts, 1999; Roberts et al.,
2003, Indermaur, 1990). Little is known, however, on how judges themselves
perceive public opinion and how sensitive they are to resulting pressures. How
do judges believe that such competing values as responsiveness on the one hand
and formal independence and professional distance on the other may be balanced?
Below, we address these questions using data obtained through an in-depth
qualitative study with a sample of 28 judges working in the criminal law divisions of Dutch district courts and courts of appeal. After a brief description of the
research methodology, i.e. group interviews, we proceed to report the findings
organized according to the three main topics of the group interviews: judges
own perceptions of public opinion and societal pressures, judges ways of dealing with perceived pressures and conflicting values, and possibilities and limitations for improving the way judges ought to deal with the pressures of public
opinion.
2. Methodology: group interviews with judges

Little is known on (Dutch) judges personal views concerning fundamental values underlying their work and professional role orientation. To uncover and ex-
474
plore existing attitudes and perceptions in some detail, a relatively unstructured

method of interviewing was considered most suitable. The methodology of group
interviews (also known as focus groups) seemed particularly suitable (cf. Bloor,
Frankland, Thomas & Robson, 2001; Morgan, 1993). It enables in-depth exploration of prevailing points of view in a specialist professional group such as criminal judges and uncovering (unanticipated) differences of opinion between subjects.
The actual group interviews were loosely structured along the lines of a topical
list that was derived from three general questions. Do judges perceive public
opinion and criticism, and its ensuing pressures as a problem? How do judges
deal with perceived pressures related to public opinion, and how do they deal
with the conflicting values of independence versus responsiveness? How can
judges be better equipped in the future in order to appropriately deal with the
pressures of public opinion and their responsiveness to those pressures?
The discussion leader (one of the authors) kept track of the topics that were
discussed, and introduced new topics from the list when the discussion either
seemed to need new input or when the discussion appeared to wander into irrelevant matters. Discussion among participants was stimulated as much as possible,
rather than putting emphasis on a question-response situation between discussion
leader and participants. A pilot session with six part-time criminal judges was
used as a training session for the discussion leader and for fine tuning the topical
list.
Five group interview sessions were organized; one in each of the five court of
appeal jurisdictions in the Netherlands. 3 Within these five court of appeal regions, a total of 19 district courts reside. 4 The aim was to have six participants in
each session, including three court of appeal justices and three judges from the
district court that resides in the same city as that court of appeal. Our strategy for
recruitment of participants was an indirect one. First, the Dutch Council for the
Administration of Justice kindly agreed to send a letter of recommendation pertaining to this research project to the chairpersons of the criminal law divisions in
the selected courts. We subsequently contacted these chairpersons, explained the
research objectives in some more detail, and requested their cooperation for recruiting participants. These chairpersons and their supporting staff were further
asked to make the necessary planning arrangements for the actual sessions to
take place in situ. Owing to the generous cooperation of the chairpersons and
3
The five court of appeal locations are: Amsterdam, The Hague, Den Bosch, Arnhem, Leeuwarden.
For more details on the organization of the Dutch court system, see Tak (2003).
475
their supporting staff, a total of 28 judges were recruited from the courts of appeal and district courts to participate in our group interview sessions. All five
group interview sessions took place in April 2003. Each session took about two
hours to complete, and was recorded on audio tape and subsequently transcribed.
Transcriptions were qualitatively analyzed by each researcher separately. Differences of interpretation were discussed in detail until agreement was reached.
From the wealth of data obtained during the course of the five group interviews
the main threads that were distilled are discussed in the next section.
3. Results
3.1
Media pressure
In each of our group interview sessions, judges indicated that they were very well
aware of the existing societal criticism on their functioning. However, they rejected this criticism largely as unfair, complaining about the dismal role of the
media. In fact, they feel truly misunderstood and misrepresented. The media are
accused of selective attention and selective reporting; a disproportionate amount
of media attention is said to be paid to the relatively rare and especially serious
cases, whilst disregarding the bulk of the cases that are dealt with to everyones
satisfaction. The judges in our interview sessions expect from the media to report
in a factual and correct way, representing the true considerations and difficulties
that underlie the cases at hand. However, on top of being selective and in contrast
to what judges expect from the media, the overall quality of the media operating
in the field of criminal justice is perceived to be waning. As one of our participants put it:
In our experience oversimplified and incorrect media coverage is all too
common.
All participants appeared to suppose that as a result of misinformed, selective and

distorted media coverage, the general public has an incorrect and unnecessarily
negative image of the functioning of the criminal court system. Judges see themselves being portrayed by the media as formalistic officials who are not responsive to the expectations of society, if at all aware of societys wishes. As a result
a gap has grown between judges and society.
One should realize that our image in the outside world is not so much the
result of what the judge really does, but rather of the way that others, especially the media, represent it.
Discussing the media and the image that judges believe is being portrayed there,
further reveals that this professional group feels quite defenseless against the
476
power of the media. Most judges reluctantly (some almost apathetically) resign
themselves to this situation. Others, however, respond by placing the situation in
a wider societal perspective in which all institutions related to state authority are
being publicly scrutinized and criticized:
There is a development to be noted: first the police force was placed in a
house of glass, then the prosecution service was placed there. And now a
magnifying glass is held over judges, also due to general media attention.
A factor that further strengthens judges sense of defenselessness is the role

played by populist politicians. On a regular basis individual members of parliament publicly comment on specific decisions taken by criminal judges, or even
comment on cases that are still being considered by the courts. Judges perceive
this to be an infringement on the formal separation of powers and reject such incidences with utmost vigor.
3.2
The gap
A supposed gap between judges and society was often and spontaneously addressed by participants in our study, often as demonstrated in the previous section, thought to be the product of one-sided media coverage. What is the nature
of this supposed gap yawning between judges and the general public?
At first glance judges appear to be referring to strong differences of opinion
between themselves and the public on matters of just and effective punishment,
or, put more direct; on punitiveness. However, through the course of the discussions, it became clear that judges perceive the gap to be primarily a cognitive
gap, rather than a normative one. Judges believe that decisions and sentences will
in fact be supported once the public is informed adequately and has an open mind
for explanation and instruction. In other words, according to judges the gap is the
result of a lack of knowledge among the public on crime and punishment in general, and on details of concrete court cases as well, both of them in turn connected to the role of the media. 5 Thinking out loud, a participant wonders
whether the public and judges really do have such profound differences of opinion as is often claimed in the media:
I dont know if such large groups in society believe that things are so
wrong. In my personal experience, not in professional circles but privately
amongst friends and acquaintances, when you speak and people confront
you on something, you explain and individualize matters () then, like a
leaf on a tree people change their opinion and when you subsequently ask
what should be done, they say precisely what has been done.
5
See, e.g. Roberts et al. (2003, mainly Ch. 2 and Ch. 6), but also Cullen et al. (2000) for discussions on the relation between public opinion and media reporting.
477
Judges further differentiate between two types of audiences with whom a gap
may exist. The first type of audience, the directly involved audience, encompass
the accused, victims, their respective families and friends, witnesses, as well as
other persons attending the public gallery. The second type of audience judges
from a distance: the general public, journalists, but also politicians. An eventual
gap with the first type of audience is not one that causes a great deal of concern
among judges. They feel that good communication in court can bridge that gap,
and most judges feel capable and eager to explain their decisions in the courtroom in such a way that even those negatively affected by these decisions will be
able to understand and eventually accept them. An important way of doing so
according to our participants is by showing to those involved that they as judges
are well aware of all relevant facts, opinions and expectations and to give account of these in explaining a particular decision. As such, it appears that according to judges their decisions with respect to those who are directly involved can
be justified and made intelligible, primarily by using procedural justice aspects
(cf. Malsch, 2002; Sherman, 2002; Tyler, 1990), and to a lesser extent, or less
explicated, through responsiveness as legitimizing or bridging mechanism. The
main instrument that judges see available to them is the way how they verbally
interact with those involved in court. One of our participants summarizes it as
follows:
The keyword is communication in court.
But also in personal contacts that judges have with lay-persons outside the courts,
judges feel confident to be able to counter any criticism and take away skepticism on the functioning of the criminal justice system in general and the courts in
particular: as long as they have the chance to explain and give reasons, they are
happy to confront criticism.
Such face-to-face opportunities for explaining and giving reasons is what
judges say to lack when attempting to convince the general public of the quality
and correctness of their judicial decisions. Hence they indeed experience a gap
with the general public, which they feel to be extremely difficult, if not impossible, to bridge. Judges complain about lacking a forum at a macro-level that is
capable to exercise counterpressure to one-sided media coverage. But what about
improving the formal motivation of decisions and sentences in court; and what
about improving relations with the media? When asked such questions, most of
our participants respond skeptical. They do not believe that a solution will be
found there:
Recently I read out three pages for a television network and only two sentences were broadcasted. Or:
I have never seen a quote from a verdict in any newspaper.
478
When asked to speculate about the future nature and development of this gap
with the general public, judges fear that it will only widen and deepen. Media
and populist politicians will urge on one another. As a result, judges fear that
public discontent with the criminal justice system will further spread like an inkstain.
3.3
Threatened independence?
In the introduction we discussed that the modern and responsive judge is expected to
take account of new forces and pressures in his environment while retaining a grasp
on what is essential to his integrity. Can independence be retained when confronted
with such public discontent as it is perceived by our participants?
To our surprise, when specifically asked about the potential conflict between
responsiveness and independence, none of our participants sees any problem regarding the pressures of public opinion. Indeed, public dismay about criminal
justice is considered a problem, and an especially annoying one as it is believed
to be of a primarily cognitive, not normative nature. However, it is not at all perceived by judges as a true threat to their judicial independence. They feel completely free and able to either ignore what reaches them as not relevant for their
decision, or on the contrary to take it into account as relevant fact or aspect when
dealing with a case.
You are simply a professional.
The most common reaction in our group interviews on this topic was that judges
consider it one of their specific professional tasks to be open to all kinds of societal developments and critique, without shielding themselves a priori from society. Perhaps naively, Dutch judges consider themselves intellectually and professionally very well equipped and capable to filter and select only what they consider relevant and without becoming unduly influenced. On a marginal note, it is
interesting to observe that in the U.S. judges have been shown to believe that also
lay persons (potential jurors) are quite capable of filtering and disregarding potentially prejudicial information that has appeared in the media about a case (cf.
Studebaker & Penrod, 1997).
In order to keep in touch with societal developments and with public opinion,
our participants strongly dismiss any kind of selfchosen isolation from society.
After all, according to the participants:
Showing societal involvement is not equivalent to giving in to societal
pressure.
When specifically asked, none of the judges in our group sessions says (or dares
to admit) having ever been led by societal pressure in an individual case. Political
479
turbulence, media criticism and signs of societal commotion are first and foremost seen as an exhortation to work as careful and painstaking as possible One of
the respondents, who judged in a very prominent and public case of manslaughter
told that the court reacted on public upheaval and clamoring for a harsh by metaphorically building a wall between public emotions and their own decision
making.
all commotion provided us with the power to stick with the case as
such. Most important are and remain the case file as well as the impression
the suspect makes on us during the public court procedure
Despite judges strong believe in their own filtering capabilities regarding external pressures, there is another reality that they are faced with. This reality is
the judge as integral part of a complex and multi-layered judicial organization.
Here judges are also confronted with pressure: pressure from within. This type of
pressure can often be traced back to societal developments and public opinion,
but is much more indirect and generally takes longer to percolate through different channels in the organization. When discussing pressures from within, it soon
becomes clear that it is considered a weak spot by the participants.
We laugh about populist notions in the media that are here today and gone
tomorrow () Effective societal pressure is always indirect. Once you manage to enter the system, you will produce effect. Through peer pressure societal pressures get to you.
One of the most noticeable and inescapable channels for Dutch judges through
which pressure oozes through the system is through the prosecutors office. The
prosecutor is dominus litis: the prosecutors office decides upon priorities in police investigations, priorities in prosecution, and has the discretion to decide
which individual cases to prosecute and which not, and to formulate the charge.
Through the prosecutor, who is more closely linked to politics, public opinion on
crime and punishment reaches judges in an inevitable way. Our participants remark that in that sense the judge is on the leash of the prosecutor.
Being a judge there is not much you can do, once the prosecutor puts cases
before the courts as an implementation of parliaments policy. In certain respects this is the tragedy of being a judge.
The prosecution service canalizes public opinion through the system. This is a
legal procedural matter of fact and an integral feature of the Dutch system, as
such acknowledged by judges. However, simultaneously, judges expect or hope
from the prosecution service that it acts as a buffer against far reaching and direct
political influence on day to day practice of the courts. During the group discussions recent prosecution policies regarding a specific and massively frequent type
480
of drugs trafficking 6 is frequently mentioned as evidence of the prosecution service being too closely related to daily politics. Because of political pressure to
catch and prosecute all of these cases, the whole court system subsequently experienced an enormous capacity shortage. It is exactly these type of processes
that cause judges great concern:
The danger lurks that you become a mere validation machine for the prosecution service.
3.4
Limited options for improvement
While Dutch judges feel very uncomfortable with the present public criticism on
their functioning, and report to be worried about the existence of a gap between
society and the judiciary, they are not at all inclined to support radical changes in
criminal justice procedures. They are outright antipathetic towards a change in
the position of the judge: their enjoying of wide discretionary powers is valued
and thought to be essential, and the total absence of lay influence is heartily endorsed. Indeed, introducing lay elements in our criminal procedure is an idea that
meets general cynicism among our participants. Judges fear that non-professional
participation would only clog up the system. Moreover, judges say that they do
not need lay involvement in order to stay in touch with society.
Also with respect to giving more elaborate motivations (i.e. presenting
grounds, reasons, explanations) for decisions as an instrument to bridge a cognitive gap, judges are skeptical. While Dutch judges have the legal obligation to
give reasons for their legal decisions, the explanations given of judicial decisions
are generally quite rudimentary. Only when it becomes known that one of the
process participants will appeal, the court will write out its motivations. But still
then, motivations are largely written to be appeal proof, rather than to create a
better understanding among the wider public (cf. Nijboer & Sennef, 1999). Nevertheless, judges do believe that more and better motivation of decisions is something where much is to be gained. If nothing else, it would constitute an intrinsic
quality improvement.
When discussing reducing judges sentencing discretion (e.g., by legally prescribing minimum sentences for specific types of cases), again great reluctance
could be observed. One segment of our participants takes a formal point of view:
judges always operate within the formal limits as delineated by the legislator.
Even in the (hypothetical) situation of reduced discretion, within those new legal
constraints judges remain completely independent in their judgment. However,
and all agree on this:
6
I.e., swallowing drugs in order to pass through customs at Schiphol airport.
481
That is a different system with a different penal climate, and, accordingly,

with a different type of judge.
The other segment of our group of participants does see a threat to their independence as a result of such legislation. They equate reduction of judicial discretion to an infraction on their independence. By legally tightening the margins for
choosing the appropriate sentence, the very essence of sentencing is impaired:
judges would no longer be able to weigh and account for all relevant circumstances in sentencing individual cases.
The freedom of determining the level of punishment is essential to your
possibilities to take a just decision. With minimum sentences that is taken
away from you.
One judge explains the magic words that should convince people why judicial
discretion should be cherished:
In dealing with your case, it is in your favor to have a judge who has an
open mind and is not constrained by legislation to such an extent that he will
no longer be able to serve you with a just decision.
Thus, for this second group of judges, judicial independence is inextricably connected to judicial discretion. They therefore make fundamental objections to legal reduction of sentencing discretion.
The most pressing problem that Dutch judges perceive in relation to society, is
their negative image with the general public. Judges strongly believe that their
image is undeservedly negative. In their view, the gap between judges and society is a corollary of distorted information about judicial decision making. As a
result, when asked about future improvements, judges mention a variety of methods which all aim at improving public knowledge of criminal justice. These include television broadcasting of court sessions, and more comprehensive attention to criminal law and sentencing in primary and secondary school. With respect to the cognitive gap, most judges see but a tiny role for judges who are specialized in instructing the media. All Dutch courts have at least one such
spokesman judge who is trained to deal with the media. Especially in serious
and complex cases that attract a lot of media attention, in their hunt for short and
spicy exclamations, journalists would sooner go to the attorney of the accused
than listen to a nuanced juridical expos of a judge specialized in explicating
cases to the media. As such judges are quite skeptical about the role spokesman
judges could play vis a vis the media. Moreover, as one of our participants puts it:
I have no confidence at all in elaborate information and instruction to the
media in large cases that attract a lot of publicity. In those cases attorneys
are like wizards who do their magic with all kinds of legal defenses. As a
judge youll need to exorcize those defenses with all sorts of spells. That
482
game is something that no-one can understand. Explain that to the average
newspaperreader! Inevitably everything you say will be earmarked humbug.
Moreover, another discussant explains the vulnerable position of judges when

dealing with news media:
We all know that when we give an interview, despite the nuance we put
into it, all we get is a couple of quotes, and on that basis its your neck ()
Every time it gets distorted in such a way that you think to yourself: What
did I get myself into, and you start defending yourself against things that
you didnt even say.
Nevertheless one of our participants dissents in this respect and stresses that:
We behave way too reserved in our media performance, especially where it
concerns cases in which a distorted picture is being portrayed. () Then I
think: Yes, if you have the opportunity, you should tell them off. We are
too reluctant in that respect.
In summary, judges do not harbor high hopes of bridging the (cognitive) gap between them and society using different types of communication. In the eyes of
our participants everything bleaches against the backdrop of mass media power,
which, in general, does not portray a correct and nuanced representation of judicial decisions.
4. Discussion
There is an important distinction that Dutch judges make between the matter of
responsiveness (i.e., dealing with and accounting for societal developments and
influences) on the one hand, and societal criticisms after judicial decisions have
been taken on the other hand. Judges do not perceive the issue of responsiveness
to be much of a problem: they consider it their duty to take notice of all kinds of
societal developments and pressures and believe that they are quite capable of
filtering and selecting what they, in their professional opinion, find relevant and
admissible. It is predominantly pressure from within the criminal justice system,
i.e. via the prosecutors office, that is cause for some concern among judges.
Societal criticism after decisions in individual cases are taken, is considered
quite disturbing by judges. They experience a gap between themselves and the
general public. This gap is believed to be a cognitive one that could be bridged
(or rather narrowed) through better education of and communication with the
general public, a task for which judges themselves lack the means. Consequently
judges rely on the mass media to play such a role. However, in the eyes of judges
mass media actually turn out to frustrate instead of support a better public understanding. This is a sensitive issue for judges: they feel themselves forsaken by
483
who they believe should have been their prime partner in communicating with
society.
In summary judges feel the need to be responsive. They sincerely believe that
they are responsive. Judges also think they are well able to filter out relevant
from irrelevant influences from public opinion and society. Being responsive
should therefore not be threat to their independence. The actual perceived threat
comes in through peer pressure within the judicial system. Judges further feel
sadly misunderstood by a general public that is ill-informed, even actively misinformed by the press and by some populist politicians. They feel at a loss with
respect to the issue of trying to get the media to do a better job in informing the
public on their decisions and on how to account for public pressure.
This study was tailored to investigate the opinions and beliefs of Dutch judges on
issues related to public opinion, independence, responsiveness, and the gap with
society. Delving, as we have done, into the way judges perceive themselves and
their relation to society sprouts four important questions that await answering.
The first question relates to the apparent conviction among Dutch judges that
shielding themselves a priori from all kinds of societal influences is not at all
necessary. They feel very capable of selecting and filtering what they professionally believe to be important and relevant for the case at hand, without becoming
unduly influenced. From a cognitive perspective we are inclined to find such an
attitude nave. Even amongst a group of exclusively professional and well-trained
magistrates such as Dutch judges, recent studies have undeniably shown that the
process of judicial decision making is not as rational as one might expect or wish
(e.g., De Keijser, 2000).
The second question relates to judges perception of the gap with society: one
that they believe is of cognitive nature, not normative. Is this a matter of fact or is
it, rather, wishful thinking? Judges believe that if only the general public would
be better educated about the criminal justice system and the nature of and reasons
for legal decisions in specific cases, the gap would disappear. Indeed, there is
ample evidence from research abroad that public sentiments in general, and decisions in specific criminal cases approach actual decisions taken by judges as a
function of the nature and extent of information that is provided (cf. Chapman,
Mirrlees-Black & Brawn, 2002; Doob & Roberts, 1988; Hough & Roberts, 1999;
Seidman-Diamond, 1990; Stalans, 1993; Kuhn, 2002; Hough & Park, 2002).
However, the question remains whether the effect of information suffices to
completely bridge any gap between Dutch citizens and their judges. Obviously
more research on the nature of the gap between Dutch judges and the general
public is needed, in which normative aspects should be addressed explicitly.
484
The third question, related to the second one, concerns judges knowledge of
public opinion. The gap, in the opinion of our participants is caused to a large
extent by the media. The media are accused of selective and distorted reporting,
with a disproportionate amount of attention being paid to the relatively rare and
especially serious cases. As a result judges have an unnecessarily negative image
with the general public. However, it is our feeling that judges assume too easily
that public opinion is on the leash of mass media. While media coverage undoubtedly has an important role to play in the formation of opinion, one can justifiably cast doubts on the extent to which this is the case. While Dutch judges assume that the public is directly and highly influenced by media reporting, the
connection between media reporting and public sentiment has been shown to be
less obvious, if not mushy (cf. Ditton, Chadee, Farrall & Gilchrist, 2004). Moreover, if not through what they read and hear in the media, how do judges obtain
their impression of public opinion? Can they sense it (Fitzmaurice & Pease,
1986; Roberts et al., 2003, Ch. 5)? Consequently, if judges perception of public
opinion is misinformed, their desire to be responsive may have unexpected and
unwarranted results (cf. Ashworth & Hough, 1996).7 Judges, therefore, may be
said to be at least as much in need of more, objective, and reliable information on
public opinion as the other way around. Related to the nature of the gap is off
course also the question whether, despite a possible gap, the public actually desires judges to respond to public opinion. In a study by Walker, Hough and
Lewis (1988), it was shown for instance that no less than a fifth of the respondents from the British general public said that judges should not pay attention to
what the public thinks when making up their minds about sentences.
The fourth and final question that needs to be addressed is a juridical
normative one concerning the role that the judiciary is supposed to play vis a vis
media and society. Judges feel defenseless against the power of the media. They
want to provide a counterweight but do not know how. What is clear to them,
however, is that a proactive approach towards the media should be pedagogical
in its ambition (cf. Gies, 2005). Naturally this results from their view on the gap
as resulting from a lack of knowledge among the general public (described as
cognitive gap above). Nevertheless, all currently available options simply
bleach against the backdrop of mass media power. The important question is
whether it should be up to judges themselves to provide a (pro)active counterweight to media reporting. Is the judiciary equipped to be an actor of importance
and impact the world of mass media? Do professional magistrates have enough
knowledge of and proficiency in media operations? We believe not. Moreover,
7
In a similar vein, see also Ouimet & Coyle, 1991.
485
we believe that judges should not even attempt to become a media force to be
reckoned with. Chances are it will only backfire, examples of which have been
documented and analyzed for media policies of the Dutch prosecution service
(cf. Brants & Brants, 2002). Especially in the 1980s the prosecution service
started to actively seek media attention, for instance through (deliberate) leaking
of information to the media, contacts with friendly journalists, and press conferences. Despite the strategic intentions of such active media policies, it also
made the prosecution service more transparent; inadvertently exposing its weaknesses and internal conflicts. Moreover, instead of acting as obedient messengers
(friendly) journalists no longer took the prosecution services information for
granted, critically investigated matters and unmasked mistakes that were made
(Brants & Brants, 2002). Rather, judges should focus on those aspects of their
primary tasks that enable and facilitate correct public representation of what they
do, how they do it and why they do it. Faced with massmedia coverage and their
frustrations about it, Dutch judges appear to be inclined in an almost apathetic
mood to marginalize exactly all those aspects that enable such correct representation. For instance, recent Dutch research showed that on some basic levels the
interaction between courts communication advisors and press judges on the one
hand and journalists on the other can easily be improved. Also rationales (reasons) given for sentences remain far too rudimentary and formal for public comprehension. Furthermore, the formal principle of open justice in the Dutch criminal justice system has been shown to be one that is put to practice to a lesser extent than one might expect or whish (Malsch, Van den Berg, De Bruijn, De Keijser & Nijboer, 2003). And finally, the Dutch system and procedures rely highly
on written and largely non-public case files. As a result Dutch procedures can
hardly be called transparent from an outsiders point of view (cf. Tak, 2003).
We must admit that we raise more questions than we can currently answer.
Moreover, the unresolved and uneasy relationship between the media and the
judiciary remains unsatisfactory. What does seem clear, however, is that this uneasy relation cannot be solved by altering the nature or role-conception of judges
in a substantial way. Rather a role may be played by politics (cf. Indermauer &
Hough, 2002). This does seem far fetched at times when much of Western politics regarding crime and justice may be characterized as populist punitiveness
(Bottoms, 1995), or penal populism (Roberts et al., 2003). On the other hand encouraging initiatives such as one taken by the British Home Office in the 1980s
may be a first step in the right direction. It involved discouraging the media,
through a specially formed commission including media personalities, from fostering undue cynicism about the criminal justice system (Roberts et al., 2003).
486
While some difference of opinion is possible as to what constitutes undue cynicism, and the effects of such commissions may not (as indeed reported by Roberts et al.) be huge, immediate, and lasting, to us this does seem an important step
forward that deserves to get a more permanent basis in criminal politics.
References
Ashworth, A. & Hough, M. (1996). Sentencing and the climate of opinion. Criminal
Law Review, 776-787.
Beyens, K. (2000). Straffen als Sociale Praktijk: Een Penologisch Onderzoek naar
Straftoemeting. Brussel: VUBpress.
Bloor, M., Frankland, J., Thomas, M. & Robson, K. (2001). Focus Groups in social research. London: Sage.
Bottoms, A.E. (1995). The philosophy and politics of punishment and sentencing. In C.
M. V. Clarkson & R. Morgan (Eds.), The politics of sentencing reform (pp. 17-50).
Oxford: Clarendon Press.
Brants, C. (1994). Justice done and seen to be done? The institutionalized relationship
between the press and the criminal justice system in the Netherlands. International
Criminal Justice Review, 60 (3), 324-333.
Brants, C. & Brants, K. (2002). Vertrouwen en achterdocht: De driehoeksverhouding
justitie-media-burger. Justitile Verkenningen, 28 (6), 8-28.
Chapman, B., Mirrlees-Black, C. & Brawn, C. (2002). Home Office Research Study
245: Improving public attitudes to the criminal justice system. The impact of information.
Cleiren, C.P. M. & De Roos, T.A. (2002). Democratisering van de strafrechtspleging. In
K. Boonen & C. P. M. Cleiren & R. Foqu & T. A. De Roos (Eds.), De weging van 't
Hart: Idealen, waarden en taken van het strafrecht (pp. 171-188). Deventer: Kluwer.
Cullen, F.T., Fisher, B.S. & Applegate, B.K. (2000). Public opinion about punishment
and corrections. In M. Tonry (Ed.), Crime and Justice: A review of research (Vol. 27,
pp. 1-79). Chicago: University of Chicago Press.
De Keijser, J.W. (2000). Punishment and purpose: From moral theory to punishment in
action. Amsterdam: Thela Thesis.
Ditton, J., Chadee, D., Farrall, S. & Gilchrist, E. (2004). From imitation to intimidation: A note on the curious and changing relationship between the media, crime and
fear of crime. British Journal of Criminology, 44, 595-610.
Doob, A.N. & Roberts, J. (1988). Public punitiveness and public knowledge of the facts:
Some Canadian surveys. In N. Walker & M. Hough (Eds.), Public Attitudes to Sentencing: Surveys from Five Countries (pp. 111-133). Aldershot: Gower.
Fitzmaurice, C. & Pease, K. (1986). The psychology of judicial sentencing. Manchester: Manchester University Press.
487
Gies, L. (2005). The empire strikes back: Press judges and communication advisers in
Dutch courts. Journal of law and society, 32 (3), 450-472.
Hoekema, A.J. (1971). Vertrouwen in justitie. Alphen a/d Rijn: Samson.
Hogarth, J. (1971). Sentencing as a human process. Toronto: University of Toronto Press.
Hough, M. & Park, A. (2002). How malleable are attitudes to crime and punihsment?
Findings from a British deliberative poll. In J. V. Roberts & M. Hough (Eds.),
Changing attitudes to punishment: Public opinion, crime and justice (pp. 163-183).
Cullompton, Devon: Willan.
Hough, M. & Roberts, J.V. (1999). Sentencing trends in Brittain. Punishment and Society, 1(1), 11-26.
Indermaur, D. (1990). Attitudes to crime and sentencing: A comparison of court practice and the perceptions of a sample of the public and judges: W.A. College of Advanced Education.
Indermauer, D. & Hough, M. (2002). Strategies for changing public attitudes to punihsment. In J. V. Roberts & M. Hough (Eds.), Changing attitudes to punishment:
Public opinion, crime and justice (pp. 198-214). Cullompton, Devon: Willan.
Jongman, R.W. (1972). Ongelijke kansen in de rechtsgang. Assen: Van Gorcum.
(1978). Klasse-elementen in de rechtsgang. Groningen: Kriminologisch Instituut,
Universiteit Groningen.
Kuhn, A. (2002). Public and judicial attitudes to punishment in Switzerland. In J. V. Roberts & M. Hough (Eds.), Changing attitudes to punishment: Public opinion, crime
and justice (pp. 115-127). Cullompton, Devon: Willan.
Lijphart, A. (1975). The Politics of Accomodation: Pluralism and Democracy in the
Netherlands. Berkely, Ca: University of California press.
Lind, E.A. & Tyler, T.R. (1988). The social psychology of procedural justice. New
York: Plenum Press.
Malsch, M. (2004). Persvoorlichting en rechtbankverslaglegging. Rechtstreeks(2), 35-71.
(2002). De burger in de rechtspraak. Den Haag: Elsevier juridisch.
Malsch, M., Van den Berg, F., De Bruijn, I., De Keijser, J.W. & Nijboer, J.F. (2003).
De ideale rechtbank: Openbaarheid en gerichtheid op de buitenstaander. In M.
Malsch (Ed.), Ervaringen en percepties van niet-professionele procesdeelnemers (pp.
117-134). Den Haag: Elsevier.
Morgan, D.L. (Ed.). (1993). Successful Focus Groups: Advancing the State of the Art.
London: Sage Publications.
Nijboer, J.F. & Sennef, A. (1999). Justification. In M. Malsch & J. F. Nijboer (Eds.),
Complex cases: Perspectives on the Netherlands criminal justice system (pp. 25112526). Amsterdam: Thela Thesis.
Nonet, P. & Selznick, P. (1978). Law and society in transition: toward responsive law.
New York: Harper & Row.
Ouimet, M. & Coyle, E.J. (1991). Fear of crime and sentencing punitiveness: Comparing the general public and court practitioners. Canadian Journal of Criminology, 33
(2), 149-162.
488
public opinion: Lessons from five countries. New York: Oxford University Press.
Seidman-Diamond, S. (1990). Revising images of public punitiveness: Sentencing by
lay and professional English magistrates. Law and Social Inquiry, 15, 191-221.
Sherman, L.W. (2002). Trust and confidence in criminal justice. NIJ Journal (248), 2231.
Stalans, L.J. (1993). Citizens' crime stereotypes, biased recall, and punishment preferences in abstract cases: The educative role of interpersonal sources. Law and Human
behavior, 17 (4), 451-470.
Studebaker, C.A. & Penrod, S.D. (1997). Pretrial publicity: The media, the law, and
ccommon sense. Psychology, public policy, and law, 3 (2/3), 428-460.
Tak, P.J. (2003). The Dutch criminal justice system: Organization and operation. Den
Haag: Boom Juridische uitgevers.
Thibaut, J. & Walker, L. (1975). Procedural justice: A psychological analysis. Hillsdale:
Erlbaum.
Tyler, T.R. (1988). What is procedural justice? Criteria used by citizens to assess the
fairness of legal procedures. Law & Society Review, 22 (1), 103-135.
(1990). Why people obey the law. New Haven: Yale University Press.
Walker, N., Hough, M. & Lewis, H. (1988). Tolerance of leniency and severity in England and Wales. In N. Walker & M. Hough (Eds.), Public Attitudes to Sentencing:
Surveys from Five Countries (pp. 178-202). Aldershot: Gower.
489
List of Authors
ELFFERS,
Henk
Prof. Dr., Senior Researcher at the Netherlands Institute for the

Study of Crime and Law Enforcement (NSCR), Professor of empirical research into criminal law enforcement at the Law Faculty
of the Free University (VU) of Amsterdam. Contact Address:
NSCR, PO Box 792, NL-2300 AT, Leiden / The Netherlands
Tel.: +31 71 5278519
Fax: +31 71 5278537
HElffers@NSCR.NL
FARRALL,
Stephen
Dr., Senior Lecturer in Criminology, School of Law, The University

of Sheffield, Crookesmoor Building, Conduit Road, Sheffield,
S10 1FL
Tel.: +44 114 222 6771
Fax: +44 114 222 6832
s.d.farrall@crim.keele.ac.uk
GRAY,
Emily
Research Fellow, Institute of Law, Politics & Justice, Keele University, Staffordshire, England, ST5 5BG.
Tel.: 447957 326433
Fax: 441782 584269
e.gray@crim.keele.ac.uk
GREEN,
David A.
Dr., John Jay College of Criminal Justice, City University of New York,
Department of Sociology, 899 Tenth Av., New York, NY 10019
Tel.: 1-212-237-8666
dgreen@jjay.cuny.edu
HARTNAGEL,
Timothy F.
Professor Emeritus, Department of Sociology, University of Alberta,

Edmonton, Alberta, Canada T6G 2H5
Tel.: (780) 492-0467 or 492-5234 (O);
(780) 435-7051 (H)
Fax: (780) 492-7196 (O); (780) 437-2925 (H)
tim.hartnagel@ualberta.ca
HIRTENLEHNER,
Helmut
Mag. Dr., Assistant Professor at the Institute for Criminal Sciences

and the Centre for Legal Psychology and Criminology of the University of Linz, Austria. Address: Universitt Linz, Altenberger
Strae 69, A-4040 Linz
Tel.: 0043/732/2468-8346
helmut.hirtenlehner@jku.at
490
List of Authors
JACKSON,
Jonathan
Dr., Lecturer in Research Methodology. Address: Methodology

Institute, London School of Economics, Houghton Street, London,
WC2A 2AE.
Tel.: +44 (0)207 955 7652
Fax: +44 (0)207 955 7639
j.p.jackson@lse.ac.uk
KAAL,
Hendrien
Dr., Researcher, Lecturer in Criminology, Department of Criminal

Law and Criminology, Faculty of Law, Leiden University, PO Box
9520, NL-2300 RA Leiden, The Netherlands
Tel.: +31 (0)71 527 8596
h.l.kaal@law.leidenuniv.nl
KEIJSER,
Jan W. de
Dr., Senior Researcher at the Netherlands Institute for the Study of

Crime and Law Enforcement (NSCR), Address: NSCR, PO Box
792, NL-2300 AT, Leiden, The Netherlands
Tel.: +31 71 5278507
Fax.: +31 71 5278537
JdeKeijser@NSCR.nl
KERNER,
Hans-Jrgen
Professor Dr., Director, Institute of Criminology, University of

Tbingen, Sand 7, D-72076 Tbingen
Tel.: 07071 - 297 2931
Fax.: 07071-295 104
hans-juergen.kerner@uni-tuebingen.de
KING,
Anna
PhD, Lecturer, Address: Keele University, Centre for Criminological Research, Staffordshire, ST5 5BG / UK
a.king@crim.keele.ac.uk and:
Center for Mental Health Services and Criminal Justice Research,
176 Ryder Lane, New Brunswick, NJ 08901 USA
Tel.: +001-732-932-2190
aking@ifh.rutgers.edu
KLAUS,
Witold
M.A., Dept. of Criminology, Institute of Law Studies, Polish Academy of Sciences, Nowy Swiat 72 st., 00-330 Warsaw, Poland
Tel.: + 48 22 826-84-84
Fax: + 48 22 657-27-51
wklaus@o2.pl
KOSSOWSKA,
Anna
Prof. Dr., Dept. of Criminology, Institute of Law Studies, Polish

Academy of Sciences, Nowy Swiat 72 st., 00-330 Warsaw, Poland
Tel.: + 48 22 826-84-84
Fax: + 48 22 657-27-51
adkossowska@o2.pl
KOVO-VUKADIN,
Associate Professor of Criminology, Faculty of Rehabilitative

Sciences, University of Zagreb, Croatia
Irma
List of Authors
491
KURY,
Helmut
Prof. Dr., Prof. h.c., Prof. h.c., Max Planck Institute for Foreign and
International Penal Law (retired), University of Freiburg. Address
(private): Waldstrasse 3, D-79194 Heuweiler
Tel.: +7666-3799; Mobile: 0175 35 27 170
helmut.kury@web.de
MARUNA,
Shadd
School of Law, Queens University Belfast, 28, University Square,

Belfast BT7 1NN, Northern Ireland
Tel.: + 44 28 90 97 3400
Fax: + 44 28 90 97 73376
s.maruna@qub.ac.uk
MATRAVERS,
Amanda
Dr., Associate Professor, Department of Justice, Law & Society, American University, Washington DC 20016, USA
Tel.: + 202.885.2618
Fax: + 202.885.2353
matraver@american.edu
MEKO,
Gorazd
Associate Professor of Criminology, Dean, Faculty of Criminal Justice and Security, University of Maribor, Slovenia
Tel.: +386 1 300 83 00
Fax: +386 1 2302 687
Gorazd.Mesko@fvv.uni-mb.si
MURATBEGOVI, Teaching Assistant, Faculty of Criminal Justice Sciences,

University of Sarajevo, Bosnia and Herzegovina
Elmedin
OBERGFELLFUCHS,
Joachim
Dr., Head of the Criminological Service, Dept. of Corrections BadenWrttemberg, Stuttgart, Germany. Address: Dr. Joachim ObergfellFuchs, Leiter Kriminologischer Dienst, Justizvollzugsschule BadenWrttemberg, Pflugfelderstr. 21, 70439 Stuttgart, Germany.
Tel.: 0711/8020-3001
Fax: 0711/8020-3049
Joachim.Obergfell-Fuchs@jvsbaden-wuerttemberg.justiz.bwl.de
RZEPLINSKA,
Irena
Prof. Dr., Dept. of Criminology, Institute of Law Studies, Polish

Academy of Sciences, Nowy Swiat 72 st., 00-330 Warsaw, Poland
Tel.: + 48 22 826-84-84
Fax: + 48 22 657-27-51
irenkarzeplinska@wp.pl
SERRANOMALLO,
Alfonso
Prof. Dr., Director and Professor Titular, Departamento de Derecho

penal y Criminologa, Universidad Nacional de Educacin a Distanzia UNED, Coordinator de los Estudios de Postgrado de Criminologa, Calle Obispo Trejo s/n Madrid / Spain
alfonsoelgallo@yahoo.es
492
List of Authors
SESSAR,
Klaus
Prof. Dr. (em.), Universitt Hamburg, Dept. of Criminology,

Schlueterstr. 28, D-20146 Hamburg. Home: Falkensteinweg 3,
79199 Kirchzarten
Tel.: +49 (0)7661-39 07 80
klaus.sessar@online.de
TEMPLETON,
Laura
PhD Candidate, Department of Sociology, University of Alberta,

Edmonton, Alberta, Canada T6G 2H4
laura.templeton@ualberta.ca
VAN DE BUNT,
Henk G.
Prof. Dr., Professor of Criminology at the Law Faculty of the Erasmus University in Rotterdam. Contact Address: Erasmus University
Rotterdam, Law Faculty, Criminology Section, PO Box 1738,
NL-3000 DR, Rotterdam / The Netherlands
Tel.: +31 10 4088802
Fax: +31 10 4089175
vandeBunt@FRG.EUR.NL
VAN OEVEREN,
Moniek
Master student in Criminology, Department of Criminal Law and

Criminology, Faculty of Law, Leiden University, PO Box 9520,
NL-2300 RA Leiden / The Netherlands
VANDERVEEN,
Gabry
Dr., Assistant Prof. of Criminology, Dept. of Criminal Law and

Criminology, Faculty of Law, Leiden University, PO Box 9520,
2300 RA Leiden, The Netherlands
Tel.: +31 (0) 71 527 7414
g.n.g.vanderveen@law.leidenuniv.nl
WOZNIAKOWSKA, M.A., Dept. of Criminology, Institute of Law Studies, Polish Academy of Sciences, Nowy Swiat 72 st., 00-330 Warsaw, Poland
Dagmara
Tel.: + 48 22 826-84-84
Fax: + 48 22 657-27-51
daga.wozniakowska@gmail.com
YOSHIDA,
Toshio
Prof. Dr., Professor for Penal Law and Criminology at the HokkaiGakuen University of Sapporo, Faculty of Law, Sapporo, Japan,
Asahimachi 4-1-40, Tyohira-Ku 062-8605 Sapporo
yoshida@wise.hokkai-s-u.ac.jp
ZARAFONITOU,
Christina
Dr., Professor of Criminology, Dept. of Sociology, Panteion University,

136 Syngrou ave, 176 71 Athens / Greece.
Tel.: +210 82 14 019
Fax: +210 82 38 680
chrizara@panteion.gr

Book Kury 08 PDF

Hochgeladen von

Dokumentinformationen

Originalbeschreibung:

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Book Kury 08 PDF

Hochgeladen von

Copyright:

Verfügbare Formate

Fear of Crime Punitivity

New Developments in Theory and Research

Fear of Crime Punitivity

Helmut Kury (Editor)

Fear of Crime .................................................................................................. 1

Punitivity ....................................................................................................... 197

Part I: Fear of Crime

Researching Everyday Emotions: Towards a Multidisciplinary Investigation of the Fear of Crime

EMILY GRAY, JONATHAN JACKSON

Emily Gray, Jonathan Jackson and Stephen Farrall

ology of emotions (Kemper, 1990; Wouters, 1992; Williams and Bendelow,

The utility and validity of fear of crime research

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

tive security (Taylor et al., 1986) and an expression of powerlessness and

Towards an interdisciplinary approach to fear of crime

Emily Gray, Jonathan Jackson and Stephen Farrall

As highlighted, the relative neglect of emotions in criminological work has not

Research on everyday emotions a complex picture

Towards a Multi-disciplinary Investigation of the Fear of Crime

the complexity and enormous diversity of human emotion in normal day-to-day

How often do we experience emotions?

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

What does it mean to be fearful?

Emotional digestion and appraisal theory

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

New directions in everyday emotions research: methodological

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

such, cognitive interviewing techniques might be a particularly valuable method

Towards a Multi-disciplinary Investigation of the Fear of Crime

12 months have you worried about? Nevertheless, retrievals are influenced by

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

Towards a Multi-disciplinary Investigation of the Fear of Crime

Emily Gray, Jonathan Jackson and Stephen Farrall

Wouters, C. (1986). Formalization and Informalization; Changing Tension Balances in

Fear of Crime or Fear of Risk?

Fear of Crime or Fear of Risk? Some Considerations

3. Governing through fear of crime or fear of risk

Fear of Crime or Fear of Risk? Some Considerations

sively; that this information operates to identify fear as a legitimate object of

4. Some recent developments

Fear of Crime or Fear of Risk? Some Considerations

How to Measure Fear of Crime

How to Measure Fear of Crime in a Valid and Reliable Way

Background of this paper: conceptualisation and operationalisation

How to Measure Fear of Crime in a Valid and Reliable Way

Outline of this paper

How to Measure Fear of Crime in a Valid and Reliable Way

Figure 1: Outline of this paper

2. Two basic problems: concept and measurement

In the introduction, the operationalisation and measurement of fear of crime

3. Improved conceptualisation and operationalisation

How to Measure Fear of Crime in a Valid and Reliable Way