Sie sind auf Seite 1von 21

Comparing Media Systems

A Response to Critics
Daniel C. Hallin and Paolo Mancini

In the preface to Comparing Media Systems we wrote that we would be satisfied if our book were
able to spark a more developed discussion of theory and method in the comparative study of media
and politics, and inspire more comparative studies in that area. Several years after the publication of
the book, we can officially declare that we are satisfied with the results. Comparing Media
Systems has been widely discussed, and the framework we propose has been used extensively in both
theoretical reflection and empirical research. Of course, not all the reactions to our book have been
positive, even if those who have commented on it have almost always acknowledged its importance
and its role in fostering comparative research on media and politics. In this paper we want to respond
to some of the criticisms that have been advanced and some of the issues that have arisen as our work
has been analyzed, applied, and tested. We will be selective: the range of responses is wide and we
do not have space to address all the issues that have arisen.
Our objective here is not so much to defend ourselves as to deepen the discussion about the topics
addressed in the book which, as we have said, was our main enterprise to begin with. Actually, many
observations about the limitations of our book, which are advanced as criticisms, are points we are
happy to endorse. Our book does have limitationsmany of them. Some of them we intended, some
arose out of the limits of our own knowledge, and some reflected the state of the field. We consider it
important to clarify these limits, and in some cases we have been as much troubled by endorsements
of our book that push our analysis beyond what it can reasonably be expected to do as by critical
comment on its limits. In what is probably the most critical commentary on the book we have seen,
Pippa Norris (2009) concludes that the Hallin and Mancini framework suffers from several major
shortcomings that need to be addressed before we can conclude that this provides an appropriate
conceptual typology for the subfield (p. 331). This is a conclusion we can certainly endorse, as we
never intended our book to be taken as providing a comprehensive conceptual framework for the
comparative analysis of media systems. This is most obviously so because it dealt only with a narrow
range of cases 18 wealthy capitalist democracies in North America and Western Europe. But also
because that would be too much for a single work to pretend, particularly in an emerging field where
there is a limited literature, both empirical and theoretical, on which to build. We intended our book
to begin a process of developing an adequate framework for comparative analysis in this area, not to
end it.
We will discuss five main clusters of issues here: (1) problems related to classification of media
systems according to our three models; (2) the issue of media system as a unit of analysis; (3) the
question of omitted variables in our analysis; (4) questions regarding the empirical validation of the
typology; and (5) the issue of convergence or homogenization.


Starting with the first discussions of Comparing Media Systems (Couldry, 2005; McQuail, 2005;
Patterson, 2007) reviewers have raised questions related to the classification of media systems using
the typology provided by our three models. Three kinds of criticisms have been raised: the first, and
in our view the most important, is related to fundamental question of whether classification of media
systems according to such a typology is really a useful way to approach comparative analysis; a
second set of criticisms has to do with issues concerning whether we correctly classified particular
cases we were analyzing; and a third, with the applicability of our typology to cases outside the scope
of our original analysis. We will not elaborate on this last point here, which is extensively discussed
in a forthcoming edited book (Hallin & Mancini, 2012). We will only underscore the point we make
in Comparing Media Systems, that, although we think the models we propose may be of some use as
points of comparison, particularly given their influence on other parts of the world, our models are
intended to identify particular patterns of media system development in Western Europe and North
America, and we do not think they should be applied in any direct way to any but perhaps a very
few cases1 outside the scope of our analysis. In the concluding lines of our book we write:
substantial modifications would need to be made to our models to apply them beyond their original
scope, and that they would be useful primarily as an inspiration for creating new models based on
detailed research into specific media and political systems (Hallin & Mancini, 2004, p. 306).
On the second point too, we can be relatively brief. Our three models were conceived as ideal
types: we were well aware that individual cases did not fit them exactly, and that scholars in
particular countries might raise issues about the classification of particular cases. As Norris notes,
the classification was not based on standardized indicators or a set of explicit decision rules, since,
as we noted in the book, the field has not developed and collected such indicators, nor is there an
agreed-upon theoretical basis for weighting them2 We proposed what we described as tentative
judgments about the similarity [of the 18 cases] to the ideal types represented by our three models
(p. 71). In the absence of the kind of indicators Norris refers to, we sent the manuscript to
distinguished colleagues in all countries involved and asked if the classification of the particular
cases they studied seemed reasonable. No criticism was raised at this regard. Of course, this is not
evidence for the correctness of the analysis; we bring it up only to make clear how conscious we
were about the risks involved in this kind of classification. Despite this procedure we fully expected
issues to be raised after the publication of the book, both because it is dificult to master the
literatures on so many cases and because there are tensions inherent in the kinds of generalizations we
were putting forward, which can never do justice to all the specificities of particular cases. We have
actually been pleasantly surprised that the objections have been relatively limited. Introducing an
edited collection in which scholars from the Nordic countries debate the extent to which our
framework can illuminate the media systems of their region, Strmback, 0rsten, and Aalberg (2008)
focus on this tension, arguing that, while it might be reasonable from a global or international
perspective, our classification of those cases as relatively pure exemplars of a single Democratic
Corporatist model triggered the question as to whether the Nordic countries really are so similar to

each other as suggested by Hallin and Mancini. Certainly from an everyday Swedish, Finnish or
Danish perspective, the differences between ones own country and the other Nordic countries might
often be as prominent as the similarities (p. 14). Nevertheless, they concluded that the individual
country chapters suggest that the Nordic countries, including Iceland, in many ways do fit Hallin and
Mancinis description of the democratic corporatist model (p. 268).
Others have disagreed more strongly. Some of the criticisms seem to us to ignore our plea that the
grouping of cases in relation to models is not meant to substitute for the more complex discussion
we introduce in the second half of the book (Hallin & Mancini, 2004, p. 71). Norris (2009), Curran
(2011), and Humphreys (2011), for example, complain that we lump Britain and the U.S. together
under the category of the Liberal model despite obvious and important differences like the existence
of strong media partisanship and strong public service broadcasting in the UK. In introducing our
three models, however, we write although the United States and Britain ... are often lumped together
with good justification up to a pointas liberal systems we shall try to show that they are very
different in important ways and that the common idea of an Anglo-American model of journalism is
in part a myth (Hallin & Mancini, 2004, p. 69). We reiterate this point in the first paragraph of our
chapter on the Liberal model, and go on to describe Britain as a mixed case combining elements of
Liberal and Democratic Corporatist models; we considered this point about the heterogeneity of the
countries we grouped under the Liberal model as a central point of that section of our book. Still, the
claim of Norris (2009, p. 334) that media systems in Britain and the United States seem to have
almost nothing in common seems hyperbolic. As Strombck, 0rsten, and Aalberg (2008) note, one
of the risks that comparative research has to avoid is that of taking ones own country, or the
particular cases one knows, as a sort of universal reference, where distinctive features of those cases
come to take on exaggerated importance, preventing generalizations that might be more useful in
broader comparative analysis.
Certainly, however, many observations of the ways in which particular cases deviate from our
ideal types are likely to prove important and to raise interesting theoretical questions. Scholars in
Portugal, for example (a country about which we had relatively limited information), have observed
that although Portugal may at one time have resembled our Polarized Pluralist model fairly closely, it
has in recent years diverged significantly from it, moving away, for example, from the political
parallelism that characterizes Spain, Italy, or Greece. If this is right, it raises questions about how to
account for a shift that our relatively path-dependent model would seem not to predict. Humphreys
(2011) points to the decentralized television system in Germany as a clear and important difference
with other systems we classify under the Democratic Corporatist model and also to the different
reasons that gave birth to commercialization in Greece if compared with the other Mediterranean
countries (also noted in Kogen, 2010). Hardy points to a number of other issues with our
classifications, both in his book (Hardy, 2008) and his chapter (Chapter 11) in this Handbook.
Beyond the question of whether we have classified particular cases correctly, there is a more
general question about whether typologies of the sort represented by our three models are really the
way forward in comparative analysis of media systems. Humphreys (2011), for example, argues:
rather than expend time and energy on producing neat typologies, it is much more important to

explore in depth a more comprehensive range of variables that bear on the complex media-politics
relationship (p. 154). The fact is that since Max Weber as Patterson (2007) notes in a review of
Comparing Media Systems classifications have been used to study society. Classifications allow
generalization: they serve to list and order features that are common to different realities, phenomena,
and situations. These observed features can be extended to cases beyond those that are under
observation, allowing the generalization that is one of the main aims of scientific observation. We
proposed our three ideal types as a way of pointing out broad patterns in the development of the
relationship between media and political systems, which seemed to us an essential step in a field in
which most scholars were still confined within the specificity of particular systems they studied. We
suspect that other scholars are likely to follow our example and propose other such ideal types that
prove useful in identifying patterns that characterize significant numbers of cases.
We must confess, at the same time, that we are not happy in important ways with the use that has
been made of our three models, and we think that Humphreys fear about neat typologies
substituting for detailed analysis is legitimate. Our three models were intended to illustrate the results
that could be accomplished through the application of the suggested interpretive schema based on the
proposed list of variables. Instead, many scholars have taken our models to be, in themselves, the
interpretive framework. We believe this is a reductive reading of our book, and we think it is in part
because of this preferred reading that many scholars have found our classifications unsatisfactory
and limited. We have noticed, for example, that our typology is often used for selecting cases, that is,
scholars will do a study of some phenomenon related to political communication and pick a case
corresponding to each of our three models. To some extent, this makes sense at least it is an
advance that the field is more aware now that systems differ, and that one may get different results
picking cases that belong to different categories. Often, however, it seems to us that relatively little
thought has been devoted to the question of how systems belonging to our different models could be
expected to differ with respect to the particular phenomenon in question, or to the ways the specific
cases chosenwhich after all will not correspond to the ideal type in every detailmight or might
not be distinctive with respect to those phenomena. We do not think that classification is an end in
itself, and we did not intend our three models to be used mainly as a classification system We
intended them to be used to think about patterns and why they exist and, precisely, for comparing how
a particular case fits or does not fit a pattern that characterizes other cases, and why? We do not
advocate that comparative analysis be built primarily around such typologies.


The second issue has to do with the scope of our analysis. The book is called Comparing Media
Systems, but, as many have noted, it does not deal with media systems in their entirety. It does not
deal with the film or television or music industries as such, for example. It is concerned with the
relationship between media and politics, and therefore focuses on news media and to some extent on
media regulation. Many have questioned whether the scope is clearly and properly defined, and those
are legitimate questions. Kraidy (2012) has argued that more attention should be paid to entertainment
as a factor in the relationship between media and politics, and Hardy makes a similar point in this
volume in calling for more attention to the cultural domain beyond news and political journalism.
Humphreys (2011) argues that if we separated the analysis of media policy specifically from the
analysis of journalism we would get different results. Norris (2009) suggests comparing systems of
political communication rather than media systems. All of these are reasonable suggestions, and to
us it seems to be an open question to ask which elements of this domain would make sense to combine
within an analysis and for what purposes. Moving to narrow levels of analysis, as Humphreys
suggests, is likely to make sense for many purposes, but we are likely to want to ask at some point
whether a broader context needs to be taken into account. Whatever system we deal with is also
likely to prove complex and heterogeneous. Norris (2009), for example, worries how we can define
a media system across such disparate phenomena (p. 328) as broadsheet newspapers, tabloid
newspapers, broadcasting, new media, etc. But would a system of political communication really
involve less disparate phenomena? Would we not have to take into account all those same media
phenomenaplus political parties, election campaigns, social movements, corporate public relations,
and, as Kraidy says, entertainment? No doubt comparative analysis will have to proceed using a
number of different levels of analysis, with an awareness that no one delimitation of the scope of the
field will be adequate for all purposes.
A related issue has to do with the concept of system, and the question of whether thinking in terms
of media systems involves assumptions of unity, coherence, and stability that distorts the variability
of relations between media and politics (McCargo, 2012). We adopted this idea from the tradition of
comparative politics and comparative sociology as it seemed to offer the possibility to analyze a
whole set of interconnections. As Almond & Powell (1966) state in their seminal work on
comparative politics, the concept of systems directs attention to the entire scope of political
activities within a society, regardless of where in the society such activities may be located (p. 17).
One reason we consider the comparative method important has to do with the fact that it offers the
possibility to deal with the impact of institutions and of culture at the macro-level, and in that sense is
a correction to the dominance of the behaviorist tradition, with its individual-level focus, in empirical
research in communication. We do not mean to bring in the assumptions of structural-functionalism,
which we would consider only one of many ways of thinking about social or political systems. Nor
do we believe the concept of system needs to imply either stability or unity, but only pattern and
interconnection. Bastiansen (2008) makes a good case for the use of the concept of media systems in
historical studies of system change:

The purpose of such studies will ... be to investigate the totality that emerges when one looks
at the various interconnected mass media ... and focuses on the relations between them. This
approach is usually based on the principle that the totality of such a system is more than the
sum of its parts: the interdependence that exists between different mass media is dificult to
identify as long as each medium is investigated separately. (p. 104)
We would add that the relevant system may also include non-media elements which also have
relations of interdependence with mass media: political parties, for example, or the state. It is true,
however, that other units of comparison are possible and may make sense in many cases, including
comparisons of particular kinds of events (political crises or decisions) or of processes (Roudakova,
2011), and we would not argue that all comparative research in media studies or political
communication need focus on media systems.

a n a l y t i c v a r ia b l e s

The third issue has to do with the question of omitted variables. Our analysis is based on two sets of
analytical variables or dimensions, summarized by Hardy (Chapter 11, in this volume), which we use
for comparing media systems and for analyzing the political context that affects their development.
Critics have raised many issues both about the conceptualization of the variables included in our
analysis and about omitted variables. One reason we followed a most similar systems design was
to limit the number of variables we had to deal with, which we considered important in part because
we believed many of the concepts involved had not been thought through very fully in the existing
literature and needed extensive reconceptualization. This implies, of course, that the list of variables
we consider, and the conceptualization of these variablesthe particular values of them we discuss
is tied to the cases we analyzed. As a consequence, scholars studying other parts of the world often
complain about a bias toward mature democracies (Hadland, 2012) and with good reason, as that
is in fact the scope of our analysis. Very clearly, new variables and new conceptualization are
required to extend the kind of analysis we carried out beyond the Western democracies that are its
focus. A project we carried out together with scholars working on a number of non-Western
systems deals more fully with some of the issues that have been raised in this context (Hallin &
Mancini, 2012). To give one example of the kinds of conclusions that emerge from that project, it is
clear that the discussion of political parallelism provided in Comparing Media Systems is closely
tied to a particular Western history of mass political parties strongly rooted in social and economic
interests and competing in a pluralist political order. Once we expand the range of cases we clearly
have to deal with other kinds of party systems, including non-competitive ones, with systems in which
parties are weak or non-existent, and with systems in which contending political, social, and cultural
forces are organized outside the framework of Western-style parties and interest groups, according to
factions of an authoritarian elite, for example, or around racial or ethnic groups. This underscores
Hardys call (Chapter 11, in this volume) for more attention to the structure of civil society and for a
broader consideration of social actors and their communications.
Even within the scope of our analysis, many legitimate issues have been raised. One of the most
obvious has to do with new media: Hardy (2008, and in this volume), Norris (2009), and a number of
other critics have noted the fact that new media are absent from our analysis. The first media system
dimension considered in Comparing Media Systems is the structure of media markets, and we focus
there primarily on newspaper markets, which differ sharply among the cases we considered, in ways
we believe are deeply rooted and strongly connected with differences in the cultural and political
roles of the media. There are many other aspects of media markets that might be considered, but the
most obvious element missing from our analysis is clearly an engagement with new media. The
reason is simple enough: we didnt know of any significant body of comparative research at the time
that addressed new media. Obviously, however, this has to be a priority for the field.
Another issue related to the structure of media markets, raised by numerous scholars including
Humphreys (2011), McQuail (2005), and Puppis (2009), has to do with market size. The argument is,
in part, that the media systems of small countries are likely to be affected by neighboring or by richer

and bigger countries, which may penetrate their media markets in various ways. Puppis (2009) argues
that this reality has important implications for media policy in small states, and refers to
Katzensteins (1985) hypothesis on corporatism in small countries stressing how corporatism may
affect also the structure and the regulation of the mass media.
McQuail and Humphreys have noted that we did not include regional differences in the structure
of media systems in our analysis. Humphreys is specifically pointing out the necessity to include
centralization vs. decentralization as a variable, referring particularly to the German case. This
example illustrates the tension between generalization and specificity that affects much comparative
work, and most obviously a broad synthetic analysis of the sort we had undertaken. In both Germany
and Spain the television system is organized on a regional basis, but this seemed to us to represent a
very specific feature of these two countries, important to understanding those cases, but less relevant
to others we were considering. Perhaps, however, this should be seen as a special case of the
broader point Hardy makes (Chapter 11, in this volume) about geo-cultural patterns which deviate
from a nation-centric framework and which would also include the kinds of transnational flows
which, as Puppis and others argue, affect small states, or which are clearly extremely important in the
Arab world (e.g., Kraidy, 2012; Lynch, 2006).
Norris (2009; see also Norris, Chapter 22, in this volume) argues that press freedom should be
foregrounded as a criterion for distinguishing media systems, particularly since cross-national
quantitative measures are available from a variety of institutions. In our analysis, press freedom was
discussed as a component of the role of the state, as well as in the historical discussion of the
development of media markets. But it did not play a major role in distinguishing the three models,
since the differences among them in this regard are small. As Norris notes, it would obviously be
more important in a wider comparative analysis (Hadland, 2012). It is worth signaling some cautions
here, however. In the first place, the quantitative measures Norris refers to, and relies on (in Chapter
22, in this volume), are not based on systematic theoretical reflection rooted in comparative analysis,
and many issues remam to be resolved about how to compare press freedom across systems.3 Second,
the concept of press freedom has a strong normative inflection and as it has been used in media
studies over many years is essentially a measure of distance from a norm rooted in North America
and Northern Europe. That doesnt mean we should avoid it as a tool of comparative analysis, but it
does mean that a comparative analysis that foregrounds this dimension too much is likely to produce
the kind of dichotomy that characterized the analysis of Four Theories o f the Press, in which nonliberal systems are understood primarily by what they are not. Furthermore, we may never get to the
point of analyzing in more detail the full range of social roles the state and media actually play in
those systems (Zhao, 2012).

The issue of measuring press freedom leads directly to a fourth fundamental issue. One of the most
important critiques made by Norris (2009) is the charge that our analysis is impressionistic and
lacks empirical validation. She writes, it remains unclear whether the concepts proposed by
Hallin and Mancini can be clearly related to broader theoretical concerns in mass communications as
well as operationalizable when applied to research (Norris, 2009, p. 335). This is a point that
deserves some discussion because it touches a central problem in comparative research in media
studies. At one level we can say that Norris is clearly right. We describe our analysis as a tentative
and exploratory (Hallin & Mancini, 2004, p. 302) one that would need to be tested by empirical
research, and it remains very much open to question to what extent the tentative generalizations we
make will stand up to empirical research or the conceptual framework more generally will prove
useful for carrying it forward. Some elements of Norriss (2009) discussion are troubling to us,
however, and seem to suggest differences about how the field is likely to advance.
One thing that is worth clarifying at the outset is that because our book was not primarily focused
on setting forth hypotheses, with much of our conceptual framework it will make less sense to ask,
Can it be verified? than to ask, Is it useful? As Lakatos (1970, p. 137) argues, it often makes
sense to appraise what he calls research programs not so much on the basis of whether they have
been verifiedmany valuable ones never will be but on the basis of their heuristicpower.'4- In
some cases, this will mean precisely the power of the framework to generate testable hypotheses. In
the case of our framework, this will probably be true less of the three models than of many embedded
observations about relationships among variables; we certainly agree with Norris (2009) in the
observation that it is important to test rigorously how far the different dimensions actually cluster
together in meaningful ways (p. 334). For example, in the case of the difference in media
partisanship between the U.S. and Britainwhich has generated so much angst among those who feel
that if they are both classified as liberal, they shouldnt differ in such waywe can put forward the
hypothesis that the degree of media partisanship depends not on the fact of commercialization (as has
often been assumed) but on political culture and the nature of the party system, or (more likely, we
suspect) on market structure, and in particular on whether there are multiple competitors in the same
market (as in the national British market) or few (as in the local U.S. one). In other cases, asking Is it
useful may mean asking, for example, whether our three models are useful as points of comparison
whether it is useful to compare East European media systems with Southern European ones, as a way
of beginning to ask, What is distinctive about the East European pattern and what is not? Are there
common patterns that might be generalizable? Are there differences that suggest new variables to
introduce? Or it may mean asking, Is it useful to think of political parallelism as a defining
dimension of a media system? Does it produce meaningful comparisons across systems, or sectors of
a media system? Does it help us to conceptualize processes of change or the effects of new media?
How far could we generalize the concept beyond the multiparty democracies of Western Europe?
Would we learn something if we tried to?
The fact that we presented the analysis of Comparing Media Systems for the most part without

empirical validation had to do with two considerations. First, as we argued in the introduction to the
book, there is clearly a shortage of data on all the aspects we consider essential to comparative
analysis of the relationships between mass media and political institutions. More fundamentally, and
lying behind this vacuum of data, is the fact that for the most part comparative research in media
studies has been carried out within a theoretical vacuum: data are collected in relationship to
different aspects of the structure and the performance of the mass media, but these data are not put
within a more comprehensive interpretive framework. In our view, it is this lack of theoretical
reflection that ultimately lies behind the Babelian experience which Norris (2009) sees in
comparative research in media studies and very legitimately would like to rectify, or the frustration
that Livingstone (2003) notes as characteristic of many comparative projects, as each member of a
research team represents and works on his or her own country, and it is never clear whether they are
really working within a common framework. We believed that conceptual clarification and theorybuilding had to come first, before empirical research could advance further, and that was the
emphasis of Comparing Media Systems. Norriss discussion seems to us to imply a kind of
measurement first, conceptualization after, approacha rush to produce standardized measures for
hypothesis testing across large numbers of cases before we are really sure what we want to measure,
and why, and an impatience with the kinds of interpretive research that would take the first steps in
raising these issuesthat makes us uncomfortable. Our hesitations about using existing measures of
press freedom without thinking through the conceptual basis of what we would consider a very tricky
concept is one example. Similarly, Norris (2009) writes about measuring professionalization of
journalism through proxy indicators such as existence of journalism training departments and
accreditation procedures. Thinking through the concept ofjournalistic professionalism more fully was
a principal focus of Comparing Media Systems, and, although our conceptualization of the concept is
of course open to debate, we would consider these very superficial indicators which might or might
not be correlated with the things we conceive as defining the concept. (This evidence suggests that
accreditation of journalistswhich in Western Europe would be characteristic particularly of Italy
and France is negatively correlated with professionalization as we define it.)
Norriss discussion also seems to us to identify empirical research too narrowly with
quantification. This is ironic in a way, because Norris spends a lot of time talking about how difficult
it is likely to be to quantify the things we care about in the study of media and politicswe make this
point also in Comparing Media Systems so much so that she comes close to convincing us that if
we insist on quantification the field will never be able to advance! We certainly do believe that
quantitative data will be important to operationalizing many of the things we want to study in
comparative analysis of media systems (see Benson & Hallin, 2007). But we also suspect that a lot of
the progress in empirical research in our field will have to come from qualitative or partly
qualitative studies of various kinds, using ethnographic, historical, or case-study methods, within
which the process of structured, focused comparison (George & McKeowan, 1985) could be
carried out. There are ersatz forms of quantification that are often used in large-N studies in
comparative politics and comparative political sociology, of course, in which specialists on various
cases are asked to rate them on particular dimensions (press freedom ratings actually rely heavily on

this kind of procedure). This approach could also be attacked as impressionistic, but it has validity
in these fields in part because they have long traditions of detailed research on particular cases and,
very importantly, common conceptual frameworks that are the result of considerable theoretical
reflection. It will be a gradual process to get to that point in the comparative study of media systems.


The fifth issue concerns convergence. In the last chapter of Comparing Media Systems we observe
that in 1970s the differences among the three groups of countries characterized by our three models
were quite dramatic; a generation later, by the beginning of the 21st century the differences were
eroded to the point that it is reasonable to ask whether a single global media model is displacing the
national variation of the past, at least among the advanced capitalist democracies discussed in this
book (Hallin & Mancini, 2004, p. 251). We went on to offer an analysis of the forces of change
within European societies which had produced significant erosion of the national differences which
are the main focus of our book. This shift can be seen roughly as a convergence toward the liberal
model in the sense that market forces have become more dominant in European media systems, and
media forms rooted in the political world of parties and organized social groups have declined. This
chapter has been a focus of criticism, particularly by scholars who study parts of the world beyond
Western Europe (Hadland, 2012), and we have sometimes been characterized as putting forward a
kind of end of history prediction of a global triumph of the Liberal model (McCargo, 2012).
We were surprised at that interpretation, in part because that chapter had been added at the end of
our book in response to early criticisms of our unpublished manuscript which took us to task for
focusing excessively on differences among Western media systems that may have been very sharp a
generation ago, but were less so today. What we stated at the beginning of the chapterthat the
differences among the countries we studied had diminished over time was simple and we really
meant it to be taken in that way. This seems to us to clearly be true. At one point in Continental
Europe party newspapers were dominant, now commercial newspapers dominate everywhere; at one
time only some of these countries had commercial broadcasting, now commercial broadcasting has
the largest market share virtually all across the region. Some elements of this process will be found in
other regions; media have become more commercialized in much of the world, for instance, and
journalistic practices rooted in the Liberal model have had important influence (e.g., Waisbord,
2000). One of the most fundamental premises of our book, though, is the idea that media practices
take their meaning within wider structural and cultural contexts; we do not believe that media
practices or institutions can simply be transferred across contexts without being transformed, and we
did not intend the discussion of convergence in Western media systems to be projected onto the rest
of the world any more than any other part of our analysis. Even within the scope of our analysis, we
did not mean to be taken as endorsing the idea that a single global media model would indeed
displace all national differences. The subtitle of the chapter was The forces and limits of
homogenization [emphasis added], and we concluded that differences among media systems remain
substantial and are likely to prevent complete homogenization of media systems for the foreseeable
future (p. 295). Our own research (Benson & Hallin, 2007) and research by others (Aalberg, van
Aelst, & Curran, 2010) has subsequently confirmed the persistence of the kinds of differences that are
the main focus of our analysis. We also observed that convergence was not a one-way street, and that
the Liberal model was also in a process of flux, with journalistic professionalism declining in the
U.S., especially with respect to information-centered journalism giving way to opinion-centered

journalism and a rise in political parallelism.5 Finally, we offered a critique of straight-line

evolutionist conceptions of change in media systems which assume an inevitable movement toward
differentiation of media from other subsystems.
An aside here about differentiation and the Liberal Model in our analysis is worth adding. In a
book published just as our chapter was going to press, Curran (2011) argues that we present an
overly rosy view of the history of the Liberal or, specifically the American media system, which is
his real focus as one of increasing differentiation of the media from the political system His
primary focus is on U.S. media coverage of foreign policy. He reviews the extensive literature (some
of it our own!) documenting the history of media deference to political elites as the U.S. has perused
an imperial foreign policy in the Post-World war II period. And he concludes (p. 34) American
media may be independent of political parties, but they are tethered to Americas political class (and
are sometimes strongly influenced by government). But this is exactly what we argue. Our section on
the role of the state in liberal systems concludes by developing the argument that structural limits on
the role of the state in Liberal systems do not mean that the state has less influence on the content of
news, in part, though not exclusively, because of the imperial histories of the U.S. and Britain. We
cite the same literature Curran reviews, and conclude that news content in liberal systems is
powerfully shaped by interpretive frameworks originating within the institutions of the state (p.
234). Later, in our discussion of differentiation theory, we make the argument that although media
may have become more differentiated from political parties and social groups, it can be argued they
became more integrated with the state, in different ways in the different systems we discuss, including
the Liberal. Curran goes on to charge us with exaggerating the way [commercialization] has
encouraged media independence from political power (p. 46). Commercialization has obviously
decreased the role of the state in Western Europe in important ways, as funder and regulator, for
example, and has increased the centrality of media relative to parties. That does not mean, however,
that it increases the independence of media from structures of power in any way, especially economic
power. We try to make this clear in our critique of differentiation theory. We do, of course, consider
the effect of commercialization on media and social power an empirical question, and we would
consider this a key area for comparative research.
In his discussion of our convergence chapter, Humphreys (2011) contrasts the convergence
hypothesis with what he calls a historical institutionalist perspective which in the study of media
policy has shown considerable continuity in national policy styles. We are very sympathetic to this
approach, and indeed the main body of our book relies quite heavily on it. But it does seem there is a
danger as critics of our initial manuscript had argued in advancing an argument that is too rigidly
path-dependent and does not offer the possibility to account for change. There is not a rigid choice
between the two perspectives. Change does take place, both through internal processes and through
exogenous forces, but it may be modified or take different routes in different systems. Professional
practices and attitudes, or legislative or structural arrangements adopted from other countries, for
example, may be delayed or transformed by the limits that each government may put to this adoption,
or by cultural resistance based on past experience, producing not pure continuity or convergence, but
hybrid systems and multiple routes of change.

We are gratified by the wide use of our analysis, and at the same time well aware that it has many
limitations. We have tried here to clarify our intent on several points. First, we want to underscore
the point that our analysis is tied to the 18 cases we analyzed, and we did not intend it to be applied
beyond those cases; we believe that new theory applying to other media systems will have to arise, as
our framework did, out of concrete analysis of those systems. Second, we did not imagine that our
analysis provided a fully comprehensive and elaborated framework for comparative analysis of
media systems. We saw it as a beginning, and we hope that other scholars will try to build on it and
not, once again, simply apply it. Third, many commentators have expressed reservations about the
use of our three models for classifying media systems, fearing that it leads to overly broad and
general (and hence simplistic) analysis. And in truth we share that concern. We believe our models
are useful as a starting point for thinking about similarities and differences in patterns of media
system development, and the extremely wide use of those models seems to confirm that we were right
about their heuristic power. But we share the concern that an overemphasis on these models and on
classification of media systems as an end in itself is a potential danger. Third, and related to this
point, we want to underscore our plea in the book that the categorization of media systems under one
of the three models was not intended to substitute for the more detailed discussion of their similarities
and differences and the reasons for these. Finally, we want to clarify that our chapter on convergence
was not intended to be taken as a claim that in fact we were headed to a complete elimination of
differences among media systems.
We would like to conclude with some reflections about how we see the route ahead in
comparative analysis of media systems or media and politics. The first thing to say is simply that we
assume methodological pluralism is legitimate and indeed essential. We expect that there will be
many styles of comparative analysis, with different units of analysis, for example, or different scopes
ranging from large N-studies to single case studies, some quantitative and some qualitative, some
more focused on hypothesis-testing and some on conceptual clarification or rethinking. This is
normal; this is how a field should develop.
We have a particular interest in analysis at the broad level of the system as a whole, and in
understanding how the development of media systems can be understood in the context of social,
political, and cultural development more generally. In this sense, we would see our work as very
much in the tradition of what Humphreys calls a historical-institutional approach. Our use of media
system models much imitated and much criticized, as we have discussed above is related to this
concern, and we do think that if approached carefully this can be very valuable. We are particularly
interested to see this kind of analysis carried forward on a wider range of systems not, as we have
stressed, by scholars trying to fit other systems into our categories, but by scholars trying to think
through what other patterns of media systems exist and what social and historical context explains
them At the same time, we are well aware that models can become reified and do not see
classification per se as particularly valuable for advancing comparative analysis. We are also
sympathetic to those who have argued that constructing typologies should not distract us from going

forward with the analysis of relations among variables.

One aspect of our approach we would like to stress is its focus on interdisciplinary theorizing.
Journalism and media studies have generally, in our view, remained much too isolated from other
branches of social science, and too often we see scholars trying to write about the role of journalism
in politics in a certain region of the world without any reference to the existing literatures on the
nature of the state in that region, the nature of civil society, the development of political culture, or the
pattern of economic development. To us, this is crucial: comparative analysis of media systems is
about understanding those systems in the context of history, culture, and social and political structure
more generally, and this means that scholars of communication need to be in dialogue with other
fields, including comparative politics, political sociology, and political anthropology. We are very
interested, for example, in rational-legal authority and its relationship to the major variables in our
study, especially journalistic professionalism and political parallelism. We suspect that thinking
through this relationship would be very useful for understanding democratic transitions (and nontransitions or hybridizations). Rational-legal authority is something that is simultaneously structural
and cultural and has been theorized in a number of different branches of social science; addressing
this kind of issue requires moving across disciplinary boundaries.
The question of quantification, finally, leads us to some important observations about ways
forward in comparative analysis. It is clear that better quantitative data, across a wider range of
systems, are crucial to advancing the field. They offer possibilities, especially for exploring more
precisely the relationship between variables, as for example, in Aalberg, van Aelst, and Currans
(2010) study of the relationship between media commercialization and political content in the news.
It is true, of course, that many of the things political communication and media scholars are interested
in are likely to prove difficult to quantify, including our dimensions of political parallelism and
journalistic professionalization, which certainly present challenges in operationalization. But in our
view, it is often precisely those things that are difficult to quantify that are important and interesting,
and we need to figure out ways to study them, quantitatively or not. In the case of political parallelism
there are some kinds of quantitative data already available on many systems that are very useful, and
there could certainly be more of this sort of data produced on a wider range of systems. This includes
survey data on the partisan affinities of media audiences as well as the very valuable data produced
by Donsbach and Patterson (2004) on such things as the relationship between partisanship and
journalists career paths. As Norris (2009, pp. 335-336) observes, however, we can only go so far in
studying media partisanship without looking at content, a massive undertaking, she worries, and one
clearly open to problems of interpretation. It is true enough that content analysis across media
systems poses many methodological problems, but it can clearly be done, and has been in a number of
fine studies (e.g., Semetko et al., 1991; Ferree et al., 2002). Content analysis across systems, guided
by comparative theory, is in our view one of the most fundamental needs in our field. The undertaking
does not have to be massive: unless we move immediately to the level of large-N studies but we
believe a lot can be accomplished with more modest, small-N comparisons (Mancini & Hallin,
There will, of course, still be problems of interpretation including the subjectivity and context-

dependent nature (Mancini & Hallin, 2012) of assessments of partisanship in media content. To us, it
is precisely in thinking through these problems of interpretation that comparative analysis is
interesting and valuable. We would like to see the field directly engage with, among other things,
partisanship and professionalism and confront them both conceptually and methodologically. This
may require putting quantitative data in context using other methods, looking not only, for example, at
content but also at the interaction of journalists with other social actors in the production of that
content, or at audience reception. In the case of journalistic professionalism, it would mean not only
looking at things we might quantifypatterns of recruitment of journalists, or attitudes we might
measure in a surveybut also doing field observation and case studies to observe the professional
routines and patterns of interaction of reporters and politicians. We believe, in other words, that
detailed, multi-method case studies, informed by comparative theory and designed to common issues
with other such studies, will be fundamental to the progress of the field.

(NI r o l

1 One might be Australia, as discussed in Jones and Pusey (2010).

We make this point in relation to the placement of points on our triangle diagram, Hallin and Mancini (2004, p. 71).
Hallin has served as a ratings reviewer for Freedom House.
Lakatos also writes, A m o d e l is a set of initial conditions (possibly together with some of the observational theories) which one
knows is bo u n d to be replaced during the further development of the [research] program. Lakatos is thinking about a different
kind of model than ours, based in the natural sciences. But we share his view of science in the sense that we believe that a
conceptual framework has value as it is used in research, and does not acquire value only after it has been verified.
5 Hallin (1992, 2000, 2006) has written a series of articles on the crisis of the American model of journalistic professionalism, once
widely seen as the end-product of the natural evolution of media systems.

Aalberg, T., van Aelst, P., & Curran, J. (2010). Media systems and the political information environment: A cross-national comparison.
International Journal o f Press/Politics, 15(3), 255-271.
Almond, G., & Powell, P. (1966). Comparative politics: A developm ental approach. Boston, MA: Little, Brown & Company.
Bastiansen, H. (2008). Media history and the study of media systems. M edia H istory, 14(1), 95-112.
Benson, R., & Hallin, D. C. (2007). How states, markets and globalization shape the news: The French and U.S. national press, 1965
1997. E uropean Journal o f Communication, 22(1), 27-48.
Couldry, N. (2005). Comparing media systems: Three models of media and politics. Political Studies Review, 3(2), 308.
Donsbach, W., & Patterson, T (2004). Political news journalists: Partisanship, professionalism and political roles in five countries. In F.
Esser & B. Pfetsch (Eds.), Com paring political communication: Theories, cases a n d challenges (pp. 251-270). Cambridge,
UK: Cambridge University Press.
Ferree, M., Gamson, W. A., Gerhards, J., & Rucht, D. (2002). Shaping abortion discourse: D em ocracy a n d the pu b lic sphere in
Germany a n d the United States. N ew York, NY: Cambridge University Press.
George, A. L., & McKeown, T (1985). Case studies and theories of organizational decision-making. A dvances in Inform ation
P rocessing in O rganizations, 2, 21-58.
Hadland, A. (2012). Africanizing three models of media and politics: The South African experience. In D. Hallin &P.Mancini(Eds.),
C om paring media systems b eyond the Western world. Cambridge, UK: Cambridge University Press.
Hallin, D. C. (1992). The passing of the high modernism of American journalism. J o u rn a l o f Communication, 42(3), 14-25.
Hallin, D. C. (2000). Commercialism and professionalism in the American news media. In J. Curran & M. Gurevitch (Eds.), M ass
media a n d society (3rd edition, pp. 218-237). London: Arnold.
Hallin, D. C. (2006). The passing of the high modernism of American journalism revisited, Political Communication Report, 16(1),
Hallin, D., & Mancini P. (2004). Com paring media systems: Three models o f media a n d p o litic s. Cambridge, UK: Cambridge
University Press.
Hallin, D., & Mancini, P. (2010). Preface. In B. Dobek-Ostrowska, M. Glowacki, K. Jakubowicz, & M. Skosd (Eds.), Comparative
media systems: European a n d global perspectives (pp. xi-xiv). Budapest: Central European University Press.
Hallin, D., & Mancini P. (Eds.) (2012). Com paring media systems b eyond the Western w o rld . Cambridge, UK: Cambridge University
Hardy, J. (2008). Western media systems. London: Routledge.
Humphreys, P. (2012). A political scientists contribution to the comparative study of media systems in Europe: A response to Hallin and
Mancini. In N. Just & M. Puppis (Eds.), Trends in communication p o licy research: New theories, methods, subjects (pp. 141
158). Bristol: Intelect.
Jones, P. K., & Pusey, M. (2010). Political communication and media system: The Australian canary. M edia, Culture & S o ciety,
32(3), 451-471.
Katzenstein, P. J. (1985). Small states in w orld markets: Industrial p o licy in Europe. Ithaca, NY: Cornell University Press.
Kogen, L. (2010). Savage deregulation in Thailand: Expanding Hallin and Mancinis European model. M edia, Culture a n d S o ciety,
3 2(2), 333-345.
Kraidy, M. (2012). The rise of transnational media systems: Implications of pan-Arab media for comparative research. In D. Hallin & P.
Mancini (Eds.), Com paring media systems b eyond the Western world. Cambridge, UK: Cambridge University Press.
Lakatos, I. (1970). Falsification and the methodology of scientific research programs. In I. Lakatos & A. Musgrave (Eds.), Criticism
a n d the grow th o f know ledge (pp. 91-196). London: Cambridge University Press.
Livingstone, S. (2003). On the challenges of cross-national comparative research. European Jo u rn a l o f Com m unication, 1<3(1), 477
Lynch, M. (2006). Voices o f the new A rab public: Iraq, Al-Jazeera a n d M iddle E ast politics to d a y . N ew York, NY: Columbia
University Press.
Mancini, P., & Hallin, D. C. (2012, forthcoming). Some caveats about comparative research in media studies. In H. A. Semetko & M.
Scammell (Eds.), SAG E h a n d b o o k o f political communication. Thousand Oaks, CA: SAGE.
McCargo, D. (2012). Partisan polyvalence: Characterizing the political role of Asian media. In D. Hallin & P. Mancini (Eds.),
Com paring media systems b eyond the Western world. Cambridge, UK: Cambridge University Press.
McQuail, D. (2005). Comparing media systems: Three models of media and politics. European Jo u rn a l o f C om m unication, 20(2),
Norris, P. (2009). Comparative political communications: Common frameworks or Babelian confusion? G overnm ent a n d Opposition,
4 4(3), 321-340.
Patterson, T. (2007). Comparing media systems: Three models of media and politics. Political Communication, 24(3), 329-331.
Puppis, M. (2009). Introduction. Media regulation in small countries. International Communication Gazette, 71(1-2), 7-17.
Roudakova, N. (2012). Comparing processes: Media, transitions, and historical change. In D. Hallin & P. Mancini (Eds.), Comparing

media systems b eyond the Western world. Cambridge, UK: Cambridge University Press.
Semetko, H. A., Blumler, J. G., Gurevitch, M., & Weaver, D. H. (1991). The form ation o f cam paign agendas. A comparative
analysis o f p a rty a n d media roles in recent Am erican a n d British elections. Hillsdale, NJ: Erlbaum Associates.
Strombck, J., 0rsten, M., & Aalberg, T. (Eds.) (2008). C om m unicatingpolitics: Political communication in the Nordic co u n tries.
Gotenborg: Nordicom.
Waisbord, S. (2000). Watchdog jou rn a lism in Latin America: News, accountability a n d dem ocracy. N ew York, NY: Columbia
University Press.
Zhao, Y. (2012). Understanding Chinas media system in a world historical context. In D. Hallin & P. Mancini (Eds.), Comparing
media systems b eyond the Western world. Cambridge, UK: Cambridge University Press.

Edited by
Frank Esser
Thomas Hanitzsch

R Routledge

Taylof & Franci Group


First published 2012

by Routledge
711 Third Avenue, N ew York, N Y 10017
Simultaneously published in the UK
by Routledge
2 Park Square, Milton Park, Abingdon, Oxon OX14 4RN
Routledge is an imprint o f the Taylor & Francis Group, an informa business
2012 Taylor & Francis
The right of Frank Esser and Thomas Hanitzsch to be identified as the authors of the editorial material, and of the authors for their
individual chapters, has been asserted in accordance with sections 77 and 78 of the Copyright, Designs and Patents Act 1988.
All rights reserved. N o part of this book may be reprinted or reproduced or utilized in any form or by any electronic, mechanical, or other
means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without
permission in writing from the publishers.
Trademark notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and
explanation without intent to infringe.
Library o f Congress Cataloging in Publication Data
Handbook of comparative communication research / edited by Frank Esser and Thomas Hanitzsch.
p. cm.
Includes bibliographical references.
1. Communication, International. 2. Communication Research. I. Esser, Frank, 1966- II. Hanitzsch, Thomas,
1969P96.I5.H368 2012
302.2 dc23
ISBN: 978-0-415-80271-0 (hb)
ISBN: 978-0-415-80275-8 (pb)
ISBN: 978-0-203-14910-2 (eb)
Printed and bound in the United States of America on acid-free paper.