(Springer Series in Social Psychology) Barbara a. Winstead, Valerian J. Derlega (Auth.), Valerian J. Derlega, Barbara a. Winstead (Eds.)-Friendship and Social Interaction-Springer-Verlag New York (198 (1)

SSSP
Springer
Series in
Social
Psychology
SSSP
Friendship and
Social Interaction
Edited by
Valerian J. Derlega and
Barbara A. Winstead
Springer-Verlag
New York Berlin Heidelberg Tokyo
Valerian J. Derlega
Department of Psychology
Old Dominion University
Norfolk, Virginia 23508
U.SA
Barbara A. Winstead
Department of Psychology
Old Dominion University
Norfolk, Virginia 23508
U.SA
With 6 Figures
Library of Congress Cataloging in Publication Data
Main entry under title:
Friendship and social interaction.
(Springer series in social psychology)
Bibliography: p.
Includes index.
1. Friendship. 2. Interpersonal attraction.
3. Social interaction. I. Derlega, Valerian J.
II. Winstead, Barbara A III. Series.
HM132.F72 1986
158'.25
85-26059
1986 by Springer-Verlag New York Inc.
Softcover reprint of the hardcover 1st edition 1986
All rights reserved. No part of this book may be translated or reproduced in any form
without written permission from Springer-Verlag, 175 Fifth Avenue, New York, New
York 10010, U.SA
The use of general descriptive names, trade names, trademarks, etc. in this
publication, even if the former are not especially identified, is not to be taken as a
sign that such names, as understood by the Trade Marks and Merchandise Marks
Act, may accordingly be used freely by anyone.
Typeset by Ampersand Publisher Services, Inc., Rutland, Vermont.
987 6 5 432 1
ISBN-13: 978-1-4612-9344-6
DOl: 10.107/978-1-4612-4880-4
e-ISBN-13: 978-1-4612-4880-4
To Wally and Stella Derlega, Bob and Lois Winstead,

William Brenner, Irene Solet, John Derlega, Ann Tekla Winstead Derlega
Preface
A neglected topic in the field of personal relationships has been the study of
friendships. Social psychologists have studied how and why individuals are
attracted to one another and the processes of interaction during initial
encounters, but they have not paid much attention to ongoing friendships.
A major goal of the present volume is to develop theories and integrate
research on the development and maintenance of friendships. Another
major goal is to build bridges between social psychologists and other social
scientists by presenting an interdisciplinary approach. Although a majority
of the contributors are social psychologists, other authors include sociologists as well as developmental, personality, and clinical psychologists. The
chapters also present research on friendship based on a wide range of
research methodologies, including laboratory research as well as longitudinal, naturalistic, and clinical studies. Hence, the book incorporates a
variety of conceptual and methodological approaches that should contribute to a cross-fertilization of ideas among disciplines.
The first chapter, by Barbara A. Winstead and Valerian J. Derlega,
provides an overview of theory and research on friendship. The second
chapter, by Daniel Perlman and Beverley Fehr, provides a summary and
conceptual critique of social psychological theories of social attraction that
are relevant to the study of friendship.
Adopting a developmental approach, Duane Buhrmester and Wyndol
Furman, in Chapter 3, demonstrate the particular importance of friendship
during middle childhood and adolescence in fulfilling interpersonal needs.
Buhrmester and Furman draw heavily on the theoretical work of Harry
Stack Sullivan concerning social-personality development. The fourth
chapter by Mark Snyder and Dave Smith demonstrates how the personality dimension self-monitoring affects individuals' definitions of friendship
and the activities and interests that they might want to share with others. In
Chapter 5, Barbara A. Winstead provides a thorough review of studies of
viii
Preface
adult same-sex friendships, indicating how gender and sex roles affect men's
and women's friendships.
Chapter 6, by John H. Berg and Margaret S. Clark, provides a critical
discussion of when close friendships are likely to develop gradually (the
traditional view held by social psychologists) or quickly. Berg and Clark
indicate that people's decisions about which relationships they want to have
are often made early in or at the beginning of a relationship. In Chapter 7,
Dorothy Miell and Steve Duck explore the communication strategies used
by individuals to form friendships. They consider how controlling the flow
of one's own and the other person's self-disclosures affects relationship
development.
Chapter 8, by Robert M. Milardo, and Chapter 9, by Maureen T.
Hallinan, concern organizational or structural factors that affect friendships. Milardo presents a theoretical model indicating how characteristics
of social networks affect the development of relationships. Hallinan
analyzes organizational determinants of interracial friendship in desegregated schools. She shows how classroom organization affects the availability and stability of interracial friendships.
In Chapter 10, Gary Alan Fine explores important features of friendships
at work. Drawing on in-depth interviews and participant observation, Fine
shows how the distinctive culture that emerges in work settings (with its own
rules, values, and norms) influences friendships.
Chapter 11, by Raymond Fleming and Andrew Baum, examines the
relationship between friendship, social support, and stress. Different models
of the role of friendship in coping with stressful events are critically
reviewed.
Although most people report having one or more friends, what is it like
for people who do not have friends? Cecilia H. Solano, in Chapter 12,
examines how people conduct their lives without friends. Are people
without friends inevitably lonely or do they have substitutes for friends that
can fulfill their interpersonal needs? Chapter 13 focuses on individuals who
experience difficulties in developing friendship. Jeffrey E. Young presents a
cognitive-behavioral model for understanding and treating patients who
have long-term and chronic difficulties in establishing friendships.
The editors express their gratitude to the individuals who contributed to
the successful completion of this book. Special thanks go to the authors for
their enthusiastic participation in the project, their patience, and their
thought-provoking chapters. We gratefully acknowledge the assistance of
the excellent staff at Springer-Verlag/New York for encouraging us in our
work on this project. We also benefited from the helpful advice of our
colleagues and friends-John Berg, William Brenner, Warren Jones,
Stephen T. Margulis, and Daniel Perlman-at various stages in the book's
progress. The original idea for the volume emerged from our participation
in the first and second International Conferences on Personal Relationships
held at Madison, Wisconsin. We and other social scientists interested in
Preface
ix
studying personal relationships owe appreciation to Steve Duck and Robin

Gilmour for their leadership in organizing these important meetings. We
are also delighted to recognize the significant part played by two important
persons in our lives, John Derlega and Ann Tekla Winstead Derlega. John
and Ann have not encouraged us to finish the book (in fact, they have
encouraged us to do the opposite on many occasions) nor have they
critically read or typed any chapters. But they have enriched our lives in
numerous ways, and we love them for it.
Norfolk, Virginia
Valerian 1. Derlega
Barbara A. Winstead
Contents
1. Friendship and Social Interaction: An Introduction .............

Barbara A. Winstead and Valerian J. Derlega
Theoretical Issues in the Study of Friendship ... . . . . . . . . . . . . . .

Methodological Issues in the Study of Friendship .............
Applications of Friendship Theory and Research. . . . . . . . . . . . . .
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2
3
5
5
2. Theories of Friendship: The Analysis of Interpersonal Attraction. . . .

Daniel Perlman and Beverley Fehr
Reinforcement Theories ....................................

Exchange and Equity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Cognitive Consistency Theories. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Developmental Theories . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
A Comparison of Theories of Attraction. . . . . . . . . . . . . . . . . . . . . .
An Evaluation of Theories of Attraction. . . . . . . . . . . . . . . . . . . . . .
Conclusion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
12
16
19
24
27
37
3. The Changing Functions of Friends in Childhood:

A N eo-Sullivanian Perspective ...............................
Duane Buhrmester and Wyndol Furman
41
Sullivan's Theory of Social-Personality Development. . . . . . . . . .

Empirical Evidence ........................................
Summary and Conclusions .................................
41
51
59
4. Personality and Friendship: The Friendship Worlds of

Self-Monitoring . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Mark Snyder and Dave Smith
63
A Scientific Approach to the Study of Friendship .............

Conceptions of Friendship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
65
67
Contents
xii
Behavioral Manifestations of Friendship .....................

The Population of Friendship Worlds. . . . . . . . . . . . . . . . . . . . . . . .
Understanding the Nature of Friendship .....................
The Paradox of Friendship .................................
69
71
73
78
5. Sex Differences in Same-Sex Friendships .....................

Barbara A. Winstead
81
Sex Differences in Adult Same-Sex Friendships. . . . . . . . . . . . . . .

A Developmental Perspective on Sex Differences in Same-Sex
Friendships ...............................................
Intrasex Differences in Same-Sex Friendships. . . . . . . . . . . . . . . . .
Understanding Sex Differences in Interpersonal Behavior. . . . . .
Sex as a Contextual Variable. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
81
6. Differences in Social Exchange Between Intimate and Other

Relationships: Gradually Evolving or Quickly Apparent? .........
John H Berg and Margaret S. Clark
85
88
91
93
95
101
Some Differences in Exchange Between Close and Casual

Relationships. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
When Are Differences in Exchange Detectable? . . . . . . . . . . . . . ..
Concluding Comments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
102
105
124
7. Strategies in Developing Friendships ..........................
129
Purposive Communication. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Information Gathering and Exchange. . . . . . . . . . . . . . . . . . . . . . ..
Accounts of the Role and Function of Personal Information
Exchange ............................................. . . ..
The Level of Awareness and Strategic Planning in Interactions. ..
130
131
Dorothy Miell and Steve Duck
132
141
8. Personal Choice and Social Constraint in Close Relationships:

Applications of Network Analysis ............................
Robert M Milardo
145
Conceptualizing Social Networks. . . . . . . . . . . . . . . . . . . . . . . . . . ..

Structural Attributes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Toward a Theory of Structural Interdependence. . . . . . . . . . . . . ..
Concluding Commentary. . ..... .. . .. .... .... . . ... . ... . ... ..
146
149
156
162
9. School Organization and Interracial Friendships ...............

Maureen T. Hallinan
167
Theoretical and Empirical Research on Interracial

Friendliness ..............................................
168
Contents
Organizational Characteristics of Schools and Interracial

Friendships ...............................................
Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
10. Friendships in the Work Place ...............................
Gary Alan Fine
xiii
172
180
185
Culture and Friendship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 185

Types of Work Relations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 187
The Workplace as Setting. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 188
The Content of Work. . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 190
Organizational Friendships and Organizational Culture ....... 192
Men and Women Together. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 197
Friendship Amidst the Hierarchy. . . . . . . . . . . . . . . . . . . . . . . . . . .. 199
Conclusions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 202
11. Social Support and Stress: The Buffering Effects
of Friendship ..............................................
Raymond Fleming and Andrew Baum
207
Stress. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Friendship Formation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Social Support and Friendship. . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Strategies for Future Research. . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
209
211
213
222
12. People Without Friends: Loneliness and Its Alternatives .........

Cecilia H Solano
227
Definitional Issues . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Loneliness ................................................
Function of Friends. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Alternatives to Friends .....................................
Therapeutic Interventions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ..
Summary .................................................
229
231
233
236
241
242
13. A Cognitive-Behavioral Approach to Friendship Disorders .......

Jeffrey E. Young
247
Cognitive-Developmental View of Friendships. . . . . . . . . . . . . . ..

Classification of Friendship Disorders .......................
Cognitive Therapy for Friendship Disorders. . . . . . . . . . . . . . . . ..
Summary.................................................
247
252
261
275
Author Index ................................................... 277

Subject Index .................................................. 285
Contributors
Andrew Baum, Department of Medical Psychology, School of Medicine,

Uniformed Services University of the Health Sciences, Bethesda, Maryland 20814, U.S.A.
John H Berg, Department of Psychology, University of Mississippi, University, Mississippi 38677, U.S.A.
Duane Buhrmester, Department of Psychology, University of California at
Los Angeles, Los Angeles, California 90024, U.S.A.
Margaret S. Clark, Department of Psychology, Carnegie-Mellon University,
Pittsburgh, Pennsylvania 15213, U.S.A.
Valerian 1. Derlega, Department of Psychology, Old Dominion University,
Norfolk, Virginia 23508, U.S.A.
Steve Duck, Department of Communications Studies, University of Iowa,
Iowa City, Iowa 52242, U.S.A.
Beverley Fehr, Division of Family Sciences, University of British Columbia,
Vancouver, British Columbia, Canada V6T IW5
Gary Alan Fine, Department of Sociology, University of Minnesota, Minneapolis, Minnesota 55455, U.S.A.
Raymond Fleming, Department of Behavioral Medicine, Mount Sinai
Medical Center, Milwaukee, Wisconsin 53201, U.S.A.
Wyndol Furman, Department of Psychology-Child Study Center, University
of Denver, University Park, Denver, Colorado 80208, U.S.A.
Maureen T. Hallinan, Department of Sociology, University of Notre Dame,
Notre Dame, Indiana 46556, U.S.A.
xvi
Contributors
Dorothy Miell, Psychology Discipline, Faculty of Social Sciences, The Open
University, Walton Hall, Milton Keynes MK7 6AA, England
Robert M Milardo, Department of Child Development and Family Relation-
ships, School of Human Development, University of Maine at Orono,

Orono, Maine 04469, U.S.A.
Daniel Perlman, Division of Family Sciences, University of British
Columbia, Vancouver, British Columbia, Canada V6T lW5
Dave Smith, Department of Psychology, University of Minnesota, Min-
neapolis, Minnesota 55455, U.S.A.
Mark Snyder, Department of Psychology, University of Minnesota, Min-
neapolis, Minnesota 55455, U.s.A.
Cecilia H Solano, Department of Psychology, Wake Forest University,
Winston-Salem, North Carolina 27109, U.S.A.
Barbara A. Winstead, Department of Psychology, Old Dominion University,
Norfolk, Virginia 23508, U.S.A.
Jeffrey E. Young, Department of Psychiatry, College of Physicians and
Surgeons, Columbia University, New York, New York 10027, U.S.A.
Chapter 1
Friendship and Social

Interaction: An Introduction
Barbara A. Winstead and Valerian J. Derlega
In the last decade there has been an upsurge of scholarly work on personal
relationships, focusing on how people establish and maintain interpersonal
closeness with others (e.g., between parents and children, spouses, lovers,
therapists, and clients). Among the various relationships being studied,
friendship has emerged as a major topic.
Surprisingly, although social psychologists have studied in great detail
how persons are attracted to one another (Berscheid & Walster, 1978; Byrne,
1971; Newcomb, 1961), they have paid relatively little attention to ongoing
friendships. Also, far more attention has been paid to romantic and/or
marital relationships (Gottman, 1979; Hill, Rubin, & Peplau, 1976; Rubin,
1970) than to platonic relationships. Most people will probably have one or
more close friendships that they will consider to be important in their lives;
yet psychologists have not until recently investigated the nature and
development of friendships, how individual differences affect friendships,
or how the presence or absence of friendships affect an individual's
behavior.
This volume focuses primarily on the social psychology of friendships,
and we hope that the chapters (emphasizing theory development and
integration of research and theory) will help expand our knowledge about
the social processes involved in friendship. Despite diverse research on
friendship (see Duck & Gilmour, 1981), this area has been somewhat
deficient in the development of theories and conceptually based research
and applications. Published research has been mainly in the area of
children's friendships (see Asher & Gottman, 1981; Foot, Chapman, &
Smith, 1980; Rubin, 1980).
Our book presents a major opportunity to expand understanding of the
nature of friendship from a social psychological perspective. We hope that
the chapters will serve as a major impetus for future research and theory
development about friendship.
Barbara A Winstead and Valerian J. Dedega
In the introduction we address ourselves to three aspects of the study of

friendship: theory, research, and application. Regarding theory, we discuss
the issue of defining friendship and the current debate concerning the need
for a field of personal relationships. Although we do not consider reviewing
research to be the main purpose of the chapters presented here, there are
certainly a great many studies reported or referred to in these pages. Thus,
we comment on the methodolgy of studying friendship. Finally, we consider
possible applications of research and theory on friendship.
Theoretical Issues in the Study of Friendship

In order to have an adequate theory of friendship, we must have a definition
of friendship. Curiously, in reviewing the chapters in this volume, we find
that the issue of definition was not addressed. There is consideration of
different types of relationships (e.g., communal vs. exchange in Berg and
Clark, Chapter 6), but when authors refer to "friendship," they seem to rely
on a consensual, but unspecified, idea of what a friendship is. When doing
research, investigators may specify that the friendship be with a member of
the subject's own sex (see Miell and Duck, Chapter 7, and Buhrmester and
Furman, Chapter 3), but subjects are clearly expected to know what is meant
by a friend or a friendship.
When defining characteristics of friendship are specified, they tend to
concern (a) age: friends should be peers (McAdams, Healy, & Krause, 1984);
(b) kinship: friends should be nonkin (Solano, Chapter 12, in this volume),
although subjects sometimes mention family members as friends (Fischer,
1982); and (c) sexual involvement: friends should be platonic, not romantic
or sexual partners, although again spouses may be regarded as "best
friends." (Snyder and Smith's research, summarized in Chapter 4 of this
volume, indicates that personality differences on the self-monitoring
dimension also influence how people define their friendships and the
activities and interests they might want to share.)
Several investigators have attempted to capture the qualitative differences
between friends and lovers. Wright (1985) found that friendships are less
exclusive, less intense in emotional expression, less permanent, and less
regulated by social rules and expectations than are love relationships. Using
interviews with subjects regarding their reactions to paradigm cases of
friends or lovers, Davis (Davis, 1985; Davis & Todd, 1982) found that these
relationships are perceived to differ in terms of a passion cluster and a
caring cluster. Love but not friendship connotes fascination, sexual desire,
exclusivity, and a willingness to give the utmost. Their subjects, however,
regarded friendships as more stable than love relationships. On other
variables (e.g., trust, enjoyment, acceptance), friends and lovers were similar.
These findings indicate that in friendships there will be less physical
affection and passion (almost by definition) and also less emotional
Friendship and Social Interaction: An Introduction
involvement. Friendships are free of the expectation of exclusivity, but also

represent less commitment. It may be easier to contact a friend with whom
one has fallen out of touch than to contact a former lover, giving friendship
a durability that love relationships do not have. But, although love
relationships may be more breakable, when they last they are apparently
stronger. Friends may also be more forgiving in terms of failures to meet
expectations of the relationship. They provide somewhat less in terms of
caring, but they also demand less.
Given these qualitative differences between friendship and a love
relationship, we expect that the process of becoming, remaining, and losing
friends is different from that of becoming, remaining, and losing lovers or
spouses. Theories of interpersonal relationships tend to be generic, i.e.,
presumed to be applicable to any relationship (see Perlman and Fehr,
Chapter 2). We are not recommending a different theory for each type of
relationship, but we do need theories with concepts that can explain how
different relationships evolve and devolve.
Currently, there is a lively debate among social scientists about whether or
not a need exists for a separate field of personal relationships (Hinde, 1985;
Surra, 1985). Berkowitz (interviewed by Surra, 1985) argued that we should
use an application of basic psychological processes to study interpersonal
relationships. This approach suggests that we can understand an interpersonal interaction by understanding the psychology of the individual and
the stimulus value of the situation (including another individual). On the
other hand, Kelley (interviewed by Surra, 1985) stated that: " ... personal
relationships stand between society and persons. And personal relationships have their own internal dynamics. They are not only affected by
persons and by social organizations, but they also playa causal role. That is,
personal relationships affect social organizations and persons" (p. 4).
Similarly, Hinde (1985) agrees that there is a need for concepts for studying
phenomena at one level of complexity (e.g., relationship) that may be
irrelevant to the study of psychological processes at another level of
complexity (e.g., the individual). In this volume Buhrmester and Furman
(Chapter 3), Berg and Clark (Chapter 6), and Miell and Duck (Chapter 7)
present the study of friendships at a level of complexity beyond individuals
or situations. Other contributors approach the understanding of friendship
at an even higher level of complexity, based on sociostructural variables
such as networks (Milardo, Chapter 8), school systems (Hallinan, Chapter
9), and work places (Fine, Chapter 10).
Methodological Issues in the Study of Friendship

Because of its rigor and convenience, social psychologists have preferred
laboratory research to naturalistic studies of intact relationships. For
example, the extensive research on social attraction has focused on what
Barbara A Winstead and Valerian J. Derlega
factors lead to initial attraction rather than on ebbs and flows in attraction
in ongoing relationships. It is important to understand why individuals are
attracted to one another, but it is also important to know, for instance,
whether similar factors predict attraction (and relationship maintenance)
after individuals have interacted extensively. (Berg and Clark in Chapter 6
provide a thorough discussion of how early predictors of attraction may also
operate later in the relationship.)
Conducting longitudinal research on relationship development, maintenance, and dissolution will make an important contribution to the study
of friendship (as well as other personal relationships). Longitudinal studies
being conducted by Hays (1984, 1985) and Berg (1984) show the developmental changes in a relationship over time. This research also provides data
on the antecedents of attraction early and later in a relationship. The use of
longitudinal designs will provide important descriptive data on typical
changes in the course of relationships as well as permit tests of hypotheses
about the nature of changes over time, a topic of considerable theorizing,
particularly in Altman and Taylor's (1973) theory of social penetration
processes and Levinger and Snoek's (1972) incremental exchange theory.
When behaviors that occur outside the laboratory are studied, concerns
about measurement naturally arise. Can subjects provide accurate reports of
these behaviors? Many findings concerning friendship have been obtained
using questionnaires asking subjects about their friends or their reactions to
or behaviors towards friends. Inconsistent findings from studies using selfreport and behavioral measures of self-disclosure (Winstead, Derlega, &
Wong, 1984) raise questions about basing our knowledge of friendship on
self-report measures. The diary technique, where subjects record information about their social interactions daily (Wheeler & Nezlek, 1977), is an
important improvement over general questionnaires. In a further attempt to
avoid distortion caused by time lapses, McAdams and Constantian (1983)
used beepers to obtain reports of subjects' current social interaction.
Although longitudinal research and improvements in naturalistic techniques will play an important role in future research on friendship,
laboratory studies emphasizing the experimental test of hypotheses can
continue to play a major role. For instance, Clark and Mills (1979)
developed a major distinction between communal and exchange relationships. If we assume that close friends maintain a communal orientation
(where individuals are concerned about one another's needs as well as
having a general concern for one another), it is possible experimentally to
manipulate this "set" for individuals in a laboratory setting and see the
effects on their behavior. (See Berg and Clark's Chapter 6 of this volume for
a review of these studies on communal and exchange relationships.)
Laboratory studies in which friends are the subjects can be conducted.
For example, how do subjects respond to a fear or anxiety induction when
they are with friends or with strangers? We found that subjects who were
told that they would be handling a snake experienced reductions in
depression and hostility while waiting with a same-sex friend but not while
waiting with a same-sex stranger (Winstead & Derlega, 1985). In this way the
effects of friendship on individual behavior can be investigated.
Applications of Friendship Theory and Research

Social psychologists need to build more bridges between their discipline
and clinical/counseling psychology. In part, the benefits of friendship can
be understood by examining the problems suffered by individuals who are
friendless, lonely, and/or shy (e.g., Margulis, Derlega, & Winstead, 1984). In
a nationwide survey, 26% of respondents reported that they have felt "lonely
or remote from other people ... during the past few weeks" (Bradburn,
1969). Certainly some of these persons seek help because of their loneliness.
In Chapter 13, Young describes a cognitive approach to changing patients'
self-perceptions and expectations of others in ways that increase their
chances of making friends. Theories of how friendships are typically
initiated and maintained may also be used to generate techniques for
enabling individuals without friends to establish friendships. It may also be
possible to make friendships more beneficial or to avoid problems in
friendships that may lead to their dissolution. For example, if we advocate
social penetration theory (Altman & Taylor, 1973), we might counsel clients
to increase gradually the level of intimacy of their self-disclosures in order to
deepen a relationship. If we adopt social interdependence theory (Kelley &
Thibaut, 1978), we might stress the importance of exchanging valued but not
isomorphic rewards. Finally, if we consider Berg and Clark's argument
(Chapter 6) that early evaluations of potential friends predict later
relationship development, we may focus on the importance of selecting
friends wisely.
On the other hand, some people probably do quite well without friends
and on occasion friends may generate more problems than they help solve.
Understanding the elements that make friendships beneficial will be as
useful as understanding how to have friends. Furthermore, when friendships are lacking either temporarily (e.g., when a person has moved to a new
location) or long-term (e.g., when a person lives in a remote area or chooses
not to have friends), alternative methods for coping with problems that
friends typically help with are needed. Solano (in Chapter 12) addresses the
issues offriendlessness (as opposed to loneliness) and ways in which people
cope with it.
Conclusion
In this chapter we have highlighted some of the major theoretical,
methodological, and applied issues encountered in research on friendship.
Friendship research has come a long way in its relatively short history, and
Barbara A Winstead and Valerian I Derlega
we can expect that new findings will be added to the literature at a rapid
rate. We hope that the following chapters will provide theoretical and
conceptual frameworks that will be useful to social psychologists and other
social scientists in understanding friendships and in suggesting directions
for new research.
References
AItman, I., & Taylor, D.A (1973). Social penetration: The development of interpersonal
relationships. New York: Holt, Rinehart & Winston.
Asher, S.R, & Gottman, J. (Eds.) (1981). The development of friendship. Cambridge:
Cambridge University Press.
Berg, IH. (1984). The development of friendship between roommates. Journal of
Personality and Social Psychology, 46,346-556.
Berscheid, E., & Walster, E. (1978). Interpersonal attraction (2nd ed.). Reading, MA:
Addison-Wesley.
Bradburn, N. (1969). The structure of psychological well-being. Chicago: AIdine.
Byrne, D. (1971). The attraction paradigm. New York: Academic.
Clark, M.S., & Mills, I (1979). Interpersonal attraction in exchange and communal
relationships. Journal of Personality and Social Psychology, 37, 12-24.
Davis, KE. (1985). Near and dear: Friendship and love compared. Psychology Today,
19(2), 22-30.
Davis, KE., & Todd, M.J. (1982). Friendship and love relationships. In KE. Davis &
T.O. Mitchell (Eds.), Advances in descriptive psychology (Vol. 2, pp. 79-112).
Greenwich, CT: JAI Press.
Duck, S., & Gilmour, R (1981). Personal relationships 2: Developing personal
relationships. London: Academic.
Fischer, C.S. (1982). What do we mean by "friend"? An inductive study. Social
Network, 3,287-306.
Foot, H.C., Chapman, AJ., & Smith, J.R. (1980). Friendship and social relations in
children. London: Wiley.
Gottman, J.M. (1979). Marital interaction. New York: Academic.
Hays, RB. (1984). The development and maintenance offriendship.Journal of Social
and Personal Relationships, 1, 75-97.
Hays, RB. (1985). A longitudinal study of friendship development. Journal of
Personality and Social Psychology, 48, 909-924.
Hill, c.T., Rubin, Z., & Peplau, L. A (1976). Breakups before marriage: The end of
103 affairs. Journal of Social Issues, 32, 147-68.
Hinde, R (1985). Toward a science of personal relationships. ISSPR News, 1(2), 35.
Kelley, H.H., & Thibaut, IW. (1978). Interpersonal relations: A theory of interdependence. New York: Wiley.
Levinger, G., & Snoek, IE. (1972). Attraction in relationships: A new look at
interpersonal attraction. Morristown, NJ: General Learning Press.
Margulis, S.T., Derlega, V.J., & Winstead, B.A. (1984). Implications of social
psychological concepts for a theory of loneliness. In V.J. Derlega (Ed.), Communication, intimacy, and close relationships (pp. 133-160). Orlando, FL: Academic
Press.
McAdams, D.P., & Constantian, C.A. (1983). Intimacy and affiliation motives in
daily living: An experience sampling analysis. Journal of Personality and Social
Psychology, 45,851-861.
McAdams, D.P., Healy, S., & Krause, S. (1984). Social motives and patterns of
friendship. Journal of Personality and Social Psychology, 47, 828-838.
Newcomb, T.M. (1961). The acquaintance process. New York: Holt, Rinehart &
Winston.
Rubin, Z. (1970). Measurement of romantic love. Journal of Personality and Social
Rubin, Z. (1980). Children's friendships. Cambridge, MA: Harvard University Press.
Surra, C. (1985). Conversations with Professors Leonard Berkowitz and Harold
Kelley: Is there a need for a field of personal relationships? ISSPR News, 1(1),28.
Wheeler, L., & Nezlek, 1. (1977). Sex differences in social participation. Journal of
Winstead, B.A, & Derlega, V.J. (1985), Benefits of same-sex friendships in a stressful
situation. Journal of Social and Clinical Psychology, 3,378-384.
Winstead, B.A, Derlega, V.l., & Wong, P.T.P. (1984). Effects of sex-role orientation
on behavioral self-disclosure. Journal of Research in Personality, 18, 541-553.
Wright, P.H. (1985). The acquaintance description form. In S.W. Duck and D.
Perlman (Eds.), Sage Series in Personal Relationships (Vol. 1, pp. 39-62). London:
Sage.
Chapter 2
Theories of Friendship: The Analysis

of Interpersonal Attraction
According to the popular song, "Love and marriage go together like a horse
and carriage, you can't have one without the other." Literally, of course, this
is not true; but, as the song implies, most people believe that love and
marriage augment each other. Both are better together than either is
separately.
Similarly, in the social sciences most scholars believe that theory and data
go together; each enhances the other. One of the prime functions of data is
to check the validity of theoretical predictions. If a theory cannot properly
predict actual outcomes, we typically either make revisions in the theory or
lower our evaluation of it.
Virtually all data collection rests on at least a few covert hunches and/or
implicit biases, but some psychologists begin studies without a wellarticulated, elaborated conceptual framework. All too often, the meaning
and significance of their findings remain unclear. Frequently, their results
seem contradictory and lack coherence. At best, they can predict behavior
under a limited set of conditions, but they have no basis for making
predictions as conditions change. Thus, their results contribute very little
toward the long-term, general solution of problems. In our opinion, such
investigations would benefit from a more explicitly formulated theoretical
base.
Theories serve several functions. They provide a vantage point for
determining what aspects of a situation are relevent and important to
consider. They contain constructs that help us label and classify phenomena. They indicate how phenomena are interrelated. Theories suggest
questions worth researching. As evidence builds, theories can summarize a
multitude of observations and facts into a few generalizations. Theories help
us to understand and explain. As relevant circumstances change, theories
provide guidelines as to how predictions may also need to be altered.
Finally, for both therapist and laypersons, theories can provide an
important basis for formulating strategies for promoting our well-being.
10
The purpose of this chapter is to present an overview of theories of

friendship. In launching into the chapter, we mention an important caveat:
Social scientists have not been very interested in theories ofJriendship. They
have concentrated instead on theories of interpersonal attraction. Although
the distinctions between friendship, personal relationships, and interpersonal attraction may initially seem small, they are important to the field
(Duck & Perlman, 1985). Research on interpersonal attraction focuses on
how much we like other people, especially during the early stages of a
relationship. Friendship involves many more components than does simply
liking (Davis & Todd, 1985). Indeed, liking may only be one of several
factors that determine such central aspects of relationships as frequency of
contact.
In investigating friendships, social psychologists have been shifting their
attention from interpersonal attraction to a broader focus on personal
relations. However, the existing theories, which date back to the 1950s, have
concentrated on earlier concerns in the field. Thus, as a synthesis of the
extant literature, in this chapter we primarily provide conceptual analyses of
the question: How well do we like various acquaintances?
Several specific theories of attraction have been formulated. In this
chapter we classify these conceptual formulations into four major groups:
reinforcement, exchange (or equity), cognitive consistency, and developmenta1. Each major perspective is reviewed in turn. We indicate how each
theoretical approach explains attraction and briefly cite illustrative research. In the final sections of the chapter, we compare and evaluate
theories. We also pose some additional questions that these theorists must
address if they are to provide a more comprehensive understanding of
friendships.
In the past 15 years, three edited books (Duck, 1977; Huston, 1974;
Murstein, 1971) and countless other publications have concentrated on
proposing or testing theoretical analyses of attraction. Obviously, in
preparing this chapter, we have had to be selective about which theories to
include. We were guided by what we see as the currently most influential
social psychological models. We have ignored general social psychological
theories (i.e., dissonance) that are less concerned with interpersonal
attraction as well as the more sociologically oriented perspectives on
attraction offered by individuals such as George Homans and George
McCall. Similarly, we have ignored the writing of scholars such as Robert
Selman and James Youniss, who are more concerned with people's
understanding of relationships than with friendship choice. We have
bypassed the notion of complementarity of needs because we do not believe
it has fared well in empirical tests. Finally, we have neglected the
contributions of a set of colleagues including Ajzen and Duck (see Duck,
1977, Chapters 3 and 19), who share an interest in the role of information
processing in attraction. At present, we see this line of theoretical development as less influentia1.
Theories of Friendship
11
Reinforcement Theories
The major theorists who have articulated the role of reinforcement in
interpersonal attraction are Byrne and Clore (Bryne & Clore, 1970; Clore &
Byrne, 1974) and Lott and Lott (1960,1974). The basic premise of reinforcemeht theories of attraction is that we like people who provide us with
rewards. Further, reinforcement theorists believe that we are attracted to
people who are merely associated with our receiving rewards, even if that
association is solely a product of chance circumstances.
Lott and Lott's Classical Conditioning Approach
Lott and Lott's (1960, 1974) formulation is based on principles of classical
conditioning, especially the Hull-Spence tradition. The basic proposition
underlying Lott and Lott's work is that liking for a person will result under
those conditions in which an individual experiences reward in the presence of
that person, regardless of the relationship between the other person and the
rewarding event or state of affairs (Lott & Lott, 1974).
Four main propositions underlie this view of attraction (Lott & Lott,
1960). First, persons are discriminable stimuli to which responses may be
learned. Second, a person who experiences a reward will react, presumably
pleasurably, to it. This leads to the third proposition, namely, that the
responses to reward (i.e., the sense of pleasure) will become conditioned to
all discriminable stimuli present at the time of reinforcement. Finally, a
person who was a discriminable stimulus at the time of reward achieves the
capacity to evoke the conditioned response. Thus, any person who is present
when we are rewarded comes to be liked.
The Lotts offer several other postulates to explain the strength of our
liking for other people. These postulates are based on the concepts of
frequency of reward, delay of reward, vicarious reinforcement, and drive or
incentive values. In slightly less technical terms, some of their predictions
can be expressed as follows. The more often we are rewarded in the presence
of another person, the more we come to like him or her. We are more
attracted to people who immediately reinforce our behaviors than to people
who pause before reinforcing our actions. The more we value a reward, the
more we like those people present when we receive it.
Byrne and Clore's Reinforcement-Affect Model
The basic premises of Byrne and Clore's model are very similar to those of
Lott and Lott. Byrne and Clore's emphasis is on the positive affect that
accompanies reinforcement, which is why this formulation is called a
reinforcement-affect model. From Byrne and Clore's perspective, reinforcement can be conceptualized as an unconditioned stimulus leading to
an unconditioned response, positive feelings. Any person associated with
the unconditioned stimulus comes to elicit the unconditioned response.
12
Thus, again through the process of classical conditioning, we come to like

those people who are present when we receive awards.
Besides merely being present, others can themselves reward or punish us.
Through the same conditioning mechanism, we come to like those who
reward us and dislike those who punish us. Of course, other people are not
always consistent in how they act toward us. To deal with this problem,
Clore and Byrne (1974) stated that: "Attraction toward a person is a positive
linear function of the sum of the weighted positive reinforcement associated
with him, divided by the total number of weighted positive and negative
reinforcements associated with him" (p. 152).
Representative Research
To test their hypothesis that we are attracted to people who are associated
with our experience of reward, Lott and Lott (1961) formed three-member
groups of children. Some were rewarded while playing a game, whereas
others were not. Later, each child was asked to select the two classmates he
or she would most prefer to take along on a family vacation. As expected,
they found that children who had been rewarded chose members of their
three-person group significantly more often than did the children who had
not been rewarded. The authors concluded that the reward had become
conditioned to the other group members, resulting in increased attraction.
Byrne (1971) reasoned that perceived similarity and dissimilarity of
attitudes in a dyad is one avenue of meting out rewards and punishments
because presumably it is reinforcing to interact with someone who shares
one's attitudes and punishing to interact with someone who does not.
Specifically, he predicted that the greater the proportion of a stranger's
attitudes that were allegedly similar to those of an experimental subject, the
more the subject would be attracted to the stranger. In a typical experiment,
Byrne would present subjects with a fictitious attitude questionnaire in
which the other's responses deviated to a preprogrammed degree from the
subjects' self-reported attitudes. Byrne repeatedly found that the proportion
of similar attitudes held by the hypothetical other exhibited a direct linear
relationship to the amount of attraction the subject reported feeling toward
the stranger.
Exchange and Equity

Reinforcement theorists are primarily concerned with the rewards we
receive from others. Equity and exchange theorists have a similar interest in
the benefits derived from our social bonds, but they go beyond mere
rewards. Equity and exchange theorists are also concerned with what we
invest in relationships and how our outcomes compare with our past
outcomes, our partner's outcomes, or the other friendships available to us. A
13
major exchange theory is Thibaut and Kelley's interdependence theory

(Kelley & Thibaut, 1978; Thibaut & Kelley, 1959).1 Another important theory
is equity theory (Berscheid & Wa1ster, 1978; Hatfield & Traupmann, 1981;
Hatfield, Utne, & Traupmann, 1979; Walster, Walster, & Berscheid, 1978).
Both of these theories, as well as Rusbult's (1980a, 1980b) extension of Kelley
and Thibaut's ideas, are briefly considered here.
Kelley and Thibaut's Interdependence Theory
Thibaut and Kelley believe that the essence of any relationship is
interaction. When people interact, their behaviors have consequences for
each other. Person A's actions influence Person B's behaviors and outcomes.
A basic premise of exchange theories is that for a relationship to have a
satisfactory outcome, the rewards must outweigh the costs incurred by each
participant. In considering how a participant assesses relationships,
Thibaut and Kelley discuss the individual's comparison level for alternatives. The comparison level is the standard used in evaluating the
absolute attractiveness of a relationship. It reflects the quality of outcomes
persons feel they deserve. The outcomes associated with the friendship can
be above or below this level.
The comparison level for alternatives is the lowest level of outcomes a
participant will accept in light of available alternative relationships. This is
the standard by which a person determines whether or not to remain in the
relationship. Thibaut and Kelley use the term dependency to refer to the
likelihood ofleaving a relationship. The more likely a person is to leave, the
less dependent the person is on that relationship. Dependency is a function
of outcomes minus the comparison level for alternatives: If our outcomes
are lower than our comparison level for alternatives, we will exit.
Having high expectations for a given friendship is often associated with
having a number of other rewarding relationships available, but this is not
always the case. Sometimes people have high expectations for a friendship,
yet they do not have gratifying alternatives. In these circumstances, people
may remain in a relationship even though they do not derive much
satisfaction from it. In essence, they have no better options available to
them.
Rusbult's Investment Model
A recent extension of Thibaut and Kelley's work is Rusbult's (1980a, 1980b)
investment model. Rusbult treats commitment and satisfaction as separate
lIn classifying Thibaut and Kelley's formulation as an exchange theory, we are
following common practice. Thibaut and Kelley, however, resist this classification.
They point out that interdependence involves a variety of other forms of coordination beyond simple exchanges.
14
concepts. In the interdependence tradition, Rusbult asserts that satisfaction

increases as the rewards associated with the relationship become greater, as
the costs decline, and as the person's expectations (comparison level)
become less. Outcome values (both rewards and costs) also influence
commitment, but here one's alternatives are more important than one's
expectations. Like Thibaut and Kelley, Rusbult predicts that individuals
become more committed to a relationship if they perceive poor alternatives
to their current relationship.
Rusbult then goes beyond exchange theory to suggest that commitment is
also influenced by investments, i.e., the more heavily invested in a
relationship, the more the individual is committed to it. Investments are
defined as resources such as time, money, and emotional energy. They are
distinguished from ordinary rewards and costs in that investments cannot
be extricated from the relationship as easily and will be lost, or at least
decline in value drastically, upon dissolution of the relationship. Thus,
investments increase commitment and serve to secure a relationship by
increasing the cost of dissolution: To abandon the relationship is to lose the
resources invested in it.
Hatfield's Equity Theory
Like interdependence theory, a basic assumption of equity theory is that
individuals try to maximize their outcomes (Walster, Walster, & Berscheid,
1978). Both interdependence and equity theory indicate that comparison
processes are important in evaluating outcomes. They differ, however, in the
nature of the comparisons that they consider most essential. In interdependence theory, Person A compares the outcomes of a given relationship
with the outcomes available from alternative relations. In equity theory,
people are seen as interested in comparing their own outcomes with those of
their partners.
Equity theorists believe egocentric outcome maximization would cause
suffering and shortages for others. Therefore, they suggest that society plays
a regulating function, creating rules so that rewards are distributed justly
among participants. An equitable relationship is one in which individuals
perceive their input (I)/outcome (0) ratio to be comparable to that of their
partner. Letting the subscripts refer to Person A and B, respectively, equity
exists when
=
Here, IIJ and IIBI designate the absolute value-irrespective of sign-of A
and B's inputs.
When relationships are inequitable, we become distressed and try to
restore equity. We can do this in various ways. Most obviously, we can alter
15
our inputs or try to change our outcomes. We can also use psychological
mechanisms to convince ourselves that equity exists even though the
objective outcome ratios remain the same. Finally, of considerable importance in the context of this chapter, we can leave the relationship.
Inequity arises regardless of the direction of the imbalance in outcome
ratios. Common sense would predict that an underbenefitted individual,
one whose outcomes are insufficient, should be distressed, but equity
theorists also maintain that overbenefitted individuals, those whose outcomes are disproportionately positive, also experience distress.
Much of the research generated by Thibaut and Kelley's ideas has involved
outcome matrices in game-like situations, the most common being the
Prisoner's Dilemma Game. Despite the importance of this work, in the
context of the present chapter, Rusbult's research serves better for illustrative purposes.
In a test of the investment model, Rusbult (1980b) asked subjects to think
of a friendship in which they had been involved. They were asked to
indicate how rewarding the friendship was, what costs were incurred, and
what alternative friendships they perceived as available to them. Respondents also estimated their degree of investment in the friendship, their
satisfaction with it, and how committed they felt to the relationship. Rusbult
found that, as predicted, satisfaction with the friendship was a simple
function of the rewards and costs associated with it, whereas commitment to
the friendship was determined by a combination of satisfaction, perceived
alternatives, and investments. In other words, people are happy with their
friendships as long as the rewards exceed the costs. Further, people will
remain committed to their friends to the extent that they are satisfied, a
more attractive alternative is not available, and they have invested heavily in
the relationship.
A study by Hatfield and her associates (Walster, Walster, & Traupmann,
1978) typifies equity research. Dating couples were asked to assess their
inputs, outputs, and outcomes as well as those of their partners by
responding to questions such as, "Considering what you put into your
dating relationship, compared to what you get out of it ... and what your
partner puts in compared to what s/he gets out of it, how well does your
dating relationship stack up?" Based on their responses, the participants
were categorized as either overbenefitted, underbenefitted, or equitably
treated. Walster et at. predicted that equitable relationships would be more
stable over time and would be characterized by greater sexual involvement.
Also, they expected that participants would feel happier and more contented
and less angry or guilty in equitable relationships.
These predictions were supported. Results for participants' feelings are
shown in Figure 2-1. One noteworthy feature of these data is that
16
Satisfaction
o = Content
o=Happy
Dissatisfaction
.=Anger
A = Guilt
3.75
3.50
Z
3.25
i=
3.00
o
o
~ 2.75
en
i=
<
en
c..
l:
en
Z
2.50
2.25
2.00
Q 1.75
l-
S
W
a:
1.50
1.25
/~A
.......
.~/
...-------..------...------.-~::::::::.f:.-..::::-::::-:::. ............
1.00
0.75
-.----L-~!:----'---~!
1L....--.::--J-!
Greatly
Underbenefitted
Equitable
Greatly
Overbenefitted
DEGREE OF EQUITY
Figure 2-1. The effects of equity on satisfaction and dissatisfaction among dating
couples. Higher numbers indicate more contentment, happiness, anger, and guilt.
(Adapted from Hatfield et aI., 1979, p. 117 e Academic Press. Reprinted by
permission.)
participants who felt slightly overbenefitted enjoyed high satisfaction in

their relationships; however, substantially overbenefitted individuals, as
predicted by equity theorists, were not satisfied with their courtship.
Cognitive Consistency Theories

Cognitive consistency theories are based on the assumption that we possess
a basic need for balance or consistency in our lives. Balance theorists most
often conceptualize cognitive balance or imbalance in terms of attitudes
(either positive or negative) between persons or objects in a triadic relation.
Although the theorists differ slightly on their views of what constitutes a
balanced system, each would agree that human beings are motivated to
maintain balance or consistency, because balanced relations are said to be
17
stable relations. An unbalanced triad, however, is hypothesized to be

unstable. A person attending to the imbalance will presumably be mobilized toward implementing change in a balancing direction. It is the
perception of imbalance that is crucial in determining whether or not an
individual will be motivated to restore or achieve consistency. Two classic
theories are considered here: Heider's and Newcomb's. The latter is
discussed in slightly more detail, as it pertains directly to interpersonal
attraction.
Heider's Theory of Cognitive Organization
Heider's (1958) theory deals with the relationships between the perceiver (P),
another person (0), and an object (X). This system is often represented in a
triangular diagram with P and at the top and X in the middle below. Lines
are drawn between P, 0, and X to represent the relationships among them.
Positive relations are marked with a plus (+) sign, negative relationships
with a minus (-) sign. A balanced state is said to exist if all possible
relations between P, 0, and X are positive, or if two of the relations are
negative and one is positive. This is equivalent to saying that the triad is
balanced if the algebraic product of the three signs is positive, but
imbalanced if the product is negative. For example, Heider postulated that if
P likes 0, and
likes X, P should have a positive attitude toward X
Similarly, if P dislikes 0, and
dislikes X, P should like X in order to
achieve a balanced state.
Heider's theory assumes that a person's attitudes to other people and his
or her attitudes to objects have similar properties. He also suggested that we
"fill in" incomplete triads to make balanced structures. Consider the case
where P likes 0, and
likes X, but P has no attitude toward X In this
situation, P would be apt to develop a positive attitude toward X so as to
achieve symmetry.
Newcomb's Balance Theory

Newcomb (1961, 1971), unlike Heider, assumed that a person's attitudes to
people and to objects have unique properties. In Newcomb's view, Person
A's attitude to Person B is affected by his view of B's attitude to A, an issue
that is not relevant when discussing A's attitude toward an object (X). In
fact, Newcomb preferred to call A's relationship to B, "attraction," and A's
relationship to X an "attitude." Thus, in Newcomb's model, the components
of the triad are uniquely defined. Furthermore, Newcomb was concerned
not only with the sign, but also the intensity of attitudes (and attraction).
Newcomb dealt mainly with the attitudes two people have toward
physical and social objects in the world. He suggested that if two people like
each other and perceive each other to have similar attitudes towards an
object, a balanced state will exist. If two people like each other, but perceive
themselves as having dissimilar attitudes, imbalance exists. In this case, if
18
the individuals wish to continue their relationship, Newcomb hypothesized

that one or both will restore balance by either changing his or her feelings
toward the object or toward the other person.
Heider and Newcomb differed in how they treated different types of
minus signs. For Heider, minus signs are equivalent, regardless of where
they occur in the system. He believes that interpersonal disliking and
negative attitudes have comparable effects. If disliking is accompanied by
attitudinal disagreement, the system is in equilibrium.
Newcomb treated interpersonal disliking differently than negative attitudes. He postulated that if two persons dislike each other, yet have
dissimilar attitudes, the situation constitutes a special case. Such an A-B-X
system is neither balanced nor imbalanced; it is an intermediate state that
Newcomb called "nonbalanced." In his view, if just two negative relations
exist, they must be confined to Person A's and Person B's attitude toward X
in order for balance to exist. Consistent with this view, Newcomb and his
associates (Price, Harburg, & Newcomb, 1966) obtained results showing the
consequences of nonbalanced systems are different than those of imbalanced systems.
Newcomb discussed the degree of imbalance (or strain), the importance
of communication, and various ways of reducing imbalance. For example,
the greater the discrepancy in A and B's attitudes as well as the greater the
importance of X (the object), the more strain toward symmetry the person
will experience. The greater the asymmetry in the system, the greater the
probability that A will communicate to B regarding X. Imbalance can be
alleviated, according to Newcomb (1959), in any of seven ways:
(a) by reduction in the strength of attraction; (b) by reduction of objectrelevance; (c) by reduction of perceived ("other's") object-relevance; (d) by
reduction of importance of the object of communication; (e) by reduction of
perceived ("other's") importance of the object of communication; (f) by
changes in cathexis or in cognitive structuring of own attitudes, such that
there is increased similarity with the other's perceived attitudes; and (g) by
changes in perceived attitudes (cathectic or cognitive) of the other, such that
there is increased similarity with own attitudes. (p. 403)
In his classic study, Newcomb (1961) recruited students, who were initially
strangers to one another, to live in a shared house. At the beginning of the
study, the students filled out an extensive battery of attitude questionnaires,
personality inventories, etc. At various time periods, the students were asked
to estimate the attitudes of others living in the house, to indicate the other
students' friendship choices, and to indicate their own friendship preferences. Consistent with balance theory, Newcomb found that the students
were attracted to those who shared similar values and liked the same people
they did.
19
A noteworthy feature of Newcomb's study was that even at the beginning

of the term, before students knew each other well, balance processes were
clearly manifest in friendship preferences. Over the term, the students' own
attitudes and values remained fairly constant, but both friendship choices
and perceptions of housemates' attitudes changed substantially. Despite
these changes, balance was still evident at the end of the term. Thus,
symmetry in relationships is initially based primarily on autistic, often
incorrect perceptions of other people's views. As time goes on, perceptions of
other people's attitudes become more accurate. As this happens, friendship
preferences change in accordance with strains toward symmetry. In essence,
balance persists, but the concordance between our intrapsychic model and
reality becomes greater.
Developmental Theories
What happens to interpersonal attraction as relationships develop over
time? The three theories discussed so far have not been especially concerned
with this issue. Presumably, reinforcement, equity, and balance all operate
in the early, the middle, and the late stages of a relationship. Vicissitudes in
liking can always be explained in terms of the vicissitudes of its everconstant key antecedent (reinforcement, equity, or balance).
A reinforcement theorist might wonder whether we satiate on our
partner's rewards. Newcomb, as previously mentioned, realized the underlying basis of balance may change over time. These theorists, however,
expressed little interest in the stages of relationships. Exchange theories
have shown at least moderate interest in relationship development.
As we discuss subsequently, experts on equity have debated whether
equity principles operate the same way in casual and intimate relationships.
Hatfield has also been concerned with the role of equity in the development
of intimacy. In essence, she has claimed: equitable dating relationships are
more likely than inequitable ones to progress to higher levels of intimacy;
equitable relationships are more apt to persist; and following transitions,
couples will either work to reestablish equity or they will move toward
breaking up (Hatfield et ai., 1985).
Finally, Thibaut and Kelley discuss the formation of relationships in
terms of sampling outcomes so as to make inferences concerning future
trends. They describe the life of a dyad as a progression through different
types of relationships from initial acquaintance to termination. They
suggest that a relationship has moved to a new stage or level when a marked
change in any relationship property is evidenced. The kinds of properties
involved include the patterns, strength, frequency, and diversity of interconnections; duration of interactions; and duration of the relationship.
Changes in one or more of these properties must be exhibited for a
20
considerable amount of time in order to conclude that the relationship has

moved to a new stage. From our perspective, another essential question is:
Do the determinants of liking change over the course of the friendship?
Several stage theories (Kerchoff & Davis, 1962; Lewis, 1972; Murstein,
1971, Chapter 6) have been proposed. Four such models are shown in Table
2-1. Most of these models, however, deal primarily with courtship development. The most influential developmental analysis of friendship has
been advanced by Levinger. Levinger and Snoek (1972) formulated a
developmental theory that begins with zero contact and ends with mutuality. We briefly present this model, followed by a discussion of Levinger's
(1980, 1983) most recent formulation: a five-phase model that extends from
initial acquaintance to relationship dissolution.
Altman and Taylor's (1973) theory of social penetration is included in
some discussions of friendship development. According to their conceptualization, as relationships develop, the breadth and depth of intimacy
expands. This growth of relationships is driven by exchange processes: We
evaluate our outcomes and forecast our future benefits. Relationship growth
is initially rapid; but at later stages, the penetration process slows down.
Altman and Taylor identified four stages in relationship development:
orientation, exploratory affective exchanges, affective exchange, and stable
exchange. In other words, interaction begins with superficial, public aspects
of the participants' personalities and progresses until interaction is relatively
free and touches on the most central areas of personality. In stable
relationships the partners continue their rich, spontaneous interaction;
however, because the members of the dyad know each other well, they can
readily interpret and predict their partners' feelings and probable behaviors.
Altman and Taylor conceptualized the dissolution of relationships as a
reversal of the penetration process.
This theory has been well-received and influential. Nonetheless, it has a
slightly different and more specific focus than the other theories covered in
this chapter. Instead of being a theory of interpersonal attraction, it is a
theory of self-disclosure and nonverbal behavior. Naturally, these are key
ingredients in the development of friendships.
Levinger and Snoek's Levels of Pair Relatedness
Levinger and Snoek (1972) discussed four levels (0, 1, 2, and 3) of pair
relatedness: zero contact, unilateral awareness, surface contact, and mutuality. The three higher levels of relatedness can be distinguished in terms of
seven attributes, including the knowledge the partners have of each other,
the communication between them, the importance of maintaining the
relationship, and the bases of the partner's attraction to one another.
Levinger and Snoek did not dwell on the attributes of the lowest level of
relatedness, zero contact. In that level, the two parties are unaware of each
21
Table 2-1. Stage Models of Relationship Development
Stage in the
Relationship
Kerchoff &
Davis' Filter
Theory of
Mate Selection
Murstein's
Stimulus-ValueLewis'Six
Role (SVR)
Processes in
Theory of
Premarital
Dyadic Formation Mate Selection
Early
Similarity in
values
Need complementarity
Similarities
Attraction based
on external
stimulus attributes (e.g.,
physical
attractiveness)
Value similarity
Rapport
Mutual selfdisclosure
Emphatic understanding of the
other person
Interpersonal rolefit (e.g., need
complementarity)
Successful performance of
roles in the
relationship
(e.g., wife,
husband)
Late
Levinger &
Snoek's Levels
of Pair
Relatedness'
Unilateral awareness-attraction
based on a's
image and
reward potential
Surface contactattraction based
on satisfaction
with experienced
outcomes, determined considerably by the adequacy of a's
role enactment
Mutualitypersonal disclosure may
lead to feelings
of "we-ness"
Dyadic crystallization (e.g.,

commitment,
identity as a
couple)
With Levels of Pair Relatedness, Unilateral Awareness occurs prior to a relationship being established.
Mutuality can extend over a long time span, from the early to the late stages of a relationship.
Note. Adapted from Brehm, S. S., Intimate Relationships (p. 145). iCl1985 by Random House, Inc. Reprinted
by permission of Random House, Inc.
other, so they have not yet developed communication patterns and the
like.
At Levell, unilateral awareness, the person is aware of the other, but a
relationship has still not developed. The two parties have not yet developed
interaction patterns, they have not yet begun to develop a store of shared
22
knowledge, and they are not concerned about maintaining their bonds.
Further progression of their relatedness depends on how the partners
evaluate each other, as well as other factors (Le., their need for affiliation, the
ease of their getting together, and how satisfied the parties each are in their
current relations). At this stage, appraisals are made largely on the basis of
external factors, such as physical appearance and unilateral estimates of
other people's potential for rewarding us. If the evaluation is positive, the
two individuals may go on to develop a relationship.
Surface contact is characterized by a superficial exchange of information
between partners. The costs of terminating the relationship are low, so
maintenance is still relatively unimportant. Knowledge of the other person
is based on that individual's public self-presentation. Interaction is governed by the dictates of socially defined roles. There is, however, sufficient
interaction so that people can begin assessing the outcomes of the relationship. During the surface contact level, Levinger and Snoek believe that the
principles of attraction advanced by Thibaut and Kelley seem "especially
relevant" (p. 6). At this point, our impressions of other people are beginning
to coagulate. If these, as well as our behavioral outcomes are positive, we are
apt to pursue the relationship further.
The third level, mutuality, depicts the gradual intersection of the lives of
the partners. Each partner possesses a significant amount of information
about the other. Moreover, this level is characterized by increasing
interdependence in Thibaut and Kelley's sense; that is, each partner
assumes responsibility for fostering satisfactory outcomes for the other and
for the couple as a unit. Interaction is spontaneous and free flowing; the
partners have constructed their own patterns. Cultural norms are less
important as a factor in regulating interaction. Both parties have a stake in
maintaining and enhancing the relationship. Interpersonal discovery via
self-disclosure promotes feelings of affection for the other as a unique
person. In essence, Levinger (1974) believes that Level 3 is a continuum: "its
baseline is a Level 2 surface contact; its ultimate realization is the total
interpenetration of two human beings, as defined by their joint attitudes,
joint behavior, and joint property" (p. 101).
Levinger's ABCDE Model

More recently, Levinger (1980) has formulated an ABCDE sequence of
relationship development. A stands for acquaintance; B, for buildup; C,
for continuation; D, for deterioration; and E, for ending. Whereas Levinger
and Snoek's model primarily dealt with attraction in the formation phases
of friendships, Levinger's more recent formulation focuses on close
relationships. Levinger's new model places greater emphasis on the
transition from one stage to the next. He advocates studying not only the
stages per se, but also the transitions between the stages.
23
Given the attention we have previously devoted to the formation of

relationships, for us the most interesting aspect of the ABCDE model is
Levinger's attention to the breakdown of relationships. With regard to
deterioration, Levinger (1980) states that a drop in involvement may result
from negative changes in one or more of the following: the diversity,
frequency, or duration of interaction; the type (positive vs. negative) and
intensity of affect; the openness of communication; the synchrony of
behavioral accomodation; trust; and the communality of the partners' plans
and outcomes. Levinger believes that internal pair processes, personal
characteristics of the partners, and external forces can all lead to deterioration of relationships.
Deteriorated relationships can continue indefinitely. Levinger maintains
that the decision to end a close relationship depends heavily on one's
attraction to an alternative relationship. In his earlier writing (Levinger,
1979), he offered an analysis of marital dissolution in terms of the forces
driving a person toward and away from their partner. The relevant field of
forces would, of course, include the attractiveness and disadvantages of
alternative relations. Levinger also noted that barriers develop around a
relationship. For example, many individuals think of marriage as a life-time
commitment; others cannot afford getting divorced even if they wanted to.
Such barriers can act to keep people together.
Levinger (1980) described an unpublished study he conducted with
DeLamarter in which they assessed pair relatedness using a behavioral
measure. Specifically, they predicted that if asked to paint a picture on a
large sheet of paper, pairs of strangers would paint two separate pictures,
each on his or her own side of the paper. On the other hand, they expected
that established couples would paint a joint picture, freely crossing over into
each other's territory. Although they did not obtain unequivocal support for
their hypothesis, DeLamarter and Levinger found that in doing their
pictures attached pairs exhibited many more behaviors such as touching,
trading places at the table, and borrowing paint without asking. Levinger
argued that each of these behaviors can be regarded as an index of mutual
interdependence.
Other studies in this area (see Levinger, 1980) consist of presenting
subjects with a list of behaviors varying in degree of intimacy and asking
them to judge how appropriate each behavior is for different relationship
types varying in closeness. Not surprisingly, as the relationships manifest
greater closeness, the expectation that the pair would engage in intimate
activities increased. Behaviors involving physical contact and self-disclosure were most affected by increases in closeness, whereas less intimate
behaviors, such as going for a walk or doing an errand for the person, were
least affected by the degree of intimacy.
24
A Comparison of Theories of Attraction

To compare theories, we first consider three prototypical relationships to see
how the theories differ in their predictions and explanations of these cases.
Then we provide a comparative summary of how the theories differ along
eight dimensions. These comparisons are largely descriptive; evaluative
comparisons are contained in the next main section of the chapter.
Case 1: Distressed and Chronically Conflicted Relationships
All of us can think of relationships in which the partners do nothing but
argue. Spousal abuse has become a major focus of research. Although one
might expect habitually conflicted, abusive relationships to break-up, they
do not always do so. How can different theories explain the persistence of
such apparently unpleasant relationships?
Neither the Lotts nor Byrne have concentrated their attention on such
phenomena. intuitively, it seems directly contradictory to a reinforcement
analysis to like someone who punishes us. Yet, by invoking a broader set of
reinforcement concepts than those of classical conditioning per se, one can
readily explain the continuation of distressed relationships. A reinforcement theorist might begin by claiming that people who stay in distressed
relationships are masochists who get their rewards in strange ways.
Dutton (1984) explained the bonding of women to abusive husbands as
due in part to intermittent positive-negative treatment (p. 292). After fights,
especially in the early stages of marriage, these couples make up. The
husbands may shower attention and affection on their partners. This
intermittent positive reinforcement serves to enhance liking and the
acceptance of the antecedent behavior, namely, the violence. Also, when the
fighting stops, this removes aversive arousal. The cessation of aversive
stimuli constitutes a second form of reinforcement (negative reinforcement),
which comes to have value for the abused person. Ergo, the wife remains.
Equity and interdependence theories provide concepts with more obvious
relevance for explaining distressed relations. According to equity theory, if
the partners are getting equally little, equity can nonetheless exist. To ensure
that one's partner is getting next to nothing, one can invest very little into the
relationship or try to create losses for the other person. Scanzoni (1979) has
considered abuse from an interdependence perspective. Thibaut and Kelley
(as well as Levinger) might ask: What are the person's alternatives? If they
are not very good, one might stay in the present relationship, lowering one's
comparison level. Levinger would also look for barriers around the
relationship serving to sustain the couple's bonds.
It is less obvious how a cognitive consistency theorist would explain
argumentative dyads. Newcomb might wonder whether the pair argue about
unimportant topics but agree on the more important (if less frequently
discussed) objects of communication. In analyzing P-O-X relationships,
25
Heider (1958) typically presupposes P likes himself. If this assumption is

incorrect, Heider might explain unpleasant relations as consistent with selfhatred (or low self-esteem). Both P and 0 have a negative attitude toward a
common object (P), and this forms the basis for liking.
Case 2: The Consequences of Being Spoiled
Let us consider a second class of relationships by introducing Pat and Chris.
They are single adults living in the same apartment complex. Pat repeatedly
does nice things for Chris. For example, she makes great bread and often
gives him hot, fresh loaves just out of the oven. Pat also is willing to take care
of his cat. Chris, on the other hand, does not do much in return. Pat is lucky
to even get a perfunctory "thank you."
We will ignore the issue of why Pat continues being so nice. Instead, we
address the question: How well does Chris, the beneficiary of all these
favors, like Pat? Reinforcement and exchange theories both have relevance
here, although they seemingly make different predictions. Reinforcement
theorists, who concentrate on the amount of reinforcement a person
receives, would expect Chris to like Pat very much. Pat has constantly been
providing Chris with what can be considered as reinforcers. Interdependence theorists (Thibaut and Kelley) as well as Levinger would
probably agree with this prediction. For Chris, the outcomes of the
relationship exceed the costs, so satisfaction should be high.
Equity theorists, on the other hand, would have a different expectation.
They are concerned with fairness. In their terms, Chris is overbenefitted.
This should lead Chris to feel guilty and try to leave the relationship. The
research cited previously is supportive of the equity prediction on this
matter.
Case 3: The Impact of Third Parties on Friendships
The third paradigmatic situation concerns the role of third parties. Again,
different theorists deal with third parties in varying ways. Classical
conditioning theorists say that we like whoever is present when we receive
rewards. The relationship between the onlookers is, however, of little or no
concern to these theorists. Interdependence and developmental theorists say
that third parties are important in that they provide a basis for evaluating
the relative attractiveness of a given relationship. Research conducted
within a developmental framework has also shown how our friendship
networks constrict as courtships become more intimate (Milardo, Johnson,
& Huston, 1983). Equity theorists focus on our rewards in comparison to our
partner's without extensive analysis of the role of third parties.
Several previous investigators (see Griffitt & Guay, 1969) have studied
interpersonal attraction in what was described to subjects as a creativity
experiment. The basic procedure consists of having subjects make up stories
in response to (Thematic Apperception Test) type cards. A confederate of
26
the experimenter listens to the stories, rating them as either highly creative
or very mundane. These ratings have been conceptualized as a form of
reward. Consistent with reinforcement theory expectations, the more of their
stories that are rated positively, the better the storytellers like the rater.
An onlooker can be introduced into this situation. The relationship of the
onlooker to the rater can be left ambiguous or specified for the subject.
From a reinforcement perspective, the storyteller-onlooker bond is of little
importance. As long as the storyteller is reinforced, anyone present when the
reinforcement is administered (including the onlooker) Shbuld come to be
liked.
The situation can also be analyzed from a cognitive consistency
perspective. According to balance theory, this is an A-B-X situation, where
the onlooker is X From a balance theory perspective, however, the system of
relationships is crucial. In the situation where the rater evaluated the stories
positively and the rater liked the onlooker, the storyteller could achieve
balance by liking the onlooker. (Here, balance theory makes the same
prediction as reinforcement theory.)
A more crucial situation exists, however, when the rater positively
evaluates the subject's stories but dislikes the onlooker. As noted previously,
a reinforcement theorist would still predict that the subject would like the
onlooker. Balance theorists would disagree; they would say that to have a
balanced system, the subject should dislike the objects their friend, the
storyteller, dislikes.
Murray (1973) tested these predictions. Unfortunately, the results of his
study were not definitive. Nonetheless, they provided some support for the
balance theory position, especially among persons who are high in selfesteem.
A Comparative Summary of Five Theories
In thinking about these theories of friendship, we have identified eight
dimensions on which they can be compared. In Table 2-2, we described the
theories along these dimensions. Naturally, constructing this table involved
making jUdgments and simplifying details. In large measure, the table is a
summary; much of the information has already been presented or is selfexplanatory. The table organizes the information so that the similarities and
differences between the models become apparent.
Indeed, the construction of Table 2-2 is revealing. Various reinforcement
(and various cognitive consistency) formulations bear enough resemblance
to one another so that each tradition can be considered as a unit. The same
is not true for interdependence and equity as variations within the exchange
camp. They differ sufficiently to warrant separate treatment. Finally,
especially in Levels 2 and 3 of relationships, Levinger draws on Thibaut and
Kelley's ideas, so some overlap in their positions emerges.
27
The last four dimensions on which the theories are evaluated are
interrelated. Concern with relationship stages, third parties, and deterioration of dyadic bonds are all aspects of how many relationship phenomena
a theory addresses. With regard to our use of the breadth dimension, readers
should note two points. First, we are usually comparing specific theorists in
terms of the breadth of the relationship phenomena they address. This is not
the same as comparing the various types of theories on how many things in
general they can explain. Readers would undoubtedly agree that reinforcement theories can account for a broad range of phenomena, but we
have classified this as a narrow theory because Lott and Lott focused
primarily on the question of attraction. Second, we have compared the
theories. Later in the chapter we return to this issue, commenting on all five
theories collectively in a more absolute sense.
The five theories summarized in Table 2-2 vary somewhat in the extent to
which they emphasize past experiences rather than current influences as
determinants of liking. None of the five theories, however, places a heavy
emphasis on past experiences. Although psychodynamic theories are not
widely used in studying attraction, they place considerably more importance
on early childhood determinants of adult behavior. Attachment theorists
such as Bowlby claim that an individual's general orientation toward other
people is largely determined by parent-child relations in infancy (ages 0 to 2
years). If the individual's early relationships are unsatisfactory, the person
will then presumably have difficulty establishing peer relations in childhood and intimate ties in adulthood. Similarly, psychodynamic theorists
(professional and amateur alike) often explain mate selection in terms of the
delayed gratification of early childhood Oedipal (or Electra) wishes.
Finally, the dimension "concern with stages" deserves special mention.
Nearly a decade ago, Berscheid (1977) published an encyclopedia review of
theories of interpersonal attraction. At that time, she noted that "relationship development has been relatively ignored by general social psychological theories of interpersonal attraction" (p. 203). Since then, Levinger
and others have made significant strides toward filling the earlier gap.
In our opinion, these efforts have been very important. The developmental perspective raises new questions (e.g., concerning transitions, the life
course of relationships) and has served as a force in redirecting the
traditional concerns of interpersonal attraction scholars. We feel that this
impetus has been beneficial, but such an assessment is jumping into the
next section of our chapter, the evaluation of theories.
An Evaluation of Theories of Attraction

In evaluating the theories, we briefly critique the basic concepts of each
position. Related to this, we ask: How universal is the operation of
reinforcement, equity, and balance? What limitations in terms of situations
People can be
molded via
reinforcement,
Emphasis on current levels of
reinforcement,
histories & past
pairings of CS
with UCS
Largely external
(rewards), but in-
View of human
nature
Emphasis on
internal
Determinants of
attmction,
historical vs.
contempomry
Reward
Central concepts
Reinforcement
Equity
Both: depends on
rewards received
Both: depends on
rewards received
Behavioral
Fairness of input/
sequences,
outcome mtios
rewards, costs,
CL, CLAlt
People seek to
People are conmaximize
cerned with
outcomes
fairness
Emphasis on conContempomry
temporary forces;
some concern with
past investments
Interdependence
Five Approaches to Attraction: A Compamtive Analysis
Dimension
Table 2-2.
People are intellectual, seek

consistency
Contempomry, but
established cognitive structures
can influence
information
processing
Internal
Balance among
attitudes
Cognitive
Stages, positive,
negative forces,
barriers, behavioral
outcomes
Needs change;
emphasis on maximizing outcomes
Contemporary yet
implies change in
attmction is based
on progression
through earlier
stages
Emphasis first on
perceptual pro-
Developmental
tv
::L
::r
..,
'<
'""
::L
'""
<:
'""
1:1:1
0-
pO
S
::s
pO
::s
a
pO
::s
00
ternal factors can

influence what is
rewarding
Very little
Note. CL
Some
plus judgment of
their fairness
Broad (satisfaction, Moderate

love, commitment,
power, etc.)
Concern with
Focus mostly on
alternatives
dyad
available
Specific considerImplicitly the
ation
reverse of
attraction
Some
plus comparison
processes
Moderate
Concern with
alternatives
available
Specific consideration
Third persons
treated like
attitude object
Decline ignored;
disliking conside red a special
case
cesses and then on

behavioral
outcomes
Considerable
Narrow (liking &

attitudes)
Very little
= comparison level, Alt = alternative, CS = conditioned stimulus, UCS = unconditioned stimulus.
Concern with
stages
Breadth of relation- Narrow
ship phenomena
considered
Concern with conVery little
stellation of
relationships
Treatment of
Implicitly the
reverse of
decline: implicitly the reverse
attraction
of attraction or
object of specific
consideration
processes vs.
external forces
::J"
0
...,
tv
\0
>8'
'"
::l
0::J"
'"rI
::l.
'".....,
0
;:1.
30
and types of people are there in the operation of these forces? Intuitively, it
seems that the more exceptions there are to these phenomena, the less
essential that force is in promoting friendship.
We then turn to a general comparison of the theories. In the past,
Murstein (1971, Chapter 1) and Hinde (1979) have assessed several theories
and others (Rubin & Levinger, 1974; West & Wicklund, 1980, Chapter 3;
Wright, 1971) have evaluated specific models. Our critique is guided in part
by our assumption (see Duck & Perlman, 1985) that the focus of the field is
shifting from the specific question of attraction per se to a longer list of
questions concerned with additional aspects of relationships. Many criteria
(i.e., parsimony, logical consistency, testability, stimulation of research,
practical utility, and empirical support) exist for evaluating theories. We
briefly discuss the heuristic value and practical utility of each theory. Yet,
because of the directions in which personal relationships research is
headed, we focus particular attention on the breadth of the theories as a
criterion in their evaluation.
The Critique of Basic Concepts
Reinforcement. A variety of criticisms can be leveled at the reinforcement
positions. Reinforcement theorists are often criticized because definitions of

reward are circular (reward changes behavior and whatever changes
behavior is rewarding). It is also difficult, if not impossible, to ascertain what
constitutes a reward for a particular individual. If a researcher cannot tell
what is reinforcing, can data disconfirming the investigator's hypothesis
ever be obtained?
Byrne and Clore have focused on attitude similarity as a form of
reinforcement. Golightly and Byrne (1964) did a clever, preliminary study
to demonstrate that attitude similarity does operate as a reinforcer. But
Byrne and Clore can be chastised for attending primarily to a very narrow
form of reinforcement. In our everyday lives, many other forms of
reinforcement are equally or more important.
Murstein (1971) has challenged whether classical conditioning is really
occurring in the Byrne paradigm:
In classical conditioning, the unconditioned response is supposed to occur
rapidly and inevitably as a result of the appearance of the unconditioned
stimulus. Obviously, this does not occur in attraction because the unconditioned stimulus must first be searched for meaning. (p. 15)
Although Byrne and Clore's analysis is clearly rooted in a classical

conditioning framework, Byrne (1971) admitted that he made rather liberal
use of existing reinforcement formulations. He indicated that the driving
force behind much of his research was "the desire to establish an empirical
law." According to Byrne,"the (theoretical) language and concepts remained
at a more general level (i.e., Thorndike & Pavlov) and, hence, not necessarily
committed to the demands of any current theoretical framework" (pp. 268,
31
269). Thus, some critics have questioned whether Bryne's ideas are based on
operant conditioning, classical conditioning, or both. The latter appears to
be the case. This, however, causes conceptual ambiguity and can lead to the
generation of contradictory predictions.
Exchanges. Equity and interdependence theories have commonly been
criticized on two grounds. The first is the problem of measurement. Hinde

(1979) expressed this dilemma as follows:
If one is dealing with personal relationships where social approval,
understanding, or intangible services are at issue, how can one make
measurements adequate for scientific inquiry? The problem is exacerbated
by the facts that the value of a particular reward is different for different
individuals, and changes with time for anyone-not just as a consequence
of satiation, but through changing contextual factors. (p. 220)
In an economic exchange, $10 given is the same as $10 received; but in the
social sphere this is not necessarily so. Is making dinner for a friend today
equivalent to their treating you to dinner at a restaurant next Friday? Thus,
for both laypersons and researchers alike, making accurate assessments of
equity in complex social situations is inherently difficult.
We anticipate that Kelley et al.'s (1983) Close Relationships will become one
of the seminal publications of the 1980s. The UCLA group, however, places
increasing emphasis on analyzing the thoughts, feelings, and behaviors that
occur in chains of interpersonal interaction. We believe collecting data of
this sort is typically difficult, time consuming, and expensive. Thus, we
expect the very kind of data required to test Kelley's theory will cause
impediments for graduate students and unfunded researchers who wish to
confirm his ideas.
A second frequent criticism directed at exchange theories is that they
provide a cold, calculating view of human nature. It is not especially
flattering to think that we spend our lives checking to see whether our
current friends are as good as the alternatives. Nor is it flattering to think
that we are constantly checking to ensure that our outcome ratios are at least
comparable to our partner's. Are there no times in life when we gladly do
nice things for other people, with no strings attached? We return to this issue
later.
Balance. The cognitive consistency theorist's caricature of human nature is
that we are all thinking individuals trying to maintain logical (or at least
psychological) consistency. Critics have questioned whether our thoughts
are all that important. They have also argued that we often seek diversity,
stimulation, and inconsistency.
A more technical feature of Newcomb's theory is that he postulated
several ways of reducing imbalance. Although this adds richness to his
theory, it creates problems for researchers. If one's predictions about liking
do not work out, it may be because imbalance was resolved in some other
32
way. Thus, researchers are faced with either preventing or measuring

alternative forms of reducing strain. (A similar problem exists for equity
theorists.)
Developmental theories. The fundamental assumptions of a developmental
theory are that relationships change over time, perhaps in invariant regular
ways, and that the determinants of attraction are different in different stages.
Developmental theorists (see Table 2-1) generally agree that external
stimulus attributes, such as physical attrativeness, are important in early
stages of relationships. They also frequently talk of the formation of dyadic
identity ("we-ness," crystallization) in the more advanced stages of relationship development. Inbetween, however, there is not agreement on the stages,
the crucial antecedents of attraction, or the exact order in which predictor
variables are important.
Several other concerns can be raised about Levinger's model. First, his
original model of pair relatedness was a general theory of relationship
development, but from our perspective, it is of concern that Levinger's
ABCDE model (and the other models of relationship development) focus
on intimate heterosexual relationships. One wonders: Are these developmental processes the same for friendship pairs? Presumably, there are some
differences.
Second, Hinde (1979) claimed that the division of any continuous process
into stages is bound to be arbitrary. It implies the occurrence of sudden
changes that do not actually occur in relationships. Levinger (1983) was
sensitive to this problem. He noted that transitions do not necessarily occur
at a single point in time and acknowledged that processes begun in one
stage can carry over into later stages.
Levinger and Snoek have focused considerable attention on the early
stages of relationship development, but they lump a much broader set of
phenomena together in the mutuality stage. We wish that Levinger had
devoted more attention to this part of his analysis. Although we agree that
alternatives are important to the breakup of relationships, we nonetheless
believe that many individuals (especially older married persons) separate
with little expectation of finding a new partner.
Exceptions to the Rule
Exceptions seem to exist for every rule. Some exceptions are, of course, due
to chance circumstances or the overriding influence of additional factors,
but others are systematic and limit the application of basic principles.
Theorists can explain these exceptions with existing concepts, they can add
new concepts to provide an explanation, or they can simply note the
limitations of their general principles. If the exceptions go unnoticed or
unexplained, they cause embarrassment for the theorists.
33
Reinforcement. Several exceptions to the operation of reinforcement, equity,
and balance have been identified. We have already noted that contrary to a
reinforcement view, overbenefitted individuals experience guilt and discomfort rather than enhanced attraction. Aronson and Linder's (1965) work
provides another interesting exception to Clore and Byrne's (1974) proposition that "attraction toward a person is a positive linear function of the
sum of the weighted positive reinforcements associated with him" (p. 152).
Aronson and Linder (1965) demonstrated a so-called gain/loss effect.
Namely, a stranger who began with a negative evaluation of a subject and
then became repeatedly positive was better liked than a stranger who gave a
comparable number of all positive evaluations.
Equity. Several exceptions to the operation of equity principles exist. In

reviewing more than 50 studies, Hook and Cook (1979) concluded that
young children use an equality rather than an equity rule in distributing
rewards. Kahn, O'Leary, Krulewitz, and Lamm (1980) provided evidence
from two studies supporting their claim of sex differences in adherence to
distributive justice norms. Although men lean more toward the equity norm,
women endorse the equality norm more strongly. Walster, Walster, and
Bescheid (1978) identified four situational factors (time constraints, high
communication costs, small outcomes, and no implications for future
decisions) that all reduce people's willingness to apportion resources
equitably.
An even more crucial issue in terms of our interests is whether equity
considerations are applicable to all or just some relationships. In particular,
questions have been raised concerning whether equity applies to intimate
relationships. Walster has claimed that equity principles do apply, and
several studies have found support for the operation of equity in dating and
marital relationships. Walster, Walster, and Berscheid (1978) acknowledged,
however, that "it should be easier to calculate equity in casual relationships
than in intimate ones. Over a short span it is easy to assess who owes whom
what ... in intimate relationships it is far more difficult" (p. 147).
Mills and Clark (1982) distinguished between communal and exchange
relationships. Superficially, it may appear that communal relationships are
a form of selfless concern, where equity considerations are tossed aside. In
communal relationships, concern about the other person's welfare predominates. Benefits are given because they will be pleasing to the other
person or because they are needed. Mills and Clark conducted several
experiments to test predictions derived from their analysis of the communal
vs. exchange distinction. For example, in communal relationships they
found that if one's partner reciprocated a favor, attraction was decreased
,
(rather than enhanced).
A methodological factor (linked to conceptual points implied in the
writing of equity theorists) may help unravel some of the ambiguity
34
concerning the operation of equity in intimate relationships. Three ways of

studying equity have been used: an assessment of the resources (e.g., money,
physical attractiveness) that people bring to relationships, global assessments by respondents concerning the equity in the outcomes of their
relationships, and exchanges in specific situations. It is in studies using the
first two procedures that we believe equity principles are generally apparent.
In studies of specific situations, we would expect equity to be less apparent
for intimate partners than for casual acquaintances. In casual relationships
people exchange a limited range of resources. Tit-for-tat immediate
reciprocity is more common. Intimate and communal relationships are
long-term. Immediate, tit-for-tat reciprocity may be less important or,
according to Mills and Clark, even offensive. In intimate dyads the
members may not keep track of how or when specific benefits get repaid.
(Indeed, for parents, the mere prospect of future repayment in old age may
be enough.) Nonetheless, we suspect that experiencing an overall sense of
equity is essential to the members of most intimate relationships.
Balance. Two exceptions to balance theory have been noted. First, people
differ in their need for cognitive consistency. Newcomb (1961) provided

evidence that the tendencies toward symmetry are more noticeable among
authoritarians than among nonauthoritarians. Others have also identified
lower levels of cognitive complexity and Machiavellianism as related to
greater sensitivity to imbalance (see Murray, 1973).
A second exception to balance principles is the romantic triangle.
Suppose John and Henry both like Mary. Normally, according to balance
theory, liking the same person should lead John and Henry to like one
another, but if both John and Henry want to marry Mary, they will
undoubtedly feel jealously and rivalry toward one another. In this situation,
the negative feelings due to competition will be stronger and take
precedence over the positive feelings engendered by attitude similarity.
Newcomb (1971) used the notion of "joint relevance" to account for other
special cases. He wrote "if a given object or event has differential impacts
upon 0 and P, there is no reason why different attitudes toward it should
lead to the tensions of imbalance" (p. 40). Through this notion, Newcomb
can understand the high satisfaction between spouses with very different
attitudes toward doing housework as well as the affection between parents
who have predinner cocktails and their children who do not.
The Practical and Heuristic Value of the Theories

Two aforementioned criteria for evaluating theories are: do they stimulate
research, and ,do they contribute to the practice of clinical psychology? In
terms of guiding clinical practice or self-help activities, we feel that these
theories have had fairly limited value. Perhaps the most notable exception
to this generalization is that interdependence principles have influenced
35
some currently prominent family therapy approaches (see Jacobson &

Margolin, 1979).
As indicated in earlier portions of the chapter, all five theorists have done
research based on their theoretical position. Newcomb is best known for the
study he did of students sharing an Ann Arbor house. Lott and Lott did a
programmatic series of studies. Byrne has replicated and extended his law
of attraction in a number of studies. Exchange theorists have done a variety
of studies. Of the theorists, perhaps Levinger has done the least to provide a
large scale or programmatic test of his model. Levinger, himself, seems
aware of the need for more research. He made a series of statements
(Levinger, 1983) such as "the transition from encounter to relationship is not
well documented" and "There is little empirical literature on the process
whereby interpersonal relations become increasingly interdependent" (pp.
324-325).
During the 1960s and 1970s, a great deal of social psychological research
was done in laboratory settings with student volunteers as SUbjects. Most of
the major attraction theorists have ventured out of the laboratory in at least
some of their research. Levinger (1980), however, noted that reinforcement
theorists neglect the later stages of relationships. He stated that their
"research has concentrated on issues that are mere preliminaries to
affiliation or attachment. It has hardly informed us even about the
conditions under which initial attraction leads one to develop a relationship" (p. 524). This evaluation applies to most of the interpersonal attraction
research done in the 1960s and 1970s. We note, however, that several of the
leading attraction theorists have done at least some of their research using
partners in established relationships.
The Breadth of Relationship Phenomena Addressed
In evaluating theories on the breadth dimension, we note that the goal of
some theorists (especially Lott and Lott) was to apply a broader general
theory specifically to the problem of attraction. Thus, they have achieved
their goal. Nonetheless, we feel that past theoretical efforts, especially in the
reinforcement and cognitive traditions, have focused too narrowly on
attraction.
Kelley and his colleagues (1983) called for a new science of relationships.
We echo this call and have already indicated our view that the scope of the
field is expanding. Kelley et al. suggested that the study of relationships
should be "concerned with the interdependence between two people-with
describing the quantity and quality of that interdependence over time and
with identifying the causal factors that both affect and are affected by that
interdependence" (p. 12). Kelley et al. focused on several aspects of
relationships: interaction, emotion, power, roles, love and commitment,
conflict, and developmental changes. We feel that Kelley's group is to be
praised for the breadth of their analysis.
36
If a broader perspective is taken, what should a theory of friendship

address? First we would like to underscore a few of Kelley et al.'s concerns.
Interpersonal attraction scholars have traditionally focused on the beginnings and the positive aspects of relationships. We feel the emphasis of
Kelley, Levinger, and others on conflict, ambivalence and the deterioration
of relationships is well taken.
We would like to add five other topics to the list of concerns to be
addressed by comprehensive theories of friendship. First, friendship
theories should examine the question of the links between attitudes and
behavior. West and Wicklund (1980) criticized Byrne and Clore for failing to
"spell out precisely how attraction and behavior are related" (p. 34).
Interdependence theories deserve credit for distinguishing between satisfaction and commitment. In general, theories of friendship should address
the question: When and for whom does liking lead to relating?
Second, existing theories of attraction presume too readily that the goal of
relationships is increased liking that eventually results in intimacy. Yet
many, perhaps even most, relationships exist in relative stability without
being intimate and without much development. As Duck and Perlman
(1985) noted, we need a clearer understanding of the circumstances in which
relationships fail to develop. We need to know more about the circumstances in which low levels of intimacy are satisfactory to both
partners.
Furthermore, theories of attraction have focused on liking as a dependent
variable. As the work on the buffering effects of social support testifies, it is
important for future theories of friendship to devote more attention to the
consequences of attraction. We imitate our friends' behavior, they influence
our opinions, and involvement in relationships contributes to our sense of
well-being. Feelings of friendship influence self-disclosure patterns, our
willingness to help others, and the way we negotiate. A comprehensive
understanding of relationships will not be achieved until both the antecedents and the consequences of friendship are integrated together in one
overall framework.
In her earlier reflections on theories of attraction, Berscheid (1977) noted
that investigators are often interested in certain types of relationships. She
claimed that these investigators have found existing perspectives too loosely
fitting to be of predictive value. Since the publication of Berscheid's article,
it is fair to say more attention has been focused on types of relationships.
Exchange theorists have concentrated on the intimacy dimension as an
important factor, and researchers such as Rusbult (1980a, 1980b) have tested
exchange notions using different kinds of dyads (e.g., couples, friends). Yet,
overall, we believe the personal relationships field would benefit from
greater attention to typologies and to how the basic personal relationships
principles are modified in different types of relationships.
Finally, we are concerned over the relative neglect of life cycle issues. The
assumption of reinforcement and balance theorists appears to be that their
37
basic principles work from infancy to old age. Even so, one wonders
whether children's lower levels of cognitive complexity might make them
especially sensitive to imbalance. Traupmann and Hatfield (1983) have
demonstrated changes in marital equity over the family life cycle. In our
view, more analyses of age differences are warranted.
Conclusion
In this chapter we have presented, compared, and evaluated several theories
of interpersonal attraction. As Berscheid (1977) and others have noted, these
theories generally assume, in one way or another, that liking is based on
reinforcement. The reinforcement theorists say that liking results from
being reinforced by another person or merely receiving a reward in the other
person's presence. According to interdependence theory, we are satisfied in
relationships where our outcomes are good. Equity theorists feel that we
seek, relative to our partner, a fair outcome. Cognitive theorists claim that
we prefer consistency, but even this can be conceptualized as reinforcing. As
a key developmental theorist, Levinger draws heavily on interdependence
theory for explaining the buildup stage of relationship growth.
We have maintained that a subtle shift is occurring in the study of dyads:
The field is moving from the analysis of attraction to the study of personal
relationships. Thus, at the present time, we believe that the issue of
theoretical breadth is crucial. We used this dimension, as well as the related
questions of third parties, stages, and deterioration, in descriptively
comparing theories. In evaluating theories we again stressed the importance
of breadth as a criterion and suggested questions that we hope theorists of
friendship will address.
We now know much about how various principles work in isolation.
Another challenge for interpersonal attraction theorists is to specify how the
various forces combine. How important is each factor, in various situations,
with different kinds of people, in different types of relationships? Integrative
theories may be grandiose, yet certainly we are at the point where more
knowledge about the interplay, overlap, and relative strength of various
forces would be of value.
Reviewing theories is a bit like reviewing movies and books: There seems
to be an inherent role pressure on the reviewer to find faults. We have
performed our role. Yet we would like to end on a more positive note.
The ideas expressed in this chapter data back to Aristotle. The positions
that we have reviewed represent our selection of the most successful,
influential theories of the last 25 to 30 years. These theories have
accomplished their goal of helping to explain why particular individuals
like one another. The propositions expounded have stimulated or helped
interpret a great deal of research. Many of the research findings have turned
out as expected.
38
We feel that these theories merit respect and admiration. In thinking

about them over the years, our conflicts have typically been of the
approach-approach variety, but if one must have conflicts would you not
agree that these are the best kind to have? All that is left is to explain why we
are attracted to each position. If this chapter has succeeded, we hope that
you can, by analogy, offer some useful ideas for explaining our dilemma. Go
to it!
Dedication. As the authors finished writing this chapter, we were saddened to learn
of Professor Theodore Mead Newcomb's death. Born in 1903, he lived four score and
two. It was in Ted's graduate seminar (circa 1966) at the University of Michigan that
the senior author was first formally introduced to the question of interpersonal
attraction. Ted was an outstanding scholar, a provocative teacher, and a fine person.
He helped to establish the study of interpersonal attraction, and his legacy continues
to shape and guide much of what is done in the field. In appreciation of those
contributions, this chapter is dedicated to Ted's memory.
Acknowledgments. The contributions of the two authors of this chapter have been
relatively equal; the order of authorship reflects in part the senior author's longer
standing interest in this topic. The senior author would like to thank the students
over the last 15 years with whom he has discussed and sharpened his understanding
of attraction. In particular, the contributions of Bob Murray, Wendy Josephson, and
Chris Sexton have been important to the formulation of views and insights
expressed in this chapter. We would also like to thank Elaine Hatfield, George
Levinger, and Bernard Murstein for their comments on an earlier draft of the
manuscript.
References
Altman, I., & Taylor, D.A (1973). Social penetration: The development oj interpersonal
relationships. New York: Holt, Rinehart & Winston.
Aronson, E., & Linder, D. (1965). Gain and loss of esteem as determinants of
interpersonal attractiveness. Journal oj Experimental Social Psychology, 1, 156-171.
Berscheid, E. (1977). Interpersonal attraction. In B.B. Wolman & L.R. Pomeroy
(Eds.), International encyclopedia oj neurology, psychiatry, psychoanalysis, and psychology (Vol. 2, pp. 201-204). New York: Van Nostrand Reinhold.
Berscheid, E., & Wa1ster, E.H. (1978). Interpersonal attraction (2nd ed.). Reading, MA:
Addison-Wesley.
Brehm, S.S. (1985). Intimate relationships. New York: Random House.
Bryne, D. (1971). The attraction paradigm. New York: Academic.
Byrne, D., & Clore, G.L. (1970). A reinforcement model of evaluative responses.
Personality: An International Journal, 1, 103-128.
Clore, G.L., & Byrne, D. (1974). A reinforcement-affect model of attraction. In T.L.
Huston (Ed.), Foundations oj interpersonal attraction (pp. 143-170). New York:
Academic.
Davis, K, & Todd, MJ. (1985). Assessing friendship: Prototypes, paradigm cases and
relationship description. In S. Duck & D. Perlman (Eds.), Understanding personal
relationships (pp. 17-38). London: Sage.
39
Duck, S. (Ed.). (1977). Theory and practice in interpersonal attraction. London:

Academic.
Duck, S.W., & Perlman, D. (1985). The thousand islands of personal relationships: A
prescriptive analysis for future explorations. In S. Duck & D. Perlman (Eds.),
Understanding personal relationships (p. 1-15). London: Sage.
Dutton, D.G. (1984). Interventions into the problem of wife assault: Therapeutic,
policy and research implications. Canadian Journal of Behavioural Science, 16,281297.
Golightly, c., & Byrne, D. (1964). Attitude statements as positive and negative
reinforcers. Science, 146, 798-799.
Griffitt, W., & Guay, P. (1969). "Object" evaluation and conditioned affect. Journal of
Experimental Research in Personality, 4,964-976.
Hatfield, E., & Traupmann, 1 (1981). Intimate relationships: A perspective from

equity theory. In S. Duck & R Gilmour (Eds.), Personal relationships [. Studying
personal relationships (pp. 165-178). London: Academic.
Hatfield, E., Traupmann, J., Sprecher, S., Vtne, M., & Hay, 1 (1985). Equity and
intimate relations: Recent research. In W. Ickes (Ed.), Compatible and incompatible
relationships (pp. 91-117). New York: Springer-Verlag.
Hatfield, E., Vtne, M.K, & Traupmann, 1 (1979). Equity theory and intimate
relationships. In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing
relationships (pp. 99-133). New YorkL: Academic.
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Hinde, RA (1979). Towards understanding relationships. New York: Academic.
Hinde, RA (1981). The nature and study of personal relationships. In S. Duck & R
Gilmour (Eds.), Personal relationships I: Studying personal relationships (pp. 1-22).
London: Academic.
Hook, IG., & Cook, T.D. (1979). Equity theory and the cognitive abilities of children.
Psychological Bulletin, 86, 429-445.
Huston, T.L. (Ed.). (1974). Foundations of interpersonal attraction. New York:

Academic.
Jacobson, N.S., & Margolin, G. (1979). Marital therapy based on social learning and
behavior exchange principles. New York: Brunner/Mazel.
Kahn, A, O'Leary, v.E., Krulewitz, J.E., & Lamm, H. (1980). Equity and equality:
Male and female means to a just end. Basic and Applied Social Psychology, 1, 173197.
Kelley, H.H., Berscheid, E., Christensen, H., Harvey, IH., Huston, T.L., Levinger, G.,
McClintock, E., Peplau, L.A, & Peterson, D.R (1983). Close relationships. New
York: Freeman.
Kelley, H.H., & Thibaut, J.W. (1978). Interpersonal relations: A theory of interdependence. New York: Wiley.
Kerchoff, AC., & Davis, KE. (1962). Value consensus and need complementarity in
mate selection. American Sociological Review, 27, 295-303.
Levinger, G. (1974). A three level approach to attraction: Toward an understanding
of pair relatedness. In T.L. Huston (Ed.), Foundations of interpersonal attraction (pp.
99-120). New York: Academic Press.
Levinger, G. (1979). A social psychological perspective on marital dissolution. In G.
Levinger & O.c. Moles (Eds.), Divorce and separation: Context, causes and
consequences (pp. 37-63). New York: Basic Books.
Levinger, G. (1980). Toward the analysis of close relationships. Journal of Experimental Social Psychology, 16, 510-544.
Levinger, G. (1983). Development and change. In H.H. Kelley, E. Berscheid, A

Christensen, IH. Harvey, T.L. Huston, G. Levinger, E. McClintock, LA Peplau, &
D.R Peterson (Eds.), Close relationships (pp. 315-359). New York: Freeman.
40
Levinger, G., & Snoek, OJ. (1971). Attraction in relationships: A new look at
interpersonal attraction. Morristown, NJ: General Learning Press.
Lewis, RA (1972). A developmental framework for the analysis of premarital dyadic
formation. Family Process, 11, 17-48.
Lott, AI, & Lott, B.E. (1961). Group cohesiveness, communication level, and
conformity. Journal of Abnormal and Social Psychology, 62,408-412.
Lott, AJ., & Lott, B.E. (1974). The role of reward in the formulation of positive
interpersonal attitudes. In T.L. Huston (Ed.), Foundations of interpersonal attraction
(pp. 171-189). New York: Academic.
Lott, B.E., & Lott, Al (1960). The formation of positive attitudes toward group
members. Journal of Abnormal and Social Psychology, 61,408-412.
Milardo, RM., Johnson, M.P., & Huston, T.L. (1983). Developing close relationships:
Changing patterns of interaction between pair members and social networks.
Journal of Personality and Social Psychology, 44964-976.
Mills, I, & Clark, M.S. (1982). Exchange and communal relationships. In L. Wheeler
(Ed.), Review ofpersonality and social psychology (Vol. 3, pp. 121-144). Beverly Hills,
CA: Sage.
Murray, RB. (1973). Reinforcement and cognitive theoretical predictions of attraction
toward two strangers. Unpublished dissertation, University of Manitoba, Winnipeg,
Canada.
Murstein, B.I. (1971). A theory of marital choice and its applicability to marriage
adjustment. In B.I. Murstein (Ed.), Theories ofattraction and love (pp. 100-151). New
York: Springer.
Newcomb, T.M. (1959). Individual systems of orientation. In S. Koch (Ed.),
Psychology: A study of a science (Vol. 3, pp. 384-422). New York: McGraw-Hill.
Newcomb, T.M. (1961). The acquaintance process. New York: Holt, Rinehart &
Winston.
Newcomb, T.M. (1971). Dyadic balance as a source of clues about interpersonal
attraction. In B.I. Murstein (Ed.), Theories of attraction and love (pp. 31-45). New
York: Springer.
Price, KO., Harburg, E., & Newcomb, T.M. (1966). Psychological balance in
situations of negative interpersonal attitudes. Journal of Personality and Social
Psychology, 3265-270.
Rubin, Z., & Levinger, G. (1974). Theory and data badly matched: A critique of
Murstein's SVR and Lewis's PDF Models of mate selection. Journal of Marriage
and the Family, 36,226-231.
Rusbult, C.E. (1980a). Commitment and satisfaction in romantic associations: A test
of the investment model. Journal of Experimental Social Psychology, 16, 172-180.
Rusbult, c.E. (1980b). Satisfaction and commitment in friendships. Representative
Research in Social Psychology, 11, 96-105.
Scanzoni, J. (1979). Social exchange and behavioral interdependence. In RL.
Burgess & T.L. Huston (Eds.), Social exchange in developing relationships (pp. 61-98).
New York: Academic.
Thibaut, IW., & Kelley, H.H. (1959). The social psychology of groups. New York:
Wiley.
Traupmann, I, & Hatfield, E. (1983). How important is marital fairness over the
lifespan? International Journal of Aging and Human Development, 17,89-101.
Walster, E., Walster, G.W., & Berscheid, E. (1978). Equity: Theory and research.
Boston: Allyn & Bacon.
Walster, E., Walster, G.W., & Traupmann, I (1978). Equity and premarital sex.
Journal of Personality and Social Psychology, 36, 82-92.
West, S.G., & Wicklund, RA (1980). A primer of social psychological theories.
Monterey, CA: Brooks/Cole.
Wright, P.H. (1971). Byrne's paradigmatic approach to the study of attraction:
Misgivings and alternatives. Representative Research in Social Psychology, 2, 66-70.
Chapter 3
The Changing Functions of Friends in

Childhood: A N eo-Sullivanian Perspective
Although friendships are important throughout the lifespan, they seem to be

particularly important during middle childhood and adolescence. Prior to
this time, parents command center stage; after this period, heterosexual
relationships become the primary concern. By examining friendship as it
develops into a significant form of relationship during this period, we hope
to shed light on the functional importance of friendship. In particular, we
consider three ways that friends contribute to social development and
adjustment: (a) the fulfillment of interpersonal needs, (b) the socialization
of interpersonal competence, and (c) the provision of natural therapeutic
experiences. Our approach is developmental in that we trace the path by
which friendship grows to its mature form. It is also comparative in that we
consider the relative role of friendship within the child's broad network of
relationships.
The work of Sullivan (1953) is our theoretical point of departure. His
theory represents an ambitious attempt to account for personality development within the context of interpersonal relations. Unfortunately, his
contribution has remained largely overlooked because his written works are
poorly organized and often confusing. We begin with an overview of those
aspects of his developmental model relevant to the functions of friendship.
In order to systematize Sullivan, we often extend his reasoning to fill in gaps
left by his failure to follow-through on certain conceptual themes. Thus, our
model is really neo-Sullivanian. The discussion of the model is followed by
a selective review of research concerning the changing functions of
childhood friendships.
Sullivan's Theory of Social-Personality Development

Basic Concepts
A central concept in Sullivan's perspective on personality is that of the
interpersonal situation. He believed that people are motivated to bring about
42
Duane Buhnnester and Wyndol Funnan
certain desirable types of interpersonal situations and avoid other types. He

postulated that there are five basic social needs: tenderness, coparticipation
in playful activity, acceptance by others, interpersonal intimacy, and sexual
contact. Underlying these needs are motivational tensions roughly equivalent to the negative emotions of fear, loneliness, ostracism, boredom, and
anxiety. These tensions, which find their initial roots in human biology,
encourage people to create desired interpersonal situations. The attainment
of these situations not only eliminates unpleasant affective tensions, but
also prompts the positive emotional rewards of security, love, and feelings of
self-esteem.
Most contemporary psychologists are likely to find Sullivan's concept of
"need" to be a bit mysterious and outdated. Although emotional tensions
and social behavior may be linked, Sullivan's attempt to explain social
motivation in terms of underlying tensions is certainly an oversimplistic
account of complex motivational processes. Nevertheless, the general
concept of social needs has considerable appeal as a heuristic way to describe
certain social-motivational processes. In particular, the concept of needs
can be used to refer to preferences for one type of activity over alternative
ones. These preferences can be inferred by observing choice behavior,
whether it be verbal statements of choice or actual selections. These
preferences are not static, but they vary in relative strength as a function of
numerous factors, such as the alternative activities available and how long it
has been since the preferred interaction last occurred. The concept of need
also implies that once persons have engaged in the preferred activity, the
strength of their preference for that activity tends to decrease; that is, the
desire for the activity is temporarily satiated. In the present chapter, the term
social needs is used to refer to preferred types of interactions and need
fulfillment to the achievement of preferred activities with a resulting
reduction in preference strength.
Developmental Concepts
Sullivan's greatest contribution to the field of developmental psychology
was his description of normal social-personality development. He believed
that from birth on children's personalities are shaped by relationships with
parents, school authorities, siblings, and peers. He viewed the overriding
task of healthy psychosocial development to be the growth of patterns of
behavior and ego defenses that allow children to fulfill interpersonal needs
while keeping anxiety within bounds. This rich account of normal
development served as the basis for explaining the deviations that lead to
pathology.
Figure 3-1 outlines the major stages of Sullivan's developmental model.
Five of his developmental stages, or epochs, are listed. Because Sullivan
defined his stages in ternis of psychosocial events rather than chronological
age, the age range given for each epoch is only approximate. The beginning
The Changing Functions of Friends in Childhood
43
SEXUALITY
Opposite-sex
Partner
INTIMACY
Same-sex Friend
ACCEPTANCE
Peer Society
COMPANIONSHIP
Parents
Compeers
Parents
Friendship
Gang
Same-sex Friend
Parents
Opposite-sex
Friend/Romance
Same-sex Friend
Heterosexual
Crowd
Friendship Gang
Opposite-sex
Friend/Romance
Same-sex Friend
TENDERNESS
Parents
INFANCY
(0 to 2 yrs.)
Parents
CHILDHOOD
(2 to 6 yrs.)
Parents
JUVENILE ERA
(6 to 9 yrs.)
Same-sex Friend
Parents
PREADOLESCENCE
(9 to 12 yrs.)
Opposite-sex
Friend/Romance
Same-sex Friend
EARLY
ADOLESCENCE
(12 to 16 yrs.)
Figure 3-1. Neo-Sullivanian model of emerging social needs and key relationships.
During each developmental period (columns) a new need emerges (stairsteps) and is
fulfilled by the key relationships (rows).
of each stage is marked by an emerging social need that is added to the
expanding list of desired forms of interpersonal situations. (Even though
Sullivan described the emergence of social needs as the force that
determines the pace of social-personality development, he never provided a
coherent explanation of the factors that control the timing of the appearance of new needs.) The stair-step organization of Figure 3.1 is meant to
represent the cumulative nature of emerging social needs, in that new needs
are added to existing ones rather than replacing them.
Sullivan's description of needs as "emerging" or "arising" during development may be misleading because it implies that children move from a state
of not preferring a type of interaction to a point of suddenly desiring it. Such
discontinuities in development are seldom observed. Rather, preferences for
different social activities change over an extended period. There may,
however, be periods of rapid increase in the strength of preferences. These
periods of rapid change are probably what Sullivan described as times in
which needs emerged.
Figure 3-1 also outlines the key relationships that are most crucial for need
fulfillment during each developmental period. In some cases, a newly
emergent need prompts the establishment of a new type of relationship; for
example, in early adolescence the appearance of the sexual drive promotes
cross-sex relationships. In other cases, the satisfaction of the emergent need
is incorporated into a previously established type of relationship; for
instance, the need for playmate companionship is initially satisfied within
44
the context of the well-established parent-child relationship. In still other

cases, the satisfaction of a previously established need is met in new types of
relationships; for example, although parents start out as the key providers of
companionship in childhood, the key companions become the peer group,
same-sex friends, and finally opposite-sex peers in subsequent stages.
These key relationships are also the contexts in which particular
interpersonal competencies are learned (see Table 3-1). Although Sullivan did
not sytematically delineate this theme, we believe that social competence
develops through a process similar to the development of other skills
(Fischer, 1980). In particular, social competencies develop through experiences in interactions that require these competencies. Interactions vary in
their structural properties (Kelley & Thibaut, 1978), and, thus, different types
of interactions require different social competencies. When children
encounter a type of interaction that is structurally novel and beyond their
current skill level, they are pressed to expand their competence to the level
demanded by the new social task. Competence in each domain of social
skill grows gradually as higher levels of skill are built by accommodating
and intercoordinating existing competencies (cf. Weinstein, 1969).
Table 3-1. Structured Interpretation of Sullivan's Model of Socioemotional Development

Emerging
Needs/Key
Rela tionships
Interpersonal
Competencies
Developmental
Arrests
Focal
Emotions
Infancy
(0 to 2 yrs.)
Childhood
(2 to 6 yrs.)
Tenderness/
Coordinated
responding
Parents
Companionship/ Compliance and
Parents
assertion
Distress & fear/

security
Isolation &
boredom/
Enjoyment &
amusement
Juvenile era
(6 to 9 yrs.)
Acceptance/
Peer society
Cooperation,
compromise,
and competition
Preadolescence
(9 to 12 yrs.)
Intimacy/
Same-sex
friend
Insecure
attachment
Cycle of
isolation and
malevolent
transformation
Peer group
ostracism
and disparagement of
others
Lonely,
isolation
Collaboration:
Perspectivetaking, empathy,
and altruism
Confused
Balancing
sexuality
intimacy,
sexuality and
anxiety
Developmental
Stage
Early
Sexuality/
adolescence
Opposite-sex
(12 to 16 yrs.)
partner
Ostracism &
rejection/Social
pride &
self-worth
LonelinesslLove
Sexual
frustration/
Lust
45
Because relationships with different network members are characterized

by different types of interactions (Furman & Buhrmester, in press), each
relationship requires somewhat different social competencies. Accordingly,
we believe that different social skills are learned in different relationships.
For example, parent-child relationships are structurally asymmetrical,
whereas peer relationships are egalitarian (Hartup, 1979); Youniss, 1980). As
a result, certain skills, such as reciprocal exchange skills or mutual selfdisclosure skills, are likely to be acquired first in the context of peer
relationships. One would not expect a simple one-to-one correspondence
between social skills and relationships, because several relationships may
provide the contexts for learning the skills needed for a particular kind of
interaction. Still, certain relationships are likely to be primary contributors
to the development of particular social competencies.
According to a neo-Sullivanian theory, there should be a relatively
orderly sequence in which children master the full spectrum of social
competencies. This sequence is dictated in part by the order in which
children try to master different types of interactions, which in tum is
influenced by the sequence of ascending social needs. As children become
motivated to undertake new types of social activities, they must practice and
master the competencies demanded by those new types of interactions. This
developmental sequence is also influenced by the hierarchical nature of
social competence; the growth of some advanced skills are dependent on the
mastery and intercoordination of lesser skills. Thus, although there
probably is not a strict universal sequence of developmental stages, we
should be able to specify a normative timetable of when different social
competencies are usually first learned and the types of relationships that are
typically most instrumental to their growth.
Although Sullivan did not go so far as to advocate that there are "critical
periods" in development, he did believe that failure to acquire the requisite
skills of a stage can have lasting negative consequences. If children fail to
establish the key relationships of a period, then they are deprived of the
opportunity to master the new skills required by those relationships.
Because many advanced skills depend on the hierarchical accumulation of
basic competencies, the failure to establish key relationships leaves children
at a distinct disadvantage for gaining the skills of the subsequent stages.
In addition to providing the context for learning particular skills, key
relationships can also playa role in developmental arrests (see Table 3.1). In
almost every case, an arrest involves the child adopting a maladaptive
coping pattern in order to deal with an anxiety-provoking interpersonal
situation. For example, during the childhood period, a maladjusted parentchild relationship can set the stage for a "malevolent transformation," in
which the child comes to believe that he or she really lives among enemies.
With an arrest, development does not actually stop; instead, the course of
development is diverted to a less-than-optimal path.
Failure to form the key relationships of a period can also have emotional
46
consequences. The clearest example of this is Sullivan's account of

preadolescent loneliness, which he believed comes about when the need for
intimacy is not satisfied because of a failure to establish a close friendship.
More generally, Sullivan's description of development suggests that there
are a number of links between social life and emotional experience. Within
each stage, the emergence of a need is accompanied by the intensification of
particularfocal emotions (see Table 3-1). These include a set of positive and
negative affects, which motivate the child to acquire the social competencies
necessary to establish new forms of interaction (Weiss, 1974). For example,
during the juvenile era feelings of rejection and social self-worth intensify as
children come to desire acceptance by the peer society.
Developmental Stages
The infancy stage begins at birth and continues until the appearance of
articulate speech. In this first stage, the infant is totally dependent on adult
caregivers to satisfy bodily needs, and out of the association between the
tender ministering of the mother and the fullfillment of basic bodily needs
evolves "the needfor tenderness-for the protective care delicately adjusted to
the immediate situation" (Sullivan, 1953, p. 290). This continues throughout
life as the need for nurturance and support in times of helplessness and
distress. Sullivan argued that anxiety in mothers can be an extremely
disruptive influence on parent-infant interactions. At its extreme, maternal
anxiety may cause infants to develop primitive expectancies that interpersonal tenderness leads to feelings of anxiety and insecurity. In a related
vein, Ainsworth (1979) argued that maternal insensitivity, which perhaps
could occur because of anxiety, may lead to the developmental arrest of
"insecure attachment."
In Sullivan's theory, the childhood stage begins around age 2 and extends
to the time children begin preschool or kindergarten. Here the "need for
adult participation is added-that is, a need for the interest and participation
of a significant adult in the child's play" (Sullivan, 1953, p. 290). This need
continues throughout life as a desire for shared companionship in activities
of mutual interest and enjoyment. During childhood parents are key
playmate companions, taking on the responsibility for structuring playful
activities appropriate to the child's developmental level. Although children
of this age also play with peers and siblings, Sullivan believed that toddlers
strongly prefer interactions with parents. If the parents are negligent of their
roles as playful companions, their children may fall into a "cycle of
isolation." In this form of developmental arrest, children develop a reclusive
pattern of self-play that is accompanied by feelings of isolation and
boredom.
It is through the parent-child relationship that children first learn about
asymmetrical exchange in relationships. Compliance with powerful others
can bring tangible rewards and the fulfillment of social needs. Children
47
learn that needs are no longer unconditionally satisfied as they were in

infancy and that need fulfillment is awarded contingent upon appropriate
behavior. This is the stage in which fear is first brought to bear as a
socialization influence, with punishment and love withdrawal being used by
parents to bring about compliance with demands. Within the context of
parent-child relationships, children learn to delay gratification and comply
with demands in order to optimize rewards and cost. Docile compliance is
not the optimal strategy, however, for it is also essential to learn effective
self-assertion.
Not all children, however, come away from childhood with a positive
attitude about the exchange potential of social relationships. Some develop
what Sullivan described as the "malevolent transformation." This form of
developmental arrest occurs when overtures for tenderness and companionship are met with a mixture of indifference and rebuff from parents.
Children then come to believe that other people are enemies and that it is
best to protect oneself by striking out at others before others have a chance
to hurt you.
The juvenile era begins with the entrance to schooling, and it is then that
companionship with other children becomes increasingly desired. Being
able to engage in harmonious play relationships with peers demands that
children master yet another form of social relationship-the egalitarian
exchange relationship. In this type of relationship, both children have
relatively equal power status with neither particularly interested in or
capable of looking out for the needs of the other (Selman, 1980). Although
there is an inherent self-centeredness in children's approaches to these
relationships, they must nevertheless master the skills of cooperation,
compromise, and competition in order to be integrated successfully into
peer play groups (Piaget, 1965; Youniss, 1980). Sullivan pointed out that it is
not always easy to learn the optimal balance between prosocial and
competition 'behaviors; whereas winning is a valued goal of play, doing so
unfairly or at the expense of playmates can undermine the cooperative basis
of the relationship and reduce one's attractiveness as a play partner.
As juveniles move outside their homes into the world of peers, they
become increasingly aware of the differences among children in appearance, competence, family background, and social skill (Ruble, Parsons
& Ross, 1976). These differences serve as yardsticks by which the desirability
of children as playmates are measured. Social status hierarchies begin to
emerge as consensus grows concerning what makes children preferred
companions. Sullivan portrayed the peer society as often cruel and
insensitive, with children openly comparing their similarities and differences and in-group and out-group lines drawn to exclude those who are
different or considered inferior.
In conjunction with this peer-evaluation process, children acquire the
needfor acceptance by peers, along with its complement, the fear of ostracism
and exclusion. Juveniles learn that the acceptance and esteem of peers is
48
contingent upon how one looks, acts, and perfonns relative to agemates.
Moreover, the juvenile era is a time when one's feelings of self-worth are
increasingly influenced by status in the peer society. Sullivan believed that
ostracism by peers is one form of developmental arrest that may leave a
lasting mark on children's sense of self-esteem and deprives them of
experiences that are necessary for learning how to interact appropriately
with peers.
Another type of developmental arrest can grow out of the juvenile
concern with social comparison and self-worth. Here, some children come
to use disparagement of others as a means of evaluating their own standing.
Sullivan (1953) viewed this as a maladaptive security operation because it
has the effect of undennining the foundation of one's own sense of selfesteem:
Since you have to protect your feeling of personal worth by noting how
unworthy everybody around you is, you are not provided with any data that
are convincing evidence of your having personal worth; so it gradually
evolves into "I am not as bad as the other swine." To be the best swine, when
it would be nice to be a person, is not a particularly good way of furthering
anything except security operations. When security is achieved that way, it
strikes at the very roots of that which is essentially human-the utterly vital
role of interpersonal relations. (p. 242)
For Sullivan (1953), preadolescence is an important turning point in

development, when the "need for intimate exchange, for friendship, or for-in
its high refinement-the love for another person, with its enonnous
facilitation of consensual validation" (p. 291) emerges. Two strong motivating forces behind the need for intimacy are the experience oflove and the
avoidance of loneliness. The key target of this need is a same-sex friend or
"chum" who is similar to oneself in age, background, and interests. These
friendship bonds are characterized by an intense closeness based on
extensive self-disclosure. Although juveniles and even younger children do
have preferred playmates or friends, Sullivan felt these relationships fall
short of full-blown friendships.
These preadolescent friendships represent a significant advance in
children's mastery of fonns of personal relationships. They are not only the
prototype of adult friendships, but they are also a key foundation upon
which romantic, marital, and parenting relationships are built (Berndt,
1982; Piaget, 1965; Youniss, 1980). Sullivan (1953) thought of "chumships" as
the first relationships that could be characterized as collaborations, which
involve "clearly fonnulated adjustments of one's behavior to the expressed
needs of the other person in the pursuit of increasingly identical-that is,
more and more mutual-satisfactions" (p. 246). In contrast to the selfcentered exchange orientation of the juvenile period, collaborations are
relationship-centered exchanges in which children believe that by improving the welfare of their chums, they can raise their relationships to a
higher level of closeness and affection, which is to the benefit of each
49
participant. Whereas an attitude of "what should I do to get what I want?"

seems to characterize juvenile relationships, a collaborative orientation is
one of "what should I do to contribute to the happiness or to support the
prestige and feeling of worthfulness of my chum?" (Sullivan, 1953, p. 245).
Common experience tells us, however, that not all preadolescent friendships
are true collaborations and that even for those that are, there are times when
friends lapse into bouts of self-centeredness.
Establishing collaborative relationships fosters the growth of certain
social competencies. Although Sullivan did not clearly spell out these
competencies, we can speculate as to what they are. First, children must put
aside inhibiting feelings of insecurity in order to disclose private feelings
and secrets. Similarly, they must achieve a sense ofloyalty in order to create
an accepting and trusting atmosphere in which their chums feel free to
make themselves vulnerable. At times, children must be empathic and
compassionate supporters in order to aid their chums in times of distress. At
other times, children may be called upon to sacrifice personal needs
altruistically for the sake of their friends. To do this, children must be able to
take their chums' perspective and fully grasp their needs and then anticipate
the effects that one's own actions will have on others (Selman, 1980). Other
types of close relationships, such as those with parents and siblings, may
also provide interactions that promote the development of these same skills,
but Sullivan believed that chumships are of particular importance.
According to Sullivan, an outstanding benefit of an intimate chumship is
consensual validation. By self-disclosing, chums often learn that they are not
peculiar or different and that others have private lives similar to their own.
Through consensual agreement, they can also conclude that their shared
interests, preferences, hopes, and fears are valid and worthy. This provides
reassurance that one's outlook is truly "right." Having chums also make
children feel important simply because they are important to others. This
experience can be a great boost to one's sense of personal worth. Unlike
juvenile relationships where social value is predominantly determined by
skill at games and popularity in the group, preadolescents can feel worthy
and important because they are intensely wanted and needed by their
chums. The crucial factor in achieving the friend's esteem is being a good
relationship partner-having a loving and caring orientation that leads the
chum to love and care about you. Although preadolescents continue to
desire group acceptance, the esteem of a close friend can compensate
somewhat for lack of public status.
In fact, Sullivan thought that one extremely valuable function of intimate
friendships is the natural "psychotherapeutic possibilities" that they can
afford; that is, the supportive atmosphere of a collaborative friendship can
partially or wholly remedy certain developmental arrests resulting from
earlier misfortunes in relationships with parents and peers. Sullivan (1953)
outlined specifically how an intimate friendship can ameliorate several of
the major developmental arrests:
50
1. The malevolent child: The "drive connected with the need for intimate
association with someone else is so powerful that quite frequently
chumships are formed even by malevolent people," which can "provide
experience which definitely opens the mind anew to the possibility that
one can be treated tenderly, whereupon the malevolent transformation is
sometimes reversed, literally cured" (p. 253).
2. The isolated child: Children who are caught in the cycle of isolation
started by the failure of parents to serve as play companions may "come
out remarkably well able to handle themselves, to develop the docile
accommodation which did not really reach them in the juvenile period;
and this is because of the peculiarly intimate consensual exchange which
goes on in [preadolescent chumships]" (p. 254).
3. The disparaging child: "These folks, getting into the preadolescent
socialization, quite often gain enough in security from intimacy with
their chums to enable them to really open their minds and discuss the
other unpleasant people who don't seem to like them, in a fashion that is
illuminating, both as to the real worth of the others and as to some of
their own traits which may not be very endearing" (p. 253).
4. The ostracized child: "It is quite possible that in preadolescence that two
unfortunate juveniles thrown together by their unfortunate social status
as juveniles may, under the influence of the growing need for intimacy,
actually do each other a great deal of good. And as they show some
improvement they will become less objectionable to the prevailing
preadolescent society and may actually get to be quite well esteemed in
the gang" (p. 252).
It is difficult to overestimate the importance Sullivan gave to the
therapeutic potential of chumships. In fact, his innovative treatment for
schizophrenia involved a form of milieu therapy in which the aim was to
recreate preadolescent chumships.
Not only do preadolescent chumships have great therapeutic value, but
the failure to establish this type of relationship can cause feelings of
loneliness. Normally, the experience of loneliness acts as a strong motivating force that drives isolated children to overcome their insecurities and
establish close relationships, but if this does not happen, children suffer
frequent feelings of loneliness and isolation.
Beyond the experience of loneliness, Sullivan did not explicitly discuss
the lasting consequences of the failure to form an intimate friendship. It is
consistent with his thinking, however, to argue that chumships are necessary
for learning how to establish collaborations. Those who fail to form a
collaborative chumship may subsequently be arrested at an immature
juvenile form of relating to others in which relationships are approached
rather selfishly. Moreover, this failure may restrict the avenues for gaining a
sense of personal worth to those of gaining extrinsic approval, public status,
and disparagement of others.
51
Early adolescence begins with the "eruption of true genital interest, felt as
lust [and extends] to the patterning of sexual behavior which is the
beginning of the last phase of adolescence" (Sullivan, 1953, p. 263).
Accompanying the rise of lust is an extension of the need for intimacy to
include peers of the opposite sex. Thus, the key relationships of early
adolescence are ones with opposite sex peers, who have suddenly become
the object of two powerful needs-intimacy and sexuality.
Sullivan observed that it is usually difficult for young adolescents to
establish cross-sex relationships that can fulfill the demands of these needs.
He describes several "collisions" among the lust dynamism, the need for
intimacy and feelings of personal security that c:an interfere with forming
collaborative heterosexual relationships. Overpowering feelings oflust often
push young adolescents into actual or fantasiz(!d sexual encounters that
are accompanied by strong feelings of anxiety, shame, and guilt. These
feelings make it difficult to form relationships based on openness,
sensitivity, and mutual caring. Parents and socie:ty at large can add to the
problem by attempting to prohibit sexual experimentation and by ridiculing
or making light of early love relationships.
Sullivan also briefly described an additional stage of late adolescence,
which begins when a characteristic pattern of fuHilling intimacy and sexual
needs is established. This stage is not depicted in Figure 3.1 or Table 3.1
because Sullivan did not describe the key features of this stage. That is, he
did not specify the new social needs that (~merge nor indicate the
relationships that are of key importance.
Empirical Evidence
Sullivan's theory serves as a provocative perspective for considering the
changing functions of relationships, in particular friendship. Although few
investigators have explicitly used a Sullivanian framework, there are bodies
of empirical literature that bear on the themes that we have developed
(Berndt, 1982). In this section we critically review selected portions of this
literature concerning friendship, pointing out where the Sullivanian theory
has received support and where additional empirical and theoretical work is
needed.
Friends as Need Fulfillers
Sullivan believed that the role of friends in need fulfillment changes over
the course of development. Of particular relevance to the present chapter are
the needs for companionship and intimacy. The emergence of these two
needs captures the major shift from dependence on family to reliance on
peers that occurs between 6 and 15 years of age. During the juvenile stage
52
"compeers" become desired providers of companionship. During preadolescence the need for intimacy intensifies, fostering the formation of
close collaborative friendships.
If needs are thought of as preferences for particular types of interactions,
then we should observe developmental changes in the degree to which peers
and friends are preferred as sources of companionship and intimacy.
Unfortunately, few researchers have directly investigated with whom
children prefer to engage in different types of interactions. If we assume,
however, that children act in accordance with their preferences, then we
should be able to infer who they prefer as interaction partners on the basis
of the frequency with which they interact with different network members. If
children turn to peers more often than to parents for companionship, it is
likely that peers are the more desired source of companionship. Our
estimate of children's preference hierarchy based on interaction frequencies is only approximate, however, because children do not have full
control over how they spend their time. For example, it is. not uncommon for
children to want to spend more time playing with friends than parents will
allow. Nevertheless, in the absence of more direct measures of preference
hierarchies, interaction frequencies provide a reasonable estimate.
Several lines of evidence converge to support Sullivan's hypotheses about
the role of peers in fulfilling the desire for companionship. Wright (1967)
recorded the amount of time children ranging in age from 1.5 to 11 years
spent with different people in the course of a typical day. Consistent with
Sullivan's model, adults were the most frequent companions until about age
6, at which point time spent with other children rose to rival companionship
with adults. Companionship with adults, however, showed only a gradual
decline with age, and it was not until age 11 that companionship with peers
was more frequent than companionship with parents.
The results of one of our own studies are consistent with this observational study but broaden the picture to include older children
(Buhrmester & Furman, 1984). Children in the second, fifth, and eighth
grades were given a multi-item questionnaire asking them to indicate how
much time they spend with eight people (mother, father, closest grandparent, closest sibling, teacher, same-sex friend, opposite-sex friend, and boy
or girl friend). Figure 3-2 presents the mean ratings for mother, father, samesex friends, and opposite-sex friends. In the juvenile era (second grade),
same-sex peers are perceived as providing as much companionship as
parents. By eighth grade same-sex peers are the most frequent source of
companionship.
As Sullivan predicted, second and fifth graders also reported spending
little time with opposite-sex peers. This finding is consistent with numerous
sociometric studies showing that preschool- and elementary-school-aged
children rarely nominate opposite-sex peers as desired play companions
(Asher, Oden, & Gottman, 1977). This taboo on cross-sex relationships may
play an important role in the socialization of masculine and feminine social
53
c..
CJ)
z
Q
z
____ - - . "
...-0
................... .......
.>., " "
t1i
. Same-sell
Friend
""
u. 3
z~
~
a:
z
Mother
'-.Father
Oppositesell Friend
.~.
2
1~~__~________~__________~1_ _
GRADE
Figure 3-2. Mean ratings of companionship with mothers, fathers, same-sex friends,
and opposite-sex friends for secondo, fifth-, and eighth-grade children.
styles. As a result of keeping to themselves, members of each sex may find

reinforcement for newly emerging and fragile sex-role norms and attitudes.
By early adolescence (eighth grade), the picture has changed substantially;
between fifth and eighth grades there is a significant increase in reported
companionship with opposite-sex friends, whereas parents' roles as companions begin to decline. Thus, Sullivan's key hypotheses concerning who
children rely on for companionship are supported.
Sullivan argued that the need for intimacy first emerges during preadolescence and that same-sex friends are the key providers. This view is
generally supported by research on children's conceptions of friendship.
Between preadolescence and adolescence children's descriptions of friendships show a dramatic increase in the number of comments about sharing
intimate thoughts and feelings (Berndt, 1981; Bigelow & LaGaipa, 1980;
Furman & Bierman, 1984). Children's ratings of the level of intimacy in
ongoing friendships also increase between childhood and adolescence
(Diaz & Berndt, 1982; Hunter & Youniss, 1982; Rivenbark, 1971; Sharabany,
Gershoni, & Hoffman, 1981). The exact point at which this increase takes
place, however, is less certain. Although Sullivan argued that the need for
intimacy with friends intensified in preadolescence (around 8 to 11 years of
54
age), most investigators have not found significantly higher intimacy levels
until adolescence (12 to 15 years of age).
These findings are complicated by sex differences, a topic Sullivan did
not discuss. (In fact, Sullivan, 1953, stated that his theory best describes the
social development of boys and that it mayor may not be an accurate
description of female development.) Investigators have consistently found
that girls report greater intimacy in friendship than do boys (Foot,
Chapman, & Smith, 1980; Rivenbark, 1971). This occurs in both children's
descriptions of what friendships should be like (Berndt, 1981; Bigelow &
LaGaipa, 1980) and in their descriptions of their own friendships (Childs &
Furman, 1984; Sharabany et aI., 1981).
As part of the study described previously (Buhrmester & Furman, 1984)
we attempted to determine at what age sex differences in self-reported
intimacy in friendships first emerge. Children in the second, fifth, and
eighth grades rated how much they talk to, share their private thoughts and
feelings with, and tell secrets to their best friend. Figure 3.3 presents the
mean ratings of friends for boys and girls separately. In the second grade,
boys and girls did not significantly differ in their ratings of same-sex
friendships. By the fifth grade, however, girls rated their friendships as being
significantly more intimate than boys did, and this difference was even
greater in the eighth grade. Thus, in keeping with Sullivan's general claim, it
appears that preadolescence may be the stage in which friends become key
providers of intimacy for girls, but boys' development may lag somewhat
behind.
Figure 3-3 also indicates that Sullivan may have underestimated the
importance of parents as sources of intimacy, particularly for younger
children. As can be seen in Figure 3-3, parents are perceived as the key
providers of intimacy for second graders and continue to be significant,
though secondary, sources through early adolescence. Thus, although
friendship serves an important function in the fulfillment of the need for
intimacy, it certainly is not the only relationship that serves this function.
Sullivan's failure to consider the role of parents as intimacy providers is
indicative of a general tendency by him not to discuss the relative
contributions made by the full range of network members.
Sullivan thought that cross-sex peer relationships do not playa role in
satisfying the need for intimacy until early adolescence. This view is
supported by the results in Figure 3-3 and by data from other studies
(Rivenbark, 1971; Sharabany et aI., 1981). Whereas most young adolescents
report having opposite-sex chums with whom they disclose a great deal,
these cross-sex friendships still remain less intense than same-sex ones.
Perceived levels of intimacy are even higher in relationships with boyfriends
and girlfriends, but these too do not exceed ratings of same-sex friends. This
situation changes in late adolescence. Numerous investigators have found
that cross-sex intimacy-both in and out of the context of romantic
relationships-becomes increasingly important as adolescents approach
55
.Same-sex
/' ~ Friend Girls
f=
z
~ 3
~
z
~
a:
z
w
~
;(
.- .--
--- ..~
-----.~
,---,-""
"
. - - -
............ ...
.........
.':'<
......
"
. Same-sex
Friend Boys
Mother
Opp.-sex Friend
Father
.~.
2
GRADE
Figure 3-3. Mean ratings of intimacy with mothers, fathers, opposite-sex friends, and
same-sex friends for secondo, fifth-, and eighth-grade children. Means for boys and
girls are presented separately for ratings of same-sex friends.
young adulthood (Shaver, Furman, & Buhrmester, 1985; Wheeler, Reis, &
Nezlek, 1983). Thus, whereas Sullivan may have overestimated the importance of friends as providers of intimacy in preadolescence, it is clear
that same-sex and opposite-sex friends have become key sources of social
provisions by early adolescence.
Friendship and the Growth of Social Competence
Following Sullivan's lead, we have sketched out a model of the development
of social competence. In our view, social competence grows as children
undertake new types of interactions in the context of different relationships.
Childhood peer relations and friendships provide some particularly
important opportunities for fostering the growth of social competence. In
this section we briefly overview selected findings concerning their contribution to the growth of social competence.
It is important to distinguish between the influence of peer-group
relations and the influence of friendship (Furman & Robbins, in press).
Sullivan suggested that during the juvenile period the egalitarian exchange
basis of peer-group relationships provides an important context for
56
becoming competent at cooperation, compromise, and competition. Children who master these modes of interaction are likely to be accepted and
popular in the peer group. Preadolescent friendships, on the other hand,
represent a collaborative structure that fosters high level perspective-taking
skills, modes of empathic support, and altruistic concern for friends' needs.
These latter competencies contribute directly to the quality of close dyadic
relationships and are required less frequently in interactions involving
groups of peers than in dyadic interactions.
Although the Sullivanian hypothesis suggests that particular relationships are more critical than others to the mastery of certain skills, it does not
imply that these skills are exclusively learned in one type of relationship.
Skills in perspective-taking, empathic support, and altruistic concern are
also called upon in peer-group interactions; however, these close relationship skills are probably of greater importance in friendships than in group
interactions. Similarly, the skills of cooperation, competition, and compromise are used in friendships, but these skills are of crucial importance in
determining successful group interactions. Thus, popularity should be more
highly correlated with one set of skills, whereas the quality of friendships
should be more highly correlated with another set, even though both should
be correlated.
Investigators studying the correlates of sociometric status have found
general support for the hypothesis that children must be able to cooperate,
compromise, and compete appropriately in order to gain peer-group
acceptance. Compared to unpopular children, popular children engage
more frequently in cooperative or prosocial behaviors, such as (a) adjusting
to and cooperating with group rules (Coie, Dodge, & Coppotelli, 1982;
Lippitt, 1941; Van Hasselt, Bellack, & Hersen, 1979), (b) giving and receiving
more reinforcement (Gottman, Gonzo, Rasmussen, 1975; Hartup, Glazer,
Charlesworth, 1967; Masters & Furman, 1981), and (c) expressing kindness
(Smith, 1950). In contrast, unpopular or rejected children are likely to
engage in more agonistic and punitive interactions than are popular
children (Gottman et. aI, 1975; Hartup et al., 1967; Furman & Masters, 1980)
and are more disruptive in groups (Coie et al., 1982). The skills of
compromise and appropriate competitiveness have not been studied
directly; however, popular children have been found to be more conforming
(Masters & Furman, 1981; Moore, 1967) and better at sports (Klaus, 1959)
than are unpopular children, suggesting that they may be more adept in
these skills as well.
Few investigators have examined the hypothesis that perspective-taking
skills, empathy, and altruism are associated with the quality of preadolescent friendships. Although the evidence is not completely consistent,
some investigators have found that peer-group popularity is positively
correlated with perspective-taking ability (Gottman et al., 1975; Kurdek &
Kri1e, 1982; Rubin, 1972), expressing kindness to peers (Smith, 1950), and
using conventional rather than unconventional modes of help-giving (Ladd
57
& Oden, 1979). Popularity, however, is a measure of the number of friends

and not necessarily a measure of the quality of dyadic friendships.
In a few cases investigators have examined the influence of close

friendship independent of popularity status. Mannarino (1976, 1979)
identified two groups of preadolescents, one group who had stable
friendships and another who had no close friends. These groups were
matched on popularity status. Children with close friends scored higher on
self-report and laboratory assessments of altruism than those without close
friends. In a similarly designed study, McGuire and Weisz (1982) found that
children with close friends displayed higher levels of altruism and affective
perspective-taking than did those without close friends. By contrast,
sociometric popularity was not related to these competencies.
Other than these few exceptions, the critical tests of Sullivan's hypothesis
about the relative contributions of peer groups and close friendships have
not been conducted. One problem has been the lack of methods to assess
children's friendships. As previously mentioned, the number of friends
children have is not a good indicator of the quality of those friendships.
There is a strong need for measures that assess the qualitative features of
friendships, such as the extent of companionship, level of intimacy, sense of
loyalty, and the frequency of conflict. Such measures have recently been
developed (Furman, Adler, & Buhrmester, 1984; LaGaipa, 1981; Mannarino, 1976), but have not yet been extensively used.
The Therapeutic Benefits of Friendship

Sullivan argued that preadolescent friendships could help remediate several
types of social adjustment problems. We have been unable to find any
research designed specifically to test his claims, but there are several studies
that shed some light on the possible role of friends in overcoming
adjustment problems.
Children who have at least one close friend may be better adjusted than
those who have no close friends. Mannarino (1978) reported that among
children matched for popularity status, those who had close friendships
reported higher levels of self-esteem as compared to those who had no close
friends. Sundby and Kreyberg (1968) found that emotionally disturbed
children who had a best friend had a much better prognosis than did those
without one. The difference in prognosis was even apparent 10 to 15 years
after the initial assessment. Because these are correlational studies, it is
unclear whether children with friends are better adjusted because they have
friends or whether they have friends because they are better adjusted.
Other tests of Sullivan's claims concerning the therapeutic benefits of
friendships are intervention studies in which peers are involved as part of
the treatment program. Although a number of investigators have developed
social skills training programs that improve children's peer interactions and
sociometric status (see Furman, 1984; Hops, 1982), only a few researchers
58
have incorporated interactions with peers as part of their treatment

procedures (Furman, Rahe, & Hartup, 1979). In one program for unaccepted
preadolescents, Bierman and Furman (1984) compared the effectiveness of
three treatment conditions: (a) social skills training, (b) interactions with
peers that were structured to be positive in nature, and (c) a combination of
skills training and positive peer-group interaction. Whereas skill training
promoted the acquisition and use of social skills, the peer involvement
increased peer acceptance and children's self-perceptions of social competence. The combined treatment condition appeared to be the most
effective; changes were found in both social skills and peer acceptance. The
structured interactions enabled the children to apply their newly learned
skills in interactions that fostered the development of positive relationships
and helped change the peers' opinion toward target children (Bierman,
1984).
In light of these findings and Sullivan's theoretical arguments concerning
the therapeutic value of chumships, we believe that it may be worth
developing friendship-enhancement programs. Unlike other programs that
have been designed to improve children's relationships with peers in
general, the aim of a friendship-therapy program would be to enhance the
quality of children's friendships by fostering the competencies demanded by
close relationships. Such a program should include several elements. First,
it should teach children social skills that are important to friendship. In
addition to the conversational and other pro social skills included in
traditional programs, it might also include training components designed to
promote intimate self-disclosure, recognition of friends' needs, and emotional support. Second, children should be given the opportunity to exercise
and master these new competencies in the context of an ongoing friendship.
Children who do not have a friend should be paired with a receptive peer.
Various interaction exercises could be developed to provide opportunities to
apply the newly learned skills to their ongoing relationship. Finally, a
trained counselor should monitor the progress of their relationship over an
extended period of time to help prevent the reappearance of maladaptive
interaction patterns.
A basic tenet of our theoretical model is that close relationship skills are
mastered within the context of a collaborative relationship, such as a
friendship, in which each partner is trying to advance the relationship to a
more mutually satisfying level. By first encouraging children to learn and
apply close relationship skills to an ongoing friendship, and then nurturing
the growth of that friendship, this treatment program should promote the
growth of skills and at the same time foster the development of at least one
friendship. The program may even have a beneficial impact on other
relationships, including those with siblings, parents, and the larger peer
group.
59
Summary and Conclusions

Overall, the empirical findings provide encouraging support for a neoSullivanian model. The evidence concerning developmental changes in the
nature of friendships seems particularly strong, and the evidence concerning the links between interpersonal competencies and relationships is
promising. At the same time, many of Sullivan's hypotheses have not been
adequately tested. For example, we know remarkably little about the
therapeutic benefits of friendships.
It is also evident that some revisions in the theory will be necessary. For
example, Sullivan gave little consideration to sex differences in interpersonal relationships, and it seems likely that the course of social
development for boys and girls will differ in some respects. As noted
previously, Sullivan principally focused on the key relationships during
different developmental epochs. We need to consider also the "secondary"
roles played by other network members. For example, even though friends
appear to be the key relationship during the transition from childhood to
adolescence, parents continue to have important roles. Sullivan neglected
other potentially important relationships as well, such as those with siblings,
relatives, and teachers.
Further, Sullivan did not adequately address late adolescent and adult
development. Clearly, the functions of friendship are likely to change as
young people select mates, establish lasting marital bonds, and become
parents. Although Sullivan devoted little attention to later development, it is
likely that different social needs become salient (e.g., the need to nurture
one's own children) and that new social competencies must be mastered.
As noted earlier, the concept of social needs also should be carefully
scrutinized. Since the time that Sullivan developed his theory, many
concerns have been raised about the value of such a construct. At the same
time, we believe that some motivational component is required to understand social development.
We also should be certain that friendships are not overidealized. For
many years these relationships were neglected by developmental psychologists. Now, however, they are receiving much attention, but the focus
has almost exclusively been on the positive benefits of such relationships.
We need to recognize that friendships can be sources of strain and have
negative impacts on children as well. Children may become overly
concerned about their friends' approval, or they may be competitive with
friends (Fincham, 1978).
Despite these limitations, we believe that a neo-Sullivanian model has
considerable promise. We have tried to systematize Sullivan's model of
social development and apply it to contemporary work. The model is a
developmental one that attempts to outline the stages that children pass
60
through as they grow into mature social beings. The description hinges on
the intuitively appealing idea that children are actively motivated to
establish new types of social relations as they get older. New relationships
not only have the function of fulfilling social needs, but they contribute to
the growth of social competencies and the remediation of adjustment
problems. By specifying several functions of social relations, the model
allows us to compare and contrast the signficance of different types of
relationships at various points in development. Thus, we hope that
investigators will draw on Sullivan's insights in building a comprehensive
understanding of the role of friendships in social development and
adjustment.
Acknowledgments. Portions of the research described in this chapter were supported
by Grant No. #R01 HD 16142 from the National Institute of Child Health and
Human Development. Preparation of the chapter was facilitated by a W. T. Grant
Faculty Scholar Award to W. Furman.
References
Ainsworth, M.D.S. (1979). Infant-mother attachment. American Psychologist, 34,932937.
Asher, S.R., Oden, S., & Gottman, I (1977). Children's friendships in school settings.
In L.G. Katz (Ed.), Cu"ent topics in early childhood education (Vol. 1, pp. 33-61).
Norwood, NJ: Ablex Publishing.
Berndt, T.1. (1981). Relations between social cognition, nonsocial cognition, and
social behavior. In J.H. Flavell & L.D. Ross (Eds.), Social cognitive development (pp.
176-199). New York: Cambridge University Press.
Berndt, T.1. (1982). The features and effects of friendship in early adolescence. Child
Development 53, 1447-1460.
Bierman, KL. (1984). Process of change during social skills training with preadolescents
and its relationship to treatment outcome. Manuscript submitted for publication.
Bierman, KL., & Furman, W.e. (1984). The effects of social skill training and peer
involvement on the social adjustment of preadolescents. Child Development, 55,
151-162.
Bigelow, B.1., & LaGaipa, J.1. (1980). The development of friendship values and
choice. In H.e. Foot, Al Chapman, & IR Smith (Eds.), Friendship and social
relations in children (pp. 15-24). New York: Wiley.
Buhrmester, D., & Furman, W. (1984, July). The need fulfilling role of friendship in
children's social networks. Paper presented at the Second International Conference
on Personal Relations, Madison, WI.
Childs, M.K, & Furman, W. (1984). Temporal changes in peer relationships. Manuscript submitted for pUblication.
Coie, ID., Dodge, KA, & Coppotelli, H. (1982). Dimensions and types of social
status: A cross-age perspective. Developmental Psychology, 18,557-570.
Diaz, R.M., & Berndt, T.J. (1982). Children's knowledge of a best friend: Fact or
fancy? Developmental Psychology, 18, 787-794.
Fincham, F. (1978). Recipient characteristics and sharing behavior in the learning
disabled. Journal of Genetic Psychology, 133, 143-144.
Fischer, K W. (1980). A theory of cognitive development: The control and construction of hierarchies of skills. Psychological Review, 87,477-531.
61
Foot, H., Chapman, A, & Smith, J. (1980). Patterns of interaction in children's

friendships. In H.e. Foot, AJ. Chapman, & J.R Smith (Eds.), Friendship and social
relations in children (pp. 267-292). New York: Wiley.
Furman, W. (1984). Issues in the assessment of social skills of normal and
handicapped children. In T. Fields, J.L. Roopnario, & M. Segal (Eds.), Friendship
in nonnal and handicapped children (p. 3-30). Norwood, NJ: Ablex.
Furman, W., Adler, T., & Buhrmester, D. (1984, July). Structural aspects of relationships: A search for a common framework. Paper presented at the Second International Conference on Personal Relationships, Madison, WI.
Furman, W., & Bierman, K (1984). Children's conceptions of friendship: A
multidimensional study. Developmental Psychology, 96, 925-931.
Furman, W., & Buhrmester, D. (1985). Children's perceptions of the personal
relationships in their social networks. Developmental Psychology 21, 1016-1022.
Furman, W., & Masters, J.e. (1981). Peer interactions, sociometric status, and
resistance to deviation in young children. Developmental Psychology, 16,229-236.
Furman, W., Rahe, D.F., & Hartup, W.W. (1979). Rehabilitation of socially
withdrawn pre-school children through mixed-age and same-age socialization.
Child Development, 50,915-922.
Furman, W., & Robbins, P. (1985). What's the point? Issues in the selection of
treatment goals. In B. Schneider, K Rubin, & J. Ledingham (Eds.), Peer
relationships and social skills in childhood: Issues in assessment and intervention (pp.
41-54). New York: Springer-Verlag.
Gottman, J., Gonzo, J., & Rasmussen, B. (1975). Social interaction, social competence, and friendship in children. Child Development, 45, 709-718.
Hartup, W.W. (1979). The social worlds of childhood. American Psychologist, 34, 944950.
Hartup, W.W., Glazer, JA, & Charlesworth, R (1967). Peer reinforcement and
sociometric status. Child Development, 38, 1017-1024.
Hops, H. (1982). Social skills training for socially withdrawn/isolated children. In P.
Karoly & J. Steffen (Eds.), Advances in child behavior analysis and therapy (Vol. 2):
Intellectual and social deficiencies. New York: Gardner.
Hunter, ET., & Youniss, 1. (1982). Changes in functions of three relationships during
adolescence. Developmental Psychology, 18, 806-811.
Kelley, H.H., & Thibaut, 1.W. (1978). Interpersonal relations: A theory of interdependence. New York: Wiley.
Klaus, RA (1959). Interrelationships of attributes that accepted and rejected children
ascribe to their peers. Unpublished doctoral dissertation, George Peabody College
for Teachers.
Kurdek, LA, & Krile, D. (1982). A developmental analysis of the relation between
peer acceptance and both interpersonal understanding and perceived social selfcompetence. Child Development, 53, 1485-1491.
Ladd, G.W., & Oden, S. (1979). The relationship between peer acceptance and
children's ideas about helpfulness. Child Development, 50, 402-408.
LaGaipa, lJ. (1981). A systems approach to personal relationships. In S.W. Duck &
R Gilmour (Eds.), Personal relationships [. Studying personal relationships (pp. 6290). New York: Academic.
Lippitt, R (1941). Popularity among preschool children. Child Development, 12,305322.
Mannarino, AP. (1976). Friendship patterns and altruistic behavior in preadolescent

males. Developmental Psychology, 12,555-556.
Mannarino, AP. (1978). Friendship patterns and self-concept development in
preadolescent males. Journal of Genetic Psychology, 133, 105-110.
Mannarino, AP. (1979). The relationship between friendship and altruism in
preadolescent girls. Psychiatry, 42, 280-284.
62
Masters, J.e., & Furman, W. (1981). Popularity, individual friendship selection and
specific peer interaction among children. Developmental Pychology, 17,344-350.
McGuire, KD., & Weisz, J.R (1982). Social cognition and behavior correlates of
preadolescent chums hips. Child Development, 53, 1478-1484.
Moore, S.G. (1967). Correlates of peer acceptance in nursery school children. In
W.w. Hartup & N.L. Smothergill (Eds.), The young child: Review of research.
Washington, DC: Education of Young Children.
Piaget, J. (1965). The moral judgment ofthe child. New York: Free Press. (Original work
published 1932)
Rivenbark, W.H. (1971). Self-disclosure among adolescents. Psychological Reports, 28,
35-42.
Rubin, KH. (1972). Relationship between egocentric communication and popularity

among peers. Developmental Psychology, 7, 364.
Ruble, D.N., Parsons, J.E., & Ross, J. (1976). Self-evaluative responses of children in
an achievement setting. Child Development, 47,990-997.
Selman, RL. (1980). The growth of interpersonal understanding: Developmental and
clinical analyses. New York: Academic.
Sharabany, R, Gershoni, R, & Hoffman, J.E. (1981). Girlfriend, boyfriend: Age and
sex differences in intimate friendships. Developmental Psychology, 1, 800-808.
Shaver, P., Furman, W., & Buhrmester, D. (1985). Aspects of a life transition:
Network changes, social skills, and loneliness. In S. Duck & D. Perlman (Eds.),
The Sage Series in personal relationships (Vol. 1 pp. 193-217). London: Sage.
Smith, G.H. (1950). Sociometric study of best-liked and least-liked children.
Elementary School Journal, 51, 77-85.
Sullivan, H.S. (1953). The interpersonal theory of psychiatry. New York: Norton.
Sundby, H.S., & Kreyberg, P.e. (1968). Prognosis in child psychiatry. Baltimore:
Williams & Wilkins.
Van Hasselt, V.B., Bellack, AS., & Herson, M. (1979). The relationship between
behavioral assessment and the sociometric status of children. Paper presented at the
annual meeting of the American Association of Behavioral Therapy, San
Francisco.
Weinstein, E. (1969). The development of interpersonal competence. In D. Goslin
(Ed.), Handbook of socialization theory and research (pp. 753-778). Chicago: Rand
McNally.
Weiss, R (1974). The provisions of social relationships. In Z. Rubin (Ed.), Doing unto
others (pp. 17-26). Englewood Cliffs, NJ: Prentice-Hall.
Wheeler, L., Reis, H., & Nezlek, J. (1983). Loneliness, social interaction, and sex roles.
Wright, H.F. (1967). Recording and analyzing child behavior. New York: Harper &
Row.
Youniss, J. (1980). Parents and peers in social development: A Sullivan-Piaget perspective.
Chicago: University of Chicago Press.
Chapter 4
Personality and Friendship:

The Friendship Worlds of Self-Monitoring
"A friend may well be reckoned the masterpiece of nature."-Emerson
When one considers the amount of time people spend with their friends, the
range and diversity of their shared activities, the roles that friends play, and
the functions that they serve, there can be no denying the significance of
friendship in peoples' lives. It is perhaps because of the pervasive influence
and the diverse impact of friends that ambiguities and contradictions have
arisen from attempts to define friendship. Thus, from the earliest speculations of the ancient philosophers to the research programs of modern
scientists, people have been trying to gain an understanding of the nature of
friendship.
The origins of the classical notions of friendship can be found, for the
most part, in ancient Greek philosophy. Beginning with the works of Plato
and Aristotle, and appearing as a recurring theme in later works, is the
notion that qualitatively different kinds of friendship may exist. What
qualifies one person as a "friend" may be radically different from what
qualifies another person as a friend, and what constitutes a friendship for
one person may be entirely different from what constitutes a friendship for
another person.
Consider first the views of Plato on friendship. In the Platonic system, the
notion of differing types of friendship emerged as a distinction between
those friendships that qualified as "true" friendships and those that were
merely "illusions" and "deceptions." In that system, "true" friendships were
thought to be the result of basic human drives that were manifested in
"philosophic ecstasy" and "contemplation of the Good"; all other friendships were considered illusions and deceptions (Plato, Lysis, Phaedrus).
For Aristotle, a somewhat more elaborated typology of friendship was
necessary to encompass both the role of the friend's personality and the
functions served by the friend. As a parallel to the ideal friend envisioned by
Plato, Aristotle developed the concept of the "primary friendship." Although
closely related to Plato's "ideal," Aristotle's "primary" friendship represented
64
an ideal "type." Aristotle used this ideal type as a theoretical template

against which actual friendships could be examined (Aristotle, Ethica
Eudemia).
In describing this ideal type or "primary friendship," Aristotle proposed
three kinds of friendship. The first two are friendship for the sake of utility
and friendship for the sake of pleasure. Those who form friendships for the
sake of utility do so because they are useful for themselves. Similarly, those
who form friendships for the sake of pleasure do so because they are
pleasurable for themselves. Because of the selfish nature of these types of
friendships, Aristotle regarded them as imperfect ones. In contrast, he
viewed the third type of friendship as perfect. Perfect friendship is the
friendship of people who are good and alike in virtue. As such, the primary
friendship, or the perfect friendship includes the good, the pleasant, and
the useful (Aristotle, Ethica Eudemia; see also Aristotle, Nichomachean
Ethics).
Added perspective on Aristotle's categories offriendship is provided by an
examination of Epicurean and Stoic philosophies. The life of Epicurus is
the greatest testimony to his philosophy of friendship. According to
Diogenes Laertius, in his Lives of Eminent Philosophers, the number of
Epicurus' friends was so large that "the population of entire cities could not
measure up to it." For Epicurus, the most important means by which to gain
wisdom was the acquisition of friends. The Epicurian sage, in contrast to
previous Greek sages, found refuge in multitudinous friendships. On the
other hand, according to the Stoics, the wise man was completely selfsufficient and did not need friends. They thought this self-sufficiency
("autarky") protected the stoic sage from personal friendships that would
hinder the attainment of the detachment ("apathia") that they considered to
be the goal of life (Diogenes Laertius, Lives of Eminent Philosophers, VII).
From these Greek theorists we are led to the late Roman Republic and a
typology offered by Cicero. Cicero discussed two types of friendship,
"amancitia" and "concordia." Amancitia covered relationships marked by
similarities of character and virtue; concordia resulted from differences of
character and virtue. Cicero considered the majority of relationships to be
concordia, superficial relationships between dissimilar partners bound
together by bonds of utility and self-interest (Cicero, Laelius De Amicitia).
If there is one message that comes from these philosophical treatments of
the nature of friendship, it is that there is no single meaning to the terms
"friend" and "friendship." Whether we consider Plato's real and illusory
friendships, Aristotle's friendships of utility, pleasure, and virtue, the
Epicurean's valuing of multitudinous friendships and the Stoic's valuing of
none, or Cicero's contrast between bonds of virtue and bonds of self-interest,
we are confronted with the notion of differing conceptions of and
orientations to friendship.
Personality and Friendship
65
A Scientific Approach to the Study of Friendship

Despite its conceptual elusiveness, friendship is, we believe, still a topic
worthy of investigation. In fact, we also believe that because of this
conceptual elusiveness, friendship and similar topics may be particularly
appropriate ones for scrutiny from a scientific perspective. For only from an
objective, scientific perspective can different orientations toward friendship
be critically examined and integrated into a coherent explanatory framework.
What is an appropriate starting place for a scientific study of friendship?
Rather than continuing in the philosophical tradition that seeks the one
definition of what constitutes the ideal friendship, we propose instead to
focus on the individual: to examine individuals' personal definitions of
friendship and the influence of their conceptions of friendship on the
networks of their actual friendships and the social worlds within which they
live. Thus, our analysis will take, as its beginning point, properties of
friendship that can be specified at the level of the individual and, from
there, progress toward an understanding of the phenomena and processes of
friendship as they exist at the level of interpersonal relationships and social
structures. In short, we propose that in order to understand friendship at the
social level, we must first understand people at the individual level. Just as it
is reasonable to propose, as conventional wisdom would have it, that a
person is the product of the company he or she keeps, so too is it reasonable
to propose, as we do, that the company kept is a product of the person.
The rationale for this person-centered approach to the study of friendship
reflects a more general strategy for the study of personality and social
behavior, one outlined by Snyder and Ickes (1985). According to the logic of
this strategy, many of the phenomena and processes of concern to social
psychologists can be understood best by focusing investigative efforts on
those individuals who characteristically manifest them. In so doing, one
gains access to ideal candidates for investigating particular psychological
phenomena or processes by examining precisely those individuals who
typically manifest them.
How, then, is this strategy, which begins with the identification of
categories of individuals to serve as bases for subsequent conceptual and
empirical analyses, to be applied to the study of friendship? Such an
application requires that, in the beginning, a decision be made about the
level of abstraction at which these categories should be defined. We propose
to begin at the most basic level possible and consider categories of
individuals who differ in the value they place on their friendships. Since at
least the era of the ancient Greek and Roman philosophers, most people in
our society seem to value, in varying degrees, their friendships with others.
Thus, the strategy adopted for this analysis suggests that we seek to
66
determine whether there is a category of individuals who especially value

their friendships and a contrasting category of individuals who do not place
such a premium on especially close friendships but who might, rather, more
highly value having a greater number and variety of friendships characterized by lesser closeness. If these two categories of individuals exist, and
if it were possible to identify them with reliable and valid empirical
measures, then examination of members of these contrasting categories
could promote a better understanding of the dynamics of friendships.
How might these two hypothetical categories of individuals who differ in
orientations to friendships be identified? One approach involves the use of
an existing measure of a generalized interpersonal orientation, one whose
construct validity evidence provides reason to believe that differences in
orientation toward friendship would be a member of the set of features that
define this interpersonal orientation. In addition to identifying people who
possess differing orientations toward friendship, this approach provides a
larger explanatory framework within which to place an understanding of
friendship.
To the extent that differing orientations toward friendship are but one
specific feature of a larger and more extensive network of regular and
consistent differences in the ways members of these contrasting categories
think, feel, and act, then the differing generalized interpersonal orientations
characteristic of these contrasting categories of individuals may provide a
framework within which to understand orientations toward friendship. (For
a more elaborated discussion of this strategy of inquiry in personality and
social psychology, see Snyder and Ickes, 1985).
Current theory and evidence suggests that the psychological construct of
self-monitoring includes such a set of features. High self-monitoring individuals, those with relatively high scores on the Self-Monitoring Scale
(Snyder, 1974), typically strive to appear to be the type of person called for
by each situation in which they find themselves. These people appear to be
particularly sensitive and responsive to social and interpersonal cues to
situational appropriateness, and, accordingly, their social behavior displays
pronounced situation-to-situation specificity. Thus, they tend to claim, in
their endorsement of Self-Monitoring Scale items, that: "In different
situations and with different people, I often act like very different persons."
"I'm not always the person I appear to be." "I would probably make a good
actor." Low self-monitoring individuals, those with relatively low scores on
the Self-Monitoring Scale, on the other hand, appear to be less responsive to
situational and interpersonal specifications of behavioral appropriateness.
Rather, they seem to choose their behavior in social situations on the basis
of their own personal dispositions and attitudes, as indicated by the
characteristically substantial congruence between their social behavior and
relevant underlying attitudes and dispositions. Thus, they tend to claim, in
their endorsement of Self-Monitoring Scale items, that: "I would not change
my opinions (or the way I do things) in order to please someone or win their
67
favor." "I can only argue for ideas which I already believe." "I have trouble
changing my behavior to suit different people and different situations."
The evidence for the construct validity of self-monitoring as measured by
the Self-Monitoring Scale is extensive, involving empirical investigations of
theoretically derived propositions about the involvement of self-monitoring
propensities in the cognitive, behavioral, and interpersonal domains (for
reviews and evaluations of this construct validity evidence, see Snyder, 1979;
Snyder & Ickes, 1985; Shaw & Costanzo, 1982).'
Conceptions of Friendship
If there are categories of individuals who differ fundamentally in their
orientations to friendship and if these contrasting categories of individuals
can be identified with the psychological construct of self-monitoring, how
should we expect that these contrasting orientations will be manifested? At
the most basic oflevels, these contrasting orientations should be revealed by
individuals' conceptions of friendship, that is, in pervasive and systematic
differences in the terms that they use to define and characterize friendship,
in the thoughts they have about friendship, and in the "meaning" that they
give to friendship.
To investigate conceptions of friendship, we asked individuals to write
essays that described their relationship with a specific person whom they
considered to be a friend (Snyder & Smith, 1984). We thought that being
based on actual ongoing friendships, these essays would more accurately
represent individuals' conceptions of what friendship actually is as opposed
to what friendship should be. Two independent judges, blind to the essay
writers' self-monitoring category, then coded these essays for evidence of
five aspects of conceptions of friendship, aspects chosen on the basis of
previous research on the interactional styles of high and low self-monitoring
individuals.
The first coding dimension concerned the basis of friendship and
contrasted friendships based on situations and chance interactions (such as,
"we both work at the same store") with those that emphasize such things as
shared values and mutual respect (for example, "we are such good friends
because we both take school very seriously and appreciate that we've done
with our lives").
The second coding dimension examined the tone of the interactions
between the essay writers and their friends. This dimension was indicated
'For information on its psychometric properties, including internal consistency, testretest reliability, convergent and discriminant validity, as well as information about
administering and scoring it, see Snyder (1974); for analyses of the underlying
structure and recent psychometric refinements of the measure, see Gangestad &
Snyder (1985).
68
by the writers' tendency to be animated and emphatic in their descriptions

as well as intensely "good-time" oriented in their interactions. For example,
whereas one essay might read "we go out" another essay expressing a similar
idea might read "we go out a lot and always have an incredibly good time."
For one person, just "going out" is a sufficient description and justification
whereas for the other person the utility and pleasure of the interaction must
also be included.
The third coding dimension was based on whether the friendships
seemed shallow or deep. Here, long-term enduring friendships that had
weathered hardships and change were considered deep ones. Short-term,
spur-of-the-moment friendships, ones based on contrived situations and
convenience, were considered shallow ones. For example, the statement "he
sold me tickets to the concert when he found out he couldn't go" was coded
as evidence of a rather shallow relationship, whereas the statement "I would
do anything for her and she for me" was taken as evidence of a relatively
deep relationship.
The fourth coding dimension, sense of enduring compatibility, required the
friendship to have sufficient breadth to generalize to situations different
from the present one. This feature usually had to be inferred from such
things as the resolution of past conflicts in constructive ways, the maintenance of a friendship with someone who has moved far away, or the
adaptation to some change, such as coworkers becoming roommates. If
there was a doubt as to whether the friendship could survive in another
situation (for instance, "we're great friends as long as we don't see too much
of each other"), it was coded as having nonenduring compatibility.
The fifth coding dimension, nurfurance, dealt with friends helping each
other in times of need, listening to each others' problems, and understanding and protecting each other; these qualities were predicted to be
relatively lacking in the friendships of some individuals while being an
important and central component to others. For example, the statement "I
seem to be the only one who is able to talk her out of her depressions"
indicates a nurturant component to the relationship.
The results of this content analysis were clear. Interrater agreement was
substantial, r = .78, between raters for all essays. Based upon the summed
judgments of the two raters, the following differences in conceptions of
friendship emerged. The friendships of high-monitoring individuals were
conceptualized in terms of an activity-based orientation, an animated and
emphatic tone to their interactions, a somewhat shallow sense of friendship,
little conception of compatibility and endurance. beyond the present
context, and little conception of nurturance. The friendships of low selfmonitoring individuals were conceptualized in terms of an affect-based
orientation, a definite sense of depth of friendship, considerable conception
of compatibility and endurance beyond the present context, and much
evidence of a conception of nurturance and sympathy. Further evidence of
these differing conceptions of friendship is provided by investigators (e.g.,
Jamieson, Lydon, & Zanna, 1984) who have found that for high self-
69
monitoring individuals, liking for other people typically reflects considerations of activity preferences and that for low self-monitoring individuals, liking typically reflects considerations of similarity of attitudes.
Behavioral Manifestations of Friendship

Clearly, individuals high and low in self-monitoring differ fundamentally in
their conceptions of friendship. But are these differing conceptions of
friendship reflected in the actual behaviors that characterize these friendships? Do the same individuals who conceive of friendship in particularly
activity-oriented terms also maintain actual friendships that can be
characterized by an activity emphasis? Similarly, do the same individuals
who conceive of friendship in particularly affect-oriented terms also possess
actual friendships that can be characterized by an affective emphasis?
To assess the influence of conceptual differences on actual choices and
preferences enacted in ongoing friendships requires the identification of
features of friendship that not only reflect the overt characteristics of the
friendship but also mirror the underlying differences in conceptions of
friendship. One set of features that fulfills both of these requirements
consists of the ways in which individuals allocate their leisure time among
their friends by choosing them as partners for social activities. To investigate
these processes, Snyder, Gangestad, and Simpson (1983) presented people
with choices that involved engaging in particular leisure-time activities with
specific members of their existing social networks. These choices always
involved activities in which the participants actually and typically engaged
(and, thus, presumably were activities performed with some enjoyment and
competence). Also, they involved activity partners who actually were
members of the participants' social worlds.
These choices posed contrasts of the form, for example, of either "playing
tennis with Steve" (where Steve is a friend known to be a particularly good
tennis player but not distinguished in general likability) or "going sailing
with Jeff' (where Jeff is a friend known to be very high in general likability
to the individual but who does not excel in sailing ability), where the
individual making the choice actually does play tennis and go sailing. When
faced with such choices, high self-monitoring individuals chose to spend
time with friends who were particular "specialists" in the activity at hand,
and low self-monitoring individuals chose to spend time with individuals
who were generally well-liked by them. There was no reliable difference
between men and women in the allocation of choices to particular
specialists and to generally well-liked friends.
The consequences of these differing approaches to choosing friends, with
decisions generally made on either the basis of general likability or that of
particular skill, may have cumulative effects. By adopting one strategy or the
other, individuals may carve out a distinctive world within which to live. If,
70
in fact, high self-monitoring individuals initiate and maintain friendships

on grounds such as skill at leisure-time activities and low self-monitoring
individuals do so out of a concern for general likability, then one would
expect that the social worlds of high self-monitoring individuals would have
very compartmentalized, segmented geographies with certain people being
paired with certain activities. The prototypic high self-monitoring individual may have one friend to play tennis with, another with whom to
discuss politics, and still another with whom to attend concerts and cultural
events. The social worlds of low self-monitoring individuals, on the other
hand, would exhibit uniformity and homogeneity, with well-liked friends
serving as activity partners for many different and varied activities. The
prototypic low self-monitoring individual may have one particularly wellliked friend with whom to go shopping, attend sporting events, go on trips,
or, indeed, participate in whatever activities he or she generally engages
in.
These characterizations of segmentation versus homogeneity are supported by empirical "maps" of individuals' social worlds. In an empirical
investigation, Snyder et al. (1983) asked individuals to generate a list of the
"population" of their social "worlds"-those people with whom they
regularly spent time. Participants then selected the one specific social
activity (e.g., "going to a fancy French restaurant," "playing tennis," "going
to the ballet") that was most representative of their social life within each of
several global categories of activities (e.g., "going out to dinner," "competitive recreactional activity," "attending live entertainment"). Having
generated the lists, participants then learned that each of the cells in a
matrix (labeled with people they had listed and with activities they had
nominated) represented engaging in a particular social activity with a
specific person. For each activity, the participants then estimated how likely
it would be that they would choose each of the people listed in the matrix as
a partner for that activity and how much they would enjoy engaging in each
activity with each of these people.
To distinguish the differentiation and segmentation in the social worlds of
high self-monitoring individuals from the uniformity and homogeneity in
those of low self-monitoring individuals, the investigators constructed an
index reflecting the residual amount of variation present in each participant's ratings that could not be independently accounted for by differences attributable to targets or by differences attributable to activities. On
this index, high self-monitoring individuals showed significantly more
nonadditive variation (i.e., differentiation, partitioning, segmentation) in
both their likelihood ratings and their enjoyment ratings than did low selfmonitoring individuals. There was no reliable difference between women
and men in the amount of person-by-activity variation within their matrixes
of social preference ratings.
The consequences of these patterns of choice also may have cumulative
effects. To the extent that high self-monitoring individuals choose friends on
71
the basis of their unique qualifications for the activities and roles they will
play, they will come to live in social worlds characterized by great
partitioning, differentiation, and segmentation. And, to the extent that low
self-monitoring individuals choose friends based upon considerations of
general similarity to and liking for them, they will come to inhabit social
worlds characterized by relatively pronounced homogeneity. In all likelihood, the segmentation characteristic of the social worlds of high selfmonitoring individuals makes it easier for them to adopt different identities
with different members of their worlds and to display the many selves they
pragmatically conceive themselves to be (Snyder & Campbell, 1982). Within
this type of social world, high self-monitoring individuals can, for example,
walk through an art gallery with one friend in the morning, go to a football
game with another friend in the afternoon, and spend the evening studying
with another. In contrast, the homogeneous characteristic of the social
worlds of low self-monitoring individuals allows them to "be themselves"
with different members of their social worlds and to display the principled
self that they strive to attain (Snyder & Campbell, 1982). Within this type of
social world, low self-monitoring individuals can, for example, spend an
entire day with the same friend or group of friends doing all the varied
activities that arise.
The Population of Friendship Worlds

Having delineated some of the structural features of the friendship worlds of
high and low self-monitoring individuals, we are now in a position to ask
questions about who populates these worlds. Who are these friends? What
qualifies these individuals for inclusion in the friendship worlds of
individuals high and low in self-monitoring? Given their contrasting
orientations to interpersonal relationships and their resulting social worlds,
one might expect that the close friends of high self-monitoring individuals
typically serve different needs and fulfill different functions than do the
close friends of low self-monitoring individuals. Clearly, individuals who
live in segmented social worlds characterized by popUlations chosen on the
basis of activity considerations ought to seek out and form friendships with
individuals who can understand and appreciate these considerations while
accepting them as satisfactory terms for their own friendships. And,
conversely, individuals who live in homogeneous worlds in which friends
are selected based upon considerations of similarity and likability ought to
contract friendships with individuals who understand and can appreciate
these considerations as well as accept them as satisfactory terms for their
own friendships.
It follows that the close friends of high self-monitoring individuals should
themselves be high in self-monitoring and that the close friends of low selfmonitoring individuals should themselves be low in self-monitoring. That
72
is, to be a close friend of a high self-monitoring individual, one would not

only have to understand and appreciate the behavioral flexibility and social
adaptiveness typically displayed by these individuals, but also be equally
capable of behaving with similar flexibility and social adaptiveness if the
relationship were to evolve into a close one. In essence, one would have to
possess the qualities and attributes typically possessed by an individual who
is high in self-monitoring. Moreover, to be a close friend of a low selfmonitoring individual, one would not only have to understand and
appreciate the cross-situational and attitude-behavior consistencies characteristically displayed by these individuals, but also have to behave in a
similarly consistent manner if a close relationship were to develop. That is,
one would have to possess the qualities typically possessed by an individual
who is low in self-monitoring.
To examine this hypothesis, Snyder, Simpson, and Smith (1984) asked
individuals, known to be either high or low in self-monitoring, to have one
close same-sex friend of their choosing complete the Self-Monitoring Scale.
Based upon the participants' ratings of the closeness of the friendship and
the self-monitoring category of the friends, the analyses revealed that, as
predicted, the close friends of high self-monitoring individuals had
substantially higher self-monitoring scores than did the close friends oflow
self-monitoring individuals.
It is important to note that we invoked a criterion concerning friendship
quality in this study. By examining only "close" friends, we hoped to obtain
a relatively pure sample of people who would be particularly likely to
exhibit those qualities individuals deem important for friendship. However,
in so doing, we created another distinction apart from the general
orientations toward friendship. This distinction was between "close" and
"casual" friendships. For our purposes, we defined closeness in terms of the
frequency and diversity of the interactions between friends. Specifically, we
defined close friends as those friends whom people simply enjoy being with,
regardless of the activity at hand. That is, it is the friend rather than the
activity that matters. In contrast, we defined casual friends as those friends
with whom people do only one or a very limited number of activities. Here it
is the activity and not the friend that is of primary importance.
This distinction itself suggests an important comparison. If, in many
regards, these casual friends qualify as friends and, evidently, are acceptable
at least as activity partners, how do they differ from those friends who are
close? What accounts for their not being in, or progressing into, the realm of
close friendships? In another investigation, designed specifically to address
the role of closeness in determining the links between friends' selfmonitoring propensities, Snyder et al. (1984) had participants nominate one
close friend and one casual friend and then had each friend complete the
Self-Monitoring Scale. For high self-monitoring individuals, close friends
were significantly higher in self-monitoring than were casual friends. In
contrast, for low self-monitoring individuals, close friends were significantly
73
lower in self-monitoring than were casual friends. These results indicate

that not only do individuals appear to gravitate toward friends with selfmonitoring propensities similar to and supportive of their own orientations
when forming close friendships, but also that this orientation toward
choosing close friends does not carry over into the domain of casual
friendships.
Perhaps this difference between close and casual friends can be explained
by considering the point in time at which a casual friendship exhibits the
potential for expanding beyond the particular circumstances in which it
began. One could argue that dissimilarity of interpersonal orientations
necessarily limits the common-ground upon which to base any extension of
tho;; friendship. Alternatively, where there is similarity of interpersonal
orientation, presumably there is much more common-ground upon which
to find alternative domains for interaction. Furthermore, the "interactional
strain" that often occurs in dyadic circumstances involving one high selfmonitoring individual and one low self-monitoring individual (cf. Ickes &
Barnes, 1977) may limit the potential for such relationships to grow beyond
the realm of casual ones.
There is yet another point that deserves recognition in any analysis of
relationship formation. Both common parlance and empirical examination
lead to two competing propositions about friendship. Is it the case that
"birds of a feather flock together"? Or is it the case that "opposites attract''?
The results of our investigations seem to indicate that, at least with respect to
self-monitoring, both propositions are, in fact, correct. Within the domain of
close friendships, similarity of interpersonal orientation provides the best
characterization-the "birds of a feather" proposition; but, within the
domain of casual friendships, complementarity of interpersonal orientation
provides the best description-the "opposites attract" proposition. It may
also be that in domains other than self-monitoring, a similar state of affairs
exists. Whether friendships are characterized by similarity or complementarity may depend on whether the friendships are close or casual ones.
Within the domain of close relationships, the rule of similarity may prevail;
but within the domain of casual relationships, the rule of complementarity
may predominate. In regard to this hypothesis, a meta-analysis of previous
and existing research on similarity and complementarity in friendship may
prove informative.
Understanding the Nature of Friendship

The view of the friendship worlds of self-monitoring that emerges from the
research conducted thus far is both consistent and clear. High selfmonitoring individuals appear to live in highly partitioned, differentiated,
or compartmentalized social worlds in which they engage in specific
activities with particular people. Further, their activity partners appear to
74
be chosen because of the partners' skill, expertise, or specialization in

specific activity domains. In addition, when close friendships do exist in the
lives of high self-monitoring individuals, they seem to be with people who
are themselves relatively high in self-monitoring. Moreover, this behavioral
orientation toward choosing friends has its cognitive counterpart in the
activity-oriented conception of friendship possessed by high self-monitoring
individuals.
Low self-monitoring individuals, by contrast, appear to live in social
worlds that are relatively homogeneous and undifferentiated in terms of the
links between specific friends and particular activity domains. Rather, the
members of their social networks appear to be chosen because they have
personal attributes similar to and/or because they are globally well-liked by
these individuals. In addition, the close friends of low self-monitoring
individuals seem to be themselves relatively low in self-monitoring. Here,
too, their behavioral orientation toward friendship has its cognitive
counterpart in their particularly affect-oriented conception of its nature.
In our analysis of personality and friendship, self-monitoring has served
us as a moderating variable, One which permitted the identification of
categories of individuals who differed in their characteristic orientations to
friendship. It made it possible to identify categories of individuals who
characteristically choose their friends On the basis of the roles they play and
the functions they serve and other categories of individuals who characteristically choose their friends On the basis of general likability and
desired global tendencies. Self-monitoring, of course, may not be the only
such moderating variable. Conceptual and empirical analyses may identify
other constructs with the ability to identify categories of individuals who, as
do low self-monitoring individuals, display a relatively affect-oriented
approach to friendship, and other categories of individuals who, as do high
self-monitoring individuals, display a relatively activity-oriented approach
to friendship. The successful identification of such categories of individuals
permits specifications of when affective/emotional considerations will
determine friendship formation and when activity/instrumental considerations will be the determinant of friendship.
The utility of this approach to the study of personality and friendship
perhaps can be conveyed best by considering what happens when one does
not distinguish between these categories of individuals. Without this
distinction, the ability of either orientation to predict friendship formation
will be diminished considerably. In our own program of research, for
instance, the ability of either the activity/instrumental orientation or the
affective/emotional orientation On its own to predict friendship has been
minimal. That is, there has been nO evidence of a pervasive tendency to
conceive of friendship in one or the other form or to contract friendships On
One or the other basis. Only by considering the moderating influences of
self-monitoring has it been possible to identify one category of individuals
75
for whom affective/emotional predictions of friendship formation perform

well (i.e., low self-monitoring individuals) and a second category of
individuals for whom activity/instrumental predictions of friendship formation perform well (i.e., high self-monitoring individuals). We expect that
the same will prove true of other moderating variables that, no doubt, will be
identified by researchers conerned with friendship.
Thus, within the domain of friendships, there appears to be a whole set of
fundamental questions that yield no answers without the perspective gained
through the identification of categories of individuals for whom it is
appropriate to ask such questions and who can serve as appropriate
candidates for answering such questions. And, although the moderating
variable strategy may appear to be an indirect one (after all, the selfmonitoring construct, concerned as it is with differences between individuals in the use of situational and dispositional guides for behavior, is a
rather indirect way to identify differing orientations to friendship), the
indirectness of the strategy may prove to be one source of its strength. For to
the extent that specific features of friendship can be placed in the larger
network of theoretical propositions provided by the psychological construct
as the moderating variable, a more elaborated and more psychologically
rich understanding of friendship may emerge.
Consider again the specific case of self-monitoring. One basis for
understanding the nature of and the purposes served by activity-oriented
and affect-oriented friendships is provided by the generalized interpersonal
orientations characteristic of high and low self-monitoring individuals. That
is, the propensity to contract either type of friendship may be one feature of
the larger set of differing cognitive, motivational, and behavioral processes
that guide the behavior of high and low self-monitoring individuals in
social contexts. Contracting activity-oriented friendships may be yet another
manifestation of a chronic striving to be a pragmatic creature of one's
situations, with activity-specific partners permitting the high self-monitoring individual to playa diversity to social roles and, quite literally, act like
different persons in different situations with different people. By the same
logic, contracting affect-based friendships may be yet another manifestation
of a continuing quest to be a principled being, with congruence between
one's actions and underlying attitudes, values, and dispositions being
enhanced by associating with similar others who will provide opportunities
and support for the expression of one's attitudes, values, and dispositions
across diverse interaction contexts.
Furthermore, the availability of a larger theoretical context may prove
useful in pointing the way toward new, and as yet unexplored, territory for
friendship researchers. Consider once again the case of self-monitoring and
these fundamental features of friendships and their evolution: initiation, the
different considerations used by persons high and low in self-monitoring
when determining whether any particular individual exhibits the potential
76
for begining a friendship; maintenance, the distinguishing features of their

ongoing friendships; and dissolution, the implications of adopting one or
the other of these orientations for reactions to friendship termination.
To begin with, the initiation stage of interpersonal relationships assumes
special importance because the initial conditions under which a relationship is formed affect its nature and subsequent development. In the case of
low self-monitoring individuals, it seems that initiations derive their
importance from the potentially vast impact friendships have on their lives.
Because persons low in self-monitoring tend to invest a great deal of time
and emotion in individual friendships, there may be times when they are
simply "closed" to new friendships. They have allocated all that they can to
their present friendships and are not available to others. High selfmonitOring individuals may never be completely "closed" to new friendships. Because they are capable of friendships based strictly on activity
terms, it is not too difficult to contract new ones on very limited bases. Thus,
the initiation of friendships may look very different for high and low selfmonitoring individuals. In the case of high self-monitoring individuals, any
activity may serve as a sufficient and necessary ground for initiating a
friendship. For low self-monitoring individuals, many interactions spanning many domains and exhibiting certain characteristics deemed important to the individual involved may be necessary to constitute an initiation
of a friendship.
Once initiated, these two types of friendships may also require two
different prescriptions for their maintenance. For high self-monitoring
individuals, it seems that as long as a friend proves to be a "specialist" at an
important activity, the relationship will continue, even though liking and
likability may fluctuate. Furthermore, the domains of their friendships may
be more predictable from the initial activity bases on which the friendships
were first contracted. For example, the friendships of high self-monitoring
individuals that begin with shared skill in athletics will probably continue
to be based on athletic pursuits. This activity-focused orientation to
friendship may at times make it difficult for individuals to learn that they
share skill and expertise in activity domains other than the original ones,
thus perpetuating their particularly activity-oriented approach to friendship.
For low self-monitoring individuals, by contrast, the maintenance of the
friendship would seem to depend primarily upon likability. For them, a
friendship ought to be maintained to the extent that the partner continues to
exhibit likable qualities, even if there are considerable gaps in the partner's
repertoire of activity skills, Thus, whereas high self-monitoring individuals
maintain a relationship to the extent that they still enjoy doing the activity
and the friend involved is still a "specialist," low self-monitoring individuals
will maintain the relationship to the extent that the friend remains wellliked, regardless of either partner's change in abilities or interests. Moreover,
low self-monitoring individuals might be expected to "branch out" into new
77
activity domains when their friends develop new activity interests. In these
cases, their interest in and liking for the activity may be secondary to their
interest in and their liking for their friends. Thus, although the activities
characteristic of low self-monitoring individuals' friendships may exhibit
considerable variability over time, the affective investment ought to be
relatively stable.
The dissolution of the friendships of high and low self-monitoring
individuals can be viewed readily in terms of the same issues that seem to be
important at the stages of initiation and maintenance. Here again, two
separate sets of considerations concerning the relationship seem particularly relevant. One could expect a relationship for high self-monitoring
individuals to dissolve when either the particular skill of the partner or the
importance of the activity were diminished. Conversely, for low selfmonitoring individuals, the dissolution would appear to hinge on perceived
likability. If, for whatever reason, the partner's likability diminished, the
relationship would dissolve.
In addition to these differences in the circumstances that prompt
dissolution of friendships, there also may be differences between the
friendships of high and low self-monitoring individuals in terms of their
actual vulnerability to the threat of dissolution. Consider what would
happen if two friends had to limit their interactions (for logistical or other
reasons), changed their activity preferences (because of, e.g., maturation), or
otherwise experienced a change in the terms of their relationship. Our
analysis suggests that different individuals may experience different
reactions to these disruptive events. Specifically, the friendships of high selfmonitoring individuals would either have to change radically to fit new
circumstances or else be terminated. The friendships oflow self-monitoring
individuals, which are characterized by less domain-specific interaction,
would be less affected by new circumstances because their affective basis
ought to transcend circumstances or activity.
Furthermore, when such circumstances actually do result in the dissolution of friendships, high and low self-monitoring individuals may
construe the reaSOnS for friendship dissolution in rather different terms.
Thus, persons low in self-monitoring might offer particularly abstract,
perhaps philosophic, and all-encompassing reasons (e.g., "It just wasn't
meant to be"), and persons high in self-monitoring might offer more specific
and concrete reaSOnS (e.g., "I just got tired of doing the same old thing").
Each case reflects the differing nature of each relationship's respective
vulnerability.
Commensurate with these differences in vulnerability to dissolution, we
suspect differences in the emotional consequences of dissolution. Those
persons who adopt an affective/emotional orientation toward their friendships should be considerably more distressed emotionally upon the dissolution of a friendship than those who adopt an activity/instrumental
orientation. Moreover, the disruptive impact of the loss of a friendship that
78
has spanned many domains ought to be greater than that of the loss of a
friendship that has been confined to one, or at most a few, domains. Thus,
we would expect that low self-monitoring individuals would experience
more emotional distress upon the termination of a friendship than would
high self-monitoring individuals.
Moreover, because the friendships of low self-monitoring individuals
affect so many different areas of their lives, we would expect that these
persons wpuld be constantly reminded of their loss as they encounter these
domains subsequent to the dissolution. But because the friendships of high
self-monitoring individuals affect relatively few domains of their lives, we
would expect that these persons would be able to avoid reminders of their
terminated friendships. For these reasons, it would probably be somewhat
easier and quicker to find replacement partners for high self-monitoring
individuals (who need only find partners to serve limited activity-specific
purposes) than for low self-monitoring individuals (who need to find
partners who can function across diverse domains). Mter all, it is somewhat
easier to find a new racquetball partner than it is to find a new soul
mate.
The Paradox of Friendship

Finally, we close by mentioning what we call "the self-monitoring paradox,"
a notion that has been particularly intriguing as we have contemplated the
friendships of high and low self-monitoring individuals. Although high se1fmonitoring individuals can be thought of as being particularly oriented
toward others (witness their pronounced tendency to look to the behavior of
others to provide them with cues to the situational appropriateness of their
own behavior), within their friendship worlds they seem to exhibit no such
investment in others. Rather than reaching out to others, they seem to build
barriers by segmenting and compartmentalizing their friendships. Similarly, low self-monitoring individuals, whose general behavioral orientation
typically does not involve turning to others to judge their own behavior but
instead is one of looking inward to their own attitudes and dispositions to
guide their behavior, practice a friendship orientation characterized by an
extreme investment, both of time and emotion, in other people. At first
glance, their other-invested orientation to friendship seems to betray their
characteristic dispositonally guided behavioral otj.entation.
As is the case with so many "paradoxes," the self-monitoring paradox may
possess more in the way of appearance than reality. That high selfmonitoring individuals, who are otherwise so "social" in their orientation,
avoid forming close social bonds of friendship may reflect the fact that their
general attentiveness to others and their compartmentalization of friendships both may be operational in the service of the same goal-that of being
able, as they indicate in their endorsement of items on the Self-Monitoring
79
Scale, to act like different persons in different situations and with different
people. Similarly, that low self-monitoring individuals, who are generally so
attuned to themselves in guiding their behavior, can become so invested, in
others in the domain of friendship may also reflect the fact that both a
dispositionally guided behavioral orientation and a formation of close
attachments to friends may serve the same goal-that of maximizing their
opportunities to "be themselves" by displaying their own attitudes, traits,
and dispositions.
That the self-monitoring paradox may be more apparent than real does
not mean that it may not cause tensions and pose conflicts for individuals.
High self-monitoring individuals may experience the approach-avoidance
conflict of being drawn to others as sources of social comparison and
behavioral guidance but experience anxiety at the prospect of being drawn
too close to other people. And low self-monitoring individuals, individuals
who form close bonds of friendship in order to spend time with similar
others, may find themselves tom between their own inclinations and those
of their partners in areas where their attitudes, values, traits, and preferences
diverge. The self-monitoring paradox, although capable of a harmonious
resolution in the minds of those of us who think about it, may nonetheless
be a source of genuine concern in the lives of those who must cope with
it.
Acknowledgments. This research and the preparation of this chapter were supported
in part by Grant No. BNS 82-07632 from the National Science Foundation and by a
grant from the Graduate School of the University of Minnesota. We are grateful to
Beatrice Ellis, Peter Glick, Alana Matwychuk, Allen Omoto, Jeffry A Simpson, and
Tom Smith for their helpful comments on this chapter.
References
Gangestad, S., & Snyder, M. (1985). To carve nature at its joints: On the existence of
discrete classes in personality. Psychological Review, 92,317-349.
Ickes, W., & Barnes, RD. (1977). The role of sex and self-monitoring in unstructured
dyadic interactions. Journal of Personality and Social Psychology, 35,315-330.
Jamieson, D.W., Lydon, J.E., & Zanna, M.P. (1984, May). Similarity of attitudes versus
activity preferences: A differential basis of interpersonal attraction for low and high selfmonitors? Paper presented at the annual meeting of the Canadian Psychological
Association, Ottawa.
Shaw, M.E., & Costanzo, P.R. (1982). Theories of social psychology. New York:
McGraw-Hill.
Snyder, M. (1974). The self-monitoring of expressive behavior. Journal of Personality
and Social Psychology, 30,526-537.
Snyder, M. (1979). Self-monitoring processes. In L. Berkowitz (Ed.), Advances in
experimental social psychology (Vol. 12, pp. 85-128). New York: Academic.
Snyder, M., & Campbell, B.H. (1982). Self-monitoring: The self in action. In J. Suls
(Ed.), Psychological perspectives on the self (Vol. 1, pp. 185-208). Hillsdale, NJ:
Erlbaum.
80
Snyder, M., Gangestad, S., & Simpson, J.A. (1983). Choosing friends as activity
partners: The role of self-monitoring. Journal of Personality and Social Psychology,
45, 1061-1072.
Snyder, M., & Ickes, W. (1985). Personality and social behavior. In G. Lindzey & E.
Aronson (Eds.) Handbook of social psychology (3rd ed., Vol. 2, pp. 883-948). New
York: Random-House.
Snyder, M., Simpson, lA, & Smith, D. (1984, May). Personality and friendship: The
role of self-monitoring in choosing close and casual friends. Paper presented at the
annual meeting of the Midwestern Psychological Association, Chicago.
Snyder, M., & Smith, D. (1984, May). Self-monitoring and conceptions of friendship.
Paper presented at the annual meeting of the Midwestern Psychological Association, Chicago.
Chapter 5
Sex Differences in Same-Sex Friendships

Barbara A. Winstead
The essential difference between male and female same-sex friendships, in

words Paul Wright (1982) has used to describe it, is that female friendships
are "face to face" whereas male friendships are "side by side." These phrases
capture the frequently replicated finding that female friends like to "just
talk" and view this activity as central to their friendship. Females compared
to males also describe their talk as more intimate and more self-disclosing.
Male friends, on the other hand, prefer to do things together other than "just
talking." They share activities, such as sports, where their attention is
focused on the same goals but not on one another.
Having begun with some conclusions of studies on sex differences in
same-sex friendships, I continue in this chapter by describing more
thoroughly the results of studies of adult same-sex friendships. In an effort
to shed light on sources of these sex differences in friendship, I explore
research and theory on the development of sex differences in interpersonal
relationships and on intra sex differences in same-sex friendships. I also
consider how we can understand sex as a source of differences in an
individual's interpersonal behavior and as a stimulus that produces
differences in reactions to the individual that can, in tum, create counterreactions. Finally, I comment on values in studying friendships, methodology, and the meaning of same-sex friendships.
Sex Differences in Adult Same-Sex Friendships

From the initiation of a friendship with a same-sex other, females and males
behave differently. Since Newcomb's (1961) classic research on friendship
development, we have believed that similarity leads to attraction and the
establishment of friendship. However, the sort of similarity that is important
to females and males is apparently different. Weiss and Lowenthal (1975)
found that the factors most central to female friendships are mutual help
82
Barbara A Winstead
and support whereas males emphasize similar interests and shared

experiences. Wright and Crawford (1971) found that ratings of friendship
correlated with similarity of abstract values for females (measured by the
Allport, Vernon, and Lindzey, 1960, Scale of Values and the Rokeach, 1968,
index of "terminal values") but with similarity of preferred activities for
males (such as "watch television news programs," "write short stories").
Looking prospectively rather than concurrently, Hill and Stull (1981) found
that perceived similarity predicted which roommates would remain together
from fall to spring term for both males and females, but actual value
similarity predicted staying together only for female roommates. Thus, for
females similarity in values is an important prerequisite to friendship,
whereas males are more concerned about similarity in interests.
The type of similarity that women and men consider in choosing friends
appears to be quite rational given what we know about what female and
male same-sex friends do together. In an investigation of implicit theories of
relationships, Rands and Levinger (1979) demonstrated that female samesex friends compared to male friends are expected to be more selfdisclosing, other-enhancing, and physically affectionate but not different in
terms of likelihood of engaging in joint activities. Caldwell and Peplau
(1982) in a questionnaire study found that women like to "just talk" with
friends whereas men prefer engaging in some activity. In a diary study
(Wheeler & Nezlek, 1977) in which subjects kept a daily record of their
interactions with friends and others, females were significantly more likely
than males to study together and to talk with one another about friends,
family, and personal problems. Males were more likely than females (but
not significantly) to go to movies and play sports together.
Even when only friends' conversations are considered, sex differences of a
similar sort are found. Caldwell and Peplau (1982) reported that females are
significantly more likely than males to talk with same-sex best friends about
feelings and problems, whereas there was a trend for males to report talking
more often to same-sex best friends about activities. Similarly, in another
questionnaire study with college students (Johnson & Aries, 1983), females
conversed significantly more frequently and in greater depth about topics
involving themselves and their close relationships, whereas males conversed
more frequently and in greater depth about activity-oriented topics (e.g.,
sports, hobbies, and reminiscences about the past). In an older adult
sample, Aries and Johnson (1983) found that females exceeded males in the
frequency and depth of talking about personal problems and family
activities, whereas males exceeded females in frequency and depth of
talking only about sports. Similarly, in a sample of young, single adults,
Davidson and Duberman (1982) found that when describing their conversations with same-sex friends, females compared with males gave almost
twice as many personal and three times as many relational accounts. Sex
differences in topical items of conversation were negligible. If male friends
do things with one another, then it is sensible for them to be concerned at
83
the beginning of the friendship about similarity in activity preference;

otherwise, there may be conflict in choosing what to do together. On the
other hand, if female friends talk with one another more and talk especially
about personal topics, then shared values may make it more likely that they
will find themselves able to validate one another's views.
Females are also more willing than males to express feelings for their
same-sex friends and about their friendships. In doing intimacy scaling of
items for his Friendship Observation Checklist, Hays (1984) found that
males endorsed very few affection items, that is "expressing any sentiment
[positive or negative] towards the other, any expression of the emotional
bond between partners" (p. 78), for casual friends, whereas females were
more willing to do so. When data were collected using this instrument,
females reported engaging in more affection and casual communication
behaviors with their same-sex friends than did males. Females also express
more love (using the Liking and Loving Scales) for their same-sex friends
than males do (Rubin, 1970; Small, Gross, Erdwins, & Gessner, 1979).
Finally, nonverbal expressions of affection are more frequent between
female friends. Rands and Levinger (1979) found that expression of physical
affection is considered more characteristic of female same-sex relationships
than of male same-sex relationships. In a behavioral study pairs of female
and male same-sex friends and opposite-sex friends were instructed to
photograph themselves (a) "however you like," (b) as if"you are greeting one
another at the airport," and (c) "seated, having a conversation." Using the
Heslin and Boss (1980) ratings for level of touch intimacy, Costanza (1984)
found that male dyads used significantly less intimate touch than did
female or mixed-sex dyads across conditions. In the seated condition, males
sat further apart than female or mixed-sex pairs did. Female same-sex
friends and opposite-sex friends were virtually identical on these measures.
There is some evidence that in stressful situations female same-sex friends
are likely to be more emotionally supportive than are male friends. In a
study where pairs of female and male undergraduate same-sex friends were
told that they would be handling a large snake and then left to wait with one
another for 3 to 4 minutes, females were more likely than males to report
that they had withheld nothing while waiting with their friend (Winstead &
Derlega, 1984). Interviews with parents of dying children and their friends
also revealed that males have difficulty sharing their feelings and asking for
help or providing emotional support in this situation. Male friends of
fathers of dying children, while speaking of their own discomfort with the
intense feelings aroused in such a situation, also expressed their expectation
that their male friends would feel uncomfortable talking about their
emotions (Chesler & Barbarin, 1985).
Given these findings it is not surprising that females describe their
interactions with same-sex friends as significantly more intimate than do
males (Wheeler, Reis, & Nezlek, 1983). Also, females are more likely than
84
Barbara A Winstead
males to describe their friendships or social networks in terms of helping,

sharing, and emotional support (Burda, Vaux, & Schill, 1984; Weiss &
Lowenthal, 1975). Because subjects in these studies are required to describe
their relationships along set dimensions, females and males may differ
because they understand the terms differently. In other words, perhaps
"intimacy" means one thing to females but something else altogether to
males. For example, Solano (1981) found sex differences in subjects' ratings
of the Taylor-Altman stimuli (Taylor & Altman, 1966) used to rate intimacy
in self-disclosure research. Females compared with males rated sexual
activity as significantly higher on intimacy, whereas males rated family
history, personal habits, opinions, feelings, and tastes significantly higher. It
is probable that subjects based these ratings on what they are willing to talk
about with friends and family. If women are, in fact, more willing to talk
about self, family, and personal relationships, but, perhaps, less willing to
talk about sex, then the ratings may reflect these behavioral differences. On
the other hand, Reis, Senchak, and Solomon (1985) had female and male
subjects rate interactions that were high, medium, or low in intimacy and
found no sex differences in rating of videotapes or written transcripts. This
suggests that females and males are generally using the vocabulary of
relationship descriptors in a similar way, at least when rating the behavior
of others.
When females and males are asked to describe what they are willing to
talk about to others, females are usually found to be more willing than
males to talk about intimate topics to same-sex friends (Bender, Davis,
Glover, & Stapp, 1976; Davidson & Duberman, 1982; Fischer & Narus, 1981;
Morgan, 1976; Stokes, Fuehrer, & Childs, 1980) and more likely than males
to confide in same-sex friends (Booth, 1972). Although in these studies the
intimacy level of the topic may be decided by the experimenter rather than
the subjects, it is the subjects' reports of their conversations and not actual
behavioral data that are being studied. Reis et al. (1985) asked their samesex friend dyads to describe their last meaningful conversation and what
they talked about on the way to the laboratory and found that in both cases
conversations reported by females were significantly more intimate than
were those reported by males. Having subjects report on a specific
conversation, especially a recent one, is an improvement over having
subjects indicate their willingness to talk about particular topics with samesex friends or having them give general descriptions of their conversations
with friends. It is still, however, not an objective observation of subjects'
behavior. When Reis et al. (1985) asked their subjects to engage in a
meaningful conversation in the laboratory, they found no sex difference in
ratings of the intimacy of those conversations (either by observers or the
subjects themselves). Similarly, Derlega, Winstead, and Wong (in press)
found no sex differences in intimacy of messages written to a same-sex
friend when the subjects were asked to produce a fairly intimate disclosure.
85
The finding that female same-sex friends report sharing more that is
personal and emotionally meaningful than do male same-sex friends is so
often replicated that I hesitate to question its validity. There is, however, very
little data based on the actual recorded conversations of same-sex friends. It
is possible that male and female same-sex friends talk with one another
about very similar topics and with similar degrees of intimacy but find
different aspects of their conversations to be salient and memorable. Or,
perhaps, males are more reluctant than females to report the personal and
intimate elements of their same-sex friendships to experimenters. It is also
possible that same-sex male friends can talk as intimately as female friends,
but choose not to. Reis et al. (1985) concluded that males have as great a
capacity as females for intimacy in same-sex friendships but are less likely
to use this capacity. In other words, males know how to have an intimate
conversation with their same-sex friends and do so on occasion but prefer
sharing activities that do not require emotional self-disclosure. Of course, a
sex difference in preference requires as much of an explanation as does a
sex difference in capacity.
A Developmental Perspective on Sex Differences in

Same-Sex Friendships
When children first begin to establish friendships with other children (at
about 2 years of age), they show little preference for the sex of their
playmate. As early as 3 or 4 years of age, however, preference for same-sex
others is notable (Rubin, 1980). The first sex difference in same-sex
friendship that emerges is in the number of children in a play group. Girls
are more likely to play with one or two other children, whereas boys tend to
play in groups. This sex difference has been observed in kindergartners, and
by the elementary-school years, it is well established (Eder & Hallinan, 1978;
Laosa & Brophy, 1972; Waldrop & Halverson, 1975). Waldrop and
Halverson found that the most social boys among their 7.5-year-olds were
those with "extensive" peer relationships (play in groups), whereas the most
social girls were those with "intensive" peer relationships (play with only
one other girl). Furthermore, being peer-oriented at 2.5 years predicted
"extensiveness" at 7.5 years for boys and "intensiveness" at 7.S years for girls,
suggesting that friendly 2-year-olds have by 7 years learned gender-related
scripts for friendship. In support of this finding, fifth- and sixth-grade girls
were found to have more exclusive dyadic friendships than did fifth- and
sixth-grade boys (Eder & Hallinan, 1978). Observations of friendship pairs
at six-week intervals during a schoolyear showed that pairs of girls were less
likely than pairs of boys to include a third child in an existing friendship
(Eder & Hallinan, 1978). This sex difference did not occur in an open
classroom. Eder and Hallinan (1978) suggested that self-selection (these
86
Barbara A Winstead
children may have received less sex-role socialization) or opportunities to

socialize outside of play periods (where sex differences in game preference
may influence size of friendship groups) may have prevented the sex
difference from occurring in this setting. Feshbach and Sones (1971) also
found that adolescent female friends are less willing than adolescent male
friends to share with a third person.
In addition to involving more players, boys' play is also more complex
than that of girls and is far more likely to be structured as a competitive
game (Lever, 1976, 1978). Girls are more likely to play tum-taking games in
which there is little role differentiation (e.g., rope jumper and rope turner)
and minimal competition. Lever (1976, 1978) argued that males learn from
their type of play about team membership, fair play, and following rules.
Indeed, in a laboratory situation, preadolescent male friends preferred an
equitable outcome over an equal one when one friend has contributed more
to a positive outcome, whereas female friends preferred an equal division of
outcomes despite unequal inputs (Benton, 1971). Equitable but not equal
distribution of rewards is certainly more likely to occur on teams or in group
play than in one-to-one or tum-taking play.
The girl's preference for equal benefits may also be related to their beliefs
and feelings about friendships and their friendship behavior. Fourth- and
sixth-grade girls received significantly higher scores than did boys on
friendship motivation, measured by themes in stories told about pictures
(McAdams & Losoff, 1984). The girls also had responses to more factual
questions about their best friends than boys did; but a measure of depth of
friendship yielded no sex differences (McAdams & Losoff, 1984). Sharabany, Gershoni, and Hoffman (1981) found higher ratings for attachment,
giving and sharing, and trust and loyalty of same-sex friendships by female
than by male 5th, 7th, and 11th graders. Similarly, Burmester and Furman
(see Chapter 3) found that females in the 5th and 8th grades scored
significantly higher than males on measures of talking, sharing private
thoughts and feelings, and telling secrets. Sex differences in a 2nd grade
sample were not significant, suggesting that differences between female and
male same-sex friendships increase throughout childhood. By adolescence
the sex differences are clear: Females tend to have one-to-one same-sex
relationships that are intense, exclusive, and emotionally close, and males
tend to interact in groups and to have companionable same-sex relationships that are based on having similar interests and sharing activities
(Douvan & Adelson, 1966; Richey & Richey, 1980).
The results of developmental studies of same-sex friendships suggest that
the beginnings of the activity- or goal-orientation of male friendships and
the person-orientation of female friendships are found early in the life cycle.
Chodorow (1978) argued that sex differences in interpersonal relationships
are created in the mother-infant interaction where there is a reciprocal
identification and attachment between mothers and daughters that is not
87
experienced or less intensely experienced by boys. Thus, females emerge

from childhood with a greater capacity for empathy and a different way of
experiencing relationships. Gilligan (1982), based on her work on female
moral development, also contended that the processes of female development and male development are not comparable and that the use of male
models of psychological development to describe and understand females is
inappropriate. She observed females as young as 8 years of age using the
context of relationships in their moral reasoning, whereas males used
hierarchy and, later, rules and principles in their moral reasoning. Both
Chodorow and Gilligan argued that there are substantial differences in the
way males and females think and feel about themselves, others, and
relationships. In this context, sex differences in same-sex friendships are
merely a reflection of these differences.
Most research on same-sex friendship has been done with the preferred
social research animal, the college undergraduate. Do sex differences in
same-sex friendships vary over the course of the adult life cycle? The finding
that females have richer and more intimate same-sex friendships than
males is repeated at all ages, even among the elderly (Booth, 1972; Powers &
Bultena, 1976). In fact, there is some suggestion that males may become
more inhibited in their same-sex friendships as they grow older.
Morgan (1976), using the Jourard Self-Disclosure Questionnaire, which is
composed of ratings of self-disclosure to parents and opposite-sex friends as
well as same-sex friends, found trends for negative relationships for his
male subjects between age and self-disclosure on intimate and nonintimate
topics. (For females there was a significant positive correlation between age
and self-disclosure on nonintimate topics.) In studies of the conversations of
same-sex friends, Aries and Johnson questioned undergraduates and
parents of undergraduates about frequency and depth of discussion of
certain topics (Aries & Johnson, 1983; Johnson & Aries, 1983). Although
they did not directly compare the data from the younger and older samples,
differences in the results of the two studies are intriguing. In the college
student sample, females compared to males reported more frequent and/or
in-depth conversations about several personal topics, such as family
problems, personal problems, doubts and fears, whereas males reported
more frequent and/or in-depth conversations about sports, hobbies and
shared activities, and reminiscences about things done together in the past
(Johnson & Aries, 1983). In the older adult sample, females were again
found to exceed males in the frequency and depth of their conversations to
same-sex friends about personal topics, but in this sample males reported
more frequent and/or in-depth conversation about only sports and work. In
fact, in the older sample females reported more frequent or in-depth
conversations on the topics of hobbies and shared activities and reminiscences than did males.
These findings are suggestive. Other investigators, however, have found
88
Barbara A Winstead
very small or no differences between younger and older males in selfdisclosure to same-sex friends (Hacker, 1981; Fischer & Narus, 1981). As
young adults leave the school environment and enter the work world where
friendships are more likely to have an instrumental rather than a communal
function, there may be a decline in the intimacy of same-sex friendships.
This may not affect women in homemaking/motherhood careers, or
perhaps women are simply more inclined to develop and maintain intimacy
in same-sex friendships whether they are established in the workplace or
elsewhere. Sex differences in functions of same-sex friendships in stages of
adulthood development other than the college years need further exploration.
Intrasex Differences in Same-Sex Friendships

Given substantial sex differences in same-sex friendships, perhaps the most
obvious personality trait with potential for explaining individual differences
in same-sex friendships is sex-role orientation. The association between sexrole orientation and self-disclosure has been investigated in several studies,
with mixed results. Fischer and Narus (1981), using the Bern Sex Role
Inventory and a measure of intimacy in relation to a target person defined
as the person closest to the subject excluding relatives, found that among
women choosing women as their closest other, androgynous women had
significantly higher intimacy scores than did either sex-typed or undifferentiated women. Even so, sex-typed men with same-sex close friends
had significantly lower intimacy scores than did sex-typed women with
same-sex close friends. Using the same male subjects, Narus and Fischer
(1982) examined the relationship between sex role and ease of communication and confidence sharing. For males in same-sex relationships,
they found masculinity to be significantly related to both of these measures
of expressivity and femininity to be significantly related to confidence
sharing only. Using analysis of variance, they found androgynous men to be
higher than any other group in this male sample for both ease of
communication and confidence sharing.
In a study examining the relationship between Bern Sex Role Inventory
and the Jourard self-disclosure questionnaire, Lombardo and Lavine (1981)
found that androgynous subjects, both male and female, reported disclosing
more to their same-sex best friends than did sex-typed subjects (i.e.,
masculine males and feminine females). These differences were much more
pronounced for disclosure of intimate rather than nonintimate topics. On
the other hand, Small et al. (1979), using the Bern Sex Role Inventory and a
self-disclosure measure on which items are scaled for intimacy (Chaikin &
Derlega, 1974), found that masculine males reported sharing more than did
89
other males with same-sex friends and that sex role made very little
difference in self-disclosure in the female sample.
In a study of female and male hetereosexual and homosexual college
students, Bender et aI., (1976) found that femininity as measured by both the
Bem Sex Role Inventory and the Personal Attributes Questionnaire was
related to disclosure for same-sex friends. Furthermore, the highest total
self-disclosure score on the 10urard measure for all subjects was for
feminine sex-typed (high feminine, low masculine) subjects. Unfortunately,
these analyses were not reported for the subsamples in the study. Thus, for
males masculinity, femininity, and androgyny have been found to be related
to self-disclosure to same-sex friends, whereas androgyny and femininity
predict self-disclosure between female friends (when results are significant).
Winstead, Deriega, and Wong (1984) have questioned the results of sex
role and self-disclosure research when paper-and-pencil rather than
behavioral measures of self-disclosure are used. Subjects may say that they
would disclose with others about topics when in the actual situation they
would choose to talk about other, perhaps more superficial, things.
When friendship variables other than self-disclosure are examined,
femininity and androgyny emerge as the more important sex-role orientations. Small et ai. (1979) found main effects for sex-role orientation on
liking and loving scores for same-sex friends. Feminine subjects liked their
friends most; and androgynous subjects loved them most. There were no
interactions with sex of subject. Using a diary technique, Wheeler et ai.
(1983) confirmed the importance of femininity, espeically for male subjects.
For males femininity, as measured by the Personal Attributes Questionnaire, was significantly positively related to meaningfulness of social
interactions with both males and females. For females femininity had
positive, but not significant, correlations with meaningfulness and a
significant positive correlation with time spent with females. These studies
indicate that possessing attributes characteristic of the female stereotype
(e.g., nurturance, emotionality, expressivity) either alone or in combination
with masculine attributes is predictive of self-disclosure and meaningfulness in same-sex relationships above and beyond the predictive power of
sex per se.
Several authors writing about male same-sex friendships have argued that
one factor inhibiting males from forming close relationships with other
males is homophobia, that is, the fear of homosexuality (Lewis, 1978;
Nichols, 1975). Preferring the company of another man for "just talking"
and sharing one's thoughts and feelings may suggest a closeness to that
friend that has connotations of sexual intimacy rather than just emotional
intimacy. Males are, in fact, more reluctant than are females to express
feelings about their friend or the relationship to their same-sex friends (or at
least to report doing so to experimenters; Hays, 1984, Rands & Levinger,
90
Barbara A Winstead
1979) or to express affection to same-sex friends nonverbally (Costanza,

1984). Males also have a significantly higher degree of homophobia than
females do (Levitt & Klassen, 1974). A direct link between homophobia and
avoidance of same-sex friendship has been less well-documented. Using an
antihomosexual scale, Stokes et ai. (1980) found a negative correlation
between antihomosexual scores and self-disclosure to same-sex intimates in
their sample of female and male subjects; but they found an equally large
negative correlation between antihomosexual scores and self-disclosure to
opposite-sex intimates. They did not report correlations separately for
females and males, although one might predict a stronger correlation for
males. Homophobia is a variable worth exploring as a predictor of intra sex
variability in same-sex friendships.
Other personality variables with relevance for the study of friendship are
the intimacy and power motives (McAdams, 1982a, 1982b; McAdams &
Constantian, 1983; McAdams, Healy, & Krause, 1984). The intimacy motive
is a desire to be close to and in communication with others; the power
motive is a desire to have an impact on others and to acquire social power.
These motives correspond generally to Bakan's (1966) concepts of communion and agency. Although this dichotomy has often been used to
describe stereotypes of female and male, respectively, there may be no sex
differences in either of these motives as measured by Thematic Apperception Test stories in the undergraduate population (McAdams et aI.,
1984).
McAdams and colleagues have also found no sex differences in intimacy
motivation in several other undergraduate samples (McAdams, 1982a;
McAdams & Constantian, 1983). Females did, however, receive significantly
higher scores than males on intimacy motivation in samples drawn from
two more-traditional (church-affiliated) college campuses (McAdams,
1982b), suggesting that sex-role attitudes and socialization may affect this
personality variable. (See McAdams and Losoff, 1984, for a study in which
they found higher friendship motivation for female fifth and eighth
graders.)
Although the intimacy motive rarely shows sex differences in undergraduate samples, related behaviors sometimes do. In a study in which
subjects reported their thoughts when paged at various times during the day,
females reported nearly twice as high a mean percentage of interpersonal
thoughts as men did (McAdams & Constantian, 1983). Intimacy motivation
was highly correlated with percentage of interpersonal thoughts for females
but not for males. Intimacy motivation was also related to relationship
behaviors (conversationlIetter-writing) and positive affect in interpersonal
situations for both males and females; but there were no sex differences on
these measures.
In a study of power and intimacy motivation and patterns of friendship,
McAdams et al. (1982) found that intimacy motivation was positively related
to listening and self-disclosure in friendship episodes and to trust/concern
91
as major reasons for friendships for both females and males. For females
there was also a negative relationship between intimacy motivation and the
percentage of friendship episodes reported by subjects depicting groups.
Power motivation was positively related to "agentic striving," defined by
McAdams et al. (1982) as "an active, asserting, controlling stance," (p. 832),
in the friendship episodes for both females and males and for males only
negatively related to percentage of friendship episodes depicting dyads and
positively related to percentage depicting groups. The latter finding suggests
that power motivation is a deterrent to one-to-one relationships for males
but not for females. (Stewart and Chester, 1982, found a similar sex
difference for marital and dating relationships.) McAdams et al. found sex
differences only for agentic striving in friendship episodes (males greater
than females). The authors argued that the intimacy and power motives
clearly account for more variance in relationship behavior than does sex.
Unfortunately, friendships in their study could be sibling relationships,
same- or opposite-sex platonic relationships, or romantic relationships.
Only nonpeer relationships were excluded. Other studies have also failed to
distinguish between behaviors or feelings related to same-sex friendships
and those related to other relationships. The impact of intimacy and power
motivation may well depend on the type of relationship; and same-sex
friendships may have properties not found in other relationships.
Understanding Sex Differences in Interpersonal Behavior

Do females and males differ in social behavior in ways that would predict
sex differences in same-sex friendship? Deaux (1984) pointed out that
although sex as a variable generally accounts for little variance in attitudes
or behaviors, sex differences may occur more often for interpersonal than
for individualistic behaviors. Taking a sociobiological point of view, Hinde
(1984) argued that dyadic relationships may be of more importance to
females than to males. In heterosexual relationships the female's reproductive success benefits from having a committed mate, whereas males
can maximize their reproductive success by competing with other males and
winning access to several females. (The more the better for him, unless the
females he impregnates will not or cannot adequately care for his offspring
while he is off impregnating others.) Sex differences in females' interest in
pair bonding may influence sex differences in same-sex relationships as
well (Hinde, 1984). In this context it is interesting to note that Banikotes,
Neimeyer, and Lepkowsky (1981) found less differentiation between ideal
same- and other-sex friends for females than for males. Establishing and
maintaining a close relationship is a strategy that can work for both
heterosexual and same-sex relationships. If males, however, need to
compete with other males on the one hand, but establish close, if brief,
92
Barbara A Winstead
relationships with females on the other, then they need different strategies
for the two sexes.
In using a biological approach to understand sex differences in interpersonal behavior, Hinde (1984) made it clear that sex differences in
behavioral propensities that may result from natural selection interact with
sociocultural influences. He further stated that knowledge of biological
differences should be used to achieve social goals not to set them.
From a more psychological point of view, Deaux (1977) proposed that
many sex differences can be understood in terms of differences in selfpresentation strategies used by females and males. Females, she argued,
tend to use an affiliative or status-neutralizing style that aims to reduce
distance and encourage emotional closeness with others. Status differences
between individuals are minimized. Females' preference for equality over
equity (Benton, 1971) supports this claim. Males, on the other hand, tend to
use a status-assertive style that maintains distance between individuals and
emphasizes status differences by highlighting successes and hiding weaknesses. The findings on sex differences in same-sex friendship are clearly
compatible with Deaux's conceptualization.
According to Deaux (1977), the sex differences in style of self-presentation
are differences in choices not in capabilities. Reis et al. (1985) have drawn a
similar conclusion from their research on intimacy in same-sex friendships.
They argued that males are capable of talking as intimately as females with
their friends, but they generally choose not to. "Choice" in this context does
not mean a rational, thoughtful, or even conscious decision, but may be
an almost automatic reaction or a well-established behavior. The distinction
between sex differences in choices and sex differences in capabilities is an
important one. If females and males have different capacities, then we
would expect little change in sex differences regardless of social and
situational variation. If the sexes differ in their choices, however, then
changes in social stimuli will make a difference. For example, males who
find themselves in a setting that encourages emotional self-disclosure
between friends may form emotionally intimate bonds that equal those of
female friends. These males will be choosing to behave in ways different
from those typical for males in our culture, even if they are not aware of this
"choice" or of the influence of the setting on their behavior (Nisbett &
Wilson, 1977).
Sex differences in patterns of friendship occur at a relatively early age,
suggesting that encouragement of female and male styles of interaction
begins with the first glimmerings of social behavior. Sex differences in
structural variables, such as size of play group and type of games played
(Eder & Hallinan, 1978; Laosa & Brophy, 1972; Lever, 1976, 1978; Waldrop
& Halverson, 1975), may be particularly important in establishing sex
differences in interactional choices. Group play and complex games (the
male pattern), for example, discourage intimacy-promoting behavior. That
degree of sex-role socialization may influence interpersonal style and same-
93
sex friendship is suggested by the observations of no significant sex

differences in exclusivity of dyads in open classrooms (Eder & Hallinan,
1978) or in intimacy motivation on more liberal college campuses
(McAdams & Constantian, 1983).
Sex as a Contextual Variable

A dyad is more than the sum of the behaviors of two individuals. Each
member is actor, observer, and stimulus person. When sex differences are
found in same-sex friendships, explanations should be sought not just in sex
differences in friends' behaviors but also in responses to male and female
others. Sex operates not just as a personality variable, but also as a social
category to which others have predictable responses. Research shows, in
fact, that female others elicit more self-disclosure than do male others
(Dedega, Winstead, Wong, & Hunter, 1985). Females also receive higher
scores on the Opener's Scale, designed and validated to measure the
capacity to get others to "open up" or disclose personal information (Miller,
Berg, & Archer, 1983). Finally, spending time with females is negatively
related to loneliness for both female and male college students, whereas
spending time with males is unrelated to loneliness (Reis, 1986). In other
words, the presence of a female has a powerful effect on the social behavior
of another; it makes him or her more self-disclosing, more open, and less
lonely.
What is the source of this stimulus effect? Individuals enter relationships
with female or male others with certain expectations about their "nature." In
general, gender stereotypes predict that in relationships females will be
communal and nurturant and males will be agentic and instrumental. In
some cases, these stereotypes will receive behavioral confirmation (e.g., a
female will be an "opener," a male will not), but there need not be
confirmation in order for the stereotype to be maintained; in fact, behaviors
may be misperceived or misinterpreted to make them fit the stereotype when
they do not. A female, therefore, is expected to be relatively interested in our
personal thoughts and feelings. She is expected to be less critical than a
male, and, in general, a safer recipient of personal information because she
is less likely to regard it as material that can be used to manipulate others.
Both females and males are likely to operate according to these stereotypes
when interacting with a female. Thus, the role that females and males take
while interacting with a female will be somewhat different from the role they
take while interacting with a male. In other words, self-presentation depends
on our assumptions (our stereotypes) about others.
Furthermore, given these expectations of a female, if a person is more
likely to self-disclose to her, then she, by the norm of reciprocity, is
somewhat obligated to disclose to him or her, making her what she was
expected to be, emotionally self-disclosing. Her behavior, however, will not
94
Barbara A Winstead
be completely independent of the sex of her partner and her interests in the
relationship. According to the previous argument, she should, like others do,
regard a female partner as a better candidate for disclosure of her personal
thoughts and feelings. If this is true, two females will have a far greater
chance of becoming intimate friends than will two males. This analysis is
also consistent with what is seen in male-male friendships: a lack of
emotional intimacy. In an interaction between two males, neither would
expect responsiveness to expressions of feelings, neither would feel encouraged to open up. In fact, males tend to expect other males to feel
uncomfortable telling them about their feelings (cf. Chesler & Barbarin,
1985). In some situations, efforts by one male to be closer to another may
even be viewed as homosexual, or a male may fear that his behavior will be
so viewed.
A relationship between female and male should be less intimate than a
female-female relationship and more intimate than a male-male relationship. Perhaps, not surprisingly, research does not indicate that opposite-sex
relationships are less emotionally intimate than female same-sex friendships. Booth and Hess (1974), for example, found that single females
confided more in cross-sex friends (who may have been platonic or
romantic) than in same-sex friends. (There are, however, females who say
that they share more personal information with female friends than with
their spouses.) In the realm of heterosexual relationships we have other
factors to consider, for example, sexual intimacy and the emphasis our
culture places on one monogamous heterosexual love relationship. Nevertheless, a female ought to have different motives and different expectations
when pursuing a same-sex friendship vs. a heterosexual relationship. With
another female she might expect openness, interest, and mutual selfdisclosure. If she finds her values and concerns appreciated and perhaps
shared by this woman, a close friendship that is supportive and comforting
may develop. With a man a female may hope for passion and romance and
eventually intimacy, but she may feel greater apprehension about her
partner's ability or willingness to be open. She may settle for other benefits
from the relationship, such as pleasurable activities, security, sexual
satisfaction, or the opportunity to have a family, rather than insisting on
emotional intimacy. The motivations, expectations, and experiences that
differentiate these relationships need to be investigated.
A more illustrative comparison for our purposes would be between
platonic same-sex friendships and platonic cross-sex friendships. According
to the argument presented here, close platonic cross-sex friendships ought to
be less frequent and less emotionally intimate than are close female samesex friendships but more frequent and more intimate than are male samesex friendships. A study of cross-sex friendships (Booth & Hess, 1974)
revealed that their frequency and intimacy are a function of sex and marital
status. All subjects reported having more same- than cross-sex friends,
although males tended to report more cross-sex friends than females did.
95
Among unmarried subjects males reported interacting more often with

female than with male friends; and both males and females confided more
in cross-sex than in same-sex friends. (There was no instruction to prohibit
reporting on romantic, dating relationships. Thus, the findings for unmarried subjects in particular may not represent platonic cross-sex friendships.) Married subjects reported interacting more with and confiding more
in same-sex friendships. Sex-segregated work and leisure activities are
common in our culture and same-sex friendships are also the norm. The
fact that males seem to rely on cross-sex friendships more than females do
is, however, consistent with the arguments presented previously.
Clearly, what we experience in dyadic interaction depends on our sex and
the sex of the other. A dyad is an interactive process where, for example, a
female's willingness to share and to be responsive and her expectation of
responsiveness from another female creates an initial drawing together. If
these women find that their values match, then a friendship is likely to
emerge. In general, males do not approach other males with either a
willingness to share nor an expectation that the other is interested in his
sharing. If this is true, then females should become friends more quickly
than do males and meet more potential friends (although time and
circumstances may limit the total number of friends that they have). These
ideas are clearly applicable in the initiation of a relationship, but they also
apply to later phases of the relationship. Any effort to deepen a relationship
will depend on one's beliefs about the other person. With time, the other's
unique characteristics are better known, but their gender continues to playa
role in terms of ongoing stereotyping and in terms of the effects of initial
stereotyping on the shaping of the image of the other.
A relationship cannot be neatly divided into behaviors resulting from
individual motives and dispositions and behaviors that are reactions to the
stimulus value of the other person. Still, this latter component of dyadic
interaction should not be ignored.
Concluding Remarks
Sex differences in same-sex friendships are well-documented. They are
consistent with stereotypes of differences between males and females, and
there is evidence that sex-role identity is related to friendship behavior in
the predicted manner. This suggests that sex differences in same-sex
friendships might be minimized when sex roles are minimized. Friendships
between males and females who do not engage in predominantly masculine
or feminine behavior may be more similar than those between more sex-role
stereotyped individuals. Also, given the importance of the sex of the other as
a determinant of our behavior, less stereotyping of others should also lead to
fewer sex differences in same-sex friendships. Finally, homophobia among
males may inhibit emotional intimacy in male same-sex friendships. A view
96
Barbara A Winstead
of the male role that includes experiencing and sharing feelings with other
males might increase the likelihood of close male-male friendships.
There tends to be an assumption that female same-sex friendships are
better than male same-sex friendships. If we value intimacy, this seems a
valid assumption. We might, however, be in danger of overlooking valuable
aspects of male same-sex friendships not generally present in female samesex friendships. Lever (1976, 1978) argued that playing team sports, which is
more common among males, gives males experience in leadership and in
successfully adopting roles in complex organizations. Perhaps the activities
that adult male friends engage in are a continuation of this practicing of
instrumental behavior. Certainly, males should also enjoy sharing personal
feelings with other men, but perhaps females should have the opportunity to
engage in goal-oriented and/or competitive activities in the relatively benign
setting of their friendships.
The preceding discussion also indicates that male-male and femalefemale friendships are not only different from one another but also different
from other types of relationships (e.g., cross-sex platonic friendships and
heterosexual or homosexual romantic relationships). In research same-sex
friendships are often grouped with other sorts of close relationships.
Although social support and some other functions may occur in all close
relationships, the unique characteristics and functions of different types of
relationships will remain undiscovered in such research. As investigators we
should be careful to specify and separate one type of close relationship from
any other type. The sex of the persons involved in the relationship and its
romantic vs. nonromantic nature are two dimensions that are of great
importance in defining types of relationships.
One of the unique aspects of same-sex friendships is the meeting of
someone like oneself at least in terms of sex. Indeed, women often mention
the ability of their same-sex friends to understand by their own experience
what it is like to be a woman. Males may be less likely to mention this in
describing same-sex friendships (given their tendency to avoid focusing on
the relationship itself as a topic of conversation). Their sense of freedom of
expression in "locker rooms," however, suggests that males do have an
expectation of a level of understanding and appreciation of certain behavior
by other males that may not be forthcoming from females. Miller (1983),
writing about his own experiences searching for male friendship, found the
french word complicite, a secret understanding, useful in defining the
meaning of friendship. The "secrets" of one's own sex, inside information
about what it is like to be female or male, may be one aspect of sharing in
same-sex friendships that cannot be duplicated in opposite-sex friendships
or heterosexual romantic relationships.
References
Allport, G.w., Vernon, P.E., & Lindzey, G. (1960). Manual: A study o/values (3rd ed.).
Boston: Houghton Miffiin.
97
Aries, E.J., & Johnson, F.L. (1983). Close friendship in adulthood: Conversational
content between same-sex friends. Sex Roles, 9, 1183-1196.
Banikotes, P.G., Neimeyer, G.I., & Lepkowsky, e. (1981). Gender and sex-role
orientation effects on friendship choice. Personality and Social Psychology Bulletin,
77,605-610.
Bakan, D. (1966). The duality of human existence. Boston: Beacon.
Bender, v.L., Davis, Y., Glover, 0., & Stapp, I (1976). Patterns of self-disclosure in
homosexual and heterosexual college students. Sex Roles, 2, 149-160.
Benton, AA (1971). Productivity, distributive justice, and bargaining among
children. Journal of Personality and Social Psychology, 18,68-78.
Booth, A (1972). Sex and social participation. American Sociological Review, 37, 183192.
Booth, A, & Hess, E. (1974). Cross-sex friendship. Journal ofMarriage and the Family,
36,38-47.
Burda, P.e., Vaux, A, & Schill, T. (1984). Social support resources: Variation across
sex and sex roles. Personality and Social Psychology Bulletin, 10, 119-126.
Caldwell, M.A., & Peplau, L.A. (1982). Sex differences in same-sex friendships. Sex
Roles, 8, 721-732.
Chaikin, AL., & Derlega, V.I. (1974). Self-disclosure. Morristown, NJ: General
Learning Press.
Chesler, M.A., & Barbarin, O.A. (1985). Difficulties of providing help in a crisis:
Relationships between parents of children with cancer and their friends. Journal of
Social Issues, 40(4}, 113-134.
Chodorow, N. (1978). The reproduction of mothering: Psychoanalysis and the sociology of
gender. Los Angeles: University of California Press.
Costanza R (1984). Sex differences and sex-role considerations in the use of touch and
distance between friends. Unpublished manuscript, Old Dominion University.
Davidson, L.R, & Duberman, L. (1982). Friendship: Communication and interactional patterns in same-sex dyads. Sex Roles, 8(8}, 809-822.
Deaux, K (1984). From individual differences to social categories. American
Psychologist, 39, 105-116.
Deaux, K (1977). Sex differences. In T. Blass (Ed.), Personality variables in social
behavior (pp. 357-377). Hillsdale, NJ: Erlbaum.
Derlega, V.I, Winstead, B.A, & Wong, P.T.P. (in press). An attributional analysis of
self-disclosure and relationship development. In G. Miller & M. Roloff (Eds.),
Explorations in interpersonal communication (2nd ed.) Beverly Hills: Sage.
Derlega, v.J., Winstead, B.A, Wong, P.T.P., & Hunter, S. (1985). Gender effects in an
initial encounter: A case where men exceed women in disclosure. Journal of Social
and Personal Relationships, 2,25-44.
Douvan, E., & Adelson, I (1966). The adolescent experience. New York: Wiley.
Eder, D., & Hallinan, M.T. (1978). Sex differences in children's friendships. American
Sociological Review, 43,237-250.
Feshbach, N., & Sones, G. (1971). Sex differences in adolescent reactions toward
newcomers. Developmental Psychology, 4, 381-386.
Fischer, IL., & Narus, L.R (1981). Sex roles and intimacy in same sex and other sex
relationships. Psychology of Women Quarterly, 5, 444-455.
Gilligan, e. (1982). In a different voice: Psychological theory and women's development.
Cambridge, MA: Harvard University press.
Hacker, H.M. (1981). Blabbermouths and clams: Sex differences in self-disclosure in
same-sex and cross-sex friendship dyads. Psychology of Women Quarterly, 5, 385401.
Hays, RB. (1984). The development and maintenance of friendship. Journal of Social
Heslin, R, & Boss, D. (1980). Nonverbal intimacy in airport arrival and departure.
Social Psychology Quarterly, 42,373-385.
98
Barbara A Winstead
Hill, C.T., & Stull, D.E. (1981). Sex differences in effects of social and value
similarity in same-sex friendship. Journal of Personality and Social Psychology, 41,
488-502.
Hinde, RA (1984). Why do the sexes behave differently in close relationships?
Journal of Social and Personal Relationships, 1,471-501.
Johnson, F.L., & Aries, EJ. (1983). Conversational patterns among same-sex pairs of
late-adolescent close friends. Journal of Genetic Psychology, 142,225-238.
Laosa, L.M., & Brophy, J.E. (1971). Effects of sex and birth order on sex-role
development and intelligence among kindergarten children. Developmental
Lever, J. (1976). Sex differences in the games children play. Social Problems, 32,478487.
Lever, J. (1978). Sex differences in the complexity of children's play and games.
American Sociological Review, 43, 471-483.
Levitt, E., & Klassen, A (1974). Public attitudes toward homosexuality: Part of the
1970 national survey by the Institute for Sex Research. Journal of Homosexuality, 1,
29 ...43.
Lewis, RA (1978). Emotional intimacy among men. Journal of Social Issues, 34, 108121.
Lombardo, J.P., & Lavine, L.O. (1981). Sex-role stereotyping and patterns of selfdisclosure. Sex Roles, 7,403-411.
McAdams, D.P. (1982a). Experiences in intimacy and power: Relationships between
social motives and autobiographical memory. Journal of Personality and Social
McAdams, D.P. (1982b). Intimacy motivation. In AJ. Stewart (Ed.), Motivation and
society (pp. 133-171). San Francisco: Jossey-Bass.
McAdams, D.P., & Constantian, CA (1983). Intimacy and affiliation motives in
daily living: An experience sampling analysis. Journal of Personality and Social
McAdams, D.P., Healy, S., & Krause, S. (1984). Social motives and patterns of
friendship. Journal of Personality and Social Psychology, 47, 828-838.
McAdams, D.P., & Losoff, M. (1984) Friendship motivation in fourth and sixth
graders: A thematic analysis. Journal of Social and Personal Relationships, 1, 1127.
Miller, L.c., Berg, J.H., & Archer, RL. (1983). Openers: Individuals who elicit
intimate self-disclosure. Journal of Personality and Social Psychology, 44, 12341244.
Miller, S. (1983). Men and friendsh ip. Boston: Houghton Mifflin.
Morgan, B.S. (1976). Intimacy of disclosure topics and sex differences in selfdisclosure. Sex Roles, 2, 161-166.
Narus, L.R, & Fischer, J.L. (1982). Strong but not silent: A reexamination of
expressivity in the relationships of men. Sex Roles, 8, 159-168.
Newcomb, T.M. (1961). The acquaintance process. New York: Holt.
Nichols, J. (1975). Men's liberation: A new definition of masculinity. New York: Penguin
Books.
Nisbett, RE., & Wilson, T. (1977). Telling more than we know: Verbal reports on
mental processes. Psychological Review, 84, 231-259.
Powers, EA & Bultena, G.L. (1976). Sex differences in intimate friendships in old
age. Journal of Marriage and the Family, 38, 739-747.
Rands, M., & Levinger, G. (1979). Implicit theories of relationship: An intergenerational study. Journal of Personality and Social Psychology, 37,645-661.
Reis, H.T. (1986). Gender effects in social participation: Intimacy, loneliness, and the
conduct of social interaction. In R Gilmour and S. Duck (Eds.), The emergingfield
of personal relationships (pp. 91-105). New York: Erlbaum.
99
Reis, H.T., Senchak, & Solomon, B. (1985). Sex differences in the intimacy of social
interaction: Further examination of potential explanations. Journal of Personality
and Social Psychology, 48, 1204-1217.
Richey, M.H., & Richey, H.W. (1980). The significance of best friend relationships in
adolescence. Psychology in the Schools, 17,536-546.
Rokeach, M. (1968). Beliefs, attitudes and values. San Francisco: Jossey-Bass.
Rubin, Z. (1970). Measurement of romantic love. Journal of Personality and Social
Rubin Z. (1980). Children'sfriendships. Cambridge: Harvard University Press.

Sharabany, R., Gershoni, R., & Hoffman, IE. (1981). Girlfriend, boyfriend: Age and
sex differences in intimate friendship. Developmental Psychology, 17, 800-808.
Small, A, Gross, R., Erdwins, C, & Gessner, T. (1979). Social attitude correlates of
sex role. The Journal of Psychology, 101, 115-121.
Solano, CR. (1981). Sex differences and the Taylor-Altman self-disclosure stimuli.
The Journal of Social Psychology, 115,287-288.
Stewart, AJ., & Chester, N.L. (1982). Sex differences in human social motives:
Achievement, affiliation, and power. In AJ. Stewart (Ed.), Motivation and society
(pp. 172-218). San Francisco: Jossey-Bass.
Stokes, I, Fuehrer, A, & Childs, L. (1980). Gender differences in self-disclosure to
various target persons. Journal of Counseling Psychology, 27, 192-198.
Taylor, D.A, & Altman, I. (1966). Intimacy-scaled stimuli for use in studies of
interpersonal relations. Psychological Reports, 19, 729-730.
Waldrop, M.F., & Halverson, CF., Jr. (1975). Intensive and extensive peer behavior:
Longitudinal and cross-sectional analyses. Child Development, 46, 19-26.
Weiss, L., & Lowenthal, M.F. (1975). Life-course perspectives on friendship. In M.F.
Lowenthal, M. Thurnher, & D. Chiriboga (Eds.), Four stages of life. San Francisco:
Jossey-Bass.
Wheeler, L., & Nezlek, I (1977). Sex differences in social participation. Journal of
Wheeler, L., Reis, H., & Nezlek, I (1983). Loneliness, social interaction, and sex roles.
Winstead, B.A, & Derlega, V.I (1984). The therapeutic value of same-sex friendships.
Paper presented at the annual meeting of the Southeastern Psychological Association Convention, New Orleans.
Winstead, B.A, Derlega, V.J., & Wong, P.T.P. (1984). Effects of sex-role orientation
on behavioral self-disclosure. Journal of Research in Personality, 18,541-553.
Wright, P.R. (1982). Men's friendships, women's friendships and the alleged
inferiority of the latter. Sex Roles, 8, 1-20.
Wright, P.R. & Crawford, AC (1971). Agreement and friendship: A close look and
some second thoughts. Representative Research in Social Psychology, 2, 52-69.
Chapter 6
Differences in Social Exchange Between

Intimate and Other Relationships:
Gradually Evolving or Quickly Apparent?
John H Berg and Margaret S. Clark
Interpersonal attraction has long been of considerable interest to social

psychologists. Much of the work in this area, however, has been of the "oneshot" variety, focusing on factors influencing initial attraction between
strangers. Social psychologists, for instance, have studied how proximity,
similarity, physical attractiveness, and equity influence strangers' initial
attraction toward one another. Recently, some investigators have argued
that close relationships such as friendships or romantic relationships are so
different from casual ones that little of what has been gleaned from the
study of initial attraction will be of use in understanding them (e.g., Levinger
& Snoek, 1972; Murstein, Cerreto, & MacDonald; Rubin, 1973). Such
criticisms, whether or not they tum out to be correct, highlight the
importance of social psychologists addressing themselves more fully to
understanding the dynamics of close friendships and how people become
close. If whatever it is that distinguishes close from not close relationships
develops very gradually, perhaps it is true that studying people who have
just met will be of little importance to understanding the formation and
dynamics of relationships such as friendships and romantic relationships.
If, however, people make some fairly clear decisions about the nature of
relationships early on, then one can argue that studying initial impressions
and behavior is important to the study of close relationships. If decisions are
made early, initial impressions should influence those decisions. Moreover,
important differences in behaviors that distinguish close relationships from
other relationships may be observed very early in those relationships,
perhaps even in the first hours and perhaps even in a laboratory setting.
In this chapter, we briefly discuss some of the ways in which the nature of
social exchange that takes place in relationships typically described as
"close" or "intimate" may differ from that which takes place in relationships
that are not described using such terms. Then we point out that to date little
102
John H. Berg and Margaret S. Clark
work has focused on the question of when and how these distinctions come
about. Is the decision-making a slow, gradual process, making it difficult to
pick up differences early in relationships? Several theories of relationships
seem to suggest this. Alternatively, do people make a decision quickly,
thereby producing large differences, early in the relationship, in the nature
of social exchange between relationships likely to become close and those
not so destined? Addressing these questions is the topic of major concern in
this chapter. Further, we ask questions about what factors have an impact
on the decision processes and whether what we have learned from the study
of initial attraction can be applied to understanding close relationships.
For purposes of this chapter, exchange is defined in very general terms. We
consider it to include verbal and other communicative exchanges as well as
exchanges of more tangible resources. The term close relationships refers to
relationships such as friendships or romantic relationships as contrasted
with relationships between casual acquaintances, co-workers, or between
people who do business with one another.
Some Differences in Exchange Between

Close and Casual Relationships
Before discussing how quickly or gradually the nature of social exchange in
close relationships is differentiated from social exchange in other relationships, we must ask what distinguishes social exchange in a close relationship from social exchange in one that is not close. Researchers have
suggested several possibilities.
First, in a recent book on close relationships, Kelley et al. (1983) defined
closeness as the degree to which two person's behaviors are interdependent.
To the extent that one person's behavior is dependent upon the other's, or, in
other words, to the degree that their chains of behaviors are "causally
interchained," the relationship is close. To the extent that persons' behaviors
are not intertwined, the relationship is not close. Working within this
framework Kelley et al. suggested that a relationship may be considered
closer when members frequently influence one another, when a single
behavior on the part of one can produce an intense reacton or a long chain of
responses on the other's part, and when the types of impacts are diverse. * The
idea that closeness is reflected in greater amounts of exchange is also
reflected in others' writings (e.g., Berg, 1983, 1984; Hays, 1984) and has
received some empirical support. For example, the amount of self-disclosure
*We note that Kelley et al.'s distinction between close versus not close relationships
does not correspond exactly with the distinction between relationships that is of
primary concern in the present chapter. A business relationship, for instance, might
be close in terms of how frequently members influence one another, yet most people
do not refer to business relationships as close, and for purposes of this chapter, we
would not consider most business relationships, as close.
Differences in Social Exchange Between Intimate and Other Relationships
103
occurring between friends has been shown to be related to the stability and
reported closeness of the relationship (Altman & Taylor; Berg; 1983). The
idea that a greater diversity of resources are exchanged in close versus not
close relationships has also been expressed by others (e.g., Berg, 1984; Clark,
1981; Walster, Walster & Berscheid, 1978).
The quality of resources exchanged is another basis on which some
investigators have distinguished close from not close relationships. For
example, Tornblom and Fredholm (1984) took note ofFoa and Foa's (1974)
categorization of resources into classes oflove, services, status, information,
money, and goods and suggested that the giving of some of these resources
(Le., love, status, and service) is more characteristic of friendships than is the
giving of others. Supporting this idea is their finding that descriptions of
exchanges of love and service, although not status exchanges, are more
likely to lead observers to conclude that friendship exists between two
people than are descriptions of exchanges of information, goods, or
money.
Others have pointed out that the quality of communication is different in
close as compared with other relationships. For instance, self-disclosure
researchers have argued that the closer the relationship, the greater the
intimacy or depth of self-disclosure (e.g., Altman & Taylor, 1973; Tolstedt &
Stokes, 1984). In support of these ideas, researchers examining the selfdisclosure of friends (e.g., Altman & Taylor, 1973; Berg, 1984; Hays, 1985)
and dating couples (e.g., Berg, 1983) have found that the intimacy of selfdisclosure is related to the stability and reported closeness of the relationship. In addition, Preno and Stiles (1983) demonstrated another difference
in communication behavior between close versus not close relations. They
reported that the communications of married subjects are more presumptive
and directive than are the communications of subjects encountering one
another for the first time.
Still other work suggests distinguishing close from not close relationships
on the basis of the norms governing when benefits ought to be given and
accepted. For example, several researchers have suggested and provided
some evidence that members of relationships such as friendships are more
likely to give one another benefits in response to needs and desires (in
Miller & Berg's, 1984, terms, selective resources) than are people not involved
in close relationships (e.g., Clark & Mills, 1979; Clark, Ouellete & Milberg,
1984; Kelley, 1979; Lerner, Miller, & Holmes, 1976; Mills & Clark, 1982),
whereas members of business relationships or strangers are more likely to
give benefits in response to past benefits, with the expectation of receiving
comparable benefits in return (e.g. Clark & Mills, 1979) or simply on the
basis of their own anticipated outcomes (Kelley, 1979).
Arguments that the circumstances under which exchanges take place
differentiate close relationships from other relationships also come from the
self-disclosure literature. Investigators have found that the immediate
reciprocity of self-disclosures may be greater between strangers than
between established friends (Derlega, Wilson, & Chaikin, 1976; Won-
104
Doornik, 1979) or spouses (Morton, 1978). An explanation of these findings

offered by Miller and Berg (1984), however, suggests some caution in
interpreting these results as indicating a general decline in reciprocity per se.
Instead, they may reflect differences in when people in close versus not close
relationships reciprocate. They suggest that strangers, particularly those
who wished to develop friendships with the other felt more urgency to
demonstrate interest in a relationship and to reciprocate. As a friendship
becomes established, feelings of urgency may decline. Consequently,
immediate reciprocity may decline as well (Miller & Berg, 1984). This
explanation, if correct, suggests that a general conclusion that high
reciprocity of self-disclosure is more characteristic of strangers than of
friends or spouses may be misleading. Instead, high reciprocity of selfdisclosure may be more characteristic of people destined to become close
than either (a) people destined not to become close or (b) those who have
already established closeness.
Finally, there are ways of distinguishing close from other relationships
that do not deal directly with the nature of social exchange but may be of
importance to understanding the nature of social exchange in close as
compared to relationships that are not close. For instance, close relationships have been distinguished from ones that are not close on the basis of
whether the relationship tends to be intrinsically satisfying (B1au, 1964), the
expected length of the relationship (Mills & Clark, 1982; Walster et aI., 1978),
and whether the people involved in the relationship think of themselves as a
unit (e.g., Hatfield & Sprecher, 1983; Levinger, 1979; Walster et aI. 1978;
Wegner & Guiliano, 1982). These variables may be of importance in
understanding exchange in relationships for a variety of reasons. For
instance, the more intrinsically rewarding participants find a relationship,
the less concerned they may feel about specific repayments for benefits
given. Expected endurance of a relationship may be important because it
makes sense to follow a need-based norm for giving benefits only when a
relationship is expected to endure long enough for one's own needs to be
met by the other, whereas benefits may be given in response to a past benefit
or with the expectation of receiving a benefit in return in a relationship of
either a short or a long duration (Mills & Clark, 1982). Alternatively,
duration may be important because it allows one to work out complicated
exchanges involving the giving and receiving of a variety of benefits over a
long time span (Walster et aI., 1978). Finally, perception of oneself and the
other as a "unit" may be important because if members of a relationship
think of themselves as a unit, they may not feel as if they are giving anything
away when they give something to the other person.
To summarize, various authors have claimed that social exchange in close
relationships differs from social exchange in other relationships in terms of
the interdependence of people's behavior, the amount and variety of
resources exchanged, the quality of resources exchanged, the circumstances
under which benefits are given, as well as in terms of such things as whether
the people involved see themselves as a unit, whether the relationship tends
105
to be intrinsically rewarding and the length of the relationship. The nature

of social exchange in close as compared with not as close relationships may
differ in other ways as well. Nonetheless, this brief discussion should
provide readers with an idea of the types of changes in the nature of
exchange between not as close and close relationships.
These differences raise another question that, as we pointed out previously, has received almost no attention to date: Given that differences
such as these exist, when do they arise?
When Are Differences in Exchange Detectable?

Is the transition in the nature of social exchange between casual acquaintances and close friends gradual? Or does it occur early in the
relationship's development? To consider still another possibility, might a
decision about closeness be made upon first meeting, such that, at least for
some aspects of exchange, there really is no transition period? Below we
briefly review theory and research relevant to each view. Our own view, as
will become evident, is that decisions about what relationship type people
wish to pursue are often made very early in a relationship, often right at the
beginning. As a result, many differences in the nature of social exchange
between casual acquaintances and close friends will occur right at the
beginning of the relationship. We do not push this view too far, however.
Some differences in social exchange, we suggest, do emerge gradually.
Before expressing our own views though, consider the positions on this issue
that some other relationship researchers have explicitly or implicitly
taken.
The Transition Is a Continuous and Gradual Process
The suggestion that the transition from a casual to a close relationship such
as a friendship or romantic relationship is gradual is exemplified in the
theoretical writings of Levinger and Snoek (1972) and Altman and Taylor
(1973) as well as in some "filter" theories of relationship formation (e.g.,
Kerckhoff & Davis, 1962; Murstein, 1970). In brief, Levinger and Snoek hold
that friendship development proceeds from a stage of unilateral awareness
(in which only one person is aware of the other) through a stage of surface
contact (in which there is interacton, but it is formal and/or governed by
social roles and norms), to a stage of mutuality. In this last stage, the life
space of the parties begins to overlap. Interaction becomes less formal and
the parties develop their own particular interaction rules and a large
number of shared experiences. Two processes are supposedly critical to this
development. One is the process of interpersonal discovery and disclosure.
The other is a process of mutual investment wherein each person
increasingly coordinates behavior with the other and cares emotionally
about the other. In the process the people become increasingly inter-
106
dependent: "their interdependence expands not only with the increased

frequency and diversity of interchain connections, but also with increases in
the affective strength of those connections" (Levinger, 1983, p. 325). There
would appear to be no substitute for the passage of time and repeated
interactions in this process, although Levinger did point out that the rate of
"progress" is not always constant.
Social penetration theory (Altman & Taylor, 1973) also suggests that the
development of close friendship is a gradual process. According to this
theory, after an interaction occurs, a person evaluates his or her outcomes in
light of preinteraction expectations and available interaction alternatives.
Then the person forecasts the likely outcome of future interactions. Both the
evaluations and the projected forecasts are stored in memory for later use. If
they are favorable, future interaction is initiated and is again evaluated in
terms of how it agrees with the previous forecast. Then revised predictions
are made. If these new forecasts are favorable, the person is likely to decide
to begin interacting at a slightly more intimate level. If the forecast is
uncertain, the decision is likely to be to slow down the course of penetration
or to hold it at the current intimacy level. Finally, if the forecast is clearly for
stormy seas, the decision is likely to be to terminate the relationship. In
other words, the model proposes that friendship development is a continuous process of evaluation and reforecasting. Because the outcome of any
given interaction can be evaluated in terms of a past history of positive or
negative encounters, with time the importance of any single event diminishes. For Altman and Taylor, the most stable and closest relationships are
predicted to be those that have developed slowly over a large number of
interactions.
Altman, Vinsel, and Brown (1981) noted that the studies done to test the
theory (e.g., Altman & Haythorn, 1965; Keiser & Altman, 1976; Morton,
1978; Taylor, 1968; Taylor & Altman, 1975; Taylor, Altman, & Sorrentino,
1969) "have consistently demonstrated that the growth of relationships
follows the hypothesized course of development from peripheral, superficial
aspects of personality to more intimate ones. The disclosure of superficial
information usually takes place rapidly during the early states of a
relationship, whereas exposure of intimate aspects of the self occurs only
gradually [italics added] and at later states of a relationship" (p. 110). Recent
research &lso supports this view (Hays, 1985). Although Altman et al. (1981)
recently added that "relationships can exhibit cyclical, reversible and
nonlinear processes" (p. 109) as well as the unidirectional, cumulative
processes implied earlier, their hypothesized course of close relationship
development may still be described as gradual (see pp. 142-145 of Altman et
aI., 1981).
Finally, filter theories in their various forms (e.g., Kerckhoff & Davis,
1962; Murstein, 1970) also seem to postulate gradual relationship formation.
These theories, developed primarily in the context of mate selection, hold
that potential partners pass through a successive series of "filters." Initially,
107
external characteristics (such as similarity in physical attractiveness) will be

of primary importance. Later issues of value consensus come to the fore,
and, finally, if a person has survived these earlier filterings, the issue of how
his or her role fits with one's own becomes the primary determinant of
attraction. The empirical evidence for filter theories is rather weak, however
(e.g., Levinger, Senn, & Jorgensen, 1970; Rubin & Levinger, 1974).
In addition to theories that suggest gradual development of friendships
and romantic relationships, some empirical work done outside the context
of these theories would also seem to suggest that development of "closeness"
is a gradual process. In studies by Bolton (1961) and Braiker and Kelley
(1979), respondents' retrospective reports of the development of their
relationships were gathered. Despite reports of some "turning points,"
relationship development was recalled as having been gradual. Braiker and
Kelley (1979), for instance, identified four content dimensions that couples
used to describe their courtship: (a) love, or the extent of belongingness or
attachment; (b) conflict, or the degree of negative affect and overt argument;
(c) maintenance, or the degree of self-disclosure between partners about the
relationship; and (d) ambivalence, or the extent of confusion or hesitancy
about continuing the relationship. Then a questionnaire based on these
dimensions was given to young married couples who completed it once for
each of four stages of their relationship's development: casual dating,
serious dating, engagement, and the first 6 months of marriage. Scores for
three of the four dimensions changed fairly gradually over each of the four
time periods. Specifically, love and maintenance increased over all four
stages, whereas ambivalence decreased over the four stages. Conflict was the
only dimension not showing continuous, gradual changes over all stages. It
increased from casual to serious dating and then leveled off.
Decisions About the Nature of Relationships
and Corresponding Differentiation Occur Quickly
In contrast to the theories and empirical evidence just described, some
recent work suggests that decisions about the nature of relationships, close
versus not close, and corresponding differences in at least some types of
exchange behaviors may be reached quite early. Before reviewing this work
however, consider the types of evidence that would indicate decisions about
the desired closeness of relationships and, consequently, differentiation in
social exchange occur quite early. One type of evidence, indicating that
decisions about the nature of social exchange are made early, comes from
studies showing that measures of attraction and of social exchange variables
obtained early in the course of a relationship predict the state of the
relationship at a later time almost as well or as well as the same measures
taken much later. This is consistent with the idea that people are making
decisions about whether a relationship will be close early and are sticking
by these decisions. It does not fit as well with the view that persons gradually
108
John H.Berg and Margaret S. Clark
and continuously explore the pros and cons of a relationship, slowly

becoming attracted and committed to each other (or gradually choosing to
drop out of the relationship). If that were the case, measures taken early
should be considerably poorer predictors of the relationship's eventual
outcome than are measures taken later. Of course, although such findings
are more consistent with the idea that decisions are made early than with
the idea that there is constant reevaluation and new decisions, it does not
necessarily mean a conscious decision has been made early. A relationship
may be initially set on a particular course; actually calling the relationship a
friendship or romantic relationship may come later. Furthermore, this type
of evidence does not demonstrate that differentiation in the nature of actual
social exchange comes early.
What type of evidence is needed to show more clearly that not only may
implicit decisions about the desired nature of a relationship be made early
but also that clear differentiation in the nature of social exchange occurs
soon afterwards? In answering this question we note first why the theories
and empirical data cited previously supporting the gradual view were not
optimal for detecting possible early decision-making and differentiation.
First, in order to detect very early changes, we must examine initial
interactions between people. Many of the studies supporting the "gradual"
view of relationship development have focused on what happens in close
relationships after the very initial states of those relationships have passed.
For instance, Braiker and Kelley (1979) started their investigation by asking
participants about a point in their relationship at which they were already
dating. By excluding first meetings, researchers may miss important differentiation in the nature of social exchange occurring very early in
relationships. Second, and more important, in order to detect early
differentiation between close and not close relationships, we must make
comparisons between the nature of social exchange in relationships likely
and those unlikely to become friendships or romantic relationships early in
the formation of each type of relationship. If decisions are made early,
perhaps even immediately upon first meeting, such direct comparisons are
crucial to demonstrate this differentiation. One cannot simply examine
differentiation in close relationships over time, assuming that they started
out in the initial few meetings exhibiting the same type and amount of social
exchange that would occur in relationships not destined to become close.
The theories just reviewed have simply not addressed the issue of whether
close relationships may be differentiated from ones that will not become
close from the very beginning. Moreover, many of the empirical studies
suggesting that close relationships are slowly differentiated from other
relationships (e.g., Bolton, 1961; Braiker & Kelley, 1979) have not examined
behavior in relationships unlikely to become close. Thus, data from these
studies could not be used to make the comparisons central to our present
concern.
109
Recent studies exist, however, that do examine relationships early in their

formation and do compare relationships destined to become close with
those not so destined. Although none of these studies were designed
specifically to address the issues we raise in the chapter, their results clearly
support the idea that decisions about whether or not a friendship will be
close and that at least some aspects of social exchange that distinguish close
from other relationships are differentiated at or near the beginning of
relationships. Some studies show differences in the quantity and quality of
exchange in relationships destined to become close as opposed to other
relationships as early as 2 weeks into those relationships. Other studies
show that simple experimental manipulations are capable of immediately
producing expressions of desire for different types of relationships (Le.,
acquaintanceship versus friendship) as well as striking differences in certain
aspects of social exchange. In other words, these studies suggest that people
know certain "rules" governing behavior in close versus other relationships
and can very quickly follow the appropriate rules for the relationship type
they desire. In addition, some evidence exists showing that these immediate
differences in the social exchange norms parallel differences that exist
between casual relationships and established close relationships.
We now review this recent evidence, beginning with studies that suggest
the course of relationships i's "established" fairly early. Then we tum to the
evidence that at least implicit decisions about the desired nature of a
relationship are made quickly and that clear differentiation in some aspects
of social exchange may immediately follow the decision.
Predicting relationship outcome from measures taken early versus late. In two
studies, one investigating dating couples (Berg, 1983) and the other
investigating college roommates (Berg, 1984), predictions of the final
outcomes of reiationships were found to be just about as accurate or as
accurate using measures obtained near the beginning of the relationship as
they were using measures obtained approximately 4 months later. Studying
dating couples, Berg employed measures of love, degree of conflict, feelings
of ambiguity about the relationship, communication about the relationship,
and the extent to which members changed their behavior to resolve
problems (Braiker & Kelley, 1979). These measures were administered to
dating couples early in their relationships (after about five dates) and again
about 4 months later. Using the early measures as predictors in a
discriminant analysis allowed 80% of the subjects to be correctly classified
as either still dating or not. Although this rose by 10% when later measures
were used as discriminating variables, it is clear that prediction of later
dating status could be made quite accurately at an early time. Also, at both
times of measurement subjects who continued to date compared to those
who broke up felt that the relationship exceeded both their expectations
(comparison level) and their comparison level for alternatives.
110
To the extent that college roommates decide to continue to live together

because they have formed friendships, a second study by Berg (1984)
provides additional evidence for decisions about the nature of relationships
being made early. In this study initial measure of students' liking for their
roommate, satisfaction with their living arrangements, self-disclosure,
equity, benefits received and desires met were obtained 2 weeks after the
beginning of classes and again in the spring. Discriminant analysis was
once again used, with the criterion being decisions about living together the
next year (yes, no, undecided). This time, measures taken later in the year
were no better as predictors than were measures taken much earlier (58%
versus 59% correctly classified, respectively).
The evidence from these two studies is consistent with the idea that people
are making decisions about the nature of their relationships early. It does
not fit as well with the view that members of relationships are gradually
exploring the pros and cons of relationships and that decisions about
whether to stay in or to leave the relationship occur only gradually.
Evidence of differentiation early in relationships. Although accurate predictions of relationship outcome from early measures provides evidence
consistent with people making decisions about the nature of their relationships early and against the idea of gradual differentiation other evidence is
needed to demonstrate that the nature of social exchange is clearly
differentiated, early in such relationships. Additional data from recent
longitudinal studies provide such evidence. Other analyses from the Berg
(1983) study, for instance, revealed that at both early and later points in time,
subjects who continued to date differed from those who broke up not only in
love for their partner and satisfaction with their relationship, but also on a
number of social exchange dimensions as well. At both times continuing
daters reported that they engaged in more communication about the
relationship, altered their behavior more to resolve problems, received more
self-disclosure from their partner, and felt that their relationship was
superior to their expectations for it and to other alternatives.
In addition, Hays (1984,1985) recently conducted two longitudinal studies
of friendship formation. His results suggest that relationships destined to
become close not only can be accurately predicted quite early but also that
at least some aspects of the nature of social exchange in those two types of
relationships are differentiated very early. In each study at a point just 2 to 3
weeks into the fall term, Hays asked new students to select two same sex
others whom they had not known prior to the school year but whom they
thought might become good friends as the year progressed. At this time and
again about every 3 weeks throughout the fall (for a total of four
measurements), subjects indicated the amount of casual and intimate
dyadic exchanges (communication, companionship, affection, and consideration) that they had with each of these others. Relationships that at the
end of the semester were rated as close friendships were compared with
those that were not rated as close.
III
In these studies, the overall amount of behavioral exchange increased

over time among pairs rated as close at the end of the semester but not
among those rated as not-close at the end of the semester. Two findings not
emphasized, however, are of importance for the point we wish to make here.
First, inspection of Hays' plots of exchange behavior across time indicates
that in both studies pairs that developed into close friends showed more
behavioral exchange in many intimate and nonintimate areasjrom the time
exchange was first assessed. They then increased very gradually to 6 weeks and
were as high at 6 weeks as they were at 12 weeks. To reiterate, though, much
and in some cases most of the differentiation was already present at 2 weeks.
As Hays (1984) noted, "The emergence of intimate behavioral exchange was
not as gradual as expected" (p. 91). This is consistent with the idea that
decisions about whether a relationship will be close are made, either
implicitly or explicitly, at an early point in relationship development and
that the nature of at least some aspects of social exchange are clearly
differentiated early on.
Although Berg's and Hays' studies show early differentiation of social
exchange processes in relationships destined to become close verus those
not so destined, their work may not have been able to reveal just how early
such differentiation takes place. After all, the reports of behavior in their
studies were not collected upon participants' first meeting but rather
following several dates (Berg, 1983) or after the passage of 2 to 3 weeks
(Hays, 1984, 1985). Moreover, the magnitude of the differentiation indicated
by these studies undoubtedly has been underestimated because in all cases
the "target" people who were studied, including the non close targets, would
seem to be, a priori, considerably more likely than average to become the
subjects' friends or romantic partners. After all, they were either already
dating the subject (Berg, 1983), had been chosen by the subject as someone
who might become a friend (Hays, 1984, 1985), or they and the subject had
been assigned to be roommates (Berg, 1984), a situation that probably leads
to a heightened expectaton of at least liking one another (Darley &
Berscheid, 1967) and probably also of becoming friends.
Fortunately, additional evidence exists indicating that people make
decisions about relationships early that result in immediate differences in
social exchange processes that avoids these issues. This evidence, which
comes mainly from recent laboratory work on "communal" and "exchange"
relationships (Clark & Mills, 1979), compares differences in the nature of
social exchange in relationships that subjects view as likely to become
friendships (or possibly romantic relationships) with those that they view as
likely to remain casual acquaintances in the first hour of those relationships.
Clark and Mills distinguished communal from exchange relationships on
the basis of norms governing the giving and receiving of benefits (Clark &
Mills, 1979; Mills & Clark, 1982). In communal relationships people are
presumably concerned with one another's needs and benefit one another in
response to needs or to demonstrate general concern for the other.
112
Friendships and romantic relationships often exemplify communal relationships. In contrast, in exchange relationships people presumably feel
no special responsibility for the other and benefit one another in response to
past benefits or with the expectation of receiving comparable benefits in
return. Acquaintanceships and business relationships often exemplify
exchange relationships.
In most of the studies done investigating this distinction, subjects were
recruited to participate along with a friendly, attractive confederate
sometimes of the same gender, sometimes of the opposite gender. Then the
type of relationship subjects desired with this confederate was manipulated.
To produce desire for a communal relationship, investigators led subjects to
believe that the confederate was new at the university and anxious to meet
people. To produce desire for an exchange relationship, investigators led
other subjects to believe that this same confederate was married, had been at
the university for a while, and presumably was not anxious to make new
friends. As a result of these manipulations, the researchers expected that
subjects would immediately choose to follow exchange or communal norms
to guide their behavior toward the other and to evaluate the others' behavior
toward them. In other words, in originally planning this work, the
investigators simply assumed that decisions about the type of relationship
one wants with another occur very early. This technique has proven
successful. The communal and exchange manipulations have been shown
to cause subjects immediately to express preferences for distinct types of
relationships, express preferences for following distinct social exchange
norms, and behave in ways consistent with those distinct social norms.
First consider evidence that manipulations of a desire for a communal
versus an exchange relationship produce immediate differences in the type
of relationships subjects say that they prefer with the other. Clark (in press)
recruited subjects for a study on impression formation during which they
expected to have a discussion with an opposite-sex stranger. Before meeting
the "other," however, the subject had a chance to look at a picture of the
other and a questionnaire the other had supposedly filled out. The
questionnaire indicated either that the other was single, new at the
university, and anxious to meet people or that the other was married, had
been at the university for 2 years, and would be picked up after the study by
a spouse. After looking over these materials, the subject checked one of five
categories of relationships he or she would like to have with the other: (a) a
romantic relationship, (b) a friendship, (c) an acquaintanceship, (d) a
business-like relationship, or (e) no relationship at all. The "communal"
manipulation led a significantly higher proportion of subjects to indicate
that they wanted a friendship than did the "exchange" manipulation, which
instead led people to choose the acquaintanceship option. The remaining
options, with the exception of the "business-like" option that was chosen
only once (in the "exchange" condition), were not chosen at all. Here, then,
is a case in which people were deciding on the type of relationship they
wanted with the other before even having met the other.
113
Of course, finding that people choose to pursue or not to pursue a

friendship or an acquaintanceship very early does not necessarily mean that
the nature of social exchange will differ right from the beginning in
friendships versus acquaintances hips. However, other results emerging
from the same program of research demonstrate that the same manipulations also cause people to say immediately that they want to follow
distinct social exchange norms and, more importantly, to behave in ways
indicating that they actually are following these distinct norms within the
first hour of the relationship.
In the Clark (in press) study, for instance, subjects exposed to the
communal manipulation were not only more likely than subjects exposed to
the exchange manipulation to say that they desired a friendship, but also
significantly more likely to agree with statements that they would respond to
the others' needs and expect the other to respond to their needs relative to
agreeing with statements that they would expect repayment for favors and
would readily repay the other for favors. The opposite pattern of results held
for the exchange subjects. Clark and Vanderlipp (in press) report similar
findings with same sex pairs of subjects.
Even more convincing evidence comes from studies examining the nature
of subjects' actual behavior immediately following exposure to these
manipulations. For instance, in a study on helping (Clark et aI., 1984),
subjects led to expect a communal relationship immediately helped the
other significantly more than did subjects led to expect an exchange
relationship. Moreover, in the same study, subjects led to expect communal
relationships responded to a cue that the other was sad by significantly
increasing helping, whereas the other's sadness had no impact on helping
when exchange relationships were expected. In addition, Clark and Mills
(1979, Study 1) found that when people were induced to help another, those
led to expect an exchange relationship with the other liked the other more if
she repayed them than if she did not, whereas the opposite pattern occurred
when subjects were led to expect a communal relationship with the other. In
still another study, subjects received aid from another (Clark & Mills, 1979,
Study 2). Subjects who were exposed to the exchange manipUlation liked the
other more when she requested repayment than when she did not, whereas
subjects exposed to the communal manipulation liked the other less when
she requested repayment than when she did not. Finally, in a study on
keeping tract of inputs into a joint task, people led to expect an exchange
relationship with another were shown to keep track of their inputs, whereas
subjects led to expect a communal relationship did not (Clark, 1984, Study
1).
These studies employing communal and exchange relationship manipulations clearly indicate that not only can people make decisions about
what type of relationship they want with another quickly, but also that the
norms they will follow in giving and receiving benefits in these relationships
are chosen very quickly. Indeed, in the case of these studies, the nature of
social exchange was differentiated within the first hour!
114
Moreover, studies employing these particular manipulations are not the

only ones demonstrating that the nature of social exchange in relationships
likely to become close can be quickly distinguished from that in other
relationships. Studies by Sholar and Clark (1982) and by Berg, Blaylock,
Camarillo, and Steck (1985) illustrate the same point. In the Sholar and
Clark study, subjects were recruited for a research session on group
problem-solving. When they arrived, they were told that they had been
matched to four-person groups based on extensive pretests given at the
beginning of the semester. The matching supposedly had been done either
in such a way that group members were very likely to become friends or in
such a way that group members were unlikely to be friends. Group members
expecting friendship formation were found to state immediately that a group
decision rule (i.e., consensus) that would take everyone's needs into account
was more appropriate than one that would not (i.e., majority rule). In
contrast, subjects led to expect no particular type of relationship with their
group members showed no such preference.
Recently, Berg et al. (1985) found that people act in ways that are
consistent with such preferences. In this study, previously unacquainted
subjects participated in a decomposed gaming procedure that provided
varying amounts of rewards for themselves and their partner. Prior to
beginning subjects had or did not have a short "get acquainted session" and
were led to anticipate or not anticipate future interaction with their partner.
The effect of these differences in prior and anticipated interaction on the
motives represented by subject's choices in the game was examined. Results
indicated that subjects (particularly males) who had experienced prior
interaction or who anticipated future interaction made fewer choices in the
game exemplifying either the desire to "beat" the other or to maximize their
own outcome exclusively. They made more choices that reflected the goals
of maximizing the other's outcome, maximizing the total outcomes of self
arid other as a unit, and ensuring equal distribution of rewards between self
and other. To the extent that past interaction and the anticipation of future
interaction increase the anticipation of a close relationship, as one might
expect based on findings such as those reported by Darley and Berscheid
(1967), these results demonstrate immediate differences in exchange
behavior that parallel the difference in rule preference found by Sholar and
Clark (1982).
Of course, the early differentiation observed in these studies may not
generalize to differences in behaviors in ongoing friendships as compared
with relationships that would not be described as close. A recent series of
studies, however, provides some evidence for such generalizability. In a
series of studies dealing with keeping track of inputs into joint tasks, Clark
(1984) not only manipulated the type of relationship expected with the other,
but also compared the same behavior under the same circumstances
between pairs of existing friends and pairs of strangers not exposed to any
relationship manipUlation. She found that the immediate changes in
115
behavior produced by manipulation of desire for a communal relationship

as compared with an exchange relationship were paralleled by differences
in the behavior of close friends as compared with strangers. Specifically,
both subjects led to desire an exchange relationship and strangers exposed
to no relationship manipulation carefully kept tract of their individual
inputs into a joint task. In contrast, neither subjects led to desire a
communal relationship nor subjects who worked with an actual friend
showed any effort to keep track of individual inputs into joint tasks. Indeed,
each showed some tendency to avoid doing so.
Differentiation may actually be exaggerated early in the development of close
relationships. We have argued that the nature of social exchange may be
differentiated very early in relationships destined to become close relative to
that in relationships not so destined. Going one step further, we now suggest
that such differentiation may even be exaggerated early in the development
of close relationships relative to the differentiation that will continue to exist
later. This may occur because behaving in ways appropriate to a friendship
or romantic relationship and actively avoiding behavior more appropriate
to acquaintanceships or business-like relationships may serve a signalling,
in other words, a communication function. Specifically, it may serve to let
the other know what type of relationship is desired. Thus, early in
relationships people may intentionally and effortfully avoid any hint that
inappropriate norms are being followed. Once the relationship is established, this communication function is no longer important. People may
stop "bending over backwards" to avoid any hint of following inappropriate
norms, and they may also be somewhat more relaxed about following the
correct norms as well.
To date, only a small amount of data exists supporting this point. The
clearest evidence comes from the previously discussed series of studies by
Clark (1984) on record keeping during joint tasks for which there will be a
reward. To make the present point, we must describe these studies in more
detail. In the first study of this series, subjects were led to expect either a
communal or an exchange relationship. They were then assigned to work on
a joint task with the other. The task involved locating numbers in a matrix
and circling them. The other took a tum first and circled numbers either in
red or black ink. Then the matrix was placed on the subject's desk, on which
both a red and a black pen had already been placed. The subject's behavior
was observed. Clark reasoned that if more than 50% of the subjects in either
condition picked a different color pen from that used by the confederate,
that would be evidence of intentional record keeping. If fewer than 50% did
so that would be evidence of actively avoiding record keeping. The results
revealed that significantly more than 50% of the exchange subjects did
choose a different color pen. More importantly, for purposes of the present
point, significantly fewer than 50% of the communal subjects did so. It is
noteworthy that communal norms do not actually call for such avoidance,
116
but nonetheless communal subjects did "bend over backwards" to avoid

following exchange norms.
The evidence that differentiation may decrease once friendships are
firmly established comes from the second two studies in the same series. The
initial study was replicated twice with pairs of established friends and pairs
of strangers. In both cases strangers were once again significantly more
likely than chance and than actual friends to choose different color pens,
but in neither study did actual friends' choice of pens differ from chance.
Presumably, actual friends no longer feel a need to go out of their way to
avoid exchange norms in order to communicate what type of relationship
they desire.
Aspects of the previously mentioned study by Hays (1985) are also
consistent with the idea that very early in relationships that people want to
become close, they may be quite concerned with demonstrating their
interest to the other and that this concern may decline over time.
Specifically, Hays noted that "successfully progressing dyads displayed an
initial flurry of interaction at the onset of their relationships followed by a
general decline" (p. 919). He called this early phase the "relationship
building phase" and suggested that the drop off in exchange behavior that
followed might be due to subjects having less free time as the school year
progressed. However, he noted, despite the decrease in behavior, the close
dyads' attitudinal ratings of friendship consistently increased over time.
Although recognizing the reasonableness of Hays' points, we suggest that
perhaps early in the relationship dyads are very concerned about demonstrating their availability and interest in the relationship and their
responsiveness (Miller & Berg, 1984) to the other. Thus, they may have
intensely pursued interactions early on. Then, once they felt confident that
the relationship was established, the perceived need for the flurry of activity
and, consequently, the activity itself dropped off.
We have now reviewed evidence suggesting that people may reach
decisions about pursuing close friendships very early and that differentiation in the nature of social exchange may occur very quickly thereafter.
Indeed, differentiation may be exaggerated early in relationships relative to
later. This raises two new questions. First, why should people reach such
decisions early as opposed to "taking their time"? Second, what factors
influence this decision?
Why a Decision May Be Reached Early
Close social relationships are what people most often cite as giving their
lives meaning (Klinger, 1977). These relationships have the potential of
yielding great rewards, but also of being the source of great costs as well.
Given their importance, and the fact that one can only maintain a limited
number of close friendships and usually just one romantic relationship,
117
people might be expected to take a "slow and easy does it" approach.
However, we have just noted that certain aspects of behavior in close
relationships are likely to differ from that in relationships that are not close
very early. Why should differentiation take place early as opposed to more
gradually? Our analysis suggests that one reason is simply that some of the
rules governing close versus not close relationships differ in fundamental
ways. To the extent that friendships or romantic relationships involve some
qualitatively different types of social exchange rather than simply more (or
less) of what is involved in other relationships, it should be quite difficult to
build a close relationship gradually. One must choose to follow the rules for
a close relationship early in order to communicate one's desire for such a
relationship to the other and, consequently, to give it a chance to develop
further. Otherwise, the other cannot tell what sort of relationship is desired
and respond in kind if he or she so desires.
Stating the same idea in a different way, one might say that quite different
"scripts" (Abelson, 1976) exist for close versus other relationships. Such
scripts may encompass the differences between close and other relationships noted previously and others as well. Should such a script be evoked by
some aspect of an interaction, it may well be evoked in full and affect the
way subsequent information about the other and the relationship is
perceived, stored, and recalled. It may also guide one's actions. Recent
research by Davis and Todd (1982, in press) demonstrating that shared
"prototypes" of various types of relationships exist (e.g., friendships,
acquaintanceships, and romantic relationships) is consistent with this idea.
Once a decision is made to pursue a friendship or romantic relationship
and to follow the rules appropriate to the relationship, other processes may
come into play, helping to ensure that the relationship develops along the
expected lines. For instance, people may tend to recall those aspects of the
other's behavior consistent with their expectations (Zadny & Gerard, 1974),
they may selectively seek out such information (Snyder & Swann, 1978;
Swann & Read, 1981) and, as a result, elicit behavior from the other
consistent with their expectations (Snyder, Tanke, & Berscheid, 1977).
Some Factors Influencing Decisions About
Whether a Relationship Will Be Close
We have, then, both some empirical data suggesting that decisions about
close friendships are made early and a rationale for why this may be. Now
we mention a few factors that may influence such a decision or evoke a
script for close friendships. We have not constructed an exhaustive list, but
rather tried to identify a few variables suggested by past researchers and
some others that look promising and merit further study.
The assumption that close relationships, such as friendships, can be
distinguished on the basis of the norms one follows in a relationship with
118
another suggests some variables that should be important to the decision to

pursue a close relationship. First, both the self and the other must be
available for a relationship in which each can respond to the other's needs.
Beyond this, the responsiveness of the other to the person and the closely
related idea of the other's degree of adherence to communal norms ought to
influence this decision. We discuss these variables first, then turn to a brief
discussion of some traditional determinants of attraction.
Availability. Both a person's availability for a close relationship and the
perception that the other is also available should exert a strong influence on
the decision to pursue close friendship. In our view, judgments of
availability include assessments of the (a) accessibility of the other, (b) the
amount/degree of both the person's and the other's prior commitments and
(c) both the person's and the other's alternatives.
Accessibility refers to the likelihood that both parties will be able to
interact with one another on the frequent basis and in the particular ways
that close friendships often require. Certainly, physical proximity should
increase accessibility, and studies investigating the effects of physical
proximity on relationships (e.g., Festinger, Schachter, & Back, 1950; Hays,
1985; Segal, 1974) very clearly demonstrate the importance of this variable to
the development of friendships.
Accessibility also involves judgments of whether both parties have the
resources necessary for the type of exchange needed to form a close
relationship. "Do I have the resources necessary to fulfill the other's needs?"
and "does the other have the resources necessary to fulfill my needs?"
should be important questions people ask themselves. It may also be that
judgments of consensus (i.e., "Do we both know what rules we will follow in
our exchanges, and do we agree on what each of us needs from our
relationship with the other?") playa role here. In short, judgments about
accessibility will involve accessing the extent to which it will be possible for a
particular friendship to become close.
Judgments of availability also involve assessments of both one's own and
the other's prior commitments and alternatives. Even if another is physically close, each person has the resources necessary to form a close
relationship, and there is consensus, if one already has many close
friendships or if the other is so involved, one or both parties may not have
the necessary time and/or desire for an additional close friendship. Also, in
our culture, some relationships explicitly (as well as implicitly) are
exclusive. Often such exclusive relationships will be cross-sex relationships.
The most obvious examples are marriage, engagement, or "going steady."
Just as the existence of other relationships decreases the chances of the
other being perceived as available for a relationship, so too might heavy
involvement with a job lead to judgments that oneself and/or the other is
unavailable and might preclude a decision to pursue a close relationship.
119
The Clark (in press) study described previously in which exposure to

"communal" and "exchange" relationship manipulations produced immediate differences in desire for a communal as opposed to an exchange
relationship fits well with some of these ideas. The major difference between
the communal and the exchange manipulation used in that study was the
availability of the attractive confederate. The communal manipulaton led
the subject to believe that the confederate had no spouse and probably had
few friends (having just moved to campus). It also explicitly conveyed that
the other was anxious to meet new people; in other words, the other was
available. The exchange manipulation, on the other hand, led the subject to
believe that the other had a spouse and probably had friends (having been
on the campus for 2 years). Further, it contained no information suggesting
that the other wanted to meet new people; in other words, the other was
relatively unavailable.
Responsiveness. Given a judgment that the other is available, one must
determine the other's interest in the relationship. The other's general
responsiveness may be extremely important in making such judgments.
Alternatively, one may not have given a thought to the other's availability or
to one's desire for a close relationship with the other but rather begin to
receive signals from the other in the form of the other's responsiveness.
Responsiveness is a dyadic construct referring to the extent to which a
person's intentional actions address the needs, desires, and past actions of
another (Miller & Berg, 1984) and are perceived by the recipient as doing so.
Miller and Berg identify two classes of responsiveness: (a) conversational
responsiveness (acts through which a person indicates interest in and
understanding of another's communications) and (b) relational responsiveness (behaviors involving the attainment or distribution of resources
through which a person indicates concern with another's outcomes or
needs). Three aspects of action influence judgments of either type of
responsiveness: (a) content (what was done), (b) style (how it was done), and
(c) timing (when it was done). In any case, the more responsive the other, the
greater the judgment that the other is interested in a relationship. As Davis
argued, responsiveness may lead to increased attraction, maintenance of
interaction, and perception that the responsive person and person responded to have a relationship (Davis, 1982; Davis & Perkowitz, 1979).
The influence of responsiveness on development of friendship is suggested by a recent study (Miller, Berg, & Rugs, 1984) in which subjects read
about an interaction between two people and judged the likelihood of
friendship developing. They found that after one person was aided by
another, the most responsive thing the person could do was either provide a
benefit that met a specific need of the initial giver (when such a need was
know) or equally divide the reward that the initial aid had allowed the
person to earn (when no specific need was known). Although the specific
120
content of an act might vary, whatever act was seen as most responsive was
also viewed as most likely to result in friendship. Thus, when a recipient of
aid knew about another's need, friendship was seen as more likely when he
or she gave the other a resource that met the need than when he or she
offered no repayment, offered one exactly equal to the initial aid, or divided
rewards equally. When no specific need was known, friendship was seen as
more likely when rewards were equally divided than in other cases.
Other researchers have investigated the effects of variations in the
content, style, and timing of conversational responsiveness on attraction and
the perception that a close friendship will develop. Consistently, investigators have found that as the content of a reply to another's communication becomes more responsive either because it explicitly expresses
concern for the other or addresses the same subject matter as the subject's
initial communication, attraction for that respondent increases. Subjects'
perceptions that a relationship exists or will form also increase (Berg &
Archer, 1980, 1983; Davis & Perkowitz, 1979). In addition, Berg and Archer
(1982) found that when subjects were given the interaction goal of
maximizing their attractiveness to another, their replies to the communications they received became maximally responsive.
As far as stylistic differences in responsiveness go, Dabbs, Uwanna,
Evans, and Bakeman (1982) found that attraction at the conclusion of a
series of conversations was significantly related to a pairs' use of backchannel communications in their initial interactions. This type of responsiveness may also be related to relationship initiation, and some people may
be better at it than others (Miller, Berg, & Archer, 1983; Purvis, Dabbs, &
Hopper, 1984). Finally, the timing of responses may also prove to be quite
important (Berg & Miller, 1985). Responding to a person's disclosure with
concern increases observers' judgment that a friendship exists, but if the
initial disclosure is oflow intimacy, quick expressions of concern may cause
perceptions of friendship to decrease.
The other's adherence to communal norms and his or her transformations.
Although responsiveness may generally promote relationship initiation, not
all types of responsiveness may be related to the formation of friendships or
romantic relationships. One could be quite responsive to another's business
offer, for instance, yet that might contribute nothing toward initiating a
close relationship in the sense in which we are using that term in the present
chapter.
Kelley (1979) postulated that the most effective, if not the only, means of
knowing that another considers us a close friend is through the transformations he or she makes to take our needs as well as his or her own needs
into account. Similarly, the communal/exchange distinction suggests that
being responsive to another's needs and desires ought to be a good way to
initiate a relationship such as a friendship or a romantic relationship,
121
assuming that the other is available and desires such a relationship. Beyond
this, the communal/exchange distinction suggests that seeking the other's
help or advice (Clark, 1983) and intentionally avoiding following exchange
norms (e.g., saying "Don't bother to pay me back") ought to be effective ways
of initiating friendships or romantic relationships, again assuming that the
target of these tactics is available and desires such a relationship.
Other traditional antecedents of attraction. We can return now to an issue
raised at the beginning of the chapter. Is it really the case that little of what
has been gleaned from the study of initial attraction will be of use in
understanding close relationships? Wi11laboratory studies of attraction and
relationships generally prove to be not very useful? The arguments in this
chapter imply that we should not be too pessimistic about the answer. To the
extent that decisions about the nature of relationships are made very early,
knowing how factors such as physical attractiveness and similarity influence initial attraction may indeed be crucial to understanding relationship formation. Moreover, to the extent that we can influence
decisions about relationships in a laboratory and that the nature of social
exchange is immediately differentiated following such a decision, laboratory studies investigating differences between relationships destined to
become close and those not so destined may also be quite useful. In
connection with this point, it is certainly encouraging that, as we discussed
earlier, measures taken early in relationships predict later relationship
status quite well (Berg, 1983, 1984; Rusbult, 1983). It is also worth noting that
some variables shown to influence initial attraction in the laboratory, for
instance, similarity (Griffitt, 1970) and matching in physical attractiveness
(e.g., Berscheid, Dion, Walster, & Walster, 1971) seem also to predict what
relationships will become firmly established later on (Newcomb, 1961).
A Compromise View
We have argued that the transition in the nature of relationships from
stranger to friend or from acquaintance to romantic partner may not be as
gradual as some earlier theories of relationship formation might imply. We
have also argued that laboratory studies of the very beginning of relationships may be quite useful. Now we wish to soften that position a bit.
The possibility of early differentiation in relationships has been emphasized primarily because that view has not been prominent in the
literature to date. The "gradual" view has been better represented. In reality,
of course, we do not advocate adopting either the view that all differences in
social exchanges of close and not close friends result from gradual,
continuous changes in the nature of exchange over time or the view that all
such differences are present immediately. The truth undoubtedly lies
between these two extremes.
122
Recall the possible differences in social exchange between close and other
relationships that were briefly discussed at the beginning of this chapter.
Various authors have postulated that close relationships involve greater
amounts of exchange, and/or greater interdependence between people.
Investigators have also suggested that close relationships involve a greater
diversity of exchange, that the intensity of the impact of one person on the
other may be greater, that the quality of resources exchanged might be
different, and that the norms governing the giving and receiving of benefits
might differ. Finally, researchers have suggested that close relationships
might be more intrinsically satisfying and expected to last longer and that
members may begin to think of themselves as "we" instead of as two
separate individuals.
Some changes are gradual. It seems to us that clear differentiation between
relationships in terms of some of these aspects may take considerable time

to occur and may occur fairly gradually over that time. For example, it
appears obvious that the interdependence of two persons' daily behaviorthe crucial defining feature of close relationships according to Kelley et al.
(l983)-would take time to develop. Also, getting to know and to understand
another person through the pool of shared experiences and disclosures that
playa central role in Altman and Taylor's (1973) and Levinger and Snoek's
(1972) theories, presumably takes time. Furthermore, only with such
knowledge may people become skilled at judging and responding to other's
needs (Berg, 1984). To give another example, it may take considerable time
before people feel free to disclose information about themselves to the other
that might embarrass them or that they dare to be presumptive (Preno &
Stiles, 1983). More work is needed to establish whether clear differentiation
along these dimensions really does occur only gradually and just how long
it takes to appear.
In addition to acknowledging that certain types of differentiation may
occur only gradually, it is important to point out that our arguments
regarding quick differentiation between close and other relationships do not
preclude such differentiation continuing to occur at a gradual pace over time.
Hays' (1984, 1985) findings discussed earlier not only indicate clear
differentiation between relationships destined to be close and those not so
destined at the time of first measurement, but also continuing differentiation in many areas of behavioral exchange up to 6 weeks, after which
time differentiation seemed to level off. Hays (1985) also revealed continuing differentiation in feelings of friendship intensity and benefits
received up to 12 weeks, after which time the study ended. In addition,
Braiker and Kelley (1979) also discussed earlier support of the idea that love
and maintenance behaviors gradually increase over the course of romantic
relationships while ambivalence behaviors decrease gradually over even
longer periods of time (i.e., over the entire time course from casual dating to
marriage), which suggests continuing gradual differentiation in the quality
of exchange between close and other relationships.
123
As a final example, despite the evidence already cited that people very
quickly decide to follow communal or exchange norms, Berg's (1984) study
on roommates provides intriguing additional evidence that with time
relationships may either become increasingly communal or that the exact
form of responsive behavior may change over time. Berg asked subjects at
both the beginning and end of a semester to list things that their roommate
had done to help them as well as things that their roommate could do that
would help them the most, regardless of whether he or she had done them.
From the responses two indices were formed: a "Total Positive" index
obtained by summing the number of things subjects listed in response to the
first question and a "Desires Met" index, calculated by counting the number
of times roommates actually did those things that subjects said would help
most. Early in the semester, Berg found significant correlations between
roommates' scores on the first measure but not the second measure. In
contrast, at the end of the semester he found significant correlations
between roommates' scores on the second measure but not the first measure.
He pointed out that this may indicate that subjects were becoming
increasingly communally oriented over time; in other words, early in the
relationship roommates may have been at least somewhat concerned with
following exchange norms and maintaining a balance in the amount of
benefits they gave each other. By the spring, however, the relationship may
have shifted to a communal basis, concern with equal amounts of exchange
may have lessened, and in its place a "communal" concern with providing
those things the other would find most helpful may have developed.
Alternatively, subjects could have been following communal norms at both
times and attempting to be responsive. Early in the year, however,
roommates may not know exactly what one another's needs and desires
were. Consequently, the best way to indicate responsiveness may have been
to maintain equality in their exchanges. Later in the year as their knowledge
of one another's needs increased, they concentrated on fulfilling these needs
rather than on equating the total number of things they did for each other.
In either case, differentiation continued after the initial few days of the
relationship, in the adherence to communal norms and/or in the form used
to demonstrate responsiveness.
Some changes occur quite rapidly. Although some aspects of close relationships may evolve gradually, the major point of this chapter remains
the same; that is, given the right set of circumstances (for example, that the
person is available and motivated to form new close relationships and that
the other either responds to the person or has initiated an interaction
indicating his or her availability), immediate or very quick differences in the
nature of social exchange result. It seems likely that the norms governing the
giving and receiving of benefits, the amount of exchange, the diversity of
exchange, and even such things as thinking about oneself and the other as a
"unit" and the relationship being very intrinsically satisfying are types of
differentiation that can occur in the first few hours of relationship
124
formation. These types of differentiation are probably quite interdependent.

Exactly which aspect of behavior changes first may vary, but Once One
aspect changes the others may very rapidly follow. After all, once a person
acts in One way that is appropriate to a close relationship, this may indicate
to the other that a close relationship is desired. If the other responds in kind,
both members of the relationship, knowing the entire set of behaviors
appropriate to close relationships, may quickly exhibit many of those
behaviors. Consequently, it may well be that differentiation of friendships
and romantic relationships from other relationships along many dimensions occurs in the first hours, as suggested by the studies On the
communal-exchange distinction. It may then continue fairly quickly over
the next few weeks (Hays, 1984, 1985) and gradually after that (Braiker &
Kelley, 1979; Hays, 1984, 1985).
Concluding Comments
In thinking about this chapter, we felt that the small amount of existing
literature in the area of relationship formation was weighted too heavily
toward suggesting that the development of intimate or close relationships
was very gradual. In contrast to that view, our own work, as well as recent
work of others, led us to believe that many aspects of close relationships
develop quite quickly. Thus, we took as our goal for this chapter putting that
literature together and making a case for there being some quickly apparent
differences in close versus other relationships. We hope we have succeeded.
In the process we also have tried to make a case that researchers should not
dismiss previous "one-shot" research on attraction too quickly. The
variables found to affect initial attraction may be crucial determinants of
initial differentiation in the nature of social exchange. That differentiation,
in tum, may set the course for the entire relationship. Nonetheless, we hope
that this chapter will encourage rather than discourage longitudinal study of
relationships. Much more longitudinal work is needed before we can feel
confident about the claims made in this chapter as well as chart the exact
course of relationship development.
On what factors should such longitudinal research focus? Clearly,
researchers need to track all aspects of social exchange and examine the
development of relationships other than those between college students.
Mter all, college students are probably more open to forming new
friendships and/or new romantic relationships than are most other groups.
Thus, it may be that by studying primarily college students, researchers have
found evidence of friendships and romantic relationships developing more
quickly than will tum out to be the case for other popUlations. Finally, it
seems important to track the development of relationships that would
clearly not be termed close, for example, relationships between employers
and employees, as well as the development of friendships and romantic
125
relationships. To understand how relationships such as friendships and

romantic relationships are differentiated from other types of relationships
across time, the development of nonintimate as well as intimate relationships must be followed.
Acknowledgments. Preparation of this chapter was supported in part by Grant No. 1
RO 3 MH40250-0l from the National Institute of Mental Health. We thank Robert
Hays for his helpful comments on an earlier version of this chapter.
References
Abelson, RA (1976). Script processing in attitude formation and decision making.
In J. Carrol & J. Payne (Eds.), Cognition and social behavior (pp. 33-46). Hillsdale,
NJ: Erlbaum.
Altman, I., & Haythorn, W.W. (1965). Interpersonal exchange in isolation.
Sociometry, 23,411-426.
Altman, I., & Taylor, D.A (1973). Social penetration: The development o/interpersonal
relationships. New York: Holt, Rinehart, & Winston.
Altman, I., Vinsel, A, & Brown, B.B. (1981). Dialectic conceptions in social
psychology: An application to social penetration and privacy regulation. In L.
Berkowitz (Ed.), Advances in experimental social psychology (Vol. 14, pp. 107-160).
New York: Academic.
Berg, J.H. (1983). Attraction in relationships: As it begins so it goes. Paper presented at the
annual meeting of the American Psychological Association, Anaheim, Calif.
Berg, J.H. (1984). The development of friendship between roommates. Journal 0/
Berg, J.H., & Archer, RL. (1980). Disclosure or concern: A second look at liking for
the norm-breaker. Journal 0/ Personality, 48,245-257.
Berg, J.H., & Archer, RL. (1982). Responses to self-disclosure and interaction goals.
Journal 0/ Experimental Social Psychology, 18,245-257.
Berg, J.H., & Archer, RL. (1983). The disclosure-liking relationship: Effects of selfperception, order of disclosure, and topical similarity. Human Communication
Research, 10, 269-282.
Berg, J.H., Blaylock, T., Camarillo, J., & Steck, L. (1985, March). Taking another's
outcomes into consideration. Paper presented at the annual meeting of the Southeastern Psychological Association, Altanta.
Berg, J., & Miller, L. (1985). The timing o/responsive acts.. Unpublished manuscript,
University of Mississippi.
Berscheid, E., Dion, K.K., Walster, E., & Walster, G.W. (1971). A test of the matching
hypothesis. Journal 0/ Experimental Social Psychology, 7, 173-189.
Blau, RM. (1964). Exchange and power in social life. New York: Wiley.
Bolton, CD. (1961). Mate selection as the development of a relationship. Marriage
and Family Living, 23,234-240.
Braiker, H.B., & Kelley, H.H. (1979). Conflict in the development of close relationships. In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing
relationships (pp. l35-168). New York: Academic.
Clark, M.S. (1981). Noncomparability of benefits given and received: A cue to the
existence of friendship. Social Psychology Quarterly, 44, 375-381.
Clark, M.s. (1983). Some implications of close social bonds for help seeking. In B.M.
DePaulo, A Nadler, & J.D. Fisher (Eds.), New directions in helping: Vol. 2. Help
seeking (pp. 205-229). New York: Academic.
Clark, M.S. (1984). Record keeping in two types of relationships .Journal o/Personality
126
Clark, M.S. (in press). Evidence for the effectiveness of manipulations of desire for
communal versus exchange relationships. Personality and Social Psychology
Bulletin.
Clark, M.S., & Mills, l (1979). Interpersonal attraction in exchange and communal
relationships. Journal of Personality and Social Psychology, 37, 12-24.
Clark, M.S., Ouellette, R, & Milberg, S. (1984, July). Recipient mood, relationship type
and helping. Paper presented at the Second International Conference on Interpersonal Relationships, Madison, Wisconsin.
Clark, M.S., & Vanderlipp, B. (in press). Perceptions of exploitation in communal
and exchange relationships. Journal of Social and Personal Relationships.
Dabbs, lH. Uwanna, R, Evans, M., & Bakeman, R (1982). Getting acquainted:
Patterns of speech and gaze. JSAS Catalogue of Selected Documents in Psychology, 63
(MS 2090).
Darley, lM., & Berscheid, E. (1967). Increased liking as a result of anticipation of
personal contact. Human Relations. 20, 29-40.
Davis, D. (1982). Determinants of responsiveness in dyadic interactions. In W. Ickes
& E.G. Knowles (Eds.), Personality, roles and social behavior (pp. 85-140). New York:
Springer-Verlag).
Davis, D., & Perkowitz, W.T. (1979). Consequences of responsiveness in dyadic
interactions: Effects of probability of response and proportion of content-related
responses. Journal of Personality and Social Psychology, 37,534-550.
Davis, KE., & Todd, M.J. (1982). Friendship and love relationships In KE. Davis &
T.O. Mitchell (Eds.), Advances in Psychology, (Vol. 2, pp. 79-122). Greenwich, CT:
JAI Press.
Davis, KE., & Todd, M.l (in press). Assessing friendships: Prototypes, paradigm
cases, and relationship assessment. In S. Duck & 0. Perlman (Eds.), Sage Series in
Personal Relationships (Vol. I). London: Academic.
DePaulo, B.M., Leiphart, V., & Dull, W.R. (in press). Helping seeking and social
interaction: Person, situation and process considerations. In E. Staub, D. Bar-Tal,
J. Reykowski, & J. Karylowski (Eds.), The development and maintenance ofpro-social
behavior. New York: Plenum.
Derlega, V.J., Wilson, M., & Chaikin, AL. (1976). Friendship and disclosure
reciprocity. Journal of Personality and Social Psychology, 34, 578-582.
Festinger, L., Schachter, S., & Back, K (1950). Social pressures in informal groups: A
study of human factors in housing. New York: Harper.
Fisher, lD., Nadler, A, & Whitcher-Alagna, S. (1982). Recipient reaction to aid: A
conceptual review. Psychological Bulletin, 9,27-54.
Foa, u.G., & Foa, E.B. (1974). Societal structures of the mind. Springfield, IL:
Thomas.
Grifitt, W. (1970). Environmental effects on interpersonal affective behavior:
Ambient effective temperature and attraction. Journal of Personality and Social
Hatfield, E., & Sprecher, S. (1983). Equity theory and recipient reactions to aid. In
lD. Fisher, A Nadler, & M. DePaulo (Eds.), New directions in helping: Vol. 1.
Recipient reactions (pp. 113-141). New York: Academic.
Hays, RB. (1984). The development and maintenance of friendship. Journal of Social
Hays, RB. (1985). A longitudinal study of friendship development. Journal of

Keiser, G.J., & Altman, I. (1976). Relationship of nonverbal behavior to the social
penetration process. Human Communication Research, 2, 147-161.
Kelley, H.H. (1979). Close relationships: Their structures and processes. Hillsdale, NJ:
Erlbaum.
Kelley, H.H., Berscheid, E., Christensen, A, Harvey, J.H., Huston, T.L., Levinger, G.,
127
McClintock, E., Peplau, L. A, & Peterson, D.R. (1983). Close relationships. San
Francisco: Freeman.
Kerckhoff, AC., & Davis, KE. (1962). Value consensus and need complementarity in
mate selection. American Sociological Review, 27,295-303.
Klinger, E. (1977). Meaning and void: Inner experience and the incentives in people's lives.
Minneapolis: University of Minnesota Press.
Lerner, M.I, Miller, D.T., & Holmes, IG. (1976). Deserving and the emergence of
forms of justice. In L. Berkowitz & E. Wa1ster (Eds.), Advances in experimental social
psychology (Vol. 9, pp. 134-163). New York: Academic.
Levinger, G. (1979). A social exchange view on the dissolution of pair relationships.
In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing relationships (pp.
l69-196). New York: Academic.
Levinger, G., (1983). Development and change, In H.H. Kelley, E. Berscheid,
A Christensen, IH. Harvey, RL. Huston, G. Levinger, E. McClintock,
L.A Peplau, & D.R Peterson, Close Relationships (pp. 315-359). San Francisco:
Freeman.
Levinger, G., Senn, D.1., & Jorgensen, B.W. (1970). Progress toward permanence in
courtship: A test of the Kerkoff-David hypothesis. Sociometry, 33,427-443.
Levinger, G., & Snoek, 10. (1972). Attraction in relationship: A new look at interpersonal
attraction. New York: General Learning Press.
Miller, L.c., & Berg, IH. (1984). Selectivity and urgency in interpersonal exchange.
In V.I Derlega (Ed.), Communication, intimacy, and close relationships (pp. 161-205).
Orlando: Academic.
Miller, L.c., Berg, IH., & Archer, RL. (1983). Openers: Individuals who elicit
intimate self-disclosure. Journal of Personality and Social Psychology, 44, 12341244.
Miller, L.c., Berg, IH., & Rugs, D. (1984, September). Selectivity in exchange: When
you give me what I need. Paper presented at the annual meeting of the American
Psychological Association, Toronto.
Mills, 1, & Clark, M.S. (1982). Exchange and communal relationships. In L. Wheeler
(Ed.), Review ofpersonality and social psychology (Vol. 3, pp. l21-144). Beverly Hills:
Sage.
Morton, T.L. (1978). Intimacy and reciprocity of exchange: A comparison of spouses
and strangers. Journal of Personality and Social Psychology, 36, 72-81.
Murstein, B.I. (1970). Stimulus-value-role: A theory of marital choice. Journal of
Marriage and the Family, 32,465-481.
Murstein, B.I. (1972). Physical attractiveness and marital choice. Personality and
Social Psychology, 22, 8-12.
Murstein, B.I., Cerreto, M., & MacDonald, M.G. (1977). A theory and investigation of
the effect of exchange orientation on marriage and friendship. Journal of Marriage
and the Family, 39, 543-548.
Newcomb, T.M. (196l). The acquaintanceship process. New York: Holt, Rinehart, &
Winston.
Preno, B.E., & Stiles, W.B. (1983). Familiarity in verbal interactions of married
couples versus strangers. Journal of Social and Clinical Psychology, 1, 209-230.
Purvis, lA, Dabbs, 1M., & Hopper, c.H. (1984). The "opener": Skilled user of facial
expression and speech pattern. Personality and Social Psychology Bulletin, 10,6166.
Rubin, Z. (1973). Liking and loving: An invitation to social psychology. New York: Holt,
Rinehart, & Winston.
Rubin, Z., & Levinger, G. (1974). Theory and data badly mated: A critique of
Murstein's SVR theory and Lewis's PDF model of mate selection. Journal of
Rusbult, C.E. (1983). A longitudinal test of the investment model: The development
128
of satisfaction and commitment in heterosexual involvements. Journal of Personality and Social Psychology, 45, 101-117.
Segal, M.W. (1974). Alphabet and attraction: An unobtrusive measure of the effect of
propinquity in a field setting. Journal of Personality and Social Psychology, 30, 654657.
Sholar, W., & Clark, M.S. (1982). Deciding in communal relationships: By consensus or
should majority rule? Paper presented at the annual meeting of the Eastern
Psychological Association, Baltimore.
Snyder, N., & Swann, W.B. (1978). Hypothesis testing processes in social interaction.
Snyder, M., Tanke, E.D., & Berscheid, E. (1977). Social perception and interpersonal
behavior: On the self-fulfilling nature of social stereotypes. Journal of Personality
and Social Psychology, 35, 656-666.
Swann, W.B., & Read, S.H. (1981). Self-verification processes: How we sustain our
self-perceptions. Journal of Experimental Social Psychology, 17,351-372.
Taylor, D.A (1968). Some aspects of the development of interpersonal relationships:
Social penetration processes. Journal of Social Psychology, 75, 79-90.
Taylor, D.A, & Altman, I. (1975). Self-disclosure as a function of reward-cost
outcomes. Sociometry, 38, 18-31.
Taylor, D.A, Altman, I., & Sorrentino, R. (1969). Interpersonal exchange as a
function of rewards and costs and situational factors: Expectancy confirmationdisconfirmation. Journal of Experimental Social Psychology, 5, 324-339.
Tolstedt, B.E., & Stokes, IP. (1984). Self-disclosure, intimacy and the depenetration
process. Journal of Personality and Social Psychology, 46, 84-90.
Tornblom, KY., & Fredholm, E.M. (1984). Attribution of friendship: The influence
of the nature and comparability of resources given and received. Social Psychology
Quarterly, 47,50-61.
Walster, E., Walster, G.W., & Berscheid, E. (1978). Equity: Theory and research.
Boston: Allyn & Bacon.
Wegner, D.M., & Giuliano, T. (1982). The forms of social awareness. In W. I Ickes &
E.S. Knowles (Eds.), Personality, roles, and social behavior (pp. 165-198). New York:
Springer-Verlag.
Won-Doornick, M. (1979). On getting to know you: The association between stage of
relationship and reciprocity of self-disclosure. Journal of Experimental Social
Zadny, J., & Gerard, H.B. (1974). Attributed intentions of informational selectivity.
Journal of Experimental Social Psychology, 10,34-52.
Chapter 7
Strategies In Developing Friendships

People spend a relatively large percentage of their lives trying to get other
people to like them, and the inability to engender positive feelings in others
is often reported as a frustrating and unhappy experience (Bell & Daly,
1984). In initiating and developing relationships, unacquainted persons
start with intentions, hopes, needs, desires, wishes, and beliefs that guide
their actions. They also develop, so we believe, strategies for serving these
intentions. In addition, they have a range of tactical subroutines and general
styles for executing such important activities as "being polite," "being
friendly," and "encouraging engagement." Central among such strategies is
the use of self-disclosure, which we believe can be used as a tactical device to
influence relationships and is not, as previously thought, some mere
reaction to the demands of a norm of reciprocity or to exchange requirements. Thus, by varying the way in which a topic is discussed (that is, by
manipulating their communicative style), partners can influence not only
the direction a particular conversation takes but also the development of the
friendship that serves as a context to that conversation.
This strategic perspective on the study of interpersonal communication
has recently become a central concern of researchers in personal relationships; indeed, as Applegate (1983) stated "There is no more important
variable for research examining relationship development" (p. 16). However,
the necessary work in this area has only just begun. In particular, three main
areas require attention from researchers. The first is a fuller understanding
of the ways in which communicative strategies are actually used in
interactions and the goals, especially the multiple goals that underlie their
use. A second area requiring attention is that of individual differences in the
use of strategic behavior. Not all individuals will be equally strategic, and
yet we cannot accurately predict the dimensions underlying these individual
differences, although some work on unravelling the concept of selfmonitoring has produced a promising beginning (Miell & LeVoi, in press).
The third area for investigation within this strategies perspective is the need
130
for a clearer and more detailed understanding of the situational differences

in strategic behaviors. Clearly, some situations are more likely to require
strategic action and close monitoring of behavior than are others, and yet
the process of describing and categorizing the influential factors in each
situation has only just begun.
The study reported on in this chapter was designed to begin to catalogue
and describe a variety of tactics and strategies used by individuals in
developing friendships. Some of the important differences in the strategic
behavior reported by individuals as appropriate at various stages of the
development and maintenance of their friendships are identified and, thus,
several major issues in the study of communication in personal relationships are addressed from this important perspective of strategic and
functional communication.
Purposive Communication
It is assumed that people communicate for good relational reasons (Bell &
Daly, 1984). However, although the essence of this view has been recognized
and accepted, the tendency has been for researchers to focus on critical events
that edge relationships forward or backward (e.g., Planalp & Honeycutt,
1984) rather than to consider the strategic underlay that prompts communication. We believe that persons are flexible (or at least, variable) in their
communication and that this flexibility is an important feature of their
interpersonal communication. Investigators have shown that intimacy can
be communicated both by revealing different aspects of personality to
friends of different degrees (Altman & Taylor, 1973) and by conversing in a
different style with friends than with lesser acquaintances (Morton, 1978).
All this requires a high level of sensitivity to appropriate cues. Our first
assertion, therefore, is that persons have a considerable range of knowledge
concerning appropriateness and that this derives from close attention to
partner, situation and extent of shared knowledge. Thus, intimacy stems not
just from "topic intimacy" but also from manner and style of communicative behavior, in a motivational and social context.
Our present approach in attending to this social-motivational context is
far removed from more traditional conceptualizations of intimate communication and self-disclosure that often ignored the relational perspective
that we attempt to expound in this chapter. Such a perspective must
concentrate not only on the dyad (rather than the individual) as the unit of
analysis (Duck, 1985), but also on the goals that they have in mind when
they communicate. We argue that the meaning of a disclosure in a
relationship is determined mainly if not entirely by the intentions and
relational goals of the discloser and that a statement does not have any
absolute level of intimacy or meaning in itself independent of its relational
context.
Strategies in Developing Friendships
131
Altman and Taylor's (1973) model of intimacy growth has progressed

towards realizing these criteria for a relational perspective. It has led to
increased emphasis on the study of both discloser and recipient and put the
study of self-disclosure firmly on an interpersonal level, seeing it increasing
in intimacy along with a deepening relationship. In their model, however,
disclosures are made about progressively more intimate topics as the
relationship becomes closer, and this has contributed to the view that
disclosures are the product of a deepening relationship. Instead, we argue
that any item can be discussed in a more or less intimate or nonintimate
style, allowing it to be raised at any point in the relationship; for example,
an intimate item can be broached by a casual and superficial disclosure
early in the relationship. This allows a painless assessment of the likely
consequences of taking it further and it allows partners to decide how they
will define and characterize the intimacy level of their relationship rather
than being constrained by the supposed objective content of the topic
concerned (see Morton, 1978). Chelune (1976) argued that partners have to
judge whether a given disclosure is appropriate for their particular
relationship aims and motives; consequently, people have to be flexible in
disclosures.
Chelune's (1976) idea of flexibility in disclosure implies that we must
interpret behavior in its unique social context, and as such the idea of
flexibility in disclosure can be closely linked with the literature on
communication as a "game," with goals, strategies, rules, and plans for being
a successful communicator (Berlo, 1977; Higgins, 1981). These investigators
argued that a focus on the shared rules of communicating helps us to
identify various ways of "playing" the game and various possible types of
outcome. Similarly, in the work reported in this chapter, the researchers
have looked at individuals' accounts of their aims and strategies m
communicating with their partners as a relationship develops.
Information Gathering and Exchange

Perhaps the clearest aim in exchanging personal information is that such
an exchange allows both partners to gather information about each other in
a direct way. As Berger and Calabrese (1975) claimed "when strangers meet,
their primary concern is one of uncertainty reduction or increasing
predictability about the behavior of both themselves and others in the
interaction" (p. 100). This view has recently been both developed (Duck, in
press) and challenged (Planalp & Honeycutt, 1984). These authors pointed
out that some information increases uncertainty (for instance, when a new
piece of knowledge contradicts one's expectations or beliefs about the
partner) and hence showed that a given piece of knowledge does not have
absolute effects but can be significant only in the context of the relationship
in which it occurs.
132
Although the result of information-gathering should thus be recognized

to be either an increase or decrease in uncertainty about the partner, the
actual strategies used to acquire and interpret the information will be the
same. However, as yet we know very little about either of these processes.
The work begun by Berger and his coworkers to categorize and describe
these processes is a useful development. Berger and Bradac (1982) suggested
several strategies for the individual to use in order to reduce uncertainty
about the partner and increase predictability and control over the relationship. These strategies, grouped under the headings of "passive," "active,"
and "interactive," include the individual observing the other person
(especially in informal social settings), asking others about the person,
asking the person direct questions and self-disclosing (hoping that the other
person will feel pressured to conform to the norm of reciprocity and disclose
something in return). These are proactive strategies, not simple reactions to
norms of reciprocity and exchange, and can be linked to Derlega and
Grzelak's (1979) instrumental analysis of self-disclosure. Derlega and
Grzelak suggested that self-disclosure serves several functions for an
individual, such as the manipulation of a partner's behavior and the control
of the pace and direction of a relationship's development.
These functional analyses examine the ways in which self-disclosure may
be used by individuals rather than being an incidental by-product of
relational development. It is this functional approach to self-disclosure in
particular and interpersonal communication in general that is explored in
the study reported on here.
Accounts of the Role and Function of Personal

Information Exchange
The study was designed to investigate the changes in personal information
exchange over the course of friendship development. These questions were
addressed in in-depth interviews with 37 respondents (20 females and 17
males), who were first-term university students. The interviews centered on
their views of the development of relationships in general and in particular
on how they got to know others and how they chose appropriate topics of
conversation at various stages of a relationship's development. By focusing
on the initiation and early development of relationships, we hoped to elicit
responses based on the immediate personal experiences of the respondents,
who were at the time most concerned with choosing new friends and getting
to know them, being new at University.
.
The transcripts from the interviews were divided into seven sections, each
corresponding to a basic question and related discussion in the interview.
These seven questions were:
1. What is it appropriate to say to, and do with, a new partner (i.e., someone
you have only recently met, but may become acquainted with)?
2.
3.
4.
5.
6.
7.
133
How would you gather information about a new partner?

How would you decide whether or not to develop a friendship?
How would you restrict the development of a friendship?
How would you intensify the development of a friendship?
What is it appropriate to do with, and say to, a closer friend?
Why do you tell a partner personal information?
Two independent judges divided each person's responses into a number of

discrete statements, each containing one (and only one) idea in answer to
the question. A total of six independent coders then assigned each of the
statements in this pool to one of a number of categories, and only statements
assigned to the same category by four or more of these coders were retained
for further analysis. This resulted in 85.5% of statements being retained.
Further discussion and analysis was then based on the results of this coding
and referred to the categories of responses rather than to individual
statements by particular respondents. Several major findings were established by this method, which shed new light on our understanding of the
role of communication in interpersonal interactions and developing
rela tionships.
Specifically, the results reported here suggest ways in which individuals
control the flow of personal information in their dyadic interactions and
about how they define appropriate behavior at different stages of relationship development.
Four Scripts for Appropriate Behavior
Figure 7-1 summarizes the statements that individuals made about what
they judge to be appropriate behaviors at different stages of relationship
development. In this figure all the behaviors seen as appropriate to each of
four stages of a relationship's development are shown enclosed in a box
labeled with that stage of development. These four lists can thus be seen to
constitute outline "scripts" for appropriate behavior at each of the four
stages of development, which were (a) interacting with a new partner (that is,
someone who has just been met, but who could become a friend), (b)
interacting with a close friend, (c) restricting a relationship's development,
and (d) intensifying a relationship's development. When the boxes overlap
in Figure 7-1, it signifies that individuals saw these behaviors as appropriate
to both scripts. A major finding of this study was the degree of such overlap
between scripts.
In describing what they felt it was appropriate to do with and say to a new
partner, individuals listed a set of behaviors with two basic aims. The first
aim was to playa delaying, holding game with the partner, being careful not
to reveal too much about themselves by talking in too much detail or about
too wide a range of issues. As they did not yet know their partner, they also
saw it as important to behave in a rather conventionally polite manner until
I
I
I
I
I
I
against existing knowledge
Check new informa tion
Plan meetings
INTENSIFYING "SCRIPT"
See partner frequently
Intima te level of discussion
discussion topics
Wide range of appropriate
Figure 7-1. Summary of statements made by subjects in interviews: Four "scripts" for appropriate behavior.
Unrestrained, easy interactions:
knowledge
Rely on mutual trust/shared
reactions
Don't observe partner's
responsive
Don't be forthcoming/
Don't ask questions
RESTRICTING "SCRIPT"
See partner infrequently
discussion topics
Limited range of appropriate
General level of discussion
Be reserved/polite
CLOSE FRIEND "SCRIPT"
Give help and support
r--
Observe partner's reactions
disclosures
Be responsive, reciproca te
Ask questions
NEW PARTNER "SCRIPT"
(")
tJ
CIl
::I
0.
po
:;
'<
tJ
o
a
;.
.j::o.
135
they found out what their partner's attitudes and values were and what the
partner would see as acceptable behavior.
The second aim for new partners was to gather information about each
other to reduce their mutual uncertainty. As shown in Figure 7-1, the main
ways in which they gathered information about each other were by asking
direct questions, observing each other, and using reciprocal self-disclosure.
Less directly, they also inferred opinions and values from the general level
of discussion. Knowing such information would allow them not only to
have relatively smooth interactions but also to make informed decisions
about later developments and appropriate future behavior.
In these early conversations, characterized by general discussions of
rather superficial issues, new partners can assess how responsive the other
is, and the other's willingness to offer personal information and to
reciprocate disclosures, even if only at a general level. Thus, before steps are
taken towards a closer relationship, partners have gathered (and inferred)
quite a range of relevant information that will help them to decide whether
the relationship has any future and whether the partner can be trusted. This
can be done, using commonly agreed upon scripts, without appearing nosy
or intrusive.
Figure 7-1 also shows, however, that several of the behaviors believed to
be appropriate for conversing with a new partner were also suggested by
subjects as appropriate when restricting the development of a relationship.
This overlap was evident in the behaviors of limiting the range of level of
appropriate topics for discussion, seeing the partner infrequently, and
acting in a restrained, polite way. However, although there appears to be
considerable overlap between the two scripts, the commOn behaviors are
given very different relational meanings by the behavioral context in which
they are placed. When interacting with a new partner, individuals are also
actively seeking out information about the partner, reciprocating disclosures, and making inferences On the basis of their conversations about
the partner's real opinions. This context of inquiring behavior indicates that
the individual is willing to be friendly and responsive and appears to want
to get to know the partner. In this context, the polite, restricted interactions
are interpreted as playing safe, whereby the individual avoids inappropriate
intimacy but implies that this is only until they know more about the
partner.
When trying to restrict a relationship's further development, however,
these commOn behaviors are placed in a very different behavioral context
and, thus, carry a different meaning for the relationship. In these circumstances, the script suggested by subjects involves behaviors opposite to those
displayed in interactions with a new partner; that is, individuals are
expected not to ask questions, not to deliberately observe the partner's
reactions, and not to offer personal information or reciprocate disclosures.
Less directly, partners are also seen as less likely to make inferences beyond
the information disclosed in the general conversations and actually to avoid
meeting the partner, rather than infrequent meetings being the natural
136
result of circumstances. Thus, the individual's behavior does not signal a

cautious willingness to become more acquainted as it does in the context of
the "new partner script," but instead signals that the individual is not
interested in getting to know the partner and is seeking to restrain and
trivialize their interactions. One script allows for the possibility of development in a relationship, whereas the other prevents any development.
A very similar process can be seen to be at work in determining the
meaning ascribed to the behavior of close friends and those wishing to
become closer (see Figure 7-1). Here, the discussion ofa wide range of topics
at a deeper level, allied with frequent meetings, are seen as appropriate
behaviors both for those wishing to intensify a relationship and those with
an already close relationship. This set of behaviors (which is, incidentally,
in direct opposition to the set seen as appropriate for interactions with a new
partner of that used when trying to restrict a relationship's development) is
given the appropriate meaning for the type of relationship in which it occurs
by the other behaviors that are reported to accompany it.
For close friends the discussion of problems and giving of help and advice
can help them build up a very high degree of mutual trust and understanding (Samter & Burleson, 1984). Such a stock of shared understanding,
in tum, allows the partners to feel safe enough to uncover and resolve
further problems. Between partners trying to become closer, this extensive
trusting base for the relationship has not yet been built up, and, as the
subjects suggested, there is a relatively high level of interpretation, planning,
and inference in these interactions, which distinguishes them from the
rather unplanned, easy-going interactions of close friends. When intensifying a relationship's development, subjects also reported a high level of
planning in choosing when and where to meet their partner. The greater
number of meetings characteristic of this stage of a relationship was the
result of these strategic attempts to intensify the relationship. Also, by
talking about a wider range of topics, and in more detail than at earlier
stages of acquaintance, and by observing the partner's reactions in a wider
range of situations, individuals continue to gather important information
about a partner. Although the subjects did not list the same informationgathering strategies for this stage in a relationship as they did for earlier
stages (for example, asking questions), this does not mean that they no
longer sought information about their partner. Instead, individuals appeared to seek their information more subtly using more inference, more
interpretation, and more comparison of new information with the existing
store of information about the partner built up earlier in the relationship.
Thus, the behaviors common to close friends and those trying to become
closer are given their particular meaning by the unique contextual
behaviors either of sharing problems and relying on shared understanding
in the relaxed interactions between friends or of planning meetings,
inferring opinions, and checking previous information between partners
intensifying their relationship.
137
Deciding the Future of a Relationship

The subjects in the study suggested a number of criteria for making
decisions about the future development of their relationships. These criteria
were concerned with establishing several things about the partner: how
similar he or she is, how trustworthy, how easy-going, how easy to talk to,
and how available. Interestingly, these criteria involved the individuals
assessing just those aspects of their partner's behavior and opinions that
were explored in the initial phase of their acquaintance. Before they are
likely to confide really personal things, individuals need to establish
whether their partner can be trusted with less personal information; and on
the basis of how the partner treats this information, individuals make an
assessment of the partner's trustworthiness, predicting his or her behavior in
a closer relationship. These initial interactions also allow individuals to
evaluate how easy-going their partner is by observing how he or she reacts to
a variety of other people in a range of situations. The number of chances
partners have to meet is also important in the decision about the future of
their relationship. A partner who accepts invitations and is met frequently is
more likely to be seen as a potential friend.
These criteria, then, seem to be valid ones to adopt, as they draw on the
information available to the partners at this early stage of acquaintance.
Furthermore, they are sensitive to the differences between similar behaviors
that are used to begin, rather than to restrict, a relationship's development.
Thus, if a partner was seeking to prevent a relationship developing,
adherence to the appropriate script would signal this effectively. In this case,
the person would not be available for many meetings, be rather difficult to
talk to (Le., not asking questions or being forthcoming) and, by not
discussing personal information, would suggest reluctance to confide in or
trust the partner.
Gathering Information About a Partner
In initial conversations, subjects in these interviews reported somewhat
strained discussions of their immediate environment, factual aspects of their
home background, and their interests. Such discussions not only allow
partners to gather information about each other, but also keep their
conversation flowing as smoothly as possible by relying on topics generally
agreed to be appropriate. If both partners are aware of, and adhere to, the
rules of the "game" they are playing, their unfamiliarity with each other is
less likely to disrupt their interaction than if they tried to converse in a lessstereotyped manner. Incidentally, adherence to these rules also allows
partners to gather information about each other most efficiently.
Asking questions was one of the strategies that Berger and Bradac (1982)
suggested individuals use in gathering information about others. Subjects in
the present study reported using questions to find out about a new partner
and stated that the question-answer sequences helped them avoid too many
138
embarrassing silences. In a previous study of the minutiae of conversations,

Miell (1984) clearly observed a higher number of questions in the
conversations of strangers than in those of friends.
Other strategies suggested by Berger and Bradac (1982) for gathering
information in initial interactions also received support from these interviews. Subjects reported that they observed their partner, especially when he
or she was talking to others and reacting to external events and relied on
reciprocity to get their partner to tell them personal information in return
for their own disclosures. Subjects did not give examples of one of Berger
and Bradac's strategies--structuring the environment to observe the partner
in desired circumstances--but they did suggest two novel strategies not listed
by Berger and Bradac. The first was the use of provocation or argument.
Subjects felt that even when discussing relatively superficial, general topics,
a disagreement or argument allowed them to find out most directly about
their partner's true opinions, to "cut through" the polite front that they kept
up during most of their interactions. Subjects also reported that they tried to
gather information about their partner's true opinions by inferring them
from their general conversations and trying to see the conversation from
their point of view. Once a relationship had become closer, subjects
suggested that sharing problems and difficulties together not only helped
partners strengthen their bond of trust and friendship, but also allowed
them to see different aspects of each other that were hidden earlier in the
acquaintance when partners tended to project the problem-free, positive
sides of their lives and personalities.
Partners thus begin to gather information about each other once they pass
beyond the initial stages of acquaintance, as suggested by the constructive
models of relationship development (e.g., Davis, 1973; Duck, 1973, 1977).
The information they collect contributes to their body of shared knowledge,
which influences later interactions and the interpretation of their communication (Miell, 1984). However, partners not only seek information to
reduce their uncertainty of each other but also seek to reduce their
uncertainty about the (possible future for) the relationship. We had already
shown (Duck & Miell, 1982) that individuals experience a marked degree of
uncertainty about, and reviewing of, their relationships. By gathering daily
reports for 20 weeks of the early development of relationships, we showed
that subjects can assess how their partner interpreted and evaluated
everyday interactions and that they strive to make sense of the path of
development that their relationship has taken. Thus, the distinction between
the reduction of uncertainty about the other person and the reduction of
uncertainty about the relationship seems a valid and important one.
Although partners may get clearer about each other, they may not
necessarily get clearer and more secure about the relationship at the same
time. Instead, they continually need to update and rethink their picture of
both their partner and their relationship to make sense of current events.
The picture of communication and interactions in closer relationships
139
that emerges from this interview study is that closer friends talk more about
aspects of their personality, things that worry them, their hopes, fears, and
ambitions. Also, subjects reported here that the conversations between
friends would be less restrained and polite than between acquaintances,
would include more discussion and dispute, and would be able to carry on
at times either without words or with incomplete utterances as a result of the
body of shared knowledge. The greater intimacy evident in the conversations between friends was a major feature of those interactions that
individuals concentrated on developing if they wanted to become friends. In
the intensification phase of a relationship's development, subjects reported
that they sought to talk more personally, more about their feelings and
problems, and more naturally than earlier in the acquaintance. Thus,
personal disclosures are instrumental, not incidental, in intensifying a
relationship'S development, supporting the functional analysis of selfdisclosure.
Strategies for Controlling the Flow of Personal Information
A further aim of the present study was to establish how self-disclosure is
actually used to achieve the goals that people seek. It is important to note
that our emphasis in this chapter and in our research was to take into
account the necessary interplay between the partnerf': in what they each say,
how they interpret their communication, and how it affects their relationship.
Several strategies were effective in controlling the flow of information in
an interaction. In all of these, the individuals' awareness of their partner's
responses and intentions is vital for the success of the strategy; no attempt to
control an interaction or a relationship can be wholly successful if it is a
unilateral one. One of the most frequently mentioned strategies was that of
varying the level of intimacy at which a topic was discussed. By talking in a
general, joking, hypothetical, or ambiguous way, the individual can
influence a conversation in a number of ways; for example, testing their
partner's reaction to a general topic area before introducing a more personal
version of the discussion or avoiding talking in a personal way when the
partner is trying to introduce a more personal level. Alternatively, by
increasing the intimacy of a topic, an individual can intensify the
relationship. Another strategy often reported by subjects involved their
reliance on the norm of reciprocity, where they would tell their partner
something in order to get their partner to respond with a similar type and
amount of information. Subjects also suggested, however, that they might
sometimes try to get their partners to talk about themselves by revealing
rather general information but trying to make the partner respond at a more
personal level, supporting Miller and Steinberg's (1975) distinction between
"apparent" and "genuine" disclosure.
A final example of a strategy is a relatively simple one: changing the topic
140
being discussed. However, it is a deceptively simple strategy as, to be

effective, it relies on a good deal of perceptiveness by the individual, who
has to be monitoring the conversation to be aware that it is drifting towards
a "taboo" area, to be aware of the degree and type of shared knowledge
existing between the partners, to know what constitutes a "taboo" topic, and
to be able to switch topics without attracting the attention (and possibly
suspicion) of the partner. In a pilot study undertaken to examine strategic
exchanges of personal information in interactions of existing friends and
acquaintances, partners were videorecorded having a natural conversation.
They then watched this tape individually, commenting on why they said
what they did (if they felt that they had a reason for saying particular
things). This pilot study yielded a very good example of this latter strategy of
changing the topic, as an excerpt from one of the partner's commentaries on
a conversation illustrates:
Now, there I nearly went further than I intended so I stopped; that's why I
hesitated, I started to say something then I thought "No, that's something
that I don't think even my best friend knows," so therefore I wasn't about to
commit myself on that, so I just stopped and fished around for something
else that would lead into another topic, that's why there was a gap. We went
onto building a wardrobe - that was it, yes. Well, it was a problem connected
with my boyfriend, which brought my boyfriend to mind. which brought
the wardrobe into mind; it was an association of ideas really. I accept it as a
superficial relationship and it comes across on this tape as such because
every time it gets too personal the subject's changed-either by Margaret or
by myself.
[Q: "Do you think Margaret realized that you were deliberately changing
the subject?")
I think it comes across as my vaguely having lost the drift of the

conversation because I struck myself as being a bit vague when I was doing
it, but it was only because I was obviously thinking, "God, I mustn't say
that, I've got to say something else." I would have thought that Margaret
would have accepted it as a general sort of, "She's got something else on her
mind that she wants to say"-she may well not have noticed at all in
fact.
In this example, the subject was right to believe that her partner had not
noticed the strategic change of topic. The partner commented (at the same
point on the taped conversation):
It's a very general conversation, not deliberately, it's just the way it's worked
out. I could talk to anyone about this (wardrobe), although I probably
wouldn't have asked her where she bought it and how much it cost if I
hadn't genuinely wanted to know. I think here we were extending the
conversation because we didn't really know what else to talk about. The
silences seem to show we were extending it. We're repeating a few things.
In both these subjects' accounts, a high degree of monitoring of the

conversation and the relationship is apparent, influencing how they
communicated with each other. Only by examining these relational goals
and the monitoring process can communication between partners really be
141
fully understood, and the method employed in this pilot study appears to be
a promising one to access these processes.
The Level of Awareness and Strategic Planning

in Interactions
The work reviewed in this chapter paints a picture of conversations and
developing relationships as very conscious and strategic, but this is not to
suggest that partners a/ways converse and act in a deliberate way. Clearly, if
the type of reasoning cited in the previous example was going on behind
every change of topic or every disclosure, partners would not be able to
converse: The cognitive load would be too heavy. It would also be an
unnecessary load, as most everyday conversations are rather trivial and
stereotyped (Duck & Miell, 1982). Instead, as Delia (1980) suggested, many
interactions are likely to be conducted almost automatically and without
strategic planning. He also suggested that relationships develop along
trajectories determined more by the constraints of work patterns and
available time for meetings than by deliberate attempts to direct and control
their path of development. Many subjects did report that their interactions
were natural, and the process of development was spontaneous and not
consciously planned. However, individuals may come into contact with a
wide range of people but only become friends with a few of them. At some
stage, decisions must be made about restricting or intensifying a relationship, even if the relationship's form after these decisions have been made
may be shaped by external pressures and coincidence more than deliberate
strategies.
Thus, strategic planning and behavior may be evident only at particular
points in a relationship's development, rather than in every interaction in
stable, long-standing relationships. As Delia (1980) suggested, then, it may
be more fruitful to examine the proportion of time spent in conscious
planning by partners at different stages in their relationship, as indeed
recent diary studies are beginning to do (e.g., Baxter & Wilmot, 1982; Duck
& Miell, 1982).
When in the relationship are deliberate strategies likely to be employed?
From research on attention (Berlyne, 1970) and emotion (Mandler, 1975), it
appears that personally relevant novelty or interruption will be likely stimuli
for such strategies. Thus, most monitoring and planning might be expected
at change points in a relationship's development and when an unexpected
piece of behavior or information is encountered that does not fit the
individual's existing picture of either the partner or the relationship.
Presumably, if the relationship is a superficial role-governed one, it would
not be important or personally relevant enough for partners to invest much
cognitive effort in. In a closer (or potentially closer) relationship, the
partners have more to gain, and to lose, and are therefore more likely to be
142
Dorothy Mie11 and Steve Duck
attentive to novel information and to consider its implications for the future
of the relationship. In a new relationship that might grow, any information
might be relevant, any point might be a change point, and, thus, the relative
proportion of effort invested will be higher than in an established
relationship. In a long-standing relationship, although the partners will
usually have a high personal investment in the relationship, their range of
shared knowledge is likely to be large enough to allow natural, easy,
unplanned interactions in the majority of cases (Miell, 1984).
We have reviewed ways in which partners control the flow of personal
information in their interactions and how such information influences the
development of their relationships. Important rules for appropriate behavior have been identified. Most importantly, self-disclosure has been
placed on a truly relational footing, showing how the process of exchanging
personal information affects and is affected by the development of the
relationship in which it takes place. We contend that the partners' plans and
aims for their relationship are determinants of the content and style of the
communication between them. These aims, together with the unique store of
shared knowledge built up by partners, constitute the "relational context"
that sets the boundaries on what can be discussed and supplies essential
semantic and pragmatic information for the interpretation of utterances, in
particular how intimate their discussions are perceived to be.
As our review has shown, we should cast aside the passive model of
acquaintances reacting to superficial stimuli presented by an active partner.
Instead, we must credit the fact that people becoming acquainted can be
active, strategic, mentally alert, purposeful human beings who have
knowledge of social rules and can use them to develop relationships or serve
other social goals where they are needed.
References
Altman, I., & Taylor, D.A (1973), Social penetration: The development oJ interpersonal
relationships. New York: Holt, Rinehart, & Winston.
Applegate, J.L. (1983). Constructs, interaction goals and communication in relationship
development. Paper presented at the fifth International Congress on Personal
Construct Psychology, Boston.
Baxter, LA, & Wilmot, W.W. (1982, July). A longitudinal study oj communication in
same-sex and opposite-sex relationships. Paper presented at the International Conference on Personal Relationships, Madison, WI.
Bell, RA, & Daly, J.A (1984, May). Affinity seeking: Its nature and correlates. Paper
presented at the International Communication Association Conference, San
Francisco.
Berger, C.R, & Bradac, J.I. (1982). Language and social knowledge. London: Edward
Arnold.
Berger, C.R, & Calabrese, R.I. (1975). Some explorations in initial interaction and
beyond: Towards a developmental theory of interpersonal communication.
Human Communication Research, 1, 99-112.
143
Berlo, D.K (1977). Communication as process: Review and commentary. Communication Yearbook, 1, 11-27.
Berlyne, D.E. (1970). Attention as a problem in behavior theory. In D.I. Mostofsky
(Ed.), Attention: Contemporary theory and analysis (pp. 25-49). New York: AppletonCentury-Crofts.
Chelune, GJ. (1976). A multidimensional look at sex and target differences in
disclosure. Psychological Reports, 39, 259-263.
Davis, M. (1973). Intimate relations. New York: Free Press.
Delia, l (1980). Some tentative thoughts concerning the study of interpersonal
relationships and their development. Western Journal of Speech Communication, 44,
97-103.
Derlega, Y.l, & Grzelak, l (1979). Appropriateness of self-disclosure. In GJ.

Chelune (Ed.), Self-disclosure (pp. 151-176). San Francisco: Jossey Bass.
Duck, S.W. (1973). Personal relationships and personal constructs: A study offriendship
formation. London: Wiley.
Duck, S.W. (1977). Theory and practice in interpersonal attraction. New York:
Academic.
Duck, S.W. (in press). Social and personal relationships. In G.R Miller & M.L.
Knapp (Eds.), Handbook of interpersonal communication. Beverly Hills, CA: Sage.
Duck, S.W., & MieH, D.E. (1982, July). Charting the development of personal
relationships. Paper presented at the International Conference on Personal
Relationships, Madison, WI.
Higgins, E.T. (1981). The "communication game": Implications for social cognition
and persuasion. In E.T. Higgins, c.P. Herman, & M.P. Zanna (Eds.), Social
cognition: The Ontario symposium on personality and social psychology (Vol. I, pp.
343-392). Hillsdale, NJ: Erlbaum.
Mandler, G. (1975). Mind and emotion. New York: Wiley.
MieH, D.E. (1984). Cognitive and communicative strategies in developing relationships.
Unpublished doctoral thesis, University of Lancaster, United Kingdom.
MieH, D.E., & LeVoi, M.E. (in press). Self-monitoring and control in dyadic
interactions. Journal of Personality and Social Psychology.
Miller, G., & Steinberg, M. (1975). Between people: A new analysis of interpersonal
communication. Chicago: Science Research Associates.
Morton, T.L. (1978). Intimacy and reciprocity of exchange: A comparison of spouses
and strangers. Journal of Personality and Social Psychology, 36, 72-81.
Planalp, D., & Honeycutt, J.M. (1984, July). Events that undermine conceptions of
relationships. Paper presented at the second International Conference on Personal
Samter, W., & Burleson, B. (1984, May). When you're down and troubled . .. have you
got a friend? Effects of cognitive and motivational factors on spontaneous comforting in a
quasi-natural situation. Paper presented at the annual International Communication Association Conference, San Francisco.
Chapter 8
Personal Choice and Social

Constraint in Close Relationships:
Applications of Network Analysis
Robert M Milardo
Social psychologists have long subscribed to the notion that individuals are
as active in constructing their environment as they are reactive and defined
by it. Lewin (1935, 1951), for example, understood behavior to be a function
of the person and environment or life space, that is, B = f(PE). The life
space was defined rather broadly as the totality of mutually interdependent
and coexisting psychological, social, and physical facts that influence both
thought and behavior (cf. Lewin, 1951). The concept of social network
provides a means of specifying the structure and content of one aspect of the
life space, the social environment. The investigation of social networks
suggests a level of analysis distinct from the individual or dyadic levels of
analysis that link the causes of behavior to personal characteristics or
characteristics of specific relationships.
This distinction between individual and relationship properties is
illustrated in Newcomb's (1961) now classic study The Acquaintance Process.
Newcomb found interpersonal attraction to be a function of attitude
similarity. A relatively enduring trait that individuals bring into an
encounter, i.e., personal attitudes, is a determinant of the development of a
unique relationship. Moreover, once individuals establish a friendly
relationship, they become more similar in attitudes over time (Newcomb,
1961). Properties of the relationship that reflect the outcomes of interaction,
as distinct from properties of the members in it, thus influence each
individual and their interconnection to one another. But what about the
social context? In what ways might the social context of individuals
influence the initiation of a new friendship or the trajectory of an ongoing
relationship over time?
This chapter is organized around three themes or objectives. First, I
attempt to demonstrate that a network analysis is fundamentally distinct
146
Robert M. Milardo
from simply analyzing particular individuals or their relationships, and

such an analysis provides a means to link characteristics of individuals or
relationships with characteristics of the social structure in which those
individuals or relationships are embedded. This would mean that the
ongoing character of a friendship, or even whether such a friendship
develops, is determined in important ways by the structure of the social
networks in which the friends (potential or actual) are members. Second, the
influence of network structure is unique and cannot be reduced to the effects
of individual or relationship properties. Not surprisingly then, network
analysis raises a variety of new questions concerning the developmental
course of personal relationships. My third and final objective in the chapter
is to develop a theoretical model that explicitly identifies critically
important structural attributes, suggests operational definitions for those
attributes, and pinpoints the precise affects of structure on interpersonal
attraction and relationship outcomes.
Conceptualizing Social Networks

A variety of criteria has been employed to characterize social networks and
to distinguish the form of one network from that of another. These criteria
have been classified traditionally as either structural or interactional
attributes of networks. Structural attributes refer to indices of the entire set
of individuals and personal relationships that comprise a given network.
Size, composition, homogeneity, stability, degree of interconnectedness,
density, cluster or clique structure, and overlap are the major examples of
structural attributes, and each is defined briefly in Table 8-1. Interactional
attributes refer not to the network as a unit, but rather to the individual
relationships of which the network is constructed. Diversity of linkages,
transactional content, reciprocity, intensity, frequency and duration of
interaction are examples of interactional attributes (see Table 8-1). The
focus of each of these interactional attributes is upon properties of the ties
that bind individuals (their interdependence with one another) rather than
the attributes of the individual members themselves. Several detailed and
comprehensive reviews of structural and interactional attributes have
appeared recently, and there seems little reason to simply review this
material once again. Interested readers are referred to the references cited in
Table 8-1.
In the following section, I concentrate analysis on selected structural
attributes that illustrate the different ways in which networks have been
defined and the implications these definitions have on other network
attributes of interest.
Overlap
Cluster or
clique structure
Density
Interconnectedness
Stability
Homogeneity
A Structural attributes
Size
Composition
Network Attribute
Definition
Number of distinct individuals in the network

Relative proportion of varying role relations
(e.g., kin & nonkin) or relationships which vary
in intensity (e.g., close & intermediate friends)
Degree of similarity of network members
(e.g., across social class, age, or gender)
Durability and consistency of the network
membership over time
Average number of connections each individual
has with other network members
Independent of the target individual or dyad,
the ratio of actual ties linking members to the
maximum number of potential ties
Extent of network segmentation such that the
number of connections linking clique members
is greater than the number of connections
linking clique members to nonmembers
Proportion of network members held in common
by individuals in a personal relationship
Table 8-1. Attributes of Social Networks
Huston & Levinger, 1978; Milardo, 1982
(continued)
Boissevain, 1974; Sal zinger, 1982; White et aI., 1976;

Wellman, 1981b
Dickens & Perlman, 1981; Fischer et aI., 1977;

Mitchell, 1969
Boissevain, 1974; Niemeijer, 1973; Ridley &. Avery,
1979
Boissevain, 1974; Mitchell, 1969; Ridley & Avery, 1979
Fischer et aI., 1977

Fischer, 1982; Milardo, 1983
Source
Definition
B. Interactional attributes
Diversity of linkages
Variety of role relations or social activities
linking a target individual or dyad to other
network members
Transactional content Exchange of material or symbolic resources
(inclusive of social support and interference)
Symmetry of exchanges
Reciprocity
Intensity
Responsiveness of a target individual or dyad
to the expectation of network members
Frequency & duration Time investment and patterning of social
interaction
of interaction
Network Attribute
Table 8.1. Continued
Boissevain, 1974; Fehr & Perlman, 1985; Leavy, 1983;

Ridley & Avery, 1979; Saulnier, 1982; Thoits, 1982
Mitchell, 1969, Ridley & Avery, 1979
Boissevain, 1974; Fischer et aI., 1977; Mitchell, 1969
Source
g.
iii
~
(1)
::4-
Personal Choice and Social Constraint in Close Relationships
149
Structural Attributes
Size
Perhaps no other dimension of a social network is more intuitively
meaningful and yet as, conceptually obtuse as size, or simply the number of
unique individuals included in a network. The difficulty in establishing the
size of a network, and therefore the range of people who may influence the
initiation, development, or termination of a particular friendship, resides in
the variety of methodologies available for identifying a network constituency and the lack of a distinct theoretical rationale for choosing one
particular method over another. The issue becomes important when we
consider that the concept of social network is often defined implicitly or
explicitly in terms of its constituency or size.
Barnes (1954), in one of the first attempts to apply the network concept to
the study of social structure, defined network in utterly abstract and
metaphorical terms:
I find it convenient to talk of a social field of this kind as a network. The
image I have is of a set of points some of which are joined by lines. The
points of the image are people, or sometimes groups, and the lines indicate
which people interact with each other. We can, of course, think of the whole
of social life as generating a network of this kind ... ties of friendship and
acquaintance which everyone growing up in Bremnes society [a Norwegian
village] partly inherits and largely builds for [oneself]. (pp. 42-43)
More contemporary scholars have advanced the metaphor as well as the
methods available for empirically defining constituent members, a point
considered shortly. Fischer and his colleagues (1977), in particular, noted:
Individuals are linked to their society through relations with other
individuals: with kin, friends, coworkers, fellow club members, and so on.
We are each the center of a web of social bonds that radiates outward to the
people whom we know intimately, those whom we know well, those whom
we know casually, and to the wider society beyond. These are our personal
social networks. (p. vii)
To date, social scientists have centered their attention on close friends or
kindred to the exclusion of other potentially important network sectors. This
bias severely limits the size of the network identified and, more importantly,
omits from scrutiny individuals who may not be considered close but who
are nonetheless important in a variety of ways. Such peripheral associates,
for example, may help us to secure employment or prevent us from doing so,
provide us with information or guidance, criticize our decision-making, or
provide positive regard (Granovetter, 1983; Milardo, 1983; Wellman,
1981a).
Variations in the size of the network of close associates have been
examined across gender (Fischer & Oliker, 1980), class and race (Belle, 1982;
Davis, 1981), marital status (Shulman, 1975), community and neighborhood
150
Robert M. Milardo
(Fischer, 1982), as well as stage of the life-course (Dickens & Perlman, 1981;
Tesch, 1983). But certainly personal networks are composed of a variety of
relationships, only a small proportion of whom are considered close
associates, although many may be considered as friends.
In an intriguing program of research, Killworth, Bernard, and McCarty
(1984) examined the range of people known to respondents. Through a
"reverse small world" procedure, each respondent was given a dossier on
each of 500 target individuals. The task was to get a message to the target
individuals, but only people who were acquainted with one another would
be allowed to pass the message along. The dossiers on targets were fictitious,
but realistic and included a name and location in addition to information
on age, sex, occupation, organizational affiliation, and hobbies. Targets
represented a range of ages, statuses, and so on, and were distributed worldwide. For example, given a dossier on Marcelina Catalan from Buenos
Aries, a 46-year-old female who worked as an accountant and enjoyed
swimming, respondents indicated their choice of an individual who would
be most able to get a message to Marcelina.
The number of distinct choices generated by a respondent should
represent an index of that respondent's total social network. In general,
respondents identified an average of 134 distinct choices, that is, network
members. Because the number of different choices increased rather rapidly
for the first few targets, and much more slowly thereafter, Killworth et aI.
(1984) argued that a curve representing the number of distinct network
members elicited from a given set of targets "would eventually become
asymptotic to a constant value" (p. 383). This "constant value" would
represent the average size of a network for a contemporary citizen of the
United States. Extrapolation suggests a mean of approximately 250 network
members. Less than 10% of the total number of network members were
identified as kin, whereas a considerable majority (87%) were considered
friends known through a variety of contexts (see Table 8-2).
Clearly, some individuals are more socially active than others. Killworth
et aI. (1984) reported considerable variation in the size of the network
generated through the reverse small world procedure just described.
Although the mean number of network members was 134, the standard
deviation was 65. Inquiries directed at accounting for individual differences
in network size have yet to be accomplished, because researchers have
tended to center on accounting for differences in averages among groups
rather than for variation among individuals. Nonetheless, a number of
factors appear to be important candidates to account for some variation,
including levels of education, occupation, and income (Belle, 1982; Fischer,
1982); physical attributes, such as attractiveness (Reis et aI., 1982);
complexity of social skills (Bums & Farina, 1984; Cook, 1977; Fischer &
Phillips, 1982); and stage of dating or the family life-cycle (Dickens &
Perlman, 1981; Johnson & Leslie, 1982; Milardo, Johnson, & Huston,
1983).
151
Table 8-2. Distribution of Network Members

Killworth et al.
(1984)
Fischer
(1982)
Role Relation
Mean
SD
% Total
Network
Mean a
% Total
Network
Kindred
Nonkin
"Just friends"
Coworkers
Neighbors
School mates
Share hobbies
Other nonkin
Relationship
unknown
Totals
10
116
25
17
17
14
8
35
6
57
26
19
18
15
12
N/Ab
8
87
19
13
13
10
6
26
7.7
10.8
4.9
l.8
l.9
N/A
N/A
2.2
42
58
26
10
10
N/A
N/A
12
8
134
N/A
65
5
100
0
18.5
0
100
aStandard deviations unavailable.

"Not available.
Note. From Measuring Patterns of Acquaintanceships by P. D. Killworth et aI., 1984, Current
Anthropology, 25, Copyright 1984 by The University of Chicago Press. Reprinted by permission.
And from To Dwell Among Friends by C. S. Fischer, 1982, Chicago: Chicago University Press.
Copyright 1982 by The University of Chicago Press. Reprinted by permission.
Using additional methods for identifying network members, researchers

generally have found great variation in network size. Fischer and his
colleagues (Fischer, 1982; Jones & Fischer, 1978) developed a method based
on the use of 11 name-eliciting questions.' These questions are designed to
identify people with whom the respondent is likely to have significant
contacts, regardless of the frequency of interaction. Respondents are
presented with a set of social situations (for example, personal household,
work) and several categories of people defined in terms of the probability of
rewarding exchanges. Prototypes include people who the respondent values
as a confidant, for their sense of judgment, or for personal favors, including
home or child care, and so on. The name-eliciting questions are useful in
identifying a network constituency based on an explicitly defined set of
criteria that are defined similarly by most respondents. Fischer (1982)
reported an average network size of 18.5, with a total of approximately 20%
'The conceptual and psychometric properties of several major approaches to
identifying networks have been reviewed elsewhere in greater depth (see Milardo,
1983). In addition, a revision of the original Fischer procedure has been developed
for use where interest is centered on spouses, their marital relationship, and
affiliation with kith and kin. Many of the problems associated with the original
instrument, particularly the potential gender bias have been eliminated, at least in
part (see Milardo, 1984).
152
Robert M. Milardo
of the sample of 1,050 respondents listing less than 9 or more than 30

network members (see Table 8-2.).
Similarly, investigators employing measures of the "interactive network"
have found considerable individual variation in network size. Measures of
the interactive network are distinguished by the use of self-reports or
structured diaries of social activity completed daily by respondents (see
Larson, 1984; Milardo et aI., 1983; Nezlek, Wheeler, & Reis, 1983). At
minimum, this method typically requires the respondent to identify those
people with whom voluntary interactions occurred within the previous 24
hours. The network is defined in terms of the number of distinct people with
whom interactions occurred during the course of the study, usually between
7 and 18 days.
Milardo et ai. (1983) reported an average network of 26.2 members with a
standard deviation of 17.4 and a range of from 13 to 64 different interactants
over 10 consecutive days. This variation in the size of the interactive network
is equally evident when the number of interactants identified as best or close
friends is examined. For instance, although the average number of close
friends with whom interactions occurred was 6, the number of close friends
identified ranged from 0 to 19.
The variance in network size is apparent both within and between
different methodologies, and this latter variability has important consequences for additional measures of network structure. The reverse small
world procedure, for example, suggests that kin compromise a rather small
percentage of the total network, approximately 8%, whereas the exchange
network identified by the Fischer procedure is virtually dominated by kin,
comprising approximately 42% of the total network (see Table 8-2).
Reserachers to date have often described the networks of target individuals
without fully addressing the limitations of the name-eliciting procedure
utilized. One result is that networks are often characterized as being largely
composed of kin and relatively small (cf. Davis, 1981; Shulman, 1975;
Verbrugge, 1977), when in fact logic would dictate otherwise. The potential
number of kin is far more limited in scope than is the potential number of
friendly relations developed through a lifetime of work, community ties,
personal interests, and exigency.
Social networks are divergent in their constituency. They are sources of
demands and constraints as well as support and opportunity. The potential
influence of a network on the personal choices of individuals seems rather
obvious considering the sheer number of people known, but size alone is
unlikely to determine the potential for social influence.
Density, Interconnectedness, and Clique Structure
Few attributes of social networks have been investigated more intensively
than density, usually defined as the degree to which network members know
one another apart from their ties to ego. Density has been operationally
153
Figure 8-1. Networks of equivalent density but divergent structures.
defined as the number of existing ties in a network divided by the number of

potential ties. For instance, the network of Alpha is presented in Figure 8-1.
Alpha is directly connected with 10 individuals. If all of these 10 individuals
knew one another, a potential of 45 relationships, the density of the network
would be maximized at 100%.2 In actuality, only 6 ties exist between the
network members, and the density of this network is 13%.
Although intuitively meaningful as an index of network structure, a
measure of density that is based on the ratio of existing to potential ties
linking members is problematic for several reasons. The definition of the
network membership and corresponding name-eliciting procedure determines, in part, the value of the network's density. As the definition of the
network changes, both network size and density vary as well. Density is at
the same time both a defining characteristic of a network and defined by
that network (Salzinger, 1982). The strong association of density and
network size is additionally problematic because the number of potential
links increases dramatically with but a minor change in network size,
whereas the number of actual ties linking individuals is far more limited
(Boissevain, 1974). For example, if the size of a network is simply increased
from 10 to 11 members, the number of potential ties increases by to.
Density represents a measure of the average connectedness of network
members and, thereby, obscures the pattern or distribution of ties and the
presence of local cliques. Figure 8-1 depicts two networks of equivalent size
and density, but a casual perusal clearly demonstrates the limitation of the
2For the purposes of illustration, I have assumed all relationships between network
members are reciprocal; a bond between Respondents 1 and 4 is equivalent to one
between 4 and 1 (see Figure 8.1). In actuality, relationships are often asymmetric in
the sense that P may report a close tie with 0, whereas reports a peripheral tie or
perhaps no tie at all with P. For example, in large scale surveys of naturally occurring
adolescent friendships (Kandel, 1978) and adult friendships (Shulman cited in
Wellman, 1981b), only about 40% of those named as close friends and kin reciprocate
the choice. It appears that asymmetric and nonreciprocal ties are far more common
in that an individual named as a close associate is likely to choose someone other
than the respondent as a close friend or kin relation.
154
Robert M. Milardo
density measure. In Figure 8-1A, (Alpha's network), the ties that bind
network members are evenly distributed, with pairs of network members
being relatively isolated from one another. Figure 8-1A may be viewed as a
prototypical network of a husband or wife who maintain relationships with
five pairs (couples) who are basically unknown to one another. On the other
hand, Figure 8-lB (Beta's network) represents a network in which a fully
saturated local clique is present (Subjects 4, 5, 6, and 7); the remaining
network members are unknown and relatively isolated from each other.
The prototypes presented in Figure 8-1 may be, in fact, quite common and
depict structural differences with important consequences for the life-course
of target individuals, particularly in terms of the initiation and maintenance
of personal relationships. For example, Alpha's network (Figure 8-1A)
includes 5 couples who generally do not know or interact with one another.
Such a network structure would typify that of an individual (or couple)
whose network is composed largely of nonkin. In contrast, Beta's network
(Figure 8-1 B) includes a fully saturated clique, a subsection of the network
in which all members know and interact with one another. This network
structure would typify that of an individual (or couple) who maintains ties
with a number of kin relations (the clique members) as well as a number of
nonkin. One implication for this structural difference is that the members of
Beta's network, through their knowledge of each other, can coordinate aid to
Beta. Such aid might be relatively formal, as in the case of an engagement or
anniversary party, or it might be informal, as in the case of occasional visits
during an illness. The coordination of aid is prevented in Alpha's network
because the members are unknown to one another. Moreover, knowledge of
the need for support is facilitated by clique structure and inhibited for those
without highly interconnected networks. Thus, the total amount of support
available and the form that support may take is determined in part by
network structure.
As either a complement or alternative measure of density, several authors
have suggested examining the presence of cliques or clusters of individuals
within the network. Cluster was formally defined by Salzinger (1982) as
"three or more people who each have at least one friend in common and
who, in toto, have more ties within than outside the group" (p. 126). Less
stringent requirements for identifying cliques have been successfully
employed as well, and several analytical methods have been developed (see
Feger, 1981; Wellman, 1981b).
The question remains, however, under what circumstances should
researchers employ traditional measures of density or an alternative
measure such as clique structure. One resolution to this problem is to
consider each of these measures as estimates of a common underlying
theoretical construct, namely, structural interdependence, a point that i-s
explored in a later section of the chapter. It should become clear, however,
that density per se is not the sine qua non of network structure, although it
has often been viewed as such.
155
Overlap
Unlike size, density, or other measures of interconnectedness, network
overlap is fundamentally a dyadic measure referring to the proportion of
network members shared by two individuals. Prototypical would be the
overlap between the networks of a husband and wife or the overlap between
the networks of two friends. Researchers have hypothesized that as pairs
become increasingly interdependent in their personal lives, influencing
each others' thoughts and actions, they develop increasingly interdependent
social environments. Among dating couples, for example, both the absolute
number of mutual friends and the proportion of mutual, as compared with
separate friends, increases as couples become more deeply involved with
one another (Milardo, 1982). This process may be expected to generalize to
developing friendships as well, such that the networks of pair members
should become increasingly interdependent (that is, overlapping) as their
relationship deepens.
Commentary
Measures of network structure have been employed largely as a means of
describing the social environments of individuals or couples. Descriptive
analyses of network structure, including membership composition, stability
of membership over time, or even density, clique structure, and overlap, are
certain to vary with the definition of the network employed. Each of the
measures of network structure relies on the precise constituency identified.
A network defined in terms of the people considered important to a
respondent (that is, a network of significant others) will vary in its
constituency from a network defined in terms of the individuals with whom
interactions occur within a fixed period of time. In short, measures of
network structure are confounded with the conceptual and methodological
definition of the network employed. Although this confounding cannot be
fully resolved, its consequences can be minimized if the network is defined
in the broadest possible sense, including name-eliciting techniques that
sample divergent network sectors, ranging from close associates to peripheral or intermediate-level friends and acquaintances.
Expanding the sampling frame has the added advantage of avoiding the
implicit bias, present in much of the friendship literature, that close friends
(or kin) are somehow primary in the lives of individuals. This bias is
problematic because it presumes that (a) individuals have close associates,
(b) there is interindividual consistency in the interpretation of closeness,
and (c) peripheral associates are inconsequential. Each of these assumptions is arguable, and, perhaps more important for the present
discussion the structural parameters of an individual's network (for
example, its density or clique structure) are far more critical than is the
precise constituency or the relative proportion of close to peripheral
associates because composition is unlikely to reveal the constraint imposed
156
Robert M. Milardo
by a network's internal structure on individual choice. Descriptive work on

social networks is useful, but it typically underestimates the implications of
network structure as the latter attenuates personal choice.
Toward a Theory of Structural Interdependence

At the dyadic or relationship level of analysis, interdependence refers to the
effects interacting persons have on one another (cf. Kelley, 1979; Kelleyet
aI., 1983). Friends of long standing influence one another's thoughts as well
as actions, and they may alter their behavior to benefit the other. The motive
for action resides in the individual's assessment of the social psychological
field. Much of contemporary theory and research on close relationships
locates the locus of control (that is, the motive for action) within the
individual. Researchers have examined qualities such as physical attractiveness, similarity of values, and intimacy of self-disclosure to determine how they influence the individual's decision to continue a particular
relationship or perhaps initiate a new one. Friendship is thus a function of
personal choice. Such analyses, although central to theory construction,
neglect the influence of social structure on the character of personal
relationships.
When interdependence is viewed as a network quality, then essentially we
are concerned with either the form or the degree of interdependence among
an entire set of network relations. The degree of interdependence between
the network and individual reflects the potential for the social environment
to influence the selection of new friendships and the maintenance of
ongoing friendships. Two individuals, for example, may become friends not
because they are similar in attractiveness, but because they each have many
friends in common who also know one another (that is, highly interdependent networks). The network members may introduce the two
individuals and encourage them to become friends.
What is necessary, however, is a theoretical framework that links network
structure to social psychological events and processes. Such a framework
can be based upon the measures of network structure reviewed earlier,
namely, the degree of interconnectedness, density, clique structure, and
overlap. These measures are frequently considered nonequivalent and
irreducible properties of networks. They are nonequivalent in the sense that
each measure is uniquely defined, at least operationally, and each captures
a unique element or characteristic of a network's latent structure. This
nonequivalence may be rather trival from a theoretical standpoint, however.
They are irreducible and distinct qualities of networks in the sense that one
could know a great deal about the individual members of a network,
including personal attitudes and so on, as well as a great deal about the pair
relationships of members, including the frequency of interaction or patterns
157
of communication, but this knowledge would reveal nothing about the

social structure of which those individuals are a part.
The measures of interconnectedness, density, clique structure, and
overlap are conceptualized herein as unique, but intercorrelated, operational definitions of a common underlying feature of social networks
that we may refer to as structural interdependence. The structural meaning of
interdependence hinges upon the placement of pair relationships within the
network, including both the simple presence or absence of pair relationships (e.g., density) as well as the overall pattern of those relationships (e.g.,
clique structure). In contrast to measures of structural interdependence,
other measures of network structure focus on the placement of individuals,
in terms of the overall number of members (network size) or the patterning
of members in particular groups (network composition or homogeneity).
The importance of this distinction can be illustrated easily. The collective
influence of a network on an individual results from the relationships
between network members and not simply from the direct bonds linking the
target individual to each network member. If Alpha maintains relationships
with 10 individuals who are unknown to one another, the collective
influence of the network is minimal, although each network member may
independently influence Alpha. On the other hand, if Alpha maintains 10
relationships with network members who in tum know and interact with
one another, then collectively the network may have a profound impact on
Alpha's thoughts and behaviors, whereas in addition each member may
influence Alpha independently of all other members.
If we proceed from the assumption that personal relationships develop as
a result of personal choice operating within the constraint of social context,
then we can begin to look for the cause of behavior apart from inner
psychological processes and within the near environment. An upper
middle-class professional, for example, may prefer to associate with other
individuals of similar economic and occupational statuses, but if an
individual lives in a working-class neighborhood, friendship choice may be
constrained to others of somewhat dissimilar background. Huckfeldt's
(1983) investigation of the friendships of Detroit men demonstrated that
social activity is fundamentally structured by the social composition of an
individual's environment.
In short, although individuals demonstrate a preference for socially
similar others, the relative proportion of friends of similar background is
largely a function of both the opportunities and constraints imposed by the
environment. Individuals residing in a neighborhood dominated by a
particular class are more likely to develop friendships from that class,
regardless of their own class membership. The connection between liking
or, more broadly, friendship choice and similarity is mediated by contextual
factors, which can be defined in terms of the simple composition of an
individual's workplace or neighborhood or, more concretely, in terms of the
158
Robert M. Milardo
structural interdependence of the target individual's contemporary social

network. Analyses, traditional to social psychology, that center on the
probability of favorable outcomes, the specific outcomes of actual interaction, or the comparison of available opportunities are limited because
they attend to dyadic interdependencies while underestimating triadic or
group interdependencies based on the analysis of local social structure
(Wellman, 1981b).
In addition to the influence of the near environment, particularly as that
environment defines the field of potential network members, the act of
initiating a relationship creates a certain degree of social constraint in the
sense that individuals who become friends have little control over the
other's friends or the degree of interconnection between friends, either their
own or the other's. Should Alpha initiate a friendship with Beta, Alpha has
little control over the interconnectedness of his or her friends and limited
control over Beta's network, and so, too, Beta has little control over the
structure of Alpha's network. Yet all individuals are likely to be influenced
by their networks. In a sense, as Salzinger (1982) stated, "we create the
structures that constrain us" (p. 118). The concept of structural interdependence provides a means of describing and predicting the impact of
social constraint on individuals and their relationships.
Salzinger (1982) has advanced several hypotheses that capitalize on the
effects of network structure in the development of personal relationships.
These hypotheses may be restated and summarized in terms of the concept
of structural interdependence as follows:
1. The degree of structural interdependence will vary inversely with the number of
individuals each network member knows; that is, members of highly interdependent networks will know fewer individuals than members of minimally
interdependent networks. In the case of high-density networks, the ties
binding members limit the potential to develop associations with individuals outside the group, and such networks are apt to be relatively
homogeneous (Granovetter, 1983). New associations tend to develop within
the existing group structure and access to different groups is limited. In
contrast, members of low-density networks have a wider array of alternatives to develop relationships with members of nonoverlapping networks
(Granovetter, 1983; Wellman, 1981b).
In a study of the friendly associations of college-age men, members of
high-density clusters freely identified fewer network members than did
noncluster members (means = 4.95 and 7.53, respectively). This effect was
stable with regard to the total size of the network and the number of best or
intermediate friends identified (Salzinger, 1982). An inverse relationship
between size and clique structure might also be suspected to generalize to
additional measures of structural interdependence. For example, couples
with overlapping networks of associates should demonstrate smaller
networks than should couples with few mutual friends but proportionally
159
more separate friends. This connection is supported in a longitudinal study

of dating couples, where network overlap covaried inversely with the total
size of the network of each pair member (Milardo, 1982).
2. Networks relatively high in structural interdependence will become increasingly
connected over time and any new relationships formed will tend to originate within
the group. In the simplest case, this proposition suggests that networks will
become increasingly interconnected when individuals who are unknown to

one another are linked to a third party. For instance, the probability that
Alpha and Beta will develop a friendship is increased when they share a
mutual friend. If only through chance encounter, Alpha and Beta are apt to
meet and, as a result, the density of each person's network tends toward
saturation. Protypical here would be individuals who meet and subsequently develop a close relationship leading to marriage. The personal
networks of each pair member should demonstrate a tendency to merge into
a network of associations held in common, and this selection process
should be particularly evident where at least one individual is embedded in
a tightly knit network.
These interrelated hypotheses are supported by research concerning the
networks of both married and unmarried couples. The networks of married
individuals appear to be far denser than those of unmarried individuals.
Shulman (1975) reported in a cross-sectional study that 63% of the married
respondents reported high-density networks whereas only 39% of the
unmarried respondents did so. In addition, analyses based on a longitudinal study of dating individuals suggest that the relative proportion of
mutual friends held in common by pair members, as well as the absolute
number of mutual friends, increases as couples become more romantically
involved over time. Casually dating couples, for instance, who become more
deeply involved with one another, increase the proportion of mutual friends
by approximately 20% (Milardo, 1982). These effects appear to generalize to
nonromantic friendships as well. Sal zinger's (1982) study of friendship
demonstrated that members of densely knit clusters were far less likely to
develop new friendships than were individuals in loosely knit network
clusters, and the new friendships that cluster members did develop over
time tended to originate within the cluster.
3. Relationships within structurally interdependent networks will be relatively
enduring and stable, whereas relationships within loosely knit networks low in
structural interdependence will be comparatively less stable. The attraction of
pair members certainly contributes to the stability of their relationship, as

do additional properties, such as the amount of conflict or the availability of
alternative relationships. Here I argue, however, that networks exert a
structural force binding members to one another quite apart from individual or dyadic properties, such as the degree of interpersonal attraction,
conflict, or the comparison level for alternatives. Again, structural inter-
160
Robert M. Milardo
dependence is operationalized in terms of the degree of network density,

clique structure, or overlap.
In structurally interdependent networks, the interconnections binding
close friends carry over to other relationships, with the result that the close
friends are interconnected by virtue of their activity with each other in
addition to their activity with other individuals. Two friends are interdependent to the extent that they have shared in one another's life history
and expect to continue doing so in the future (Huston & Burgess, 1979;
Levinger, 1974; Plath, 1980), but they are also interdependent as a result of
the relationships held in common with network members. These individuals, either independently or collectively, may attempt to seek reconciliation between the conflicting pair. In a general sense, any change in the
network's status quo, including the termination of a relationship between
network members or the addition of a new network member, is likely to be
met with opposition from the network constituency, particularly in highly
interdependent networks (cf. Johnson & Milardo, 1984).
Johnson's (1978, 1982) theoretical analysis of commitment illustrates the
interdigitation of dyadic and structural interdependence. In addition to the
personal commitment of individuals to continue a particular relationship,
an evaluation based on personal preference and choice, Johnson (1982)
argued persuasively for the fundamental importance of social constraint as
a determinant of relationship longevity.
It is illusory that [relationship] quality explains more variance than
constraint-simply because, in most studies of dissolution, constraint does
not vary dramatically. The typical social psychological study of relationship dissolution investigates one type of relationship in one social location
at one point in history. It seems to make perfect sense that personal
inclinations rather than structural constraints apparently explain all the
variance that is explained. However, one needs to ask why friendships end
more often than kin relationships, why casually dating relationships end
more often than engagements, why divorce rates vary dramatically among
different countries, social classes, and states, and at different times in
history. One might try to argue that we like our friends less than our kin, our
fiance(es) more than our dates, and so on. Nevertheless, the more plausible
interpretations of such variations emphasize changes in constraints and
argue that it is these that produce the most dramatic effects on the social
and cognitive impact of dissolution. (p. 53)
A direct test of the stability of romantically involved couples (including

spouses) in the light of the structure of their networks has yet to be
accomplished. Prospective studies of marital dissolution are, of course,
difficult to conduct; however, several measures of dissolution potential have
been developed recently that attempt to measure a couple's likelihood of
divorce (Booth, Johnson, & Edwards, 1983; Weiss & Cerreto, 1980). I am at
the present writing in the process of gathering extensive data from couples
married between 5 and 10 years that will allow the investigation of the
interrelations of network structure and dissolution potential.
161
The effects of structural interdependence on the stability of friendships is

apparent, however. Membership in a dense clique has a stabilizing
influence in that clique members are far less likely to either lose or gain
friends than are noncluster members (Salzinger, 1982). This finding is
particularly startling because changes in friendship choice were monitored
over a relatively short period of time, approximately 3 months, and a great
deal of change might not be expected. Nonetheless, Salzinger found that
90% of the noncluster members lost or gained friends during the course of
the study. In contrast, 58% of the cluster members reported no change in
their friendship choices, and those that did change alliances tended to gain
previously unknown cluster members as friends and dissolve friendships
with noncluster members. These data represent strong support for the
contention that structure influences stability.
4. The structural interdependence of a network will influence the internal
character of personal relationships such that both the diversity and intensity of
dyadic interdependencies will be greater in tightly knit networks relative to loosely
knit networks. The degree of structural interdependence of a network
influences the intensity of the constituent personal relationships in a

number of ways. Friends have a greater opportunity to share information
about one another if the constituencies of their respective networks are
similar, in other words, overlapping. Much of what we know about our
friends is not necessarily the result of direct communication (self-disclosure)
but results from what others tell us about our friends. Interdependent
networks simply increase the potential for learning about our friends
indiretly through communications with third parties. These events may
increase the intensity of a pair relationship as well as the diversity of
knowledge on which that relationship is based.
Goodstein and Russell (1977) investigated self-disclosure patterns among
friendship and kinship dyads. College students were asked to complete
detailed reports of their self-disclosures to people whom they considered
significant and important in their lives. In a second phase of the study, the
significant others (a friend or relative) identified by subjects reported on
their perceptions of the subject's self-disclosure to them. The findings
demonstrated that subjects consistently underestimated the amount of
knowledge significant others had about them. The question of how the
discrepencies between self-reports and the reports of others originate
remains unanswered. Are subjects simply misperceiving or misreporting the
amount of disclosure to significant others, as Goodstein and Russell
suggested, or do significant others glean important knowledge from
communications with other network members who also know and interact
with the subject? The latter process seems quite tenable and perhaps
represents a rather common means by which people learn of one another's
personal dispositions, interests, and so on. This process would be far more
162
Robert M. Milardo
likely to occur in a network where many of the members know and interact
with one another. Thus, the intimacy of a particular relationship is
determined not only by what pair members share directly with one another,
but also by the proportion and absolute number of friends they have in
common, and what they intentionally or unintentionally reveal to these
individuals.
The degree of structural interdependence should influence the ongoing
character of personal relationships by facilitating the coordination of social
support and social interference to members. Within highly dense and
interdependent networks, friends are able to judge each other's circumstance and thereby coordinate aid to a third party when that individual is
thought to be in need. A classic example here is a baby shower, where an
expectant mother is lavished with gifts and advice from her network. The
amount and type of social support and social interference, as well as the
relative proportion of each, may have important links to the structure of the
target individual's network.
In addition, Sal zinger (1982) has argued that pair relationships in
structurally interdependent networks are apt to be close, multidimensional,
and reciprocal. In fact, among individuals who identified associates both
within and outside of dense clusters, the closest associate is far more likely
to originate within the cluster than outside of it, and these associations are
more likely to involve reciprocal friendship choices and serve a variety of
functions. Paradoxically, however, associations within dense clusters may
be high in conflict as well as closeness simply by virtue of the intensity and
breadth of the structural interdependencies (cf. Braiker & Kelley, 1979;
Fiske & Beattie, 1981).
Concluding Commentary
The concept of social network provides a useful vehicle through which the
environmental context of personal relationships may be examined. In the
past, researchers have generally centered on descriptive analyses of personal
networks or more broadly based analyses of social structure. When network
analysis is applied to the initiation, development, and termination of
personal relationships, a number of new avenues of research appear.
Certainly relationships are as much a function of person-to-network
interdependencies as they are person-to-person interdependencies, a
position that represents a radical departure from contemporary social
psychological theory and research.
Typically, investigators of personal relationships seek explanations of the
development of friendships in the personal attributes of the pair members
or in attributes of interaction proces&es or outcomes. Measures of personal
dispositions, attributions of causality, outcome interdependencies, and
relative and absolute comparison levels for alternatives all reflect a person-
163
or relationship-centered focus. There is little doubt that an understanding of

the formation, ongoing character and eventual termination of personal
relationships has been and will continue to be advanced by studies of
individual and dyadic properties. Individuals and their relationships,
however, are embedded within a social system, a system that profoundly
influences people, their availability to one another, the choices they make
with regard to one another, and the character of their relationships.
Network analysis provides a means to specify pertinent features of local
social structure through measures of network interconnectedness, density,
clique structure, and overlap. These structural features of networks can in
turn be thought of as measures of a common underlying dimension of
networks, namely, structural interdependence. Throughout this chapter I
have maintained that the degree of structural interdependence influences
the size of personal networks, the source of new relationships, and the
longevity of ongoing relationships as well as internal qualities, such as the
intimacy or multidimensionality of pair relationships.
Acknowledgements. Thanks are expressed to Catherine Surra and Michael Johnson
for their critical reviews of an earlier draft of this chapter. Support for the
preparation of this chapter was provided by the Faculty Development Fund of The
University of Maine and the Maine Agricultural Experiment Station (MAES Pub.
No. 1063).
References
Barnes, J.A (1954). Class and committees in a Norwegian island parish. Human
Relations. 7, 39-58.
Belle, D.E. (1982). The impact of poverty on social networks and supports. Marriage
and Family Review. 5,89-103.
Boissevain, J. (1974). Friends offriends. Oxford: Basil, Blackwell.

Booth, A, Johnson, D., & Edwards, J.N. (1983). Measuring marital instability. Journal
of Marriage and the Family. 45,387-394.
Braiker, H.B., & Kelley, H.H. (1979). Conflict in the development of close
relationships. In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing
relationships (pp. 135-168). New York: Academic.
Burns, G.L., & Farina, A (1984). Social competence and adjustment. Journal of Social
and Personal Relationships. 1, 99-113.
Cook, M. (1977). The social skill model and interpersonal attraction. In S.W. Duck
(Ed.), Theory and practice in interpersonal attraction. New York: Academic.
Davis, KE. (1981, August). Friendship thru the life-cycle: Age. gender, and race effects on
the number and quality of personal relationships. Paper presented at the annual
meeting of the American Psychological Association, Los Angeles.
Dickens, W.J., & Perlman, D. (1981). Friendship over the life-cycle. In S. Duck & R
Gilmour (Eds.), Personal relationships. 2: Developing personal relationships (pp. 91122). London: Academic.
Feger, H. (1981). Analysis of social networks. In S. Duck & R Gilmour (Eds.),
Personal relationships. 1: Studying personal relationships (pp. 91-105). London:
Academic.
Fehr, B., & Perlman, D. (1985). The family as a social network and support system. In
L. L'Abate (Ed.), Handbook offamily psychology and therapy. Vol. 1. (pp. 323-356).
Homewood, IL: Dow Jones-Irwin.
164
Robert M. Milardo
Fischer, C.S. (1982). To dwell among friends: Personal networks in town and city.
Chicago: University of Chicago Press.
Fischer, C.S., Jackson, RM., Stueve, CA, Gerson, K, & Jones, L.M. (with
Baldassare, M.). (1977). Networks and places. New York: Free Press.
Fischer, C.S., & Dliker, SJ. (1980). Friendship, sex, and the life-cycle (Working Paper
No. 318). Berkeley: University of California, Institute of Urban and Regional
Development.
Fischer, C.S., & Phillips, S.L. (1982). Who is alone? Social characteristics of people
with small networks. In L.A Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of current theory, research, and therapy (pp. 21-39). New York: Wiley
Interscience.
Fiske, S.T., & Beattie, AE. (1981, August). Affect and information: Schema-triggered
affect in the initiation of close relationships. Paper presented at the annual meeting of
the American Psychological Association, Los Angeles.
Goodstein, L.D., & Russell, S.W. (1977). Self-disclosure: A comparative study of
reports by self and others. Journal of Counseling Psychology. 24,365-369.
Granovetter, M. (1983). The strength of weak ties: A network theory revisited. In R
Collins (Ed.), Sociological theory (pp. 201-233). San Francisco: Jossey-Bass.
Huckfeldt, RR (1983). Social context, social network, and urban neighborhoods:
Environmental constraints on friendship choice. American Journal of Sociology, 89,
651-669.
Huston, T.L., & Burgess, RL. (1979). Social exchange in developing relationships:
An overview. In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing
Huston, T.L., & Levinger, G. (1978). Interpersonal attraction and relationships. In
M.R Rosenzweig & L.W. Porter (Eds.), Annual Review of Psychology, 29, 115-156.
Johnson, M.P. (1978). Personal and structural commitment: Sources of consistency in the
development of relationships. Unpublished manuscript, Pennsylvania State University, Department of Sociology.
Johnson, M.P. (1982). Social and cognitive features of the dissolution of commitment
to relationships. In S. Duck (Ed.), Personal relationships:4 (pp. 51-73). New York:
Academic.
Johnson, M.P., & Leslie, L. (1982). Couple involvement and network structure: A test
of the dyadic withdrawal hypothesis. Social Psychology Quarterly, 45, 34-43.
Johnson, M.P., & Milardo, RM. (1984). Network interference in pair relationships: A
social psychological recasting of Slater's theory of social regression. Journal of
Jones, L.M., & Fischer, C.S. (1978). Studying egocentric networks by mass survey
(Working Paper No. 284). Berkeley: University of California, Institute of Urban
and Regional Development.
Kandel, D.B. (1978). Similarity in real-life adolescent friendship pairs. Journal of
Kelley, H.H. (1979). Personal relationships: Their structure and processes. New York:
Wiley.
Kelley, H.H., Berscheid, E., Christensen, A, Harvey, IH., Huston, T.L., Levinger, G.,
McClintock, E., Peplau, LA, & Peterson, D.R (1982). Close relationships. San
Francisco: Freeman.
Killworth, P.O., Bernard, H.R, & McCarty, C. (1984). Measuring patterns of
acquaintanceship. Current Anthropology, 25,381-397.
Larson, RW. (1984, July). States of consciousness in personal relationships: A life-span
perspective. Paper presented at the second International Conference on Personal
Leavy, RL. (1983). Social support and psychological disorder: A review: Journal of
Community Psychology, 11,3-21.
165
Levinger, G. (1974). A three-level approach to attraction: Toward our understanding

of pair relatedness. In T.L. Huston (Ed.), Foundation of interpersonal attraction (pp.
100-120). New York: Academic.
Lewin, K (1935). A dynamic theory of personality: Selected papers of Kurt Lewin. New
York: McGraw-Hill.
Lewin, K (1951). Field theory in social science: Selected theoretical papers. New York:
Harper & Row.
Milardo, RM. (1982). Friendship networks in developing relationships: Converging
and diverging social environments. Social Psychology Quarterly, 45, 162-172.
Milardo, R.M. (1983). Social networks and pair relationships: A review of substantive
and measurement issues. Sociology and Social Research, 68, 1-18.
Milardo, RM. (1984). The social participation of spouses in distressed and nondistressed
marriages. Paper presented at the second International Conference on Personal
Milardo, RM., Johnson, M.P., & Huston, T.L. (1983). Developing close relationships:
Mitchell, J.e. (1969). The concept and use of social networks. In J.e. Mitchell (Ed.),
Social networks in urban situations (pp. 1-50). Manchester, England: University
Press.
Newcomb, T.M. (1961). The acquaintance process. New York: Holt, Rinehart, &
Winston.
Nezlek, IB., Wheeler, L., & Reis, H.T. (1983). Studies of social participation. In H.T.
Reis (Ed.), Naturalistic approaches to studying social interaction, No. 15 (pp. 57-73).
San Francisco: Jossey-Bass.
Niemeijer, R (1973). Some applications of the notion of density. In IF. Boissevain &
Ie. Mitchell (Eds.), Network analysis (pp. 45-66). The Hague: Mouton.
Plath, D.W. (1980). Contours of consociation: Adult development at discourse. In P.
Baltes & O. Brim (Eds.), Life-span behavior and development (Vol. 3). New York:
Academic.
Reis, H., Wheeler, L., Spiegel, N., Kernis, M., Nezlek, J., & Perri, M. (1982). Physical
attractiveness in social interaction: II. Why does appearance affect social
experience? Journal of Personality and Social Psychology, 43, 979-996.
Ridley, CA, & Avery, AW. (1979). Social network influence on the dyadic
relationship. In RL. Burgess & T.L. Huston (Eds.), Social exchange in developing
Sal zinger, L.L. (1982). The ties that bind: The effects of clustering on dyadic
relationships. Social Networks, 4, 117-145.
Saulnier, K (1982). Networks, change, and crisis: The web of support. Canadian
Journal of Community Mental Health, 1,5-23.
Shulman, N. (1975). Life-cycle variations in patterns of close relationships. Journal of
Tesch, S.A (1983). Review of friendship development across the life-span. Human
Development, 26, 226-276.
Thoits, PA (1982). Conceptual, methodological, and theoretical problems in
studying social support as a buffer against life stress. Journal of Health and Social
Behavior, 33, 145-159.
Verbrugge, L.M. (1977). The structure of adult friendship choices. Social Forces, 56,
576-597.
Weiss, RL., & Cerreto, M.e. (1980). The marital status inventory: Development of a
measure of dissolution potential. American Journal of Family Therapy 8, 80-85.
Wellman, B. (1981a). Applying network analysis to the study of social support. In
B.H. Gottlieb (Ed.), Social networks and social support (pp. 171-200). Beverly Hills,
CA: Sage.
166
Robert M. Milardo
Wellman, B. (198Ib). Network analysis from method and metaphor to theory and
substance (Working Paper No. IB). Toronto: University of Toronto, Department of
Sociology, Structural Analysis Programme.
White, H.C., Boorman, S., & Breiger, R. (1976). Social structure from multiple
networks: I. Blockmodels of roles and positions. American Journal of Sociology, 81,
740-780.
Chapter 9
School Organization and

Interracial Friendships
Maureen T. Hallinan
The primary rationale for desegregating the American public schools is to

provide equal educational opportunities for all students regardless of race or
ethnicity. A related aim is to reduce racial hostility among youth and,
eventually, members of adult society. Despite disagreement over how best to
accomplish school desegregation, lawmakers and educators generally have
remained committed to these two goals, believing that desegregation is an
effective tool to reduce racism in American society.
Unfortunately, although considerable empirical research exists on school
desegregation, much of the research is not designed to evaluate directly the
success of desegregation in attaining these goals. For example, there have
been few longitudinal studies that compare the growth in achievement of
black students in segregated and integrated schools. Nor are there many
studies that compare the racial attitudes of black and white students or the
extent to which they interact prior to and after desegregation. Moreover, the
results of existing studies, especially those concerned with racial attitudes
and behavior, are often contradictory.
The strong commitment of educators to school desegregation and the
dearth of convincing empirical research to support the desegregation effort
point to a need to rethink how desegregation is linked to the cognitive and
social outcomes of students. This is particularly true of the effects of
desegregation on the social relations of black and white students because
theoretical and empirical research in this area is even less systematic than
studies of black and white achievement in racially mixed schools.
The existing research is generallly conducted from a psychological
perspective in which individual level determinants of student behavior are
emphasized. Although understanding individual effects on interracial
friendliness is of considerable importance, it is only one way to approach
the study of black and white peer relations in a school setting. A sociological
perspective focusing on structural or organizational determinants of
interracial friendliness should be equally fruitful. The absence of such a
168
Maureen T. Hallinan
perspective in much of the literature on desegregation motivates the analysis

presented here.
In this chapter I analyze organizational determinants of student interracial friendliness in desegregated schools. Interracial friendliness is more
narrowly defined than is interracial sociability; it is seen as a measure of
persistent positive sentiment between black and white students. The
occasional friendly interaction between blacks and whites connoted by the
term interracial sociability that demonstrates the absence of strong prejudice
is not of interest here.
The organizational perspective on interracial friendliness is important for
three reasons. First, it identifies constraints and limitations on interracial
interactions imposed by school and classroom organization. Thus, it defines
the pool of potential cross-race friends available and the likelihood of
interaction with those peers. Second, the organizational perspective reveals
factors that affect the stability of interracial friendship over the school year.
In so doing, it permits predictions about the duration and possible
significance of cross-race friendship relations. Third, the organizational
perspective describes factors that influence the content of interracial
interaction, and that, in turn, are likely to affect the probability of interracial
friendliness. The understanding of interracial friendliness provided by an
organizational perspective should complement the individualistic perspective currently dominating the literature. And, of course, this organizational
or structural approach to interracial friendliness can be equally useful in
analyzing any type of friendship or social relationship.
Theoretical and Empirical Research on

Interracial Friendliness
When the massive effort to desegregate the public schools began in the
1960s, many educators and parents believed that merely placing black and
white students together in the same school would initiate a process of
interracial contact that would lessen hostility, alter negative stereotypes, and
lead to interracial friendships. This belief, referred to as the natural
progression assumption, is contradicted by a number of empirical studies.
For example, Shaw (1973) analyzed longitudinal sociometric data from
fourth-, fifth-, and sixth-grade students in a school that had recently been
desegregated and concluded that contact with other-race peers led to less
acceptance. Similarly, Gerard and Miller (1975) analyzed sociometric data
from a large sample of black, white, and Mexican students in several
desegregated schools over a 2-year period. They found little evidence that
real integration took place and concluded that ethnic and racial cleavage
became more pronounced over the period of the study.
In contrast to the natural progression assumption, Allport (1954), in his
contact theory argued that intergroup contact may reinforce previously held
School Organization and Interracial Friendships
169
stereotypes and increase intergroup hostility unless the situation is structured in such a way that it provides equal status for minority and majority
group members and provides strong institutional support for positive social
relations. Allport also emphasized the importance of cooperative interracial
interaction aimed at attaining shared goals. He claimed that these conditions must be met before intergroup relations will improve.
Although contact theory has not been systematically tested in desegregated schools, some empirical evidence provides post hoc support for its
propositions. For example, Patchen (1982), in his study of students in 12
integrated secondary schools, found that friendly contact between the races
increased when the school officials made interracial contact rewarding by
having students of both races cooperate on common tasks that lead to
common goals. Schofield and Sagar (1977) studied the seating patterns of
students in the school cafeteria of a desegregated middle school. They found
that interracial interaction increased over the school year for seventh-grade
students who were not tracked but decreased for the eigth-grade students
who were assigned to either a predominantly white accelerated track or a
predominantly black regular track. A follow-up study (Schofield & Sagar,
1977) the following year replicated the eigth-grade findings when ,the
previously integrated seventh graders became more segregated after being
tracked by ability in the eigth grade.
Pettigrew (1975) extended the conditions of Allport's contact theory by
arguing that a desegregated school should provide equal access to sources of
social status as well as to physical facilities and materials. Consequently, a
school's faculty should be integrated and positions of leadership should be
distributed equally across racial groups. In addition, he claimed that school
authorities must communicate a positive attitude toward school integration.
Virtually no empirical research is available that directly tests hypotheses
derived from Pettigrew's formulation.
In an application of expectations states theory (Berger, Cohen, &
Zelditch, 1972), Cohen (1972) argued that race is a status characteristic in
American society that creates expectations about competence that lead to
dominance and superior performance by the higher ranked groups. Unlike
Allport, Cohen argued that a cooperative equal status environment is not
sufficient to improve interracial interactions because biased expectations
for the behavior qf both blacks and whites will lead whites to continue to
dominate blacks. In order to equalize status truly, Cohen claimed, the roles
of superior and subordinate must be reversed, at least temporarily.
Several empirical studies (Cohen & Roper, 1972; Riordan & Ruggiero,
1980; Yuchtman-Yaar & Semyonov, 1979) support Cohen's arguments and
show that lower status blacks, other ethnic groups, and females participate
more in interactions with higher status white males when they have been
prepared for the interactions by being equipped with superior knowledge.
This is referred to as interaction disability treatment. The long-term effects
of this treatment have not been determined nor have creative methods of
170
Maureen T. Hallinan
equalizing status through alternate classroom practices been established

(see Cohen, 1980).
A few other theoretical approaches to the study of race relations can be
found in the literature. These include social comparison processes (Pettigrew, 1967), reinforcement theory (Hauserman, Whalen, & Behling, 1973),
attribution theory (Duncan, 1977), and social identity theory (Tajfe1, 1982;
Turner, 1981). These theories, like contact theory and expectations states
theory, have been used to explain, and occasionally to predict, interracial
friendliness in desegregated schools. All of these theories focus on
individuals interacting in groups. In some cases the context of the
interaction is seen as a key factor in explaining interracial behavior.
In general, these social psychological theories provide some important
insights into the nature of race relations in desegregated schools. At the
same time, they are limited by their failure to explain or predict when the
social psychological processes they describe can be expected to occur. The
study of students' race relations would benefit from a conceptualization that
specifies the conditions that generate the social psychological processes that
govern interracial sociability.
&. the conceptualization presented in this chapter, I adopt a sociological
perspective and aim to provide a theoretical framework that outlines the
conditions under which the social psychological processes discussed earlier
are believed to operate. The analysis is focused on structural and organizational factors of schools and classrooms that influence the formation of
interracial friendships. I argue that school and classroom racial composition and the organization of instruction affect both opportunities for
interracial interaction and the status hierarchy of the classroom. The reward
structure of a classroom also defines status similarities and differences.
These factors, in turn, influence the frequency, nature, and content of social
interaction. They shape an environment that sets in motion the social
psychological processes that govern interracial friendliness. By so doing,
they provide a context that either fosters or hinders positive social relations
between blacks and whites.
An examination of the effects of organizational characteristics of
desegregated schools on student social relations has been seriously neglected. The few empirical studies that are available are fairly atheoretical.
These generally consider the effects on race relations of one of three
organizational factors, namely, racial composition, instructional practices,
and reward structure.
In some recent empirical studies, investigators have examined the effects
of school racial composition on cross-race friendships. St. John and Lewis
(1975) found that the percentage of other race students in a class had a
positive effect on interracial popularity for boys but that black girls were
more popular with whites in majority white classrooms. In an analysis of a
large longitudinal data set obtained from 5,479 students in 12 secondary
171
schools, Patchen (1982) reported descriptive information showing that

whites had more friendly interracial contact as the percentage of blacks
increased. Blacks became less friendly toward whites (measured by
avoidance, aggression, etc.) as the proportion of black students increased
from a small to a large minority and then became more friendly as the
proportion of blacks increased to a large majority. In a sample of 20
desegregated classrooms in six schools, Hallinan (1982) found that both
blacks and whites were more integrated in their friendships in majority
white classrooms and that segregation by whites decreased over the school
year, whereas black segregation remained constant.
The results of these studies are not entirely consistent, but neither are the
studies directly comparable. St. John and Lewis (1975) ignored possible
differential levels of friendliness between blacks and whites. Hallinan (1982)
controlled for the greater friendliness of blacks but did not include
individual level characteristics of students in her model. Patchen's (1982)
study was based on observational and interview data and provides a
descriptive rather than inferential analysis.
In examining the effects of pedagogical practices on interracial sociability, Damico, Bell-Nathaniel, and Green (1981) compared the number of
cross-race friendships of students in middle-school classes that were
organized into mUlti-age, heterogeneous teams for instruction with the
friendships of students in classes without teams and found that white pupils
in team-organized schools had significantly more cross-race friends than
did whites in traditionally organized schools. Blaney et al. (1977) found that
fifth-grade students who were assigned to small peer-tutored groups liked
their other-race peers more than did students not assigned to groups. In
Schofield and Sagar's (1977) study of the seating patterns of seventh- and
eigth-grade students in a school cafeteria, racial cleavage was found to
increase over the school year for the eighth-grade students who were tracked
for instruction.
A few researchers have examined the effects of classroom reward
structure on race relations. These investigators generally looked at the effect
of cooperative learning settings on achievement and on social behavior and
tended not to define reward structure explicitly or look at it directly. They
tended to find that cooperative settings are more conducive to positive
interracial sentiment than are competitive ones. Reviews of this research are
found in Slavin (1980) and Sharan (1980).
The three structural or organizational characteristics examined in these
studies (racial composition, organization of instruction, and reward structure of the classroom) have been found to have some effect on interracial
sociability. In the remainder of this chapter, I examine how these three
factors are linked to interracial friendliness in desegregated classrooms.
These organizational characteristics create the environment in which the
social psychological processes governing interracial behavior occur.
172
Maureen T. Hallinan
Organizational Characteristics of Schools and

Interracial Friendships
Racial Composition
In early studies in the school-effects tradition, researchers examined the
effects of school racial composition, measured as the percentage of black or
white students in a school, on student achievement or aspirations (e.g.,
Armor, 1972; Cohen, Pettigrew, & Riley, 1972; St. John & Smith, 1969). These
investigators argued that the percentage of white students in a school affects
normative or reference group processes that influence student behavior.
Although these arguments are convincing, measuring racial composition at
the school level likely does not appropriately portray the racial composition
of the immediate subset of students most relevant to a black or white child.
A more reasonable measure is the racial composition of the class, especially
in elementary schools where classes are self-contained and students remain
with the same classmates all day.
The most obvious way that the racial composition of a class affects
student sociability is by defining the pool of other-race peers with whom a
student can interact within the classroom. The more cross-race peers that
are present in a class, the more likely it is for a student to interact with peers
of the other race merely by chance. Because interaction frequently leads to
friendly relations (Homans, 1950), the more other-race peers available in a
class, the more likely the formation of cross-race friendships.
Unless a classroom is racially balanced, classroom racial composition
differentially impacts on the interactions of black and white students. In a
majority white classroom, black students have more opportunities to
interact with whites by chance than whites have for black contact. Similarly,
in majority black classes, whites have more occasion for cross-race
interaction than do blacks. For example, in a class with twice as many
whites as blacks, the black students are twice as likely to interact with whites
by chance than the white students are to interact with blacks. This indicates
that, when blacks are in the minority, proportionately more blacks may
develop friendly relations with whites than whites will with blacks.
The differential impact of racial composition on black and white students
is also consequential because it can affect the development of different
norms regarding the acceptability of interracial interaction. If, for example,
in a majority white classroom, half of the blacks are friendly with whites,
then blacks may define cross-race interaction as acceptable. However, if
only 25% of the whites are involved in cross-race exchanges, whites may
continue to view interracial interaction as deviant behavior. The reverse
would be true, of course, if blacks were in the majority. If only 25% of the
blacks in a class were involved in cross-race exchanges compared to 50% of
the whites, blacks may be more likely than whites to perceive interracial
173
interaction as deviant. This effect of class composition On interracial contact

should be most extreme when One race is in the overwhelming majority and
should disappear in classes that approach racial balance.
Some recent evidence is available that supports the opportunity hypothesis, that is, the prediction that opportunity for interracial interaction
leads to positive social relations between blacks and whites. In a study of 18
desegregated elementary classrooms, Hallinan and Smith (1985) found that
percentage of blacks in a class had a significant positive effect on the
likelihood that a student would choose a cross-race peer as a best friend.
The results held for both black and white students at several points over the
school year. In addition, the study showed a weak effect of percentage of
black students On change from a weak friendship tie to a closer One for
blacks and whites.
Classroom racial composition may constrain interracial interaction in
another significant way. If teachers decide to create small groups within the
classroom for instructional or other purposes, they have the option of
making the groups homogeneous or heterogeneous with respect to race. If
racially mixed groups are formed, then the composition of the groups is
likely to reflect the classroom racial composition. Because the probability of
interacting with any other group member by chance is greater in small
groups than in the whole class, the racial composition of the group affects
the likelihood of cross-race interaction and, consequently, influences
friendship formation. Moreover, within-group norms governing interracial
contact may be influenced by the proportion of each race in the group
involved in cross-race contact. If the teacher forms groups that are racially
homogeneous, then the impact of the racial composition of the class On
interracial contact is reduced.
Another mechanism through which classroom racial composition affects
interracial interactions and sociability may be referred to as a sociability
threshold or a fund of sociability (Nelson, 1966). Due to such variables as
time constraints, social preference, and past habits, students may be
predisposed to having a certain number of acquaintances or friends and
may continue to make friends until they reach this limit. In classes that are
racially imbalanced, if students do not have enough same-race peers to
interact with or choose as friends, they may select classmates of the other
race in order to reach their sociability threshold. The more other-race peers
available, the greater the likelihood of students' finding cross-race peers
possessing the characteristics that they value in a friend. Again, this implies
that students in the minority will demonstrate greater interracial friendliness than will those in the majority because they are more likely to need to
select other-race friends to reach their friendship threshold.
These arguments all pertain to the effect of classroom racial composition
on opportunities for interracial contact. They lead to the hypothesis that as
the ratio of other-race to same-race peers in a classroom increases, the
number of cross-race interactions and friendships in which a student is
174
Maureen T. Hallinan
involved increases. I note that because most desegregated classrooms have a

white majority, an implication of the differential impact of classroom racial
composition on students' sociability is that black students are less likely to
segregate in their interaction and friendship patterns than are white
students.
Class racial composition is also linked to interracial friendliness through
social status. Two factors affect the social status of students in desegregated
classrooms. The first is the presence of racism and, in particular, prejudice
against blacks, and the second is the ratio of one race to another. The
dynamics of racism are built on prejudice and negative stereotypes. When
these beliefs are activated, the group that is discriminated against tends to be
rejected and is either ignored or becomes an object of hostility (Allport,
1954). Prejudice makes the discriminating group a social threat to the other
students, who may react by becoming aggressive and hostile themselves or
by withdrawing and relying on their same-race peers for social support. The
process is cyclical because, as Katz (1964) argued, prejudice lowers the selfesteem of the group that is discriminated against, which leads to further
social rejection. Consequently, prejudice creates a status difference between
black and white students. Because the classroom is likely to mirror the
prejudices and biases of society, white students are likely to be the
discriminators and have higher status.
The ratio of black to white students in a desegregated classroom may
affect the social status of both groups. Ehrlich (1973) argued that when a
racial minority is small, it has greater visibility, which results in greater
consensus about stereotypes. Hyman (1969) claimed that stereotypes
become more exaggerated when a minority group is very visible. Resulting
in students' inability to perceive similarities in classmates of the other race.
St. John and Lewis (1975) argued that being in the racial minority in a
classroom is a conspicuous and possibly defenseless position that increases
students' identification with their same-race peers and decreases their
interactions with the other racial group.
This reasoning suggests that in majority white classes, the lower status of
blacks is due to their being in the numerical minority, as well as to
prejudice. At the same time, the higher status of whites obtained through
discrimination is augmented by the social power and control that accompanies being in the majority. Consequently, status differences and the
resulting racial segregation should be greatest in majority white classes. An
exception might be in classes with a very small number of blacks, where
they are likely to be perceived more as individuals than as a racial group. In
majority black classes, the relative status of blacks and whites is more
complex. In this situation the higher status of whites obtained through
discrimination is diminished by their minority position, whereas the lower
status of blacks is increased by their majority position. This adjustment in
status should produce greater interracial interaction and sociability in
majority black classes than in majority white classes.
175
The opportunity hypothesis seems to contradict the social status argument. According to the former, the more other-race peers available, the
friendlier students will be toward that race. The social status argument
implies that blacks will be less friendly toward whites in majority white
classes, and, to a more limited degree, whites will be less friendly toward
blacks in majority black classes, because of the status differences between
the races. The two positions likely can be reconciled only when other factors
are taken into account. For chance interracial interactions to occur, a
teacher must allow students a certain degree of freedom to mingle, converse,
and work together on class assignments. If a teacher's pedagogical practices
are a barrier to chance contact, then the opportunity hypothesis is not likely
to be supported, whereas if teachers do not constrain student interactions,
the number of other-race peers in a classroom or within a class group may
be a factor in the initiation of cross-race interactions. At the same time, the
ratio of one race to another is expected to affect the social status of blacks
and whites only in classrooms where status differences are significant to
start with, making race an important determinant of sociability. Consequently, classroom characteristics and teacher pedagogical practices may
influence the extent to which each factor, availability and relative status,
affects cross-race interactions and friendships.
Instructional Grouping
A second organizational factor affecting the interracial friendliness of
students in desegregated classrooms is the assignment of students to
teacher-led instructional groups or peer-work groups within the classroom.
The most common basis for assignment to teacher-led groups is student
ability. The rationale for creating ability groups is to enable teachers to
direct the pace and content of instruction to the level of student aptitudes
and to minimize management and discipline problems during instruction.
I argued earlier that the racial composition of an instructional group
affects the likelihood that black and white students in a class interact by
chance. An association between ability and race can affect the racial
composition of an ability group. Because blacks generally rank lower than
whites in the achievement distribution of a class, high-ability groups tend to
be predominantly white, and low-ability groups have a larger proportion of
blacks. Consequently, ability grouping should limit opportunities for black
and white students to interact on task-related or other activities. However,
other factors lessen the potential negative impact of ability grouping on
interracial contact. Hallinan and Sorensen (1983) argued that organizational constraints (such as the physical layout of the classroom, number
of books and instructional materials, and norms governing the equal
allocation of teacher time across groups) lead teachers to form approximately three equal-sized groups rather than a larger number of smaller,
more homogeneous groups. Because large groups are more likely to be
176
Maureen T. Hallinan
racially mixed than smaller groups, these constraints on the assignment

process result in more interracial interaction than would occur if more
groups were formed.
In addition to ability, it is possible that ascribed characteristics of students
affect their assignment to ability groups. Sorensen and Hallinan (l984a)
examined the role of race in the formation of ability groups in a sample of
759 students in 32 desegregated classrooms. Their analysis showed that
although there was no direct effect of race on the assignment process, race
did influence the formation of groups in such a way that high-ability groups
were larger in desegregated classrooms. This increases the chances of at
least some black students being assigned to the high-ability group. Not only
does membership in the high-ability group improve students' growth in
achievement (Sorensen & Hallinan, 1984b), but it assigns them a high rank
in a very visible status hierarchy based on ability. This may evoke the
esteem of a student's other-race peers and result in friendly interactions
between black and white pupils. If peer-tutoring is used as a pedagogical
device in ability-grouped classes, being in the high-ability group should
increase the probability that a black student will be relied on for
instructional assistance, which provides valuable opportunities for positive
interactions on task-related activities.
In short, both organizational constraints on the number and size of
ability groups formed in a class and the effect of race on the assignment of
students to the high-ability group lessen the tendency of ability grouping to
segregate students by race for instruction. Nevertheless, despite these two
factors, ability groups will rarely be as racially mixed as the class as a whole,
so the practice places some limitations on black and white interactions and,
consequently, on interracial friendship formation. Schofield (1982) provided striking empirical evidence of the segregating effect of ability
grouping in a desegregated middle school and pointed as well to the
variance across classes in the degree to which the group racial composition
fails to mirror the racial composition of the class as a whole.
When students are assigned to the same-ability group, a number of social
psychological processes can affect their interpersonal interactions. First,
ability grouping emphasizes the existing academic similarities among those
assigned to the same group and creates new similarities as students work
together on assigned tasks and share the same learning environment.
Because similarity is one of the strongest bases of interpersonal attraction,
students in the same-ability group are more likely to become friends than
are those in different groups. Moreover, membership in the same-ability
group assigns the same academic rank to students. If status differences are a
barrier to friendship, then friendly relations are more likely to develop
within an ability group than across groups or in ungrouped classes. In a
study of 32 classes that were ability grouped for reading, Hallinan and
Sorensen (1984) found a significant positive effect of assignment to the
same-ability group on the formation of student friendships.
177
Another way of organizing students for instruction is the peer workgroup. This method is employed much less frequently at the elementary
level than are teacher-led groups. Stodolsky's (1984) exhaustive typology of
peer work-groups includes the following types: completely cooperative,
cooperative, helping obligatory, helping permitted, and peer tutoring. In the
first two types, students work together toward a common goal, whereas in
the latter three types, students may help their peers attain their goals, which
mayor may not be the same as their own. Peer work-groups are almost
always heterogeneous with respect to ability.
If peer work-groups are racially mixed, they create a situation that
requires cross-race interaction. In cooperative groups all the students are
expected to work together toward some common goal. Ideally, this should
produce interracial sociability. However, as Cohen (1982) suggested, the
status differences that exist between black and white students may preclude
this outcome. A number of pedagogical techniques (for example, teamsgames-tournament, jigsaw, student teams-achievement division) are available to equalize status differences between blacks and whites and encourage
equal participation by all group members. Slavin (1980) and Sharan (1980)
described these techniques and the research evaluating their effectiveness in
producing growth in achievement and positive race relations. In general,
these methods appear remarkably successful in improving interracial
attitudes and fostering interracial friendliness. The mechanism that produces this success seems to be either the process of working together (that is,
task interdependence) or of working toward the attainment of a team reward
(that is, reward interdependence). In the absence of special preparation
aimed at fostering participation by low-status group members, racially
mixed cooperative groups are not likely to affect change in interracial
sociability.
In the three types of helping groups, status differences are a structural
feature of the group or dyad. One person is identified as possessing more
information or greater skills than another. If the status and authority bases
of the relationship are recognized as legitimate, then both tutor and tutee
can benefit. Tutoring is seen to increase motivation and sense of responsibility, promote more positive attitudes toward school, and increase
prosocial behavior for both members of the dyad (Allen, 1976). In addition,
tutoring increases the tutor's self-esteem and improves the tutee's academic
skills.
One of the ways that tutoring promotes pro social interracial behavior is
through the mechanism of role playing. Collins and Hoyt (1972) argued that
attitude change occurs in a situation of role taking when a person feels
responsible for his or her behavior, is not tangibly rewarded for the
behavior, and believes that the behavior has important consequences.
Because tutoring satisfies these conditions, it is expected to change a tutor's
negative racial attitudes to more positive ones.
At the same time, the tutee benefits from the exchange in terms of
178
Maureen T. Hallinan
learning outcomes as well as the social dynamics of the interaction. It is

generally believed that people like those who help them attain rewards
(Berkowitz & Daniels, 1963; Johnson & Johnson, 1975). Thus, in the tutoring
relationship, the tutee should experience positive sentiment for the tutor,
which should lead to friendly cross-race relations.
Fteward Structure
A third organizational feature that affects interracial friendliness is the
reward structure of a school or classroom. A reward structure can be
conceptualized along five dimensions: the basis for allocation of rewards,
the type of tasks rewarded, the kinds of rewards assigned, the visibility of the
rewards, and the significance attached to the rewards. A student can be
rewarded for performance relative to (a) some absolute standard (such as
the national norm on a standardized achievement test), (b) the accomplishment of other students in the class, and (c) the student's previous mastery
level. These methods of assigning rewards can be referred to as absolute,
relative, and self-mastery. Because rewards can be assigned for individual or
group performance, there are basically six different bases for the distribution of rewards. Slavin (1980) focused on the most commonly used of
these methods and referred to competitive rewards (a student's success
depends on another student's failure), cooperative rewards (one student's
success increases the success of other peers in the class), and individualization (the attainment of rewards is independent of the performance of
classmates). Ftewards may be assigned for performance on academic tasks,
co-curricular activities, or extra-curricular activities.
The types of rewards most frequently used in schools are tangible objects
(such as grades, prizes, and promotions), social rewards (such as esteem,
respect, and power), and self-reinforcement through the intrinsic rewards of
success and accomplishment. Schools and classrooms vary considerably in
the prominence given to rewards. Some teachers display graded papers and
projects, whereas others treat grades as confidential. Students may vary too
in the salience they attach to rewards, with some pupils being more eager
than others to attain them.
One can argue that the type of reward structure that exists in a school or
classroom affects students' interracial sociability. Competition is the most
commonly used basis for assigning rewards. Most theorists of competition
discuss its effects on intergroup behavior. Sherif et al. (1961) demonstrated
that negative functional interdependence between groups for the achievement of their goals leads to competitive social interaction, which produces
antagonistic social attitudes among the groups. Doise (1978) claimed that
factors promoting comparative self-evaluation (such as competition) produce competitive intergroup processes that foster attitudinal biases and
prejudices. Worchel (1979) stated that explicit competition draws clear

ti:
179
boundaries between groups and accentuates we/they differences, making

group membership a salient feature of the situation.
Competitive reward structures for academic achievement generally
involve individual-level competition rather than intergroup competition. It
would be highly unusual, for example, to find formal competition between
black and white students. Nevertheless, because race and achievement are
correlated in many classes, a competitive reward structure draws attention to
academic differences between the two races and can foster covert competition between the two racial groups. This can lead to resentment on the part
of the group who receives fewer rewards and aggravate existing negative
stereotypes and prejudices or lead to new racial barriers.
Academic competition is also likely to magnify the differences that
students believe exist between blacks and whites. Ames, Ames, and Felker
(1977) argued that in a competitive setting where rewards are scarce,
students distort the salience of ability in making judgments about their
peers. If students exaggerate the importance of ability and if blacks rank
lower than whites in the achievement distribution of the class, competition
will increase the perceived status differences between blacks and whites.
Consequently, a competitive academic reward structure acts as a barrier to
interracial sociability and fosters racial cleavage.
In contrast to the negative effect of a competitive reward structure,
cooperative rewards should increase interracial friendliness. Sherif et al.
(1961) argued that cooperative social interaction to attain interdependent
goals promotes cohesive social relations. Solidarity and pride in one's own
group lead to ingroup biases that overeva1uate the characteristics and
performance of ingroup members and undereva1uate those of outgroup
members. Turner (1981) claimed that being defined as an interdependent
group produces intermember attraction that leads individuals to define their
group attributes positively and be more likely to perceive inclusive
similarities among themselves.
These evaluation processes are initiated when black and white students
are assigned common tasks that require interdependent work to attain a
reward. Intragroup cooperation produces a feeling of group solidarity that
can transcend or at least lessen the importance of racial barriers. As the
group structure emerges, members begin to identify task-related similarities
of their peers and come to value member contributions to the task; this
member approval should make racial differences less salient and member
efforts to obtain the common reward more relevant to the group members.
Considerable empirical evidence can be found of positive effects of a
cooperative academic reward structure on students' race relations (e.g.,
DeVries, Edwards, & Slavin, 1978; Slavin, 1979).
Another situation in which a cooperative reward structure is expected to
affect race relations is extra-curricular activities. This is particularly true in
the case of athletics. Intra- and intermural sports and games create
180
Maureen T. Hallinan
opportunities for students to interact and work together to win a common

reward. The intergroup competition stimulates within-group cooperation,
which fosters group cohesion. When athletic teams are racially mixed,
group cooperation should have a positive effect on interracial friendliness.
Cooperative rewards for extra-curricular achievement may indeed be a
more powerful influence on black and white friendliness than are rewards
based on academics if extra-curricular activities are more salient to students
than school work. In a re-analysis of data from a national sample of
secondary schools, Slavin and Madden (1979) found that participation in
mixed-race sports teams had a significant positive effect on social relations
between blacks and whites.
Finally, a reward structure in which rewards are allocated on the basis of
self-mastery or individualization is likely to have little or no impact on
student race relations because it fails to generate social comparison
processes and has little effect on the distribution of status in the classroom.
Conclusions
In this chapter I showed how structural and organizational characteristics
of schools and classrooms create the conditions under which social
psychological processes governing interracial friendliness occur. The need
for such a conceptualization was evident in the failure of existing social
science literature to explain inconsistencies in the empirical research on
students' cross-race relationships in desegregated settings. Stipulation of the
conditions that generate relational processes should allow better predictions
about when favorable and unfavorable effects of desegregation on interracial sociability can be expected.
Three organizational factors were related to students' interracial friendliness in school settings: classroom racial composition, instructional
practices, and reward structure. Racial composition is seen to affect
opportunities for interaction that generate the cycle of positive sentiment
predicted by Homans (1950). It also has an impact on status differences
between blacks and whites. This is crucial to understanding race relations
because almost all of the social psychological theories of interracial
attitudes and behaviors point to a status differential as the major barrier to
cross-race friendliness. If majority white classes widen status differences
between blacks and whites by augmenting the higher status of whites
through social power and control, then the theories would predict an
intensification of racial cleavage in these classrooms. If majority black
classes lessen status differences because blacks acquire status from being in
the numerical majority, then more positive interracial relations would be
predicted when whites are in the minority.
Instructional grouping also is seen to alter status differences between
blacks and whites. When students are ability grouped, a disproportionate
181
number of whites are found in the high-ability group. This intensifies status
differences between blacks and whites and leads to greater segregation.
However, when blacks and whites are assigned to the same-ability group,
the status differences between the two groups are decreased because they are
located at the same position in the academic hierarchy of the class.
According to contact theory and its extensions, this should produce
interracial sociability among group members. However, even if status
differences are eliminated, one can argue from expectations states theory
that intragroup race relations will not improve without direct intervention to
give blacks at least temporary superior status over whites. Moreover, the
social comparison processes that students engage in are not limited to
within groups, and the unequal distribution of black and white students
across ability groups creates a visible barrier to friendliness.
The reward structure of the classroom can accentuate or deemphasize
status differences between blacks and whites. A competitive academic
reward structure exacerbates differences and promotes negative attitudes by
depriving lower achieving blacks of equal access to rewards. This kind of
reward structure is inconsistent with the institutional support that is
believed to be essential for promoting positive race relations. On the other
hand, when cooperative rewards are allocated to racially mixed groups for
academic or nonacademic achievement, an environment is created where
social comparisions can have a positive influence on interracial behavior.
The conceptualization presented here supports the argument that structural and organizational characteristics of schools and classrooms have a
significant impact on interracial friendliness. By determining opportunities
for interaction and by defining bases for status and similarity between
blacks and whites, professionals responsible for the organization of a school
and classroom set into motion the social psychological processes that either
maintain or intensify interracial hostility leading to racial segregation or
improve racial attitudes and increase interracial interaction and friendliness. If the effort to improve race relations through desegregation does not
seem successful, it may be because these factors have not been taken into
account.
In the analysis in this chapter, I explain why merely placing black and
white students together in the same school is not sufficient to foster
interracial friendliness. A school or classroom must be organized in such a
way that it promotes cross-race friendships by creating opportunities for
positive interaction, reducing status differences based on achievement, and
highlighting areas of equality and reverse status. Because school and
classroom organization can be manipulated by school authorities, real
desegregation, in the form of within-school integration, can be viewed as a
challenge and a distinct possibility.
Acknowledgments. The research reported in this paper was funded by Grant No.
lROl-HD-17776 from the National Institute of Child Health and Human Development and by the Wisconsin Center for Education Research, which is supported in
182
Maureen T. Hallinan
part by Grant No. NIE-G-81-()009 from the National Institute of Education. The
opinions expressed in this chapter do not necessarily reflect the position, policy, or
endorsement of these agencies. I am grateful to Karen Shire and Stevens S. Smith for
helpful comments on this chapter.
References
Allen, V. (1976). Children as teachers: Theory and research on tutoring. New York:
Academic.
Allport, G. (1954). The nature of prejudice. Cambridge, MA: Addison-Wesley.
Ames, C, Ames, R, & Felker, D. (1977). Effects of competitive reward structure and
balance of outcome in children's achievement attributions. Journal of Educational
Research, 69, 1-8.
Armor, DJ. (1972). School and family effects on black and white achievement: A
reexamination of the USOE data. In F. Mosteller & D.P. Moynihan (Eds.), On
equality of educational opportunity (pp. 168-229). New York: Vintage.
Berger, 1., Cohen, E.G., & Zelditch, M. (1972). Status characteristics and social
interaction. American Sociological Review, 37,241-255.
Berkowitz, L., & Daniels, L.R (1963). Responsibility and dependency. Journal of
Abnormal and Social Psychology, 66, 664-669.
Blaney, NT., Stephan, C, Rosenfield, D., Aronson, E., & Sikes, 1. (1977). Interdependence in the classroom: A field study. Journal of Educational Psychology, 69,
121-128.
Cohen, D.K, Pettigrew, T.F., & Riley, RT. (1972). Race and outcome of schooling. In
F. Mosteller & D.P. Moynihan (Eds.). On equality of educational opportunity (pp.
343-370). New York: Vintage.
Cohen, E.G. (1972). Interracial interaction disability. Human Relations, 37, 648655.
Cohen, E.G. (1980). Design and redesign of the integrated school. In W.G. Stephen &
1.R Feagin (Eds.), School desegregation: Past, present and future (pp. 251-278). New
York: Plenum.
Cohen, E.G. (1982). Expectation states and interracial interaction in school settings.
Annual Review of Sociology, 8,209-235.
Cohen, E., & Roper, S. (1972). Modification of interracial interaction disability: An
application of status characteristics theory. American Sociological Review, 36, 643657.
Collins, B.E., & Hoyt, M.F. (1972). Personal responsibility-for-consequences: An
extension of the 'forced compliance' literature. Journal of Experimental Social
Damico, S.B., Bell-Nathaniel, A, & Green, C (1981). Effects of school organizational
structure on interracial friendships in middle schools. Journal of Educational
Research, 74, 388-393.
DeVries, D.L., Edwards, K1., & Slavin, RE. (1978). Biracial learning teams and race
relations in the classroom: Four field experiments using teams-garnes-tournaments. Journal of Education Psychology, 70,356-362.
Doise, W. (1978). Groups and individuals: Explanations in social psychology. Cambridge: Cambridge University Press.
Duncan, B. (1977). Differential social perceptions and attributions of intergroup
violence: Testing the power limits of stereotyping of blacks. Journal of Personality
Ehrlich, H.1. (1973). The social psychology of prejudice. New York: Wiley.
Gerard, H.B., & Miller, N. (1975). School desegregation. New York: Plenum.
183
Hallinan, M.T. (1982). Classroom racial composition and children's friendship.

Social Forces, 61, 56-72.
Hallinan, M.T., & Smith, S.S. (1985). The effects of classroom racial composition on
students interracial friendliness. Social Psychology Quarterly, 48, 3-16.
Hallinan, M.T., & Sorensen, AB. (1983). The formation and stability of ability
groups. American Sociological Review, 48,838-851.
Hallinan, M.T., & Sorensen, AB. (1984). Ability grouping and students' friendships.
Unpublished manuscript, University of Wisconsin, Madison.
Hauserman, N., Whalen, S.R., & Behling, M. (1973). Reinforced racial integration in
the first grade: A study in generalization. Journal of Applied Behavior Analysis, 6,
193-200.
Homans, G.C (1950). The human group. New York: Harcourt, Brace & Co.
Hyman, H.H. (1969). Social psychology and race relations. In I. Katz & 1. Gresin
(Eds.), Race and the social sciences (pp. 3-48). New York: Basic Books.
Johnson, D.W., & Johnson, R T. (1975). Learning together and alone. Englewood Cliffs,
NJ: Prentice-Hall.
Katz, I. (1964). Review of evidence relating to effects of desegregation in the
intellectual performance of negroes. American Psychologist, 19, 381-399.
Nelson, J.1. (1966). Clique contacts and family orientations. American Sociological
Review, 31,663-673.
Patchen, M. (1982). Black-white contact in schools: Its social-academic effects. West
Lafayette, IN: Purdue University Press.
Pettigrew, T. (1967). Social evaluation theory: Convergences and application. In D.
Levine (Ed.), Nebraska symposium on motivation (Vol. 15, pp. 93-117). Lincoln:
University of Nebraska Press.
Pettigrew, T. (1975). The racial integration of the schools. In T. Pettigrew (Ed.), Racial
discrimination in the United States. New York: Harper & Row.
Riordan, C, & Ruggiero, 1. (1980). Producing equal-status interaction: A replication.
Social Psychology Quarterly, 43, 131-136.
Schofield, 1.W. (1982). Black and white in school: Trust, tension or tolerance? New York:
Praeger.
Schofield, 1.W., & Sagar, H.A. (1977). Peer interaction patterns in a integrated middle
school. Sociometry, 40, 130-138.
Sharan, S. (1980). Cooperative learning in small groups: Recent methods and effects
on achievement, attitudes and ethnic relations. Review of Educational Research, 50,
241-271.
Shaw, M.E. (1973). Changes in sociometric choices following forced integration of an
elementary school. Journal of Social Issues, 29 (4), 143-157.
Sherif, M., Harvey, OJ., White, BJ., Hood, W.E., & Sherif, CW. (1961). Intergroup
conflict and cooperation: The robber's cave experiment. Norman: University of
Oklahoma Book Exchange.
Slavin, R (1979). Effects of biracial learning teams on cross-racial friendship and
interaction. Journal of Education Psychology, 71,381-387.
Slavin, RE. (1980). Cooperative learning. Review of Educational Research, 50, 315342.
Slavin, R, & Madden, N. (1979). School practices that improve race relations.
American Educational Research Journal, 16, 169-180.
Sorensen, AB., & Hallinan, M.T. (1984a). Effects of race on assignment to ability
groups. In P.L. Peterson, L.C Wilkenson, & M.T. Hallinan (Eds.), The social context
of interaction: Group organization and group processes (pp. 85-103). San Diego:
Academic.
Sorensen, AB., & Hallinan, M.T. (1984b). The effects of ability grouping on growth in
achievement. Unpublished manuscript, University of Wisconsin, Madison.
Maureen T. Hallinan
184
St. John, N.H., & Lewis, RG. (1975). Race and the social structure of the elementary
classroom. Sociology of Education, 48,346-368.
St. John, N.H., & Smith, M.S. (1969). School racial composition, achievement and
aspirations. Cambridge, MA: HaIVard University, Center of Educational Policy
Research.
Stodolsky, S. (1984). Frameworks for studying instructional processes in peer workgroups. In P.L. Peterson, L.e. Wilkenson, & M.T. Hallinan (Eds.), The social context
of interaction: Group organization and group processes (pp. 107-124). San Diego:
Academic.
Tajfel, H. (1982). Social psychology of intergroup relations. Annual Review of
Psychology, 33, 1-39.
Turner, J.e. (1981). The experimental social psychology of intergroup behavior. In

J.e. Turner & H. Giles (Eds.), Intergroup behavior (pp. 66-101). Chicago: University
of Chicago Press.
Worchel, S. (1979). Cooperation and the reduction of intergroup conflict: Some
determining factors. In W.G. Austin & S. Worchel (Eds.), The social psychology of
intergroup relations (pp. 262-273). Monterey, CA: Brooks/Cole.
Yuchtman-Yaar, E., Semyonov, E., & Semyonov, M. (1979). Ethnic inequality in
Israel: Schools and sports: An expectation states approach. American Journal of
Sociology, 85,576-590.
....~
Chapter 10
Friendships in the Work Place

Gary Alan Fine
In a society in which people derive much of their identity from their

occupation, the social connections made at one's place of employment are
of crucial importance. Instrumental relations occur at work, but expressive
relations are found there as well. Nowhere is this better illustrated than in
an academic department. Most professors are "friends" with the majority of
their colleagues, sometimes associating with them outside of normal
working hours and sharing activities that are not work-related. Work is
frequently punctuated with acts of friendship.
Despite the prevalence of friendships at work, there has been little
discussion of these relationships. My goal in this chapter is to describe some
of the important features of these friendships, recognizing the limitations of
generalizing from a skimpy literature, sometimes being forced to generalize
beyond what might be advisable. Occasionally, I draw upon some unpublished ethnographic research (in-depth interviews and participant
observation) that I have conducted in four restaurants in the Twin Cities
area.
Culture and Friendship

Friendship can be approached in several directions. I have selected the lens
of culture, examining the content of friendships and how the settings in
which friendships occur contribute to their creation. Although this perspective is helpful for understanding all friendships (Fine, 1980), because of
the salience of the work environment for directing action, it seems
particularly relevant to work friendships. By emphasizing the cultural
components of friendship, I try not to slight other substantive approaches
that focus on factors such as affect, similarity, information, and power.
By culture I refer to the traditions, behaviors, and assumptions that are
collectively held by a group and that permit members to recognize that they
186
Gary Alan Fine
belong to the same social unit. People participate in national cultures as well
as regional, ethnic, age, and religious subcultures. They have family
traditions and may be involved in an avocational culture as well as a
vocational culture.
Most American workers are not self-employed, and most do not work in
isolation; rather, they are employed by organizations of various sizes. These
corporate entities have organizational cultures or corporate cultures (e.g.,
Fine, 1984b) that consist of traditions characterizing the work organizations
in their entirety. Organizations develop distinctive traditions, climates,
world-views, or characters. Peters and Waterman (1982) discussed companies with extensive traditions as having "strong cultures" and suggest that
this commits workers to the company and leads to excellence.
Different cultures may emerge within a company. Gregory (1983) referred
to the existence of multiple cultures (subcultures) within an organization.
Martin and Siehl (1983), examining the conflictual nature of some of these
cultures, termed them countercultures. Using the controversial tenure of John
DeLorean at General Motors (GM) as a case study, they demonstrated how
he broke the rules at GM and, in the process, gathered around him a group
of co-workers who shared his perspective and created an alternative view in
opposition to the dominant conservative culture.
Work cultures can be even more fine-grained. Departments within an
organization have unique cultures; even small groups within a larger unit
develop "idiocultures" (Fine, 1979). Finally, each friendship creates what
might be described as a "friendship culture," a common set of assumptions,
traditions, and behaviors that emerge from the sharing of "self' among
friends (Suttles, 1970).
Beyond the organization exist "occupational subcultures." People with
similar work tasks often develop similar techniques for negotiating the
constraints of their job and identify with each other (Hebden, 1975). These
cultural traditions may emerge when separately facing common problems
but may be diffused through an occupational communications network.
Occupational groups share jargon, stories, and jokes transmitted through
job transfers, common schooling, and meetings outside the workplace, such
as conventions. As a result, an employee who switches organizations but
keeps the same occupation can easily fit into the new setting.
Studying friendship through culture requires that the meaning of social
relations be investigated. This suggests a methodology different from that
frequently used in social psychological studies of friendship. To understand
the attitudes that people have toward their friendships and to understand
what they do with their friends, researchers typically use qualitative
techniques such as in-depth interviews and participant observation (e.g.,
Bogdan & Taylor, 1975). A methodology that allows participants to respond
at length concerning their feelings about their relationships and that
involves following them into their environments and observing what they
187
do produces more detailed findings than those typically reported in

friendship studies.
My goal in this chapter is to explore the several dimensions on which a
cultural approach can broaden our perspective on friendship. I begin by
examining the types of workplaces and jobs in which friendships occur.
Obviously, the behavior setting has a great impact on the types of cultures
and friendships that can be formed. From this base, I turn to the role of
culture (especially joking, rumor, and ceremonies) in the workplace and
discuss how these forms affect the friendships of work. I conclude the
chapter by examining two important variables that constrain the types of
relationships that can be found on the job: the relationship between the
sexes and friendship within a power structure. I have no illusions that I have
covered all possible topics related to friendship and work but hope that I
will have spawned enough ideas for others to blaze the trails to which I have
directed their attention.
Types of Work Relations

Workers have the option of establishing relationships with their co-workers.
In the idealization of work as an instrumental process, people may be seen
as working "together in isolation," that is, working on identical or complementary tasks without communicating. From this perspective work relations
have little affective significance. An alternative image is to picture all
workers forming close ties. Some work relations do blossom into intimacy,
but many do not. Most relations are between these extremes: acquaintanceship, friendship, or friendly relations. However these relations are labeled,
these are people with whom one is "friendly," but with whom one is not
"close." Kurth (1970) noted that "friendly relations" develop out of formal
role relationships and may be a preliminary stage in the development of
more intense friendships. These relationships are especially characteristic of
work, where individuals begin by interacting with each other because they
must, according to formal roles, and, making a virtue of necessity, turn these
contacts into something more informal and pleasurable.
Two additional relationships should be noted: love and hatred. Although
these relations are not frequent, both occur at work. Their existence suggests
that rich emotion can be found mixed throughout society, even in its more
instrumental corners. Particularly common are the milder forms of hatred:
annoyance, vexation, and dislike, which often appear in competitive
workplaces. One justification for union seniority rules is that they relieve the
pressure on workers to compete for scarce jobs and rewards.
Love-and sex-is not a new intruder into the office but, as discussed
later, only recently have these relationships developed between equals or
near-equals. "Fishing in the secretarial pool" had benefits for both parties,
even if there was also a darker side of sexual harassment. With the increased
188
Gary Alan Fine
involvement of women in the work force, and their promotion to high-status

positions, the issue of equal-status liaisons has become more significant.
Homosexuality in the office adds an additional twist, but to my knowledge
has not been studied.
The Workplace as Setting

Although the structure of work differs widely among occupations, in general
workers are expected to remain together throughout most of their workday.
Some workers, such as cooks, typically work in close proximity, whereas
others, academics for instance, can retreat into their own offices. Still others,
including many clerical workers, work at separate desks but regularly
converge, for example, during coffee breaks (Swanson, 1978). Whatever the
structure of a workplace, common areas are typically found in which
personal transactions occur, for example, the main office, elevators, lounges,
restrooms, libraries. People feel awkward seeing the same strangers day after
day, week after week; therefore, co-workers and regular customers are
incorporated into an acquaintanceship network; in the case of co-workers,
this often occurs on the first day on the job.
Workplaces are significant for the development of friendships beyond the
pragmatic reality that they force individuals together. The culture of work
allows and, in some cases requires, friendly behavior. Many jobs (e.g.,
cooking and serving a meal) encourage friendly relations because they
assume complementary tasks. Creating a smooth working day is an
interactional achievement by the workers in that they must create the
conditions for cooperation (Nusbaum, 1978). As a consequence, workers in
industrial plants "cover" for each other, permitting the completion of
personal business (Burowoy, 1979; Jones, 1980). This practice is sometimes
referred to as the informal organization of work, a phrase that is true
enough, but perhaps sounds too instrumental. Workers continually are
asked to demonstrate to their colleagues that they can be trusted and that
they are "good guys"; then, at some later time, they can request help
themselves. The exchange value of this activity is supplemented by the fact
that it bolsters good feelings within the workplace. Because workers depend
on each other (being at different stages of the production process), the
failure of one reflects poorly on the performance of others. Although the real
tensions in such occupational relations should not be neglected (for
example, between cooks and waiters, who operate under different, conflicting pressures), there is a push toward friendliness or toward leaving the
job if this proves impossible. It is not only poor performance that create
problems; the overly effective performance of one worker may make it
difficult for others to keep up with the flow of work. These workers (whether
they be professors and secretaries, or cooks and waiters, or workers at two
assembly line stations) must negotiate their problems if work is to proceed
smoothly (see Stelling & Bucher, 1972; Strauss et aI., 1963).
189
Workers with identical or similar tasks know that the successful completion of the tasks of one does not directly imply anything about the work of
another; however, this structure also implies negotiation. Specifically, most
workers value controlling the pace of their activity. For the group to function
smoothly and happily, all should be working at nearly the same rate and
should define themselves as contributing equally to what must be done (see
Roy, 1952, 1954). Ideally, the group should collectively agree on the proper
pace (a cultural choice) and have sufficient solidarity to uphold this
decision, although typically the agreement will be implicit. Roy's (19591960) participant observation of a small factory group demonstrated how
these punch press operators developed an informal schedule. At the same
time each day, one worker playfully stole a banana from the lunch bucket of
another, shared smoked fish, or engaged in a mock fight over the opening of
a window. Roy demonstrated the effect of these rituals by showing how they
could be undermined by a breach in the group's social relations; the culture
of the work group and social relations are interdependent. In this case
interaction rituals were supported by the fact that these workers were
isolated from the rest of the factory and were infrequently observed; in other
words, these workers could control the conditions of their own work. This
desire for autonomy or control supports the establishment of friendships or
friendly relations within the workplace. If co-workers are not cooperative,
autonomy becomes difficult to achieve, particularly in those circumstances
in which workers can be observed by those who wish to control them.
The need for control over the conditions of their work leads workers to
incorporate as many levels of employees into their community as possible.
In studying restaurants I found that cooks liked their chefs to joke with
them. Such relations not only provide for a more pleasant environment, but
also integrate their supervisors into the occupational group and make
formal discipline more difficult. There are also benefits for authorities;
although they give up some control, they receive a cooperative work force of
members who attempt to be "reasonable" in their claims of autonomy.
Ecology has an effect on relationships and work culture. All the world
may be a stage, but the design of the theater constrains the dramas that can
be performed. Behaviors possible in one physical environment are not
possible in another. In those environments of physical danger (some
factories, construction sites, or mines), workers are limited in what they are
permitted to do. This does not mean that there will be no joking but that the
workers will take this danger into account. Haas (1972,1974) found that the
development of trust among high-steel ironworkers is crucial. Because of the
ever-present danger, workers must assure themselves that the new workers
can be trusted not to endanger them. As Hughes (1945) noted: "To be sure
that a new fellow will not misunderstand requires a sparring match of social
gestures. The zealot who turns the sparring match into a real battle, who
takes a friendly initiation too seriously, is not likely to be trusted with the
lighter sort of comment on one's work or with doubts and misgivings; nor
Gary Alan Fine
190
can he learn those parts of the working code which are communicated only
by hint and gesture" (p. 356). Haas (1974) described a verbal sparring match
on a 21 story skyscraper:
"Abe," the journeyman says, "These fucking apprentices don't know their
ass from a hole in the ground." The journeyman turns to me and says, "I
hope you don't think these guys are representative of the whole apprenticeship.They're a pretty sad lot" Joining in with the kidding, I say, "Yeh, I've
noticed that." Bud, an apprentice, says to me, "Don't listen to him. He's just
a fucking Indian." The journeyman responds, "Yeh, and he's a fucking
nigger." Abe then yells down to an apprentice below, "What the fuck are
you doing down there, playing with yourself? For Christ sake get up here
and bring that machine up here with you." (p. 101)
This interaction, relying on cultural stereotypes, is designed to ascertain that

the apprentice will not "fly off the handle," that is, can be trusted to keep his
poise under pressure. Socialization ensures that new employees are
"worthy" of trust and deserve to be integrated into the group. Once they have
passed, they have committed themselves to a set of close ties, at least while
on the job. These individuals need not be close friends, but on the job they
must maintain friendly relations.
The Content of Work

The content of work affects the friendships that are likely to develop, in part
because of the people that an occupation attracts and in part because of the
nature of the work. Because friendships are socially situated, to understand
the development of social ties, one must address what must be accomplished on the job. Consider restaurant kitchens. The structure of the
cook's day influences how relationships will develop. The typical restaurant
(serving food from 11 a.m. through 11 p.m.) has two periods of intense
activity: lunch from about noon to one, and dinner from seven to nine. The
rest of the day (as much as 14 to 16 hours) involves preparing for this rush
and cleaning up after it. There are short periods of time in which cooks must
work together, helping whomever is overburdened, and other periods in
which the work pressures are light and when cooks leisurely discuss topics
of interest while preparing for the next rush. The fact that restaurants vary in
their hours, in the scheduling of their work force, and in the location of this
work force (in the same area of the kitchen or separated) influences the
extent to which friendships will develop. Likewise, the composition of the
work force affects the strength and type of social relations. Some restaurants
hire only waitresses, other restaurants only waiters, and these hiring
practices determine how the cooks and "waitrons" will relate. Sexual
orientation (in some restaurants many waiters are homosexual), age, race,
and social class also contribute to their relations.
The amount of autonomy and the desire for autonomy affects social
conriections at work. If individual workers are autonomous (such as doctors
191
or professors), positive relations among peers are not crucial for completing
the work; however, such positive relations are still common because of
shared values, norms, and behaviors. These professionals require cooperation of their support staff who do the "dirty work" (e.g., doctors depend on
nurses, cooks on dishwashers, and professors on graduate assistants).
People in autonomous occupations have relatively greater opportunities to
determine the content of these relationships than subordinates do. Occupations with relatively less autonomy (factory workers, for example) have a
proportionately greater need for solidarity, expressed through unions and
through informal worker understandings. Case studies of blue-collar
occupations, such as construction workers (Riemer, 1979), miners (Vaught &
Smith, 1980), and printers (Lip set, Trow, & Coleman, 1956), point to
solidarity among those who do not control the conditions of their own
employment.
Commitment (Becker, 1960) also affects social relationships. Some jobs
are left at the end of the day with the workers barely considering them until
time to return. Most cooks, for example, report that they rarely think about
their jobs when not working. Other workers are constantly on: The work
has, in some sense, taken over their self. It is difficult for a minister to be
anything other than a minister, because laypersons expect holiness
(Kleinman, 1984); the doctor is always expected to give medical advice.
Thus, these workers may find it hard to spend time with laypersons, who
will never let them escape their role self. On the other hand, some people do
not wish to remove their role selves. College professors are the most
stereotypical example; the job represents them to themselves.
Commitment is encouraged in occupations in which the inhabitants are
forced by circumstances or choice to remain physically apart from others;
the nature of work presupposes something close to a "total institution"
(Goffman, 1961). Thus, religious order priests are particularly likely to
choose as friends those within the order (Gannon, 1981). The structure of
the priest's life and activities keeps his friendship ties within the occupation.
The priest's close friends are co-workers (although in these situations the
statuses of neighbor and co-worker are confounded), and the friendships are
necessarily multiplex (see Fine & Kleinman, 1983; Verbrugge, 1979).
Military friendships are similar (Little, 1981). Because of the extensive rules
(etiquette) of military life, the isolation of military personnel, and the norms
for mutual support, military friendships tend to be intense, particularly
among officers. The occupation of soldiering is an all-encompassing status,
and the military verges on being a total institution; both factors increase the
salience of occupation-related friendships.
As mentioned previously, different occupations recruit different types of
personnel, and because ethnic, racial, age, and sexual differences in
friendship exist, this influences the friendship on the job. Although there
are not adequate comparative studies of work friendship, male friendships
are characterized by "peer-group friendships." Generally women are more
192
Gary Alan Fine
likely to have close friendships, whereas males have extensive networks of

friends (Wright, 1982). If males have wider social connections, this may
facilitate information transfer; yet males may have a lower level of
commitment to the organization than do females whose friendships are
more intense.
Occupations that recruit youth are likely to have a greater proportion of
single persons, who have more time to spend with colleagues outside of
work. Lower class males with the tradition of spending time "with the boys"
may spend more time with some of their workmates, particularly if they are
neighbors as well. Although these workers may have more close ties, they
may also have fewer acquaintanceship ties, leading to a constricted social
network (Granovetter, 1973). These speculations about the effects of
demography on friendship must be supported by systematic research.
Organizational Friendships and

Organizational Culture
Beyond allowing someone to structure one's time, working for an organization places one in a cultural milieu. An organization with a "strong" culture
(Deal & Kennedy, 1982; Peters & Waterman, 1982; Rohlen, 1973) should
have more loyal workers and less alienation than one with a "weak" culture.
Managers must be aware of the expressive or symbolic content of their
decisions or, in Pfeffer's (1981) terms, they must see "management as
symbolic action" (p. 91). The expressive forms that occur on the managerial
and worker levels include slang (jargon), jokes, ideology, sagas and
histories, rituals and ceremonies, and stories.
How do these cultural traditions affect an employee? First, concrete
information is more readily comprehended and has more impact than
abstract information. A work-related story can have a greater effect than a
set of facts (Martin & Powers, 1983). The account that describes the outcome
of disobeying a rule is more effective in directing behavior than the rule
itself. Stories typify in a way that statements alone do not.
Ceremonies also induce commitment. Individuals "buy into" the system
through enacting organizational rituals. To the extent that behavior
convinces people of their attitudes, this may have a greater impact than
private attitudes. If individuals behave in public in a friendly way, they will
consider themselves to have friendly relations; public culture promotes
private relations. These public announcements of loyalty and good feeling
facilitate the establishment of connections among individuals. The importance of ritual is well-recognized among fraternities, retreats, and parachuting units, which use these traditions to cement socialization (Aran,
1975; Domhoff, 1974; Leemon, 1972; McCarl, 1976).
In this section I examine three types of cultural traditions that contribute
to friendly relations among workers: joking, rumor and gossip, and
ceremonies and activities outside of work.
193
Joking
Many tribal groups establish relationships in which the participants are
compelled to joke with each other. Fonnalized "joking relationships" are not
found in Western societies; we are not sufficiently structured and stratified
to require such a formalized informal relationship. Still, there are times and
places in which joking is expected, if not required. Bradney (1957), studying
sales assistants in a large department store, found that the joking relationships that develop promote cooperation in the face of competition among
salespeople for sales. The need to conform makes this joking effective for
social control and socialization.
In routine jobs a leavening agent, such as humor, is needed. Roy's (19591960) discussion of punch press operators depicts the set of humorous
break-times (such as "banana time") that these men use to give their
workday structure (Handelman, 1976). Much of what occurs within workers'
joking relationships is traditional: communication that is repeated again
and again is given significance by its repetition, and comes to characterize
the relationship. People continue to raise topics that have produced
enjoyment in the past. This humor culture makes the interaction among
workers comfortable (Leary, 1978; Nusbaum, 1978; Swanson, 1978).
The significance of humor for belonging is particularly evident in
informal initiation rites. Humor is a sign that a person can "take it." For
example, Vaught and Smith (1980) graphically described the obscene rituals
that are involved in initiating or "making" a miner, in which the new miner's
genitals are greased. This form of social control informs the miner that the
group always comes before his individuality. Such horseplay helps to create
a sense of mechanical solidarity (in Durkheimian terms) in the group.
These traditions appear to be most common and extensive where danger
is part of the work, but most occupations in which workers must depend on
each other have similar rituals, even medical training (Becker, Geer,
Hughes, & Strauss, 1961). In the kitchens I studied, new employees were
tested to see how gullible they could be made to be. Gullibility, after all, is
not entirely a trait of the person, but is an interaction between personality
and knowledge of the work situation. Cooks or busboys were sent to the
storeroom for "left-handed knives" or boxes of prunes, items that competent
workers would know were not available. On occasion, a cook would spike a
co-worker's soft drink with hot sauce to see the reaction and whether he or
she could "take it." In a cook's training program, one student was sent to the
storeroom for a can of "evaporated water," for which the student actually
searched. It is not only carrying out the joke that is important for building
rapport, but telling it, which indicates to the participants that the group is
comprised of friends who can joke, take jokes, and have a common
history.
Humor, of course, is not always conducive to the development of group
unity. Jokes involve the putting down or derogation of a target (Gruner,
1979); when humor is directed at someone present in the setting, hostility or
Gary Alan Fine
194
prejudice may result. Such humor may even constitute illegal harassment
under the Equal Employment Opportunities guidelines. Whether or not this
joking is illegal, such derogatory humor stratifies employees and makes
minority group members feel less than welcome. Because this is humor, the
joker can always retort when called on to account, "Well, I was only joking"
(Fine, 1984a). The battle against such jokes has not been won, but it is now
more common for the joker to gauge the attitudes of the audience before
speaking and to use techniques of "facework" to disclaim any malicious
intention in the name of presenting "pure wit" (see Mechling & Mechling, in
press).
Rumor and Gossip
Individuals strenuously make an "effort after meaning" (Allport & Postman,
1947). People wish to know where they stand and what to expect from the
events that surround them. It is virtually impossible to conceive of any
workplace that does not have a rich vein of rumor and gossip. Davis (1972)
has argued that organizational grapevines are an expression of the healthy
human motivation to communicate:
In fact, if employees are so uninterested in their work that they do not
engage in shoptalk about it, they are probably maladjusted. If employees
are so uninterested in their associates that they do not exchange talk about
who will likely get the next promotion or who recently had a baby, they
probably are abnormal. (pp. 261, 263)
Suppressing rumor in one location only leads it to break out elsewhere.

From this view, rumor and gossip serve as a form of collective problemsolving (Shibutani, 1966).
In situations in which facts are in short supply, rumor will proliferate,
particularly in those organizational settings in which chains of command
prevent the free flow of information (Davis, 1969). Those with power attempt
to control information, but leaks develop. Hershey (1956) found numerous
rumors in an industrial firm that related to job security, promotion, benefits,
and working conditions, a result oflocal economic difficulties. Rumors also
flourish in the stock market to such a degree that Rule 435 of the New York
Stock Exchange forbids brokers to spread rumors on the floor of the
exchange even in the absence of an intent to defraud.
Gossip, in many ways similar to rumor, is also common at work. People
are intensely curious about each other's problems and triumphs, and most
workplaces are awash in such communications. Even more than rumor,
gossip implies a social connection among the parties to it; it assumes the
existence of community. Gossip involves preserving group status, particularly according to functional theory (e.g., Gluckman, 1963). The more
exclusive or highly organized the group, the more gossip is present, for
example, in medicine, law, or academic life.
195
Being involved in gossip-as a target, an audience, or as a gossipindicates that one has been accepted (Haviland, 1977). A flourishing "gossip
culture" suggests a "community" of interest (see Spacks, 1983).
Functionalism is not the only theory by which one can explain gossip.
Gossip also can be used strategically by individuals. This transactional
approach emphasizes potential benefits that individuals gain through such
talk (Paine, 1967). People try to control the impressions that others have of
them and their acquaintances. This approach describes gossip in terms of
social networks. Information spreads according to who knows whom and
will continue to be communicated so long as the teller believes the listener
cares about the target and the substance of the talk. This model of network
diffusion through social relations even applies to hysterical contagion
(Kerckhoff & Back, 1968).
A strategic approach is congruent with understanding gossip in terms of
conflict. Cox (1970) described gossip as a "degradation ceremony," assigning
the target a lower status in the moral order. However, gossip is degradation
with an etiquette, which, if ignored, can discredit the performer. In work
settings there are times and places at which gossip is permissible (in small,
homogeneous groups when there is sufficient time to complete the
conversation) and times and places at which it is not (in front of customers,
supervisors, or other outsiders). Some gossip can leave a legacy of ill will.
Co-worker C may be angered or hurt when he or she learns what Workers A
and B have been saying, changing relations of friendship to those of enmity.
Enmity is more polluting than a friendly relation, because it undermines the
belief that there is a collective consensus and may create intolerable
bitterness, leading to collapse (as in the case of sports teams racked by
dissension). Despite the positive function that communication of information can have in promoting harmony, it can also create tension and
dissatisfaction, leading to diques and isolating outcasts.
Ceremonies and Activities Outside of Work
Some work-related ceremonies are remarkable in their elaborateness and
impressive in their scale. Peters and Waterman (1982) described one such
ceremony:
[T]he manager of a loo-person sales branch rented the Meadowlands
Stadium (New Jersey) for the evening. After work, his salesmen ran onto the
stadium's field through the players' tunnel. As each emerged, the electronic
scoreboard beamed his name to the assembled crowd. Executives from
corporate headquarters, employees from other offices, and family and
friends were present cheering loudly. (p. xxiv)
Admittedly, this IBM ceremony is more extensive than those that most
companies hold, but this and similar examples-the MatteI Thanksgiving
party (Dandridge, 1975), the beer busts at Hewlett-Packard, or the Monday
196
Gary Alan Fine
night rallies of Tupperware salespersons-serve the same function. Although some might argue that these rituals are synthetic and only
demonstrate the patriarchal control of American corporations, many
business consultants believe that such "silliness" contributes to corporate
integration. This perspective emerged from asking why Japanese corporations consistently outperformed American ones during the 1970s. One
answer was that Japanese corporations had traditions that made the
existence of the corporation real (see Rohlen, 1973).
Of course, picnics and Christmas parties have always been a part of work
culture in some companies. Companies even have corporate retreats. Today
many organizations are rushing to create such unifying rituals, a ploy that
can fail if the employees do not accept the genuineness of the feelings
behind such events or do not enjoy the substance of the ceremonies.
Perhaps the most frequent area of worker involvement in nonwork
activities is sports participation. Many companies have their own bowling,
softball, or basketball teams. These sports groups provide opportunities for
workers to get together "on their own time" and have a good time playing for
sports teams either sponsored by a company, organized by it, or organized
by the workers without direct support. Some workers invest considerable
time, energy, and even money in these teams. For example, a softball team
composed of faculty and students in the Department of Sociology at the
University of Minnesota played a team from the Department of Political
Science. For this event the players raised $1,000 to rent the Metrodome, the
stadium where the Minnesota Twins play.
Sports teams vary on the extent to which they receive organizational
support, on their competitive status, on whether the team is limited to
employees, and on whether the games are played against teams outside the
organization. These sports teams are composed of amateurs, but teams
differ in the importance that they give to winning at all costs and whether
they bet on the outcomes of their games.
Because most sports are dominated by males and are of particular interest
to males, a possible latent effect of maintaining such sports teams is to
exclude women from participation in the informal life of the corporation.
Although this may not be as significant for women in the executive suites, it
may affect low- and middle-level female managers and, of course, female
workers and work supervisors. Sports pools and talk about various athletic
subjects can separate women from the informal lives of their organizations.
I argue that the culture of the work organization shapes and directs
worker relationships. First, cultural traditions connect workers to the
organization as a whole, producing a sense of belonging, a company self,
and making personal ties within the company more likely. Second,
organizational culture expands an individual's social network, which has
benefits both for personal interactions and for communication flow in the
organization through acquaintanceship ties.
197
Men and Women Together

So far I have focused on the relations among friends or acquaintances; that
is, I have discussed liking, not loving. Although most positive relationships
within the workplace involve liking, not all do. As men and women have
become more likely to work together in organizations in which they have
the same or similar statuses, the sexes must develop friendly relations. This,
however, is difficult because of the commonly held assumption that men
and women can be lovers, but never friends. Kurth (1970) noted that
friendships between men and women are problematic in that they violate
expectations about the potentially sexualized interaction of men and
women. This "problem," Kurth stated, is compounded when the two parties
are of different ages: Can a man have a woman as a protege; how can a
woman handle a man as a mentor? Such relations frequently lead to gossip.
Many "naturally" assume that a man and a woman who spend a lot of time
together are interested in more than talk. The platonic relations that we take
for granted when two men or two women are close, become difficult for a
man and a woman. Although the workplace is one of the easiest locales for a
platonic cross-sex friendship to develop because of the forced proximity of
workers and because the relationship can legitimately start as a purely
instrumental one, dangers exist. People, after all, are sexual and often drift
toward the couch. It is the rare single female who has never experienced
sexual pressure from her platonic male friend to consummate their
relationship (Myers, 1984).
The "love" or "sexual" relationship is increasingly recognized as a
problem at work (E. Collins, 1983; Quinn, 1977). Meyers (1984) quoted a
male who speculated about the problems he and his platonic friend might
face if they became involved:
When I worked with [my friend] ... we spent forty hours a week working
together and if you start having a sexual relationship and it doesn't work
out and you are still coming to work you have that pressure and you can't go
to the boss and say "[My friend] and I were having an affair and it's just not
working out and I can't work at that [place] with her." Well, he'd think you
were crazy, so I think that is how it initially got started [as a platonic
relationship]. (p. 30)
Although an organizational romance may be positive, such intimacy

typically is viewed negatively, particularly when the relationship is visible
and affects the public behaviors of the partners. Although in one sense these
relationships are dyadic, they also may affect the behaviors of all the coworkers. If the company is a family, an organizational romance is incest.
Quinn's (1977) study of third-party accounts of organizational romance
provides the best available evidence of its dynamics and structure. He noted
that the physical proximity oflovers was a factor in 63% of the romances. An
instrumental relationship between an executive and his secretary can
198
Gary Alan Fine
blossom more easily than one between an executive and the secretary of
another. Three motivations generate these relationships: job concerns
(advancement, security, and power), ego-enhancements (excitement, ego
satisfaction, and sex), and, of course, love.
Quinn (1977) reported that the impact of these relationships varies. Power
may shift through favoritism or "spying." Some lovers find that they are
preoccupied with each other or are isolated from colleagues, particularly
when the connection is romantic rather than purely sexual. Lovers, like love,
may be blind. On the other hand, being involved in a romantic liaison may
make an employee easier to get along with, happier, or more productive.
These romantic relationships may produce hostility (gossip, resentment,
and lower productivity), they may have positive impact (a feeling of family,
increased teamwork, and positive emotions), or they may have little effect.
The specific impact of such a relationship is based on the immediate
circumstances of the work environment and the way in which the
participants handle their feelings. Still, to the extent such relationships are
exclusive and cannot be shared like friendship, they remove the lovers from
full participation in the group. Intense involvement will be favored over
friendly relations.
The work romance poses particular difficulties for the female, who
typically is the lower ranking member of the dyad. Quinn's data (1977)
suggest that women are twice as likely to be dismissed as men are. As usual,
the powerless suffer relatively more than those with power and resources.
These romantic ties can have a significantly detrimental effect on
organizational stability. By polluting the "face" of several organizational
actors, these relations threaten the organization itself. The romance
becomes a problem to be handled rather than a blissful state to be
encouraged. E. Collins (1983) suggested guidelines for managers to handle
these relationships:
1. Treat the relationship as a conflict of interest, since those outside of the
relationship will be unable to compete fairly for rewards against this
coalition.
2. Advise the couple to get outside help. The lovers need advice about
whether the relationship can or should survive.
3. Persuade the couple that either the person least essential to the company
or both must go. The high-status person is typically more crucial to the
company. This unfortunately may provide an excuse for stereotyping the
woman as a corporate Eve.
4. Help the ousted executive find a new and possibly better position. The
departure should be amicable, forced not out of hatred for love or sex, but
in the interest of organizational efficiency. (pp. 149-152)
Some sexual relationships occur without both parties agreeable or even

aware of all the factors involved. Here we become entangled in sexual
harassment, most often the relationship between a male supervisor and a
female subordinate, although occasionally the sexes are reversed. Although
199
the term sexual harassment suggests that the relationship is not one of
friendship, on occasion these relationships begin as friendship, even
romantic relationships, eventually breaking down. These relationships are
in some measure perversions of true romantic relations, causing not selftranscendence (Foote, 1953) but lowered competence and job satisfaction
among those harassed (Gruber & Bjorn, 1982). Certain instances of
harassment may derive from mistaken communication: What is given as
deference by the underling is taken as interest or acquiescence by the
superior (Collins & Blodgett, 1981). On the other hand, because of the
confounding of power and desire, many who write about this subject suggest
that romantic relationships between supervisors and those supervised are
inevitably exploitative. With the growth of affirmative action sexualharassment suits, some organizations have attempted to control sexual
activity in the workplace. Universities have moved furthest in this regard,
suggesting that a relationship voluntarily entered into by a professor and a
graduate student constitutes, on prima facie grounds, improper sexual
advances and can be redefined in this way subsequently. The University of
Minnesota has taken such a position toward its professorate in its Policy
Statement on Sexual Harassment (University of Minnesota, 1984):"Consenting romantic and sexual relationships between faculty and student, or
between supervisor and employee, while not expressly forbidden, are
generally deemed very unwise .... A faculty member who enters into a
sexual relationship with a student (or supervisor with an employee) where a
professional power differential exists, must realize that, if a charge of sexual
harassment is subsequently lodged, it will be exceedingly difficult to prove
immunity on grounds of mutual consent" (pp. 1-2). The crucial element
here is that even those relationships that appear to be voluntary and loving
are still condemned. Such a policy runs the risk of providing the less
powerful with a potent weapon to control their partner. The rules of work
relationships, at least among professionals, are being shifted, and this
makes the social arrangements between men and women more tender and
more prickly than what they might otherwise be.
Friendship Amidst the Hierarchy

Friendship, as well as love and sex, can bloom among persons at different
levels of the organizational hierarchy; but, as is true of most social
distinctions, it is far easier to develop friendships among those who are
status equals. Mter all, friendship is a transactional relationship (Schutte &
Light, 1978), and parties with approximately equal, or at least complementary, resources find it easier to establish relationships of exchange.
Those with different resources must overcome the barriers that unequal
exchange ratios cause. Status "homophilic" friendships are more common
than those in which the parties have markedly different statuses (Timms,
1967). Within an organization those who hold high-status positions may
200
Gary Alan Fine
base their friendships more on "status" characteristics, whereas those

workers with lower status, who are not as occupationally mobile, may select
their friendships on the basis of proximity (Schutte & Light, 1978).
Some writers characterize executives by their strategic use of friendships.
Packard (1959), working in the tradition of Thorstein Veblen, suggests that
friendships in the business world are highly stratified and political.
Executives (and their wives) carefully construct careers in the organization
through those they know, building social networks. Like Sinclair Lewis's
Babbitt, executives recognize that whom they associate with influences how
they will be viewed (see also Goffman, 1959). Although many desire to
hobnob with those of higher status, these individuals do not wish to be
contaminated by those with lesser status. So, for sociopolitical reasons, most
friends tend to have roughly equivalent social status. Lower status workers
who are befriended tend to be those who are viewed as oriented to success
(Corwin, Taves, & Haas, 1960).
A change in status implies a change in one's social networks. Once a
friend has been promoted, old work acquaintances tend to be dropped. In
this view, friendships are grounded in rational choices (e.g., Feld, 1984), and
a change in the perceived needs of the participants will alter patterns of
association. Blau (1956) suggested that occupational mobility, both upward
and downward, poses a dilemma for the establishment of interpersonal
relations and becoming accepted within the community. These individuals
have become "marginal men," out of tune with both their former and
current colleagues:
The upwardly mobile must choose between abandoning hope of translating
his occupational success into social acceptance by a more prestigeful group
and sacrificing valued social ties and customs in an effort to gain such
acceptance. The downwardly mobile must choose between risking rejections for failure to meet social obligations that are beyond his financial
resources and resigning himself to losing his affiliation with a more
prestigeful group. (p. 290)
The desirable position of high-status persons-their authority and background qualities, including education-leads to their central position in
networks of instrumental ties and in those of primary ties (Lincoln & Miller,
1979). This contrasts with individual characteristics, such as sex or race, that
affect workers' primary ties, but not their instrumental ties. To gain some
measure of control in a social organization, the person without much status
must obtain access to and control over persons, information, and instrumentalities (Mechanic, 1962), which is relatively more difficult than if one
has a position with explicit power.
Mentoring
Goode (1972) noted that friendship is one of the four major control
mechanisms in society, operating along with force, prestige, and wealth.
201
Friendship relationships help to control people in organizations by

providing a personal motivation to accept the world as it is, rather than to
disassociate oneself through alienation. Likewise, friendship paves the way
for the smooth organization of promotions and changes within an
organization. The influence of friendship is particularly evident in studies
of "mentors" in business and industry, especially at the higher levels of these
organizations, although similar aid from more to less experienced members
occurs in blue-collar occupations, such as trucking and printing (Lembright
& Riemer, 1982; Lipset et aI., 1956).
Promotion in an organization is facilitated by a sponsor, a fact that
managers began to recognize explicitly in the late 1970s, especially with the
publication of The Seasons of a Man's Life by Levinson and his colleagues
(1978). This bestseller included life histories of middle-aged businessmen
and scientists. Many of these individuals were helped in their careers by
others who took an interest in them (see also Zuckerman, 1977).
According to Kram (1983), the mentor relationship in business typically
develops through four stages: initiation, cultivation, separation, and redefinition. In the initiation phase, lasting about a year, the junior employee
admires and respects his or her senior for competence or capacity to provide
support and guidance. The senior manager sees the young manager as
someone who has potential and is coachable.
The cultivation phase lasts from 2 to 5 years. At this time the positive
expectations that have been developed in the initiation phase are tested
against the cold light of reality. Ideally, the mentor relationship will bloom
in this period, including both instrumental functions and interpersonal
ones. The senior manager directs work assignments to the young manager,
coaches, and provides information for surviving politically. This should
permit the young manager to increase his or her confidence. The senior
manager should gain personal satisfaction in knowing that he or she has
positively influenced a younger colleague; the younger colleague's success
will indirectly contribute to the public esteem of the mentor, who is known
by his or her proteges. Such an effect is evident in academic life, where the
established scholar is evaluated in part through his or her students.
Mter several years, the rewards to both individuals diminish; the younger
worker should be able to succeed without the help of a mentor. Whether this
separation is similar to "negative transference" or, more simply, to having
outgrown the relationship, change is necessary. Anxiety, feelings ofloss, and
personal turmoil may characterize this period. Yet the young manager
experiences new independence and autonomy.
In the final phase the relationship becomes a friendship. The two
individuals continue their contact, but it is no longer primarily a mentorprotege relationship, although mutual support may continue. The junior
manager is treated as a peer; the senior is treated with grateful recognition.
Mentoring is particularly significant in business today because of its
Gary Alan Fine
202
potential for incorporating women into the executive suite. Indeed, advice
books for businesswomen emphasize the value of having a mentor (N.
Collins, 1983). Perhaps mentoring is easier for women than for men in that
women are used to developing intensive friendships, whereas male friendships tend to be more extensive (Wright, 1982). On the other hand, until
recently men have wished to mentor other men. This may be changing
because of the emphasis on bringing women into higher positions in
business. As noted previously, there are difficulties in cross-sex mentoring
relationships because of the needs of women in the organizations, the
stereotypes about women (Kantor, 1977), and because of the sexual nature of
men and women. According to Fitt and Newton (1981):
[AJt lower levels in organization women typically need more encouragement than their male counterparts, and at higher levels mentors of women
have to spend more time "selling" their proteges ... the mentors-using the
status afforded by their high positions in the organization-strive to give
the women legitimacy in their organizations by conferring an implicit
stamp of approval. ... At this stage the men spend much time counseling
the women about managerial style, organizational culture, and personal
style-sometimes even including clothing and conduct. (p. 56)
The skills that are a "natural" part of male socialization must be learned by
women.
As Mary Cunningham (1984), formerly of Bendix Corporation, can attest,
rumors of a sexual relationship with one's mentor can have devastating
effects. Sexual possibilities are rarely absent from such relationships, even
though women are given near universal advice to avoid mixing sex with
business (N. Collins, 1983). There is little evidence that women can advance
their careers by sleeping around; career damage seems much more
common. Thus, a woman's life in a predominantly male organization is
more complex than that experienced by males.
Conclusions
I hope that this overview of friendships in the workplace will permit a
recognition of the significance of interpersonal relations in instrumental
situations. Too often researchers have treated friendships as divorced from
the locations in which they regularly occur. This situational context of
friendship-the staging area-is crucial for recognizing the limits and
content of these ties. Work, with its centrality of instrumental activity,
imposes ecological and normative constraints on friendship.
In this chapter I have emphasized the importance of culture on
friendship. This culture operates on several levels: the firm, the division, the
work group, and so forth. These cultures may be coordinated, oppositional,
or diverse. Whatever they are, they provide meaning in which a friendship
can be based. Members of a friendship dyad or group share meaning and
build a variety of traditions. Each friendship is, from this sociological
203
perspective, a mini-society with its own rules, norms, values, and ceremonies. Friendships are not only social structures, but like all structures
they are filled with meaningful behaviors. We have an obligation to examine
the content as well as the emotional and structural form of the relationship.
This chapter should demonstrate, if nothing else, the need for more
research. We need ethnographies that examine the dynamics of friendship
at work; we need surveys which examine its prevalence and its perceived
benefits and costs; we need interviews which examine the contingencies that
influence the development of these relationships; and we need historical
research to examine how community ties affected work traditions. The
workplace encompasses the range of emotions and attitudes of the tavern
and the home, although their behavioral expressions may be muted. With
individuals increasingly oriented to their work as an extension of their core
selves, as we become increasingly "professionalized," we are where we work.
To use a worn metaphor, those who labor beside us become our kith and
kin.
References
Allport, G., & Postman, LJ. (1947). The psychology of rumor. New York: Holt
Aran, G. (1975). Parachuting. American Journal of Sociology, 80, 124-152.
Becker, H.S. (1960). Notes on the concept of commitr.J.ent American Journal of
Sociology, 66, 32-40.
Becker, H.S., Geer, B., Hughes, E., & Strauss, A (1961). Boys in white: Student culture in
medical school. Chicago: University of Chicago Press.
Blau, P. (1956). Social mobility and interpersonal relations. American Sociological
Review, 21, 290-295.
Bogdan, R, & Taylor, S. (1975). Introduction to qualitative research methods. New York:
Wiley.
Bradney, P. (1957). The joking relationship in industry. Human Relations, 10, 179187.
Burowoy, M. (1979). Manufacturing consent. Chicago: University of Chicago Press.
Collins, E.G.e. & Blodgett, T.B. (1981). Sexual harassment: Some see it ... some
won't Harvard Business Review, 59, 76-95.
Collins, E.G.e. (1984). Managers and lovers. Harvard Business Review, 61, 142-153.
Collins, N.W. (1983). Professional women and their mentors. Englewood Cliffs, NJ:
Prentice Hall.
Corwin, R, Taves, MJ., & Haas, J.E. (1960). Social requirements for occupational
success: Internalized norms and friendship. Social Forces, 39, 135-140.
Cox, B. (1970). What is Hopi gossip about?: Information management and Hopi
factions. Man, 5, 88-98.
Cunningham, M. (1984). Powerplay. New York: Simon & Schuster.
Dandridge, T.e. (1975). Symbols at work: The types andfunctions of symbols in selected
organizations. Unpublished Doctoral dissertation, University of California, Los
Angeles.
Davis, K (1969). Grapevine communication among lower and middle managers.
Personnel Journal, 48,269-272.
Davis, K (1972). Human behavior at work: Human relations and organization behavior
(4th ed.). New York: McGraw-Hill.
204
Gary Alan Fine
Deal, T.E., & Kennedy, AA (1982). Corporate cultures. Reading, MA: AddisonWesley.
Domhoff, G.W. (1974). The Bohemian grove and other retreats. New York: Harper &
Row.
Feld, S.L. (1984). The structured use of personal associates. Social Forces, 62, 640652.
Fine, GA (1979). Small groups and culture creation: The idioculture of Little
League baseball teams. American Sociological Review, 44, 733-745.
Fine, GA (1980). The natural history of preadolescent male friendship groups. In H.
Foot, A Chapman, & J. Smith (Eds.), Friendship and social relations in children (pp.
293-320). Chichester: Wiley.
Fine, G.A (1984a). Humorous interaction and the social construction of meaning:
Making sense in a jocular vein. In NK Denzin (Ed.), Studies in symbolic
interaction, (Vol. 5, pp. 83-101). Greenwich, CT: JAI Press.
Fine, GA (1984b). Negotiated orders and organizational cultures. Annual Review of
Sociology, 10, 239-262.
Fine, GA, & Kleinman, S. (1983). Network and meaning: An interactionist
approach to structure. Symbolic Interaction, 6,97-110.
Fitt, L.W., & Newton, DA (1981). When the mentor is a man and the protege a
woman. Harvard Business Review, 59, 56-60.
Foote, N.N. (1953). Love. Psychiatry, 16, 245-251.
Gannon, T.M. (1981). Friendship patterns of the American Catholic clergy. In H.
Lopata & D. Maines (Eds.), Research in the interweaves of social roles: Friendship
(Vol. 2, pp. 197-217). Greenwich, CT: JAI Press.
Gluckman, M. (1963). Gossip and scandal. Current Anthropology, 4,307-316.
Goffman, E. (1959). Presentation of self in everyday life. New York: Anchor.
Goffman, E. (1961). Asylums. Garden City, NY: Anchor.
Goode, W.J. (1972). The place of force in human society. American Sociological
Review, 37, 507-519.
Granovetter, M.S. (1973). The strength of weak ties. American Journal of Sociology, 78,
1360-1380.
Gregory, KL. (1983). Native-view paradigms: Multiple cultures and culture conflicts
in organizations. Administrative Science Quarterly, 28, 359-376.
Gruber, J.E., & Bjorn, L. (1982). Blue-collar blues: The sexual harassment of women
autoworkers. Work and Occupations, 9, 271-298.
Gruner, C. (1979). Understanding laughter. Chicago: Nelson-Hall.
Haas, J. (1972). Binging: Educational control among high steel ironworkers.
American Behavioral Scientist, 16,27-34.
Haas, 1 (1974). The stages of the high-steel ironworker apprentice career. The
Sociological Quarterly, 15,93-108.
Handelman, D. (1976). Re-thinking 'banana time': Symbolic integration in a work
setting. Urban Life, 4,433-448.
Haviland, lB. (1977). Gossip, reputation and knowledge in Zinacantan. Chicago:
University of Chicago Press.
Hebden, lE. (1975). Patterns of work identification. Sociology of work and occupations, 2, 107-132.
Hershey, R (1956). Heed rumors for their meaning. Personnel Journal, 34,299-301.
Hughes, E.c. (1945). Dilemmas and contradictions of status. American Journal of
Sociology, 50, 353-359.
Jones, M.O. (1980). A feeling for form, as illustrated by people at work. In N.
Burlakoff & C. Lindahl (Eds.), Folklore on two continents: Essays in honor of Linda
Degh. Bloomington, IN: Trickster Press.
Kantor, RM. (1977). Men and women of the corporation. New York: Basic.
205
Kerckhoff, A, & Back, K (1968). The June bug: A study of hysterical contagion. New
York: Appleton-Century-Crofts.
Kleinman, S. (1984). Equals before God. Chicago: University of Chicago Press.
Kram, KE. (1983). Phases of the mentor relationship. Academy of Management
Journal, 26, 608-625.
Kurth, S.B. (1970). Friendships and friendly relations. In GJ. McCall, M.M. McCall,
N.K Denzin, G.D. Suttles, & S.B. Kurth (Eds.), Social relationships (pp. 136-170).
Chicago: AIdine.
Leary, J.P. (1978). Strategies and stories of the Omaha stockyards. Folklore Forum, 11,
29-41.
Leemon, T.A (1972). The rites of passage in a student culture. New York: Teachers
College Press.
Lembright, M.P., & Riemer, 1.W. (1982). Women truckers' problems and the impact
of sponsorship. Work and Occupations, 9,457-474.
Levinson, D., Darrow, C.N., Klein, E.B., Levinson, M., & McKee, B. (1978). The
seasons of a man's life. New York: Ballantine.
Lincoln, J.R., & Miller, 1. (1979). Work and friendship ties in organizations: A
comparative analysis of relational analysis. Administrative Science Quarterly, 24,
181-199.
Lipset, S.M., Trow, M., & Coleman 1. (1956). Union democracy. New York: Free Press.
Little, R (1981). Friendships in the military community. In H.Z. Lopata & D. Maines
(Eds.), Research in the interweave of social roles: Friendship (V01. 2 pp.22l-235).
Greenwich, CT: JAI Press.
Martin, J., & Powers, M.E. (1983). Organizational stories: More vivid and persuasive
than quantitative data. In B. Staw (Ed.), Psychological foundations of organizational
behavior (pp. l61-l68). Glenview, IL: Scott, Foresman.
Martin, 1., & Siehl, C. (1983). Organizational culture and counterculture. Organizational Dynamics, 12,52-64.
McCarl, RS., Jr. (1976). Smokejumper initiation: Ritualized communication in a
modern occupation. Journal of American Folklore, 89,49-66.
Mechanic, D. (1962). Sources of power of lower participants in complex organizations. Administrative Science Quarterly, 7,349-364.
Mechling, E., & Mechling, 1. (in press). Shock talk: From consensual to contractual
joking relationships in the bureaucratic workplace. Human Organization.
Myers, H. (1984}.Justfriends: Definition and maintenance ofplatonic cross-sexfriendship.
Unpublished manuscript, University of Minnesota.
Nusbaum, P. (1978). A conversational approach to occupational folklore: Conversation, work, play, and the workplace. Folklore Forum, 11, 18-28.
Packard, V. (1979). The status seekers. New York: McKay.
Paine, R (1967). Gossip and transaction. Man, 3,305-308.
Peters, TJ., & Waterman, RH., Jr. (1982). In search of excellence. New York: Harper &
Row.
Pfeffer, 1. (198l). Management as symbolic action: The creation and maintenance of
organizational paradigms. In L.L. Cummings & B. Straw (Eds.), Research in
organizational behavior (Vol. 3, pp. 1-52). Greenwich, CT: JAI Press.
Quinn, RE. (1977). Coping with Cupid: The formation, impact, and management of
romantic relationships in organizations. Administrative Science Quarterly, 22, 30-45.
Riemer, 1. (1979). Hard hats: The work world of construction workers. Beverly Hills:
Sage.
Rohlen, T.P. (1973). Spiritual education in a Japanese bank. American Anthropologist,
75, 1542-1562.
Roy, D. (1952). Quota restriction and goldbricking in a machine shop. American
Journal of Sociology, 57,427-442.
206
Gary Alan Fine
Roy, D. (1954). Efficiency and the fix: Informal intergroup relations in a piecework
machine shop. American Journal of Sociology, 60,255-266.
Roy, D. (1959-1960). Banana time: Job satisfaction and informal interaction. Human
Organization, 18, 156-168.
Schutte, J.G., & Light, J.M. (1978). The relative importance of proximity and status
for friendship choices in social hierarchies. Social Psychology, 41, 260-264.
Shibutani, T. (1966). Improvised news. Indianapolis: Bobbs-Merrill.
Spacks, P. (1983). Gossip: How it works. Yale Review, 72,561-580.
Stelling, J., & Bucher, R (1972). Autonomy and monitoring on hospital wards.
Sociological Quarterly, 13,431-446.
Strauss, A, Schatzman, L., Ehrlich, D., Bucher, R, & Sabshin, M. (1963). The
hospital and its negotiated order. In E. Friedson (Ed.), The hospital in modem
society (pp. 147-169). New York: Free Press.
Suttles, G. (1970). Friendship as a social institution. In GJ. McCall, M. McCall, N.K
Denzin, G.D. Suttles, & S.B. Kurth (Eds.), Social relationships (pp. 95-135).
Chicago: Aldine.
Swanson, C. (1978). Joking at the office: Coffee-break humor. Folklore Forum, 11,
42-47.
Timms, D.W.G. (1967). Occupational stratification and friendship nomination: A

study in Brisbane. Australian and New Zealand Journal of Sociology, 3, 32-43.
University of Minnesota. (1984). Policy statement on sexual harassment. Unpublished
manuscript, University of Minnesota.
Vaught, c., & Smith, D.L. (1980). Incorporation and mechanical solidarity in an
underground coal mine. Sociology of Work and Occupations, 7, 159-187.
Verbrugge, L.M. (1979). Multiplexity in adult friendships. Social Forces, 57, 12861309.
Wright, P.H. (1982). Men's friendships, women's friendship and the alleged
inferiority of the latter. Sex Roles, 8, 1-20.
Zuckerman, H. (1977). Scientific elite: Nobel laureates in the United States. New York:
Free Press.
Chapter 11
Social Support and Stress:

The Buffering Effects of Friendship
The idea that people have beneficial effects on one another is not a new one,
but scientists have only recently become aware of the extent of these effects.
Research on mental and physical health outcomes has suggested that
having friends and confidants can have a positive effect on how one copes
with stress and how vulnerable one is to a variety of ills. However, the
reasons for this are not clear, and the ways in which having friends
translates into these positive outcomes, or how loss of friendship relates to
more negative ones, is not well-established either. In this chapter, we focus
upon the relationship between friendship, social support, and stress. Several
issues are addressed, including the nature of social support and its relation
to social comparison theory, the stress response and how it is affected by
support variables, and how friendship and social support overlap and differ
from one another.
In the 1950s, Festinger (1954, 1957) proposed his theories of social
comparison and cognitive dissonance, suggesting that others can play a
substantial role in the development of opinions and needs as well as
reduction of dissonant cognitions. These theories focus on the fact that
people look to others in their environment in order to evaluate their own
opinions, beliefs, and attitudes. Through this process, one can evaluate
where one stands with reference to the opinions and attitudes of an
appropriate comparison group, composed of others similar to oneself. With
regard to abilities, one may compare performance on a given task to a
standard (e.g., how fast one can run the mile can be compared to an average
time for a runner of the same age and athletic ability or to one's own last
performance at the same distance). However, evaluation of one's opinions
or attitudes has no objective reference and yields information that is relative
only to the anchor provided by the comparison group that one chooses.
Essentially, researchers have determined that individuals evaluate their
opinions, beliefs, and attitudes through comparison with similar others and,
208
partly as a result of such findings, propositions have been made that social
support may be derived from_ the social comparison process. More
specifically, it is possible that the processes involved in social comparison
provide a basis for the beneficial action of social support.
The way that we view the environment is profoundly influenced by those
who are close to us. Festinger (1954) provided a real-life example of how
social comparison may affect one's appraisal of a situation. He cited
instances in which persons who had experienced an earthquake looked to
others in order to assess the environment and to determine whether the
worst was over or if more destruction was forthcoming. In this instance,
social interaction was seen to validate (through rumors) individuals'
estimates of what was likely within their environment, even though few, if
any, really knew. Festinger showed that two types of rumors developed:
excitatory and calming. Depending upon the type of rumor that circulated,
people were more or less frightened about the possibility of destruction in
their area. This may have led to stress for persons hearing excitatory rumors
and to a reduction in stress for persons exposed to calming rumors (the
worst of the earthquake is over). It is also possible that gaining some
validation of beliefs about the danger was calming, regardless of what the
information or belief was.
Another example is provided by Schachter (1959) in his studies of
affiliation. He showed that some people preferred to wait with similar others
when threatened (with electric shock). Schachter proposed that individuals
could deal better with the uncertainty and anxiety created by the threat by
choosing to wait with a group of similar others because the anxiety led to a
need for social comparison of one's feelings and expectations. This
"affiliative need" was postulated to lead to anxiety reduction by clarifying,
reducing, disconfirming, or otherwise influencing evaluation of the situation. Waiting in a group served as a means of support for the subject. As
Festinger (1954) noted "it seems clear that the drive for self-evaluation is an
important factor contributing to making the human being 'gregarious"'(p.
136).
These social processes highlight a history of research focused on the ways
in which people derive benefits from interacting with one another. We are
clearly a social species, and the advantages we derive from group membership are extensive. Social support refers to some of these benefits, and one
source of positive results of group membership is the constructive process of
refining one's opinions, placing them in context, and providing us with a
firm reference point with which to create a social reality. However, there are
a number of other aspects of social support, including the assistance
received from friends and emotional aid provided when one is troubled. We
discuss these issues in this chapter, exploring different hypotheses about
how support comes about and why it is effective in reducing stress and
affecting health. Before doing so, however, we briefly review the notion of
stress.
Social Support and Stress
209
Stress
During the first 2 decades of this century, the autonomic nervous system
and its component parts were being discovered and identified. In 1911,
Cannon and de la Paz first demonstrated that the adrenal gland secretes a
vasoactive substance (sympathin), which has become known as epinephrine. This discovery led Cannon to very detailed investigations into the
function(s) of the sympathetic nervous system and, finally, to his delineation
of the "fight or flight" response (Cannon, 1914, 1928, 1936).
In outlining the "fight or flight" response, Cannon (1914,1928,1936) listed
many accompanying physiological changes that were thought to facilitate
either fighting or fleeing. These changes included increased cardiovascular
response (arterial pressure, heart rate, myocardial contractility), increased
respiration, increased blood flow to the major striated muscle groups (and
concurrent decreased flow to organs not needed for rapid activity),
increased muscle strength, and increased sweating. These changes were
explained in a Darwinian framework, which posited their utility in
preparing the organism to defend itself or flee from a threat. All of these
responses to stress were autonomic changes, and were also observed during
states of pain, hunger, fear, and rage. Adrenal hormones have been shown to
evoke and prolong such autonomic fluxes and have been recognized as
being of major importance to the "fight or flight"/stress response.
Another early theory of stress was proposed by Selye (1956), who noticed
what he referred to as a "nonspecific response to injury." This discovery
came as he was studying the effects of injecting extracts of sex hormones in
rats. Following injections, Selye found that no matter how pure he made his
extracts, the rats developed a three-part pathology after exposure to the
solutions: enlargement of the adrenal glands, involution of the thymus, and
gastric ulceration and bleeding. It was soon discovered that this was a
nonspecific response to injury occurring in response to almost any noxious
or aversive event. In order to explain these effects, Selye developed the
general adaptation syndrome.
The general adaptation syndrome consists of three stages of reaction to
aversive stimuli: alarm, resistance, and exhaustion. In the alarm stage, the
organism prepares to respond to the noxious stimulus with enhanced
activity, releasing both pro- and anti-inflammatory corticosteroids. This
stage is relatively short in duration and is followed by resistance. During the
resistance stage, the organism no longer displays any overt signs of
hyperreaction to the intruding stimulus and appears to have adapted to it.
The final stage, exhaustion, is marked by a reappearance of signs of the
organism's struggle to adapt to the stimulus. It is during this stage that Selye
(1956) believed the "diseases of adaptation" (e.g., hypertension) occurred.
This stage ends in the death of the organism.
Selye (1956) presented important information about physiological response to stressors. He argued that the response to stress is nonspecific (and
210
always consists of the triad and the general adaptation syndrome), and he
provided a link between stress and disease or illness. Although the notion of
nonspecificity of response to injury has been questioned (see Mason, 1975),
the general concept of a response to stress has received enormous attention.
However, very little attention was paid to the influence of central appraisal
mechanisms or psychological mechanisms in this generalized response.
Lazarus (1966) brought psychosocial variables and the appraisal process
into the picture. He pointed out that stress has an important psychological
component. Not all stress experiences will be psychological; there is very
little that is psychological about infection. Yet, there are many stressors that
are psychological, at least in part (that is, they rely on interpretation for any
effect). The ways in which stressors, such as loss of a loved one or failing an
important examination, affect us can be mediated, to a large extent, by
psychological processes (e.g., Frankenhaeuser, 1978; Kasl & Cobb, 1970).
Appraisal is central to Lazarus' theory. Lazarus (1966) suggested that
virtually the only way for a stimulus to be a stressor is for it to be appraised
as threatening or involving harm or loss for the perceiver. Appraisals of
threat or harm/loss lead to stress and the need for a coping response to the
situation. This coping response, or secondary appraisal, can lead to two
basic forms of coping. Direct-action coping involves interaction with the
stressful stimulus in such a way as to change the situation (for example, if
something breaks, you fix it). Palliative coping involves manipulation of one's
emotional response to a threatening situation (for example, trying not to let
the threat bother you; thinking of other things). Secondary appraisal is thus
focused on adapting to the stressful situation, in other words, coping with
it.
The effects of stress may be expected to appear when coping is
inadequate. When this occurs, adaptation to the stressor is incomplete and
may be manifested behaviorally. These manifestations of reaction to the
stressor may occur during exposure to the stressor. In this case, the
responses are referred to as stress "effects." If the reaction to the stressor
continues after the stressor is no longer present, then the reaction is termed
an "aftereffect." Glass and Singer (1972) discussed the "aftereffects" of
exposure to loud noise. They found that subjects who were led to believe
that they could shut off the loud noise if it became unbearable performed
better on a concentration task (proofreading) after the cessation of the noise
than those who believed that they could do nothing to terminate the noise.
Glass and Singer hypothesized that the perceived control over the noise was
sufficient to mitigate its aftereffects.
The stress response involves not only psychological up~et but also the
physiological "readying" processes described by Cannon (1936). Research
on the stress response has become more integrative, including affective,
behavioral, and physiological manifestations of response to a given stressful
stimulus and has necessarily become more interdisciplinary (Baum,
Grunberg, & Singer, 1982). In general, we have come to a more integrative
211
notion of the stress response, which includes not only the physiological and
affective domains, but also the behavioral domain as well. It is within the
behavioral domain that the interplay between the affective response(s) and
physiological response(s) is often manifested.
Coping with a stressor has been shown to be facilitated by a number of
variables. For instance, control has been extensively studied for its effects in
mediating the stress response (Glass & Singer, 1972; Rodin, Rennert, &
Solomon, 1980). Investigators have also shown that the loss of control may
be regarded as stressful (Baum & Valins, 1977) and that even the illusion of
control is powerful in mitigating the effects of stress (Glass & Singer, 1972;
Langer, 1975).
Attitudes toward or opinions about a potential stressor have also proven
to be effective mediators of the stress response. For example, Jonsson and
Sorenson (1967) found that giving subjects either a positive or a negative
induction to airport noise can affect their perceptions of the noise as
bothersome. Also, Sundstrom, Lounsbury, DeVault, and Peele (1981) have
reported that negative attitudes toward a nuclear power plant were related to
perceptions of hazards associated with the functioning of the plant.
Evidence suggests that the relationships that one has with others also
mediate stress (Cassel, 1976; Cobb, 1976). Friends, loved ones, and
neighbors can provide aid in coping with stressful situations. This aid may
take the form of tangible assistance (e.g., direct assistance in solving the
problem) or may be manifested as emotional support and understanding.
Others may be the best resource in dealing with a stressful situation, and the
relationship of social support to stress responding is a crucial one.
Friendship Formation
In 1950, Festinger, Schachter, and Back published a monograph on the way
that social factors influence informal groups. Within this project they
examined friendship formation and found that the physical layout of a
housing project was a powerful determinant of who became friends with
whom. Their finding was striking: The functional spatial relationships
within the project (e.g., whose door was in the path between one's own door
and one's destination) determined friendship formation rather than merely
the physical distance between individuals. This finding helped to demonstrate the importance of passive social contact in determining friendship.
That is, those with whom one has the greatest amount of contact are most
likely to become one's friends. The implications of these findings were
followed over a full decade later when research into the effects of
architectural design characteristics on crime rates (Newman, 1972) and
social control (yancey, 1972) produced results suggesting that the arrangement of space within a neighborhood could affect not only vandalism and
crime, but also the likelihood of someone getting to know (and becoming
friends with) one's neighbors.
212
Yancey (1972) highlighted the necessity of appropriate space for passive

contact in the development of social control over that space. He described
the atomization of social networks and the inhibition of group formation as
a function of the design of a housing project. He argued for his point by
contrasting the Pruitt-Igoe housing project, which had provided a minimum
of appropriate space for informal social contact, with an adjacent "slum"
neighborhood, which provided more of this type of space. He argued that in
the slum neighborhood the cluttered streets and alleys provided space for
people to gather and conduct informal social contacts. This not only
supported informal social control over the neighborhood, but also facilitated the development of a social network of friendships. In contrast, the
Pruitt-Igoe high-rise apartment provided a minimum of space for informal
social contact. This design minimized passive social contact by eliminating
space that was appropriate for it, and residents retreated into their
apartments. Social networks did not really develop at Pruitt-Igoe and,
eventually, due to rampant crime, vandalism, and low occupancy, the entire
project had to be demolished.
This distinction, based on the availability of controllable space for
informal social contact, is highlighted in an anecdote told by Yancey (1972).
Although virtually no instances of social network formation were seen in the
Pruitt-Igoe project, a curious occurrence in One of the high-rises was noted.
This particular building was scheduled for maintenance and installation of
some recreational equipment. In order to protect people from injury and the
equipment from theft, authorities installed a construction fence around the
entire building. The keys were given only to the workers and residents of the
building. During the 6 months of the construction project, the appearance of
the buildings' hallways changed dramatically. Prior to the construction of
the fence, the grounds were strewn with litter and debris, and the hallways
were filled with trash, broken glass, and the stench of urine. Toward the end
of the 6 months, the building looked entirely different. Residents who had
previously remained in their apartments, not getting involved with anything
going on in the hallways or outside, now came out of their apartments,
cleaned up the hallways and the apartment grounds, and as a group
petitioned to have the construction fence remain in place. Also, the vacancy
rate of most of Pruitt-Igoe was about 70%, but this building's vacancy rate
was only 2 to 5% after the fence was erected (Newman, 1972). This is a
dramatic instance of the influence of the surrounding environment on
informal social contact and friendship (or social support). Although the
"slum" neighborhood residents complained about the internal workings of
their homes (e.g., heating), they were generally satisfied with the exterior
environment, such as the neighborhood and friendships. Yancey (1972)
described this fenced area as supporting a social network, where neighbors
looked out for the interests of one another. The addition of a construction
fence made dramatic changes in the one high-rise, which appeared to head
in the direction of a stronger social network.
213
There is a paucity of research isolating sources of social support.

Commonsense suggests that friendship formation is related to social
support, and, as a result, one can turn to determinants of friendship as a
potential source of support. However, social support is not simply a matter
of friendship, and though one would expect some index of the quality and
quantity of friendship to correlate positively with support, other factors are
no doubt involved. In Schachter's (1959) work on affiliation, more subjects
exposed to the threat of electric shock chose to wait with other subjects who
were also threatened with shock rather than with dissimilar others.
Presumably, this group of similar others provided an adequate comparison
for their own emotions and reactions to the threatening situation. Similarly,
if a person is faced with a specific stressful situation, not all of his or her
friends may be able to provide equal or appropriate support. Indeed, at such
times it may be just those one or two friends who are experiencing (or have
experienced) the same situation who are best suited to provide support. The
problem lies in whether those friends who are best suited for comparison
are available to us at the time. It seems reasonable that a pregnant woman
might receive different degrees of support from her friends who have never
been pregnant than from those friends who are now (or have been)
pregnant. Thus, although our friends may be willing to provide us with
support, they may be better equipped to help us in some situations than in
others.
Whether these dynamics affect social support is not currently known.
Logic suggests that they would, and research is presently underway to test
the possibility that environmental factors affect social support networks. In
all likelihood, the development of support depends on an interaction among
environmental, social, and personality variables. The delineation of these
influences remains an interesting research question.
Social Support and Friendship

Research on social support derives from several disciplines (for example,
anthropology, sociology, psychology) and a number of theories, including
conceptualizations of social comparison. Although there are literally
dozens of definitions of social support circulating in the literature, most
include emotional support, tangible assistance, and informational components. That is, social support is seen as providing anyone or a combination
of these three. Emotional support refers to benefits associated with having
someone to whom one can turn for reassurance or comfort. People may also
derive feelings that they are valued and cared for. Typically, individuals with
adequate emotional support feel that there is someone to whom they could
turn when troubled. Tangible support refers to the availability of physical
assistance (for example, having someone to drive one to the store, assistance
in fixing one's car). This support can be obtained from persons who are not
214
close to us as well as from those who are and includes financial assistance or
aid in solving a problem. Injonnational support may be thought of as aid in
understanding a problem. For instance, rumors that circulate in times of
disaster may help to define the situation and thereby reduce ambiguity and
uncertainty. The opportunity to discuss rumors in a group and compare
one's opinions with others may also be beneficial. Cobb (1976), in an
address to the American Psychosomatic Society, defined social support in
terms of benefits associated with feeling loved and valued and with being a
member of a "'network' of communication and mutual obligation" (p. 300).
His conception was that the encouragement, opinion validation, and
reassurance that people get from friends and family influence their response
to stress and somehow make them more resistant to its effects. Some
research has indicated that during periods of stress or life change people
manage better when they can derive support from social relationships
(Cobb, 1976; Cohen & McKay, 1984; Kaplan, Cassel, & Gore, 1977).
Social support has been categorized in a number of ways. Some
researchers have focused on the distinction between quantitative and
qualitative aspects of support (Donald, Ware, Brook, & Davies-Avery, 1978).
The difference here derives from varying operationalizations of the concept
of social support, which can be measured in terms of amount of social
contact (the number of persons one has available to interact with) or quality
of contact (the usefulness of the interactions to the person). For example,
Miller and Ingham (1976) found that women with a greater number of
acquaintances reported fewer physical symptoms than did women who
reported having fewer acquaintances.
Sources and Function of Support
Research has not directly tested hypotheses about where social support
comes from or how it develops. Most researchers treat it as a full-blown
phenomenon that either is there or is not. Little attention has been paid to
social or environmental conditions that promote or inhibit the development
of social support. Clearly, support is more than simply having friends, and
not having support is more complex than being unpopular. Because social
support ultimately derives from group and friendship formation, however, it
is possible that variables that affect these processes also partially determine
support levels. In the above instance, "counts" of social supports were
predictive of symptom reporting. Alternatively, Medalie and Goldbert
(1976) reported that among highly anxious men, those who perceived their
wives as more supportive showed a lower incidence of angina pectoris.
These researchers relied on perceived support as their predictor variable,
presumably holding the number of supports constant. This state of affairs,
where different studies present findings based on different measures, can
result in widely variant approaches to the assessment of social support and
sometimes contradictory findings.
215
It appears that a distinction can be drawn between emotional aspects of

support (e.g., reassurance, comfort) and instrumental aspects of support
(e.g., tangible assistance). Lin, Dean, and Ensel (1981) have drawn such a
comparison, and Schaefer, Coyne, and Lazarus (1981) have demonstrated
that emotional and informational support are highly correlated, whereas
tangible assistance remains distinct. Thus, at least two major types of
support-instrumental and emotional-have been delineated.
Wallston, Alagna, DeVellis, and DeVellis (1983) noted three ways in
which social support may affect stress and related changes in health. First,
social support may directly remove the threat (e.g., financial assistance might
eliminate threats associated with .loss of a job) or at least reduce it. This
suggests that social support may change features of the environment that
were threatening to the individual. Second, social support may reduce stress
by changing perceptions of threat. If something is viewed as less dangerous
because one's friends view it that way, stress may be reduced or prevented.
Third, support from others may increase one's sense of control, which could
in turn reduce the consequences of a threat (see Glass & Singer, 1972).
Cohen and McKay (1984) have also suggested that the help available by
virtue of membership in a group must be appropriate to the specific threat
and the needs of the recipient. Although financial support may benefit an
individual experiencing money-related troubles, it will likely be less
efficacious to offer money to someone who is grieving the loss of a loved
one. The reverse may also be true; lending a sympathetic ear to someone
facing bankruptcy may not be as effective in reducing the threat as lending
that person the necessary cash to solve his or her problems. The realization
that specific types of social support may be of benefit in some instances and
not in others is at least partly related to the problem of measurement.
Measuring Social Support

Several problems that characterize research on social support limit the
findings of much of the research. Many studies of the role of social support
in stress-mediated illness or health problems focus on self-reported
symptomatology as the major dependent measure. Although self-reported
symptoms may well be related to stress and sensitive to variations in
support, such measures are subject to a number of biases that limit their
generalizability (Baum et aI., 1982). Integrated assessments of several
measures of stress (i.e., combinations of self-report, behavioral, and
physiological measures) should produce more informative results.
Parallel problems exist in the measurement of social support. Some
measures of social support focus on direct counts of social contacts and
cannot account for differences in the quality of these interactions. More
frequent social contacts are interpreted as an indication of greater social
support. However, more contacts do not necessarily indicate better support;
the quality of one's social interactions may be equally or more important in
216
determining support than the raw number of interactions. Accordingly,

some studies have measured perceived social support rather than counting
contacts. There are problems with this measurement strategy as well, but
inclusion of measures of perceived support or perceived adequacy of
support may increase the utility of social support in predicting relationships
between social support and stress. Cohen and McKay (1984) referred to
these perceived social support measures as functional measures of support,
and they have had success in finding effects of social support within the
extant literature when researchers use this type of measure.
Another methodological limitation of social support studies has been a
reliance, almost exclusively, on retrospective analysis from which no causal
inferences can be drawn (Wallston et aI., 1983). Further, as Thoits (1981)
noted, the operationalization and conceptualization of studies on social
support and stress are too often inadequate. Few studies distinguish between
types of support used. Thus, reports of social support are often combinations of instrumental and emotional support or, because function is not
often specified, studies of emotional and instrumental support may be
compared, not knowing that they are different. Finally, measurement of
social support has not been standardized; few scales have been tested for
reliability and validity, and most studies have used a measure designed for a
particular study.
Despite these problems, research has suggested that having social support
is a good thing, that it seems to help us deal with stress, and that health is
also affected by it. However, there is less agreement on how this occurs or
even why. This is more than a matter of identifying function, however, and
at least two positions on the effects of social support have been proposed.
The first, which has been called the "assets-benefits hypothesis" (Fleming,
Baum, Gisriel, & Gatchel, 1982) predicts main effects for social support. The
second hypothesis predicts interaction effects between stress and social
support (Cohen & McKay, 1984; Dean & Lin, 1977). In other words, social
support is useful in coping with stress, but not as helpful when stress is low.
This is referred to as the "buffering hypothesis." However, interaction effects
do not preclude main effects, and proponents of the buffering hypothesis do
not deny that main effects of social support may occur. The difference
between the two hypotheses may be seen as one of focus.
There are several benefits that may derive from having the support of
others. We may benefit directly through instrumental aid provided by
friends or supporters. Others may serve as a reference against which we may
validate our opinions and beliefs. Also, others may be influential in
maintaining our self-esteem or providing emotional support in times of
need or distress. The question for the investigator lies in the expression of
these effects in collected data. Controversy exists within the literature on
social support as to whether one should expect main effects of social
217
support, suggesting a general assets-benefits hypothesis, or if one should

expect interaction effects between social support and level of stress,
suggesting a buffering hypothesis.
It seems reasonable from previous discussion that support derived from
friendship(s) is beneficial in more than just one way. That is, it is reasonable
to expect that some of the effects of social support or friendship are direct
and that others are indirect. Direct effects of support might include
instrumental aid received from others (for example, financial assistance),
which may either reduce or eliminate the stressor or remove the individual
from the stressful situation. In effect, this type of aid may reduce stress
responding by directly eliminating or removing the stressful stimulus rather
than directly affecting palliative coping efforts. Alternatively, direct support
could help the individual to avoid stressful encounters. This would suggest
main effects of level of support (on stress responding). Emotional support
may function in a more indirect manner in reducing stress. We may benefit
the most from the emotional reassurance and support from friends when we
are experiencing stress. If supporters are unable to assist us in avoiding,
removing, or eliminating a stressor, they may still be of benefit in helping us
with our response to the stressor. Emotional support from others may be
thought of as directed at our ability to deal with our own reactions involved
in coping with a stressful situation. In this case, one would predict
interaction effects between level of stress and level of emotional support.
Support from others also enables us to validate our opinions and beliefs
(through social comparison) and may bolster our overall levels of selfesteem. Both of these benefits of social support appear to be indirect in that
they are not aimed at changing the stressful situation. However, the effects
of these types of aid may benefit stressed individuals either by influencing
their coping with the stress (palliation) or by enabling them to cope more
actively with the stressor (direct action). Thus, either the general assetsbenefits hypothesis or the buffering hypothesis may be demonstrated for
this type of support. Overall, investigators must be careful to understand not
only the type of support being provided, but also the context within which it
is made available. To the extent that this information is determinable, better
predictions of the effects of social support and friendship may be made in
reference to stress responding.
The Assets-Benefits Hypothesis

Bradburn and Caplowitz (1965) reported that high affiliators show greater
avowed happiness than do low affiliators. This finding reflects the fact that
people affect one another all of the time. In that sense, it is consistent with
the notion that social support may be beneficial to people even when they
218
Raymond fleming and Andrew Baum
are not under stress. Researchers have also found that general benefits, such
as longer life and better health, are associated with higher levels of support
(Berkman, 1977; Cassel, 1976). These benefits may well be due to direct,
positive effects of social support or to the indirect effects of social support in
lessening the effects of stress. Alternatively, it may be that not having social
support is stressful.
This alternative interpretation of an assets-benefits hypothesis suggests
that stress may not be necessary to produce support effects because lack of
social support is bad or stressful in and of itself. Studies of bereavement
have shown devastating health effects of losing a loved one, including
depressed immune functioning (Bartrop et a!., 1977), heightened susceptibility to illness and infection (Holmes & Masuda, 1974), and greater
mortality (Kraus & Lilienfeld, 1959). Gans (1962) studied the disintegration
of neighborhoods in Boston as a result of urban renewal efforts and noted
that previously extensive support networks dissolved and left many
residents of the area experiencing a grief reaction. Fried (1963) agreed with
Gans's analysis and reported that elevated levels of depression and physical
symptomatology remained in several relocated residents up to 2 years after
the move. These consequences were attributed to grieving for the lost
neighborhood and friends rather than to aspects of the new neighborhood.
Also related to this is the notion of social breakdown syndrome noted by
Grunberg (1967) in discussing institutionalized mental patients. The focus
of this research was on how patients give up a sense of independence and
begin to adapt to institutional life after their social support networks fail
them. This formulation is similar to Seligman's (1975) description oflearned
helplessness, in which noncontingency is learned and lack of response
produced. Grunberg attributed much of the necessary institutionalization of
mental patients to the effects of losing the support of family, friends, and
community. Another instance of negative effects of loss of social support is
provided by Leighton (1959), who described needs that, when not met, could
lead to psychiatric disability. Changes in social relationships often result in
difficulty in meeting social needs and thus may facilitate psychiatric
complications. Segal, Weiss, and Sokol (1965) have noted greater psychiatric
utilization among individuals who did not affiliate as much as others.
Similarly, Langer and Michael (1960) made a distinction between having no
friends and having one or more friends; the former increased mental health
risks.
The direct relationship between level of social support and physical and
mental health has not been clearly established. Many studies suggest a
correlation between level of social support and psychological and physical
health independent of stress, but there is no work that can support causal
statements concerning the direction of this relationship.
219
The Buffering Hypothesis

Whether or not a lack of friendship and support is aversive by itself or
whether other conditions, such as stress, are required to demonstrate this
effect remains an area for future study. The buffering hypothesis proposes
that social support mitigates the effects of stress but does not necessarily
provide benefit in the absence of stress. Conversely, lack of support is not
necessarily stressful, but lack of support in the face of other stressors is
associated with poorer outcomes. There are a number of studies and reviews
suggesting that social support buffers stress, and it has been fairly wellestablished that having social support is beneficial during stress (Cassel,
1974a, 1974b; Cobb, 1976, Cohen & McKay, 1984, Dean & Lin, 1977;
Eckenrode & Gore, 1981). Space does not permit more than a selected
review of the most pertinent findings of this literature.
Two studies that bear on the relationship between social support and
stress and stress consequences have included measures of social support as
psychosocial assets. According to Nuckolls, Cassel, and Kaplan (1972),
psychosocial assets refer to factors that contribute to one's ability to adapt
during stress. This includes both direct, situation-oriented, manipulative
coping and more indirect, inwardly focused, accommodated response. In
particular, they were interested in women's ability to adapt to their first
pregnancy. They used a scale representing the adaptive potential for
pregnancy, which contains five categories of assets: self, marriage, extended
family, social resources, and definition of pregnancy. Only two of the five
subscales were related to social support; the remainder referred to psychological functioning, characteristics surrounding the marriage (e.g., concordance of age, duration of marriage), and questions concerning the women's
feelings about the pregnancy. The researchers measured complications
during pregnancy and delivery (e.g., Apgar rating of infant of less than 7,
maternal systolic blood pressure greater than 139 and/or diastolic over 89
during both the labor and the postpartum periods, stillbirth) and examined
them as a function of life change score and psychosocial assets rating.
Among the results of this study was the finding that of women with a high
life change scores for the 2 years before and during pregnancy, 91 % of
women with low psychosocial assets scores had one or more complications
with their pregnancy, whereas only 33% of women with high psychosocial
assets experienced any complications.
A second study considering psychosocial assets used the Berle Psychosocial Assets Index in a study of chronic asthma patients (DeAraujo, van
Arsdel, Holmes, & Dudley, 1973). The Berle Index includes three subscales:
patient history, patient's perceptions of family and interpersonal relations,
and physician's rating (based on the patient's personality structure and
attitudes toward their illness). One year after administering the Berle Index,
220
the researchers asked patients to complete a life change inventory and

consulted physicians' records of the patients' use of bronchodilator
medication. Results showed that patients with low life events scores used
less medication regardless of their psychosocial assets score, but among
patients with high life events scores, use of medication was only high if
psychosocial assets were low.
These studies support the buffering hypothesis, even though their
measures do not exclusively represent social support. It is conceivable that
overlap between measures of social support and psychosocial assets are
operative in buffering stress, but this cannot be inferred from the few studies
available, and no study has directly compared the two measures.
Other studies have more directly studied social support. In a study of 720
adults, Myers, Lindenthal, and Pepper (1975) focused on the relationship
between life events and symptoms. Although they did not analyze their data
in terms of social support, Eaton (1978) reanalyzed the study in just such a
manner. Eaton employed a measure of social support that consisted of
whether the subjects reported having friends, being married, and the like
and performed regression analysis to indicate the amount of variance that
these "social support" items could account for in symptom reporting.
Controlling for life events and psychiatric symptoms assessed in the
beginning of the study, Eaton found that being married or living with at
least one other person buffered symptom reporting.
Studies on the availability of a single confidant or close friend show fairly
strong support for the buffering hypothesis. Brown and Harris (1978)
reported on a study oflife change stress and psychiatric disorders in women.
In this study the measure of social support assessed the availability of a
confidant for these women. Women who reported intimate ties with their
husband or boyfriend were considered to have high social support, and
women who had the support of a friend, sister, or mother only once a week
were labeled as having low support. Results showed that women experiencing high life stress who had low support reported the greatest number of
symptoms. Women who experienced little life stress reported the fewest
symptoms, regardless of their social support level.
Lin, Simone, Ensel, and Kuo (1979) examined life events, social support,
and psychological impairment among 170 Chinese-Americans (all heads of
households). This study used a 9 item measure of social support, but the
study failed to show significant interactions between social support and life
events for psychological symptom reporting. However, Boyce (1981) argued
that the pattern of means reported by Lin et al. reveals a buffering effect.
Those individuals who were under the most stress and who reported the
highest levels of social support showed less psychological symptoms than
would be expected from the remaining means.
Cohen and McKay (1984) reviewed the results of a number of studies such
as Lin et al.'s (1979) that suggested a buffering effect of support but did not
221
report statistically significant interaction terms. In each of these studies,

there were results that fit the buffering hypothesis for social support, but the
strength of the findings varied considerably. Cohen and McKay argued that
there are actually two forms of the buffering hypothesis and that each can
be supported by current literature. The first form (the strong version)
predicts increased pathology for persons who experience high levels of
stress and who have low levels of support. Pathology is unrelated to stress
and relatively low for persons with high levels of support. The weak version
of the buffering hypothesis predicts that pathology may increase among
people with high levels of support when stress increases but that the extent
of this increase must be less than that among people with low levels of
support.
The degree to which different kinds of support or different functions of
suppport are buffered or not has not been established. Andrews, Tennant,
Hewson, and Vaillant (1978) considered several forms of social support (e.g.,
crisis support, neighborhood support) and examined their effects on
symptom reporting. Although no interactions were found, Cohen and
McKay (1984) pointed out that the mean values for the social support and
stress 2 x 2 design would yield a pattern similar to the "weak" buffering effect
Andrews et al. discussed. Similarly, Frydman (1981) presented data on a
study using the same measures as Andrews et al. and showed significant
interactions between social support and stress on "neighborhood support"
in one of the two samples used. Of the remaining comparisons that did not
show an interaction effect, three-quarters of both samples show the patterns
of means suggested by Cohen and McKay as reflecting the "weak" version of
the buffering hypothesis. Sometimes mixed findings add information rather
than confusion. In a study by Cohen and Hoberman (1983), measures of
tangible, emotional, self-esteem, and belonging support were used. Through
the use of regression analyses, the authors were able to demonstrate the
usefulness of a combination of four subscales of the support measure in
predicting depression and physical symptoms. Self-esteem and belonging
support buffered physical symptoms, whereas self-esteem and appraisal
support buffered depression. Comparison of the use of this scale with the
Inventory of Socially Supportive Behaviors on the same group showed that
the ISSB could not demonstrate the buffering effects. The ISSB is a global
support measure and may not have been sensitive enough to show the
buffering effects.
Fleming et al. (1982) showed that the perceived availability of support was
sufficient to buffer the emotional (self-reported symptoms of depression,
anxiety) and behavioral (errors on an embedded figures task) effects of
stress associated with living near the Three Mile Island nuclear power
station. Among people living near Three Mile Island, those who reported
that they had a great deal of support responded similarly to control subjects,
presumably experiencing lower levels of stress, whereas those reporting
222
lower levels of support responded with greater stress, (namely, greater

number of symptoms reported, fewer concentration errors on a timed task).
Results consistent with these, concerning the availability of a confidant,
have been reported by several researchers (Habif & Lahey, 1980; Medalie &
Goldbert, 1976; Miller & Ingham, 1976).
Strategies for Future Research

There are a number of strategies that might prove beneficial if adopted by
researchers interested in the effects of social support. As already mentioned,
researchers can start by adopting better methods of assessing stress (i.e.,
multimodal assessment). This includes prospective measurement of stress
as well as examination of discrete stressors, rather than reliance on reported
life change. Research concerning the influence of moderator variables, such
as social support, on stress responding can only be as good as the
measurement of the incidence of stressors and the stress response.
Measurement of social support also suffers from confusion; some research
focuses on social networks and measures of actual social interaction,
whereas other research focuses primarily on perceived availability and
adequacy of social support. Still other researchers use a combination of
these two approaches. Cohen and McKay (1984) suggested that social
support may only buffer the effects of stress when that support is
appropriate to the needs of the individual and to the demands of the
situation. This follows from Festinger's (1954, 1957) notion that social
support serves the function of promoting social comparison and dissonance
reduction. In this light, it may be advisable for researchers to study carefully
the demands of the specific stressful situation before deciding on an
appropriate measure of social support. Because not all forms of social
support may be operative (or even necessary) in a given situation, it would
be fruitful to measure those aspects of social support that are important to
the situation at hand.
There are some possible confounds within the notion of social support
that might be studied profitably. Heller (1979), for example, discussed a
"social competency" confound: the possibility that those who have the
highest levels of support are the most capable of eliciting support from
others, as well as being better able to cope with the vicissitudes of life in
general. Adequate controls and prospective studies could address this
directly.
It is also possible that social support affects different aspects of stress in
different ways. In Fleming et al.'s (1982) study of stress among Three Mile
Island area residents, a perceived support measure (emotional support)
showed buffering effects for self-reported symptoms, depression, and
behavioral performance on a concentration task. Main effects of stress and
223
social support rather than interactions were obtained for urinary norepinephrine levels and for self-reported somatic distress, suggesting that the
assets-benefits hypothesis might better explain this pattern of results. These
findings have not yet been replicated, but if they are replicated, more
specific hypothesizing may become the rule rather than the exception in the
literature on social support and stress.
Other issues also remain to be investigated. The role of the physical
environment in shaping, facilitating, or inhibiting the development of
friendship networks and social support is not known. We do know that
architectural arrangement of space can increase casual social contact,
facilitate the development of friendships, and provide usable space for
social activities (e.g., Festinger, Schachter, & Back, 1950). These physical
features can also atomize social networks and inhibit friendship formation
or local group development (Baum & Valins, 1977; Yancey, 1972). Whether
these effects also translate into differences in social support or whether they
are supportive or sociallly enhancing environmental arrangements has not
been studied.
The role of other mediating conditions in generating the effects associated
with social support is another area that awaits careful study. The effects of
social support, for example, are like those of perceived control; when social
support is present, stress and its consequences are less than when it is
absent. It is possible that social support has its effects by increasing
perceived control or that both of these effects are related to notions of selfefficacy. Again, work is needed to clarify the situation.
In summary, although research generally indicates that a relationship
between social support and stress exists, specifics of the relationship are not
as well-established. An overreliance on life events measures as indicators of
stress has tended to keep findings very general (for example, social support
buffers a person against the effects of increased life change). Better
techniques for measuring the stress response, including the use of multimodal assessments, are available and affordable and may allow greater
freedom in the types of stressful situations studied with respect to social
support. Reactions to disasters (Fleming et aI., 1982), job stress (La Rocco,
House, & French, 1980), and unemployment (Gore, 1978) are issues
currently being investigated for the potential role of social support in
mediating stress in these situations. Also, studies of reactions to crowding
(Baum & Valins, 1977; Paulus, McCain, & Cox, 1978) have provided some
information concerning how mediating variables (e.g., perceptions of
control) may be related to stress produced in these situations.
Acknowledgments. This chapter was facilitated by support from the Uniformed
Services University of the Health Sciences Protocol No. C07216. The opinions or
assertions contained herein are the private ones of the authors and are not to be
construed as official or reflecting the views of the Department of Defense or the
Uniformed Services University of the Health Sciences.
224
References
Andrews, G., Tennant, e., Hewson, D.M., & Vaillant, G.E. (1978). Life events stress,
social support, coping style, and risks of psychological impairment. Journal of
Nervous and Mental Disease, 166, 307-316.
Bartrop, RW., Luckhurst, E., Lazarus, L., Kiloh, L.G., & Penny, R (1977,
April). Depressed lymphocyte function after bereavement. Lancet, 16, 834836.
Baum, A, Grunberg, N.E., & Singer, J.E. (1982). The use of psychological and
neuroendocrinological measurements in the study of stress. Health Psychology, 1,
217-236.
Baum, A, & Valins, S. (1977). Architecture and social behavior: Psychological studies in
social density. Hillsdale, NJ: Erlbaum.
Berkman, B. (1977). Community mental health services for the elderly. Community
Mental Health Review, 2(3), 3-9.
Boyce, W.T. (1981). Interaction between social variables in stress research. Journal of
Health and Social Behavior. 22, 194-195.
Bradburn, N., & Caplowitz, B. (1965). Reports on happiness. Chicago: Alden.
Brown, G.W., & Harris, T. (1978). Social origins of depression: A study of psychiatric
disorders in women. London: Tavistock.
Cannon, W.B. (1914). The emergency function of the adrenal medulla in pain and
the major emotions. American Journal of Physiology, 33,356-372.
Cannon, W.B. (1928). Neural organization for emotional expression. In M.L.
Reymert (Ed.), Feelings and emotions: The Wittenberg Symposium. Worcester, MA:
Clark University Press.
Cannon, W.B. (1936). Bodily changes in pain, fear. hunger. and rage (2nd ed.). New
York: Appleton-Century.
Cannon, W.B., & de la Paz, D. (1911). Emotional stimulation of adrenal secretion.
American Journal of Physiology, 28, 64-70.
Cassel, Ie. (1974a). An epidemiological perspective of psychosocial factors in
disease etiology. American Journal of Public Health, 64, 1040-1043.
Cassel, Ie. (1974b). Psychosocial processes and "stress": Theoretical formulation.
International Journal of Health Services, 4, 471-482.
Cassel, J.e. (1976). The contribution of the social environment to host resistance.
American Journal of Epidemiology, 104, 107-123.
Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine,

38,300-314.
Cohen, S., & Hoberman, H. (1983). Positive events and social support as buffers of
life stress. Journal of Applied Social Psychology, 13(2), 99-125.
Cohen, S., & McKay, G. (1984). Social support, stress, and the buffer hypothesis. In
A Baum, S. Taylor, & J.E. Singer (Eds.), Handbook of psychology and health (Vol.
IV). Hillsdale, NJ: Erlbaum.
Dean, A, & Lin, D. (1977). The stress-buffering role of social support: Problems and
prospects for systematic investigation. Journal of Nervous and Mental Disease, 165,
403-417.
DeAraujo, G., Van Arsdel, P.P., Holmes, T.H., & Dudley, D.L. (1973). Life change,
coping ability, and chronic intrinsic asthma. Journal of Psychosomatic Research, 17,
359-363.
Donald, CA, Ware, IE., Jr., Brook, RH., & Davies-Avery, A (1978). Conceptualization and measurement of health for adults in the health insurance study: Vol. IV;
Social health. Santa Monica: Rand.
Eaton, W.W. (1978). Life events, social supports, and psychiatric symptoms: A
reanalysis of the New Haven data. Journal of Health and Social Behavior. 19, 230234.
225
Eckenrode, J., & Gore, S. (1981). Stressful events and social supports: The
significance of context. In B.H. Gottlieb (Ed.), Social networks and social support.
Beverly Hills: Sage.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 5,
117-140.
Festinger, L. (1957). A theory of cognitive dissonance. Stanford: Stanford University

Press.
Festinger, L., Schachter, S., & Back, K (1950). Social pressures in informal groups. New
York: Harper and Row.
Fleming, R, Baum, A, Gisriel, M.M., & Gatchel, RI (1982). Mediating influences of
social support at Three Mile Island. Journal of Human Stress, 8(3), 14-22.
Frankenhaeuser, M. (1978). Experimental approaches to the study ofcatecholamines and
emotion. Reports from the Psychological Laboratories, University of Stockholm.
Fried, N. (1963). Grieving for a lost home. In LJ. Dohl (Ed.), The urban condition.
New York: Basic Books.
Frydman, M.1. (1981). Social support, life events and psychiatric symptoms: A study
of direct, conditional and interaction effects. Social Psychiatry, 16, 69-78.
Gans, H. (1962). The urban villagers. New York: Free Press.
Glass, D.e., & Singer, IE. (1972). Urban stress: Experiments on noise and social stressors.
New York: Academic.
Gore, S. (1978). The effect of social support in moderating the health consequences
of unemployment. Journal of Health and Social Behavior, 19, 157-165.
Grunberg, e. (1967). The social breakdown syndrome, some observations. American
Journal of Psychiatry, 123, 12.
Habif, V.L., & Lahey, B.B. (1980). Assessment of the life stress-depression relationship: The issue of social support as a moderator variable. Journal of Behavioral
Assessment, 2, 167-173.
Heller, K (1979). The effects of social support: Prevention and treatment implications. In AP. Goldstein & P.H. Kaufos (Eds.), Maximizing treatment gains. New
York: Academic.
Holmes, T., & Masuda, M. (1974). Life change and illness susceptibility. In B.S.
Dohrenwend & B.P. Dohrenwend (Eds.), Stressful life events. New York: Wiley.
Jonsson, E., & Sorenson, S. (1967). On the influence of attitudes toward the source on
annoyance reactions to noise: An experimental study. Nordisk Hygienisk Tidskrijt,
48,35-45.
Kaplan, B.H., Cassel, Ie., & Gore, S. (1977). Social support and health. Medical Care,
15,47-58.
Kasl, S.V., & Cobb, S. (1970). Blood pressure changes in men undergoing job loss.
Psychosomatic Medicine, 32, 19-38.
Kraus, A, & Lilienfe1d, A (1959). Some epidemiological aspects of the higher

mortality rate in the young widowed group. Journal of Chronic Diseases, 10, 207217.
Langer, EJ. (1975). The illusion of control. Journal of Personality and Social
Langer, T., & Michael, S. (1960). Life stress and mental health. New York: Free
Press.
LaRocco, 1M., House, IS., & French, IP.R, Jr. (1980). Social support, occupational
stress, and health. Journal of Health and Social Behavior, 21,202-218.
Lazarus, RS. (1966). Psychological stress and the coping process. New York: McGrawHill.
Leighton, A (1959). Rapid acculturation and social disorders. In I. Galdston (Ed.),
Medicine and anthropology. New York: International University Press.
Lin, N., Dean, A, & Ensel, W.M. (1981). Social support scales: A methodological
note. Schizophrenia Bulletin, 7, 73-89.
Lin, N., Simone, RS., & Ensel, W.M., & Kuo, W. (1979). Social support, stressful life
226
events, and illness: A model and empirical test. Journal of Health and Social
Behavior, 20, 108-119.
Mason, lW. (1975). Emotion as reflected in patterns of endocrine integration. In L.
Levi (Ed.), Emotions: Their parameters and measurement. New York: Raven Press.
Medalie, lH., & Goldbert, U. (1976). Angina pectoris among 10,000 men, II:
Psychosocial and other risk factors as evidenced by multivariate analysis of a five
year incidence study. American Journal of Medicine, 60,910-920.
Miller, P.M., & Ingham, lG. (1976). Friends, confidants and symptoms. Social
Psychiatry, 11,51-58.
Myers, JoK, Lindenthal, 1.1., & Pepper, M.P. (1975). Social integration and
psychiatric symptomatology. Journal of Health and Social Behavior, 16,421-429.
Newman, O. (1972). Defensible space. New York: Macmillan.
Nuckolls, KB., Cassel, le., & Kaplan, B.H. (1972). Psychosocial assets, life crisis,
and the prognosis of pregnancy. American Journal of Epidemiology, 95, 431-441.
Paulus, P., McCain, G., & Cox, V. (1978). Death rates, psychiatric commitments,
blood pressure, and perceived crowding as a function of institutional crowding.
Environmental Psychology and Nonverbal Behavior, 3, 107-116.
Rodin, J., Rennert, K, & Solomon, S.K (1980). Intrinsic motivation for control: Fact
or fiction. In A Baum & lE. Singer (Eds.), Advances in environmental psychology:
Applications of personal control (Vol. 2). Hillsdale, NJ: Erlbaum.
Schachter, S. (1959). The psychology of affiliation. Stanford: Stanford University
Press.
Schaefer, e., Coyne, J.e., & Lazarus, RS. (1981). The health-related functions of
social support. Journal of Behavioral Medicine, 4, 381-406.
Segal, B., Weiss, E., & Sokol, R (1965). Emotional adjustments, social organization,
and psychiatric treatment. American Sociological Review, 30, 584.
Seligman, M.E.P. (1975). Helplessness: On depression, development, and death. San
Francisco: W.H. Freeman.
Selye, H. (1956). The stress of life. New York: McGraw-Hill.
Sundstrom, E., Lounsbury, J.W., DeVault, Re., & Peele, E. (1981). Acceptance of a
nuclear power plant: Applications of the expectancy-value model. In A Baum &
lE. Singer (Eds.), Advances in environmental psychology, Vol 3. Energy: Psychological
perspectives. Hillsdale, NJ: Erlbaum.
Thoits, PA (1981). Undesirable life events and psychophysiological distress: A
problem of operational confounding. American Sociological Review, 46, 97-109.
Wallston, B.S., Alagna, S.W., DeVellis, B.M., & DeVellis, RF. (1983). Social support
and physical health. Health Psychology, 2,367-391.
Yancey, W.L. (1972). Architecture, interaction, and social control: The case of a large
scale housing project. In J.F. Wohlwill & D.H. Carson (Eds.), Environment and the
social sciences: Perspectives and applications. Washington, DC: American Psychological Association.
Chapter 12
People Without Friends:

Loneliness and Its Alternatives
Cecilia H Solano
To have friends is considered to be a normal and desirable aspect of a

modern American social life. The mass media is filled with images of all
types of people working and relaxing with one or more friends. Empirical
studies support this image of friends as being an important part of the
normal social life. In one such study, Lowenthal, Thurnher, and Chiriboga
(1975) did an extensive survey of the friendship patterns of adults in the
United States. They found that, on the average, people report having
approximately six relationships that can be called friendships. However,
this number varies in predictable ways with life stage. For example,
newlyweds have the highest number of reported friends (eight). This is
higher than the average five reported by high-schoolers, the average five
reported by middle-age married persons, or the average six reported by
persons about to retire. Even with these fluctuations it is clear that
Americans typically have a substantial number of friends over the life
cycle.
Centering attention on the average number of friends, however, obscures
the fact that there is also considerable variability. Lowenthal et al. (1975)
found an average of six friends, but they also found a range of 0 to 24. The
fact that some people reported having no friends at all is particularly
intriguing, given the cultural assumption that having friends is normal. The
Lowenthal et al. data are supported by several recent studies in which
investigators have shown that the condition of being completely without
friends is actually not uncommon. In 1981, Bell did a survey of friendship in
adults and found that 10% of the males reported having no close friends. In
a review of friendship studies on children, Asher (1978) noted that in any
given sociometric study, 10% of the children were not selected as a friend by
any other child. In a study of adolescents (cited by Bell, 1981),20% reported
having no friends. In 1979, Lopata reported that 16% of elderly widows
indicated having no friends either before or after their bereavement. A
228
Cecilia H. Solano
reasonable conclusion from these data is that many people conduct their
daily lives without friends.
Are people without friends lonely? There are several reasons for assuming
that they are. First, it seems plausible that a person lacking friends in a
culture that values friendship would suffer both social and psychological
discomfort. Indeed, writers such as Lynch (1977) and Slater (1970) have
made such an assumption and have used it as a basis for their discussions of
the consequences of loneliness. Furthermore, there is a variety of evidence
to suggest that loneliness, like friendlessness, is a commOn phenomenon.
For example, while doing cross-cultural research using projective tests,
Kiefer (1980) found that fear of social isolation was an unusually frequent
theme in North American stories. Bradburn (1969) noted that 26% of his
U.S. sample surveyed by phone reported feeling very lonely in the past few
weeks. In a more recent study on college students, Cutrona (1982) found that
75% ofthe students reported occasional mild feelings ofloneliness, and 40%
reported severe feelings of loneliness. All of these data indicate that
loneliness, like friendlessness, is a familiar aspect of life for many people.
Only a few studies, however, have attempted to make a direct connection
between lack of friends and loneliness. In one such study, Russell, Peplau,
and Cutrona (1980) found that loneliness in college students was in fact
associated with listing fewer close friends. Perlman, Gerson, and Spinner
(1978) also found this to be true for elderly SUbjects. On the other hand, in
their studies on college students, Williams and Solano (1983) and McCormick and Kahn (1980) did not find this association; results of these latter
studies do not support the contention of a strong relationship between
lacking relationships and feeling lonely. Further evidence of a weak
association is found in sociological studies on social networks and
dissatisfaction, which have also produced only low correlations (see Peplau,
Bikson, Rook, & Goodchilds, 1982, for a review).
Thus, a lack of friends cannot be assumed to be directly equivalent to
feeling lonely. The purpose of this chapter is to consider the nature of this
nonequivalency in greater depth. A useful perspective for exploring this
issue has been provided by Suedfeld (1982). In a study on complete social
isolation, he commented that loneliness and isolation are overlapping, not
redundant concepts (p. 55); that is, there is a set of people who lack adequate
friendships and do feel lonely. Yet there are also people who do not have
friends and who do not feel lonely. To study this issue in this chapter, I first
consider some of the problems in defining and operationalizing both
loneliness and a lack of friends. Mter this introduction, the current research
on loneliness is reviewed to see what part of this literature refers to
friendlessness. The third part of the chapter is focused on consideration of
people who are without friends and yet are not lonely; the functions that
friends fulfill and alternative ways of satisfying these needs that mitigate
loneliness are discussed.
People Without Friends
229
Definitional Issues
When studying whether people without friends are lonely, it is useful to
begin with a definition. Peplau and Perlman (1982) stated in a widely used
definition that "loneliness is a response to a discrepancy between desired
and achieved levels of social contact" (p. 8). In considering this definition,
however, one must be aware of the definitional ambiguity of the two halves
of the equation and how that ambiguity affects the comparison. The
problems center on the way in which "achieved social contact" is defined
and the role of expectations and desires in feeling lonely.
To study people without friends would seem to require first a measure of
achieved number of friends. The simplest method is to ask people to list
their friends by name or ask them simply to indicate the number. Although
these measures seem clear-cut, they leave much room for interpretation. For
example, should the group of interest include only people without any
friends at all, or should it also include people with only a few friends? The
problem then becomes how few is few? Fischer (1982), who has studied
people with small social networks, has commented on the difficulty of
deciding what exactly should be considered a "small" social network.
In addition, there is an issue of how the label "friend" is being applied by
different people. Is it that two people differ in actual number of close
relationships or that they use the label "friend" differently? Because many
researchers allow subjects to self-define the role friend, the answer to this
question is not clear. A third problem in objectively specifying the number
of friendships is the emphasis on the perspective of just one person.
Friendship may exist from the perspective of both or only one of the
partners. Individuals might state that they have no friends but still might be
included by other people in their list of frientls. The reverse might also
occur, with people choosing others as friends but being neglected in the
friendship choices of others. This situation was actually found in a study by
Williams and Solano (1983), in which a substantial percentage of friendship
choices by college students were not reciprocated. In these situations it
becomes unclear which people should be identified as those without
friends.
Alternative approaches to measuring the existence of friendship have
emphasized the amount of social interaction rather than the number of
relationships (Jones, 1982). Amount of social interaction has been defined
both as number of social contacts with friends and as amount of time spent
with friends both based on self-report. As Jones (1982) noted, both of these
measures have produced significant and nonsignificant correlations with
feelings of loneliness. The subjective element is particularly noticeable in
yet another measure of friendship patterns. In the third method, subjects are
asked to rate the level of interpersonal intimacy for a variety of relationships. Studies using this method have also not produced a clear relationship
230
Cecilia H. Solano
with feelings of loneliness (Jones, 1982). In general, then, few studies

have provided measures of achieved friendship that can be specified
independently.
The second set of measurement issues arises from the definition of
loneliness. The Peplau and Perlman (1982) definition ofloneliness indicates
that the number of achieved relationships must be contrasted with the
number desired. It is the ratio of achieved to desired friendships that should
correlate with loneliness, not the number of friendships alone. Yet little of
the current research on loneliness directly considers the issue of desires and
expectations. This attitude is appropriate only if it is assumed that there are
few individual differences in number of desired relationships.
In fact, there are many indications in the psychological literature that
there are important variations across people in the need for others. In his
theory of privacy, Altman (1975) stated that there are individual differences
in the optimal amount of social stimulation received from other people.
Also, a large number of scales in personality psychology are designed
specifically to measure differences in desire for social contact. A partial
listing would include Murray's Need for Affiliation Scale on the Adjective
Checklist (Gough & Heilbrun, 1964), the Sociability Scale on the California
Personality Inventory (Gough, 1964), and the Privacy Preference Scale
(Marshall, 1972). Thus, outside of the field of close relationships, there is
strong support for the proposition that differential desire for social contact
should be considered.
In addition to differences in the desire for social interaction, differences
in expectations have been found. For example, Jong-Gierveld (1971) noted
that unmarried people tend to feel lonelier when they are around married
people, presumably because their expectation level for relationships is
raised. In a study on social isolation in old age, Lowenthal (1964) found that
people who had been isolated all their lives felt much less lonely than those
who had become isolated only in old age. There are also indications that
unrealistic expectations of friendship may also be involved with feeling
lonely. Jones (1982) reported that college students who feel lonely often have
unrealistic and rigid expectations of relationships. LaGaipa and Wood
(1981) found that socially maladjusted adolescents had distorted or
immature views of normal friend relationships. Hypothetically, it also seems
probable that people who fail to adjust their expectations to changing life
situations might also encounter problems. For example, the level of
expectations that adolescents develop in high school could cause serious
problems initially in the adjustment to college.
In the previous discussion, I have attempted to point out that one reason
for a lack of association between achieved relationships and perceived
loneliness could be based on measurement difficulties. Even so, in recent
years a considerable amount of research has been conducted on the topic of
loneliness. Conceivably, some of this research should relate to people who
231
lack friendships, whereas other aspects of this research might relate to

loneliness arising from other sources. It would be useful to know how much
of the current loneliness research is specifically oriented to people who lack
friends.
Loneliness
Programmatic research on the topic ofloneliness did not begin until 1977.
Much of this new research has been summarized in an influential book by
Peplau and Perlman (1982). One factor in the recent interest in this topic has
been the development of measures of loneliness. Although several scales
have been devised (e.g., Loucks, 1980; Schmidt & Sermat, 1983), only the
UCLA Loneliness Scale has been widely used (Russell et at, 1980). As a
result, the UCLA Scale has had a considerable effect on the direction of the
development of the field. Therefore, a consideration of the nature of this
scale would help determine the relevance of the current research on
loneliness to people without friends.
The UCLA Scale can be evaluated on the basis of the item content,
correlation with other scales, and answer format. An inspection of the item
content indicates that the scale is measuring feelings of having received
companionship, understanding, and acceptance from others. According to
Davis and Todd (1982), who investigated the content of different types of
close relationships, the UCLA Scale describes the core components of
friendship. They showed that romantic relationships, for example, included
additional elements, such as passion and exclusivity, that are not measured
by the UCLA Scale. In addition, a pilot study Solano and Koester (1984)
indicated that the UCLA Scale does not correlate equally with all of the
subscales on the Differential Loneliness Scale (Schmidt & Sermat, 1983).
The UCLA Scale correlated significantly only with the Friend and
Community Group subscales of the Differential Loneliness Scale, not the
Family or Romantic Partner subscales. Thus, much of the current research
which uses the UCLA Scale is oriented to identifying people who perceive a
lack of closeness with friends.
The answer format of the UCLA Scale has also had an effect on the
direction of the research on loneliness. High scores on the UCLA Scale do
not directly indicate the degree of loneliness, but rather the frequency with
which a lack of companionship is felt. This emphasis on chronic loneliness
has oriented researchers toward looking for stable internal causes. Thus, the
majority of the investigators using the UCLA Scale has assumed that people
who are frequently lonely are the cause of their own social problems, as a
result of either their poor social skills or undesirable personality traits. This
chain of logic is important because it has led to a concentration on social
skills deficits, which as a topic is most relevant to achieved relationships,
such as friends, and less relevant to ascribed relationships, such as family.
232
Cecilia H. Solano
There is reason, therefore, to assume that most of the research literature on

social skills and loneliness is directly relevant to people who lack adequate
friendships. A brief review of this literature should be instructive.
The earliest research efforts on social skills deficits were self-report
studies on the causes of loneliness. These studies showed that lonely people
personally considered social skill problems (such as shyness, a lack of selfdisclosure, and a low level of self-assertiveness) as causes of their social
problems (Horowitz & French, 1979; Maroldo, 1981; Solano, Batten, &
Parish, 1982). Given the emphasis of the UCLA Scale on companionship, it
is not surprising to find that loneliness as measured on this scale was related
to a lack of self-reported disclosure to same-age peers but not to parents
(Solano et aI., 1982).
A number of studies have also been done on the actual verbal behaviors
of lonely people (Hansson & Jones, 1981; Solano et aI., 1982). The majority
of this research has been conducted with college students, using either
same-sex or opposite-sex strangers. The same-sex studies are particularly
interesting because they would be most relevant to the issues of friendship
problems and loneliness. These researchers have found that with actual
same-sex strangers, lonely persons tend to speak less, use inappropriately
high levels of self-disclosure, and offer advice more than do nonlonely
students. Only one study (Sloan & Solano, 1984) has been conducted on
interaction with well-known same-sex others (roommates). This study
showed that lonely persons tend to speak less and use "familiar" modes of
conversation less often. For example, subjects were more apt to acknowledge
just that the roommate had spoken rather than to agree or offer advice.
In addition to these studies on verbal social skills, there are also a few
investigations of problems with nonverbal behaviors. For example, Gerson
and Perlman (1979) showed that elderly persons who are lonely are less
accurate in their perceptions of the nonverbal behavior of others than are
the nonlonely elderly. With adolescent subjects, Putallaz and Gottman
(1981) found that social maladjustment was associated with inappropriate
nonverbal behaviors with peers, such as turning their backs on partners and
using body-focused behaviors to lower their anxiety.
In addition to studies of problems with verbal and nonverbal social skills,
there has been research on the cognitions of londy people. Most of these
investigators have emphasized the cognitions that result from failure to
achieve a social relationship, again a topic more relevant to friends than
family relations. In a typical study, Peplau, Russell, and Heim (1979) found
that external attributions for a perceived lack of close relationships (such as
coldness of people) were associated with hostility and coping, but internal
attributions (such as own worthlessness) were associated with depression
and withdrawal. In a study of children's relationships, Dweck and Goetz
(1978) found that internal attributions for being rebuffed were particularly
associated with giving up the effort to make friends. In addition, lonely
people tend to be more initially hostile and distrustful of others and more
233
sensitive to rejection (e.g., Jones, Freemon, & Goswick, 1981; Jones, Sansone,
& Helm, 1983; LaGaipa & Wood, 1981).
All of this research demonstrates that chronically lonely people suffer

from a perceived lack of close relationships, probably caused by their selfreported and empirically observed inability to interact in a socially
appropriate manner. Clearly, this inability could lead to difficulties in
obtaining and maintaining relationships and, hence, friendlessness. In
addition to the large number of social skills and cognition studies, there are
also a few investigations concerned with loneliness resulting from temporary, situationally based causes. Even these studies, however, are relevant,
specifically with regard to a lack of friendship relationships. Virtually all of
these investigations have been focused on first-semester freshmen attempting to adjust to college life (Cheek & Busch, 1981; Cutrona, 1982;
Ponzetti & Cate, 1981). Again, these studies emphasize the lack of peer
relations and support the primary conclusion that much of the current
research on loneliness is specifically oriented to studying people without
friends.
Function of Friends
The current researchers on loneliness appear to be investigating the overlap
area that Suedfeld (1982) described as existing between feeling lonely and
lacking friend relationships. Yet an emphasis on this overlap area may
obscure the nonoverlapping area in which a person may not have any
friends and yet not feel lonely at all. A broader perspective needs to be taken
to explain this apparent anomaly. Friendship relations are obviously only
one of many types of relationships. To equate loneliness simply with a lack
of friends ignores the important role that other relationships (such as
family, romantic partners, and acquaintances) play. It also overlooks the
possibility that satisfaction can be derived from nonsocial as well as social
activities.
One straightforward way to study the general issue of loneliness and
friendship is to consider the rewards that people receive from friendships
and how these rewards might be replaced. The wide range of needs that a
friend may fulfill can be broken down into three general categories: the
material needs, the cognitive needs, and the social-emotional needs. The
first set deals with the material advantages of aid and support that can be
derived from friendship. The second set of needs is related to stimulation,
frames of reference, and a sense of meaning. Finally, there are the needs
associated with love and self-validation. Each set of needs is considered
separately.
Material Needs
The willingness to give help has been identified as one of the major
characteristics of a friend (LaGaipa, 1977). Indeed, sociobiologists, such as
234
Cecilia H. Solano
Trivers (1971), have maintained that the reciprocal exchange of help is one
of the primary bases of all social relationships. Thus, a friend is one who
can be expected to provide aid and support in times of need. Specifically, a
friend may be called upon to furnish goods, services, money, or information
(Foa & Foa, 1974). For example, friends can provide information in the form
of news about a job opening. Friends might lend money to buy a soft drink
or handle a financial emergency. Services provided can be as diverse as
picking up mail to helping a friend move across town. Goods can be a loan
of a lawn mower or car. Thus, the relationship of friend involves substantial
expectations of receiving and giving aid in both large and small matters.
In addition to providing needed resources, friends are an important asset
in terms of safety. A common piece of advice to women traveling at night is
to take a friend. The use of friends for protection is analogous to a point
made by Lorenz (1966). He noted that animals often prefer to stay with
herds because predators are more likely to pick off single animals. The
protection can be more indirect, however, as when a friend arranges a blind
date. In this situation, the friend is assuring both parties that this new other
person is both safe and similar. The usefulness of friends as a form of
protection has been shown explicitly in a study by Schwartz (1972). He
found that children were much more willing to explore a strange environment with a friend than when alone.
The preceding points demonstrate that a very important function of
friendship is providing aid. As a result, people with few friends will feel the
lack of easy access to these services. In fact, Rubenstein and Shaver (1980)
found that one of the correlates of feeling lonely was feeling that there was
no one to help in an emergency.
Cognitive Needs
The second set of functions includes both the need for stimulation and the
need to create a meaningful frame of reference for social behavior. With
regard to the need for stimulation, both classic studies (Spitz & Wolf, 1946)
and more recent ones (Suomi & Harlow, 1978) have shown the importance
of stimulation for normal emotional and social development in infants.
There also appears to be an active stimulation-seeking drive in adults.
The stimulation sought can either be from another person or from some
impersonal source. Friends are definitely a source of stimulation. In the
form of conversation, they provide gossip, ideas, and jokes. They also allow
us to engage in activities (such as sports) that would be difficult to do alone.
They can provide experiences such as parties, dinners, and vacations.
Obviously, having friends can make life a great deal more interesting and
stimulating. Again, the research on loneliness provides an insight into a
problem of not having friends. Several researchers have found that people
who lack social companionship report strong feelings of boredom (e.g.,
Moore & Schultz, 1983; Rubenstein & Shaver, 1980).
235
Another important function that friends serve is to give a frame of

reference or meaning to events. Sociological theories of the self-concept
have described how interacting with parents and peers allows children to
find labels for their actions (e.g., Meade, 1934). Other sociologists, such as
Berger and Luckmann (1966), have theorized about the importance of
interpersonal interaction in the social construction of reality. Social
psychologists have described how this process of using others to label our
own acts and events in the world continues into adulthood (e.g., Festinger,
1954; Schachter & Singer, 1962).
A chain of logic demonstrates why friends are of major importance in the
process of constructing social meaning. Festinger (1954) stated that people
who are similar provide the most useful comparison when judging our
abilities. Similarity, as many studies have shown (Byrne, 1971), is a basic
component of friendship. Therefore, friends who are highly similar in
interests and activities will be particularly useful in interpreting new
information about oneself and the world.
Emotional Needs
The third set of needs is related to the giving and receiving of love. An
important aspect of this need for love is the desire to be able to disclose to a
trustworthy other and to feel understood. As many researchers and
theoreticians have stated, the need to be loved is one of the most basic
human needs.
Friends clearly are one of the important ways of fulfilling the need for
love. This point is made in the study by LaGaipa (1977), cited under the
section on Material Needs. He asked subjects to characterize the components
of friendship. In addition to helpfulness, he found several components that
emphasize the importance of love in friendship: being sincere, being
accepting of the friend, being willing to self-disclose, providing understanding, and giving positive regard. By asking subjects to differentiate
between the characteristics of a "real" friend versus other types of
relationships, Davis and Todd (1982) obtained a set of characteristics very
similar to LaGaipa's list.
Friendship, in addition to the need to fee110ved, satisfies a need for social
status. The status component is usually considered relevant to friendship in
the case of a prominent person elevating a friend's social status by
association (Foa & Foa, 1974). Harre (1977), however, made the point that
initiating and maintaining a friendship is a task accomplishment. Having a
friend of any status level, therefore, means that a person has successfully
completed a social task. This success provides status and validates the
person as a worthwhile and socially competent individual (Gordon, 1976).
The point is that having a friend per se is a form of social status.
The importance of friends to status and self-esteem is also supported by
research from the literature on initial attraction. Jacobs, Berscheid, and
236
Cecilia H. Solano
Walster (1971), for example, demonstrated that when a person's status on

other dimensions (such as ability) has been lowered, he or she may seek to
compensate with the status of being loved. The perspective of using social
relationships to bolster self-esteem fits into the more recent research on
social support networks. This research has shown that social networks are
important buffers against stress. Gottlieb (1981) has suggested that social
networks may help a person handle stress either by providing aid directly to
the problem or (as is relevant to the current point) by helping a person's
general self-esteem.
Alternatives to Friends
Other Relationships
Although friends can fulfill all these needs, obviously other types of people
can also fulfill them. The roles of family, romantic partners, and acquaintances immediately come to mind. These roles to some degree can be used
interchangeably with friendship because the nature of friendship is not that
different from these types of relationships. LaGaipa (1977) found substantial overlap in the characteristics of acquaintances, casual friends, and best
friends. The only difference was in the amount and pattern of the
characteristics. As noted earlier, Davis and Todd (1982) found that there was
considerable overlap in the characteristics of romantic partners and friends,
with only a few dimensions differentiating the two roles. Family relationships are probably the most different from friendships, being ascribed
rather than achieved and involving people of different ages.
The conclusion from these studies is that friendship serves many of the
same purposes that other close relationships do. Therefore, a person without
friends might well be able to fill these needs by using relationships other
than friends. For example, a person might have a number of close
acquaintanceships in which he or she receives some understanding, but
none of these relationships are quite intimate enough to be labeled
"friendship." Here, the solution to a need forintimacy is solved by having
many friendly relations rather than single friendships (Kurth, 1970).
Advantages of this approach are that some of the rewards of friendship
are received while many of the costs are avoided. Also, these relationships
may be easier to replace after moving to a new place than are deeper
friendships. Disadvantages would be that the lower level of intimacy in
these casual relationships may not be enough to fill the need for
understanding. Indeed, Weiss (1973) described social loneliness as resulting
from having a large shallow social network without any specific deep
relationships. Although some authors, such as Brain (1976), have considered an extensive, but less-intensive network to be a sign of the failure of
modern society, it is equally possible that this form of "friendship" is an
237
adaptive reaction to a particular type of social environment, for example, a

highly mobile one.
A more likely alternative to friends is interaction with a romantic partner.
Brain (1976) stated that close opposite-sex relationships are to be found in
many different types of cultures. These romantic relationships can exist
alone or be integrated into a social network. A major problem in having
only a single romantic partner is that reliance on a single relationship to
fulfill all emotional needs must place an enormous strain on that one
relationship. Also, if that one relationship should fail, there are few
alternatives to replace it. Weiss (1973) described a second type ofloneliness
having only a single close other and no general social network.
An appropriate alternative, then, would seem to be to have a romantic
relationship that is integrated into a friendship network. Although this is
obviously a common pattern, there are indications that maintaining friends
and a romantic partner is not always easy. Gordon (1976) endorsed this
point by describing how there is a hierarchy of importance of close
relationships in American culture that places friendship below romance. As
an example she noted that American women place greater value on a
possible date with a male than on firm plans with a female friend. Several
empirical studies (Johnson & Leslie, 1982; Milardo, Johnson, & Huston,
1983) have reported that when people begin to develop a close romantic
relationship, they typically withdraw, at least temporarily, from their
friendship networks. For instance, Weiss (1975) noted that people who relied
on friends after a divorce often neglected these friends once they found a
new romance.
People who are married and have children may also neglect their
friendship networks. As was shown by Lowenthal et al. (1975), married
people in their middle years tend to have fewer friends than do younger,
unmarried people. One reason might be the difficulty in finding compatible
friends for both spouses. Researchers, such as Bott (1957), have found that
in a middle-class marriage, friendships are usually based on the couple as a
unit. Individual friendships for a single spouse are less common. A
reasonable inference is that it is harder for a couple to find another couple
that is jointly acceptable than it is for a single person to find individual new
friends. In addition to the spouse, children can also occupy the time that
might be spent with friends. Thus, close relationships within a marriage can
limit the time and desire for additional social interaction with friends.
Spouses and romantic partners are still not the only social means of
fulfilling the need for love and belonging. Another group, which is often
neglected in the psychological literature on adults, is family. An impression
that one might get from the current literature is that if an American adult is
divorced or friendless, he or she is alone. This view is reinforced by writers,
such as Brain (1976), who suggest that kin relationships are no longer
important in modern urban societies. This may be only partially true.
Fischer (1982) and others have shown that kin are important in modern
238
Cecilia H. Solano
America, but mainly for specific groups. Fischer stated that kin are not
important as adult relationships for men, higher income people, those who
have moved away from the family of origin, and those who live in cities. It is
interesting that the segment of the population that Fischer described as not
valuing kin relationships probably includes most social psychologists. Allan
(1979) criticized the tendency of psychologists to assume that their own
social patterns are the norm for the population in general.
Given Fischer's (1982) findings, it appears that kin relationships are
relatively more important to females, lower income groups, those who live
l~lear their family of origin, and nonurban dwellers. This includes a fairly
large segment of the American population. Various classic studies from
sociology support this point. Komarovsky (1967) showed that for workingclass women, over 63% of the people listed as "friend" were actually relatives.
Gans (1962), in his study of the working-class, Italian West End of Boston,
noted that the most important social relationships for young people were inlaws, siblings, and cousins. Researchers in Britain (e.g., Wilmott & Young,
1960) found that "Mum" was considered to be the most important
relationship for young women.
Fischer (1982) interpreted his data to mean that in traditional societies,
people have close relationships with family, fellow church members, and
neighbors. Only the special urban group that he studied relied heavily on
friends, who were usually either coworkers, fellow members of secular
organizations, or "free friends" (people who once held one or the other
position). He suggested that the family is normally the last resort for both
material and emotional aid and he implied that friends are a luxury, only
found in times of affluence and security. A further implication is that a
reliance on friends may lead to ignoring the kin network, which then cannot
be reactivated in times of crisis. In this context, Kiefer's (1980) work on
affiliation in the United States and Japan is interesting. The assumption in
the United States is that loneliness is related to a lack of friends. In Japan,
however, Kiefer found that loneliness was often related to a lack of contact
with family rather than with friends.
Nonsocial Alternatives
Even a person without any form of close relationships will have resources
for covering the three needs previously outlined. For example, regarding
material needs, both government agencies and private businesses can
effectively cushion the lack of this facet of friendship. Loans can be
obtained from banks instead of family, dating services can replace the blind
date referral services of a friend, employment agencies locate jobs, moving
companies pack households, and night guards provide protection.
One implication of the use of agencies and businesses for help is that
wealthier and more knowledgeable people will feel the lack of friends less
than will poor people. For example, Myers and Roberts (1959) found
239
relationships among poverty, alienation, and unsatisfactory social relationships. One reasonable inference is that poor persons have neither the money
to replace the aid of friends nor the knowledge of how to use impersonal
resources to their benefit. Another inference is that richer and more
knowledgeable people perhaps are called upon less often to demonstrate the
helping aspect of friendship, which may have consequences of its own, such
as a lack of interdependency.
Beyond material needs, people require stimulation and a frame of
reference. Whereas friends can be important sources of satisfaction of these
needs, there are obviously many types of impersonal activities that can also
serve these purposes, either in addition to friends or as substitutes.
Rubenstein, Shaver, and Pep1au (1979), who found boredom to be a problem
of socially isolated persons, also found that calling up a friend was only one
of a large variety of ways to handle this boredom. The subjects in their study
provided a list of alternative nonsocial activities, such as working, watching
television, eating, taking drugs, exercising, reading, listening to music,
shopping, going to movies, or working on a hobby. In confirmation of these
results, Moore and Schultz (1983) found that adolescents reported that when
they felt lonely they turned to listening to music and watching television for
stimulation. Watching television was also a preferred tactic for the elderly
(Perlman et ai., 1978).
Television, in particular, is important in that it provides not just
stimulation, but also a frame of reference (e.g., Frank & Greenberg, 1979).
Viewers can use fictional characters as a substitute for friends for social
comparison. Frank and Greenberg did express concern that the information regarding the nature of society received from television may be
somewhat skewed. Whereas this does present problems regarding the
accuracy of the framework, television viewing still provides a usable set of
information.
Other important sources of stimulation and involvement that might
replace friendship are careers or religion. A career can provide both
material security, stimulation, and a frame of reference for many life
activities. Religious affiliation can also satisfy these needs. Pa10utzian and
Ellison (1982) found that people who felt that they had a more personal
relationship with Jesus Christ were less likely to fee110ne1y even if socially
isolated than were people who had a more abstract image of God.
This brief list of nonsocial ways of replacing the stimulation and
meaningfulness of friends, however, raises a question. Why might a person
choose an impersonal source versus a social source of satisfaction? Thibaut
and Kelley's (1959) theory of social exchange provides a framework that can
be used to answer this question. They postulated that the act of associating
with anyone person or situation involves both rewards and costs. The
balance of these rewards and costs is compared to those expected for this
person or activity. In addition, this ratio is also compared with the ratio for
alternative activities or persons. Thus, a person who decides to go out
240
Cecilia H. Solano
jogging rather than to meet a friend has made a decision based on the
relative costs and rewards of the two activities.
Using the concept of comparison level for alternatives, some people may
find nonpersonal experiences more rewarding than interaction with friends
or any other form of social interaction. In this case, it may well be that a lack
of friends would not be felt as a problem. A person who finds television
more interesting than the conversation of their friends will not miss visits
from these friends. A person who has devoted his or her time and energy to
being a good athlete, musician, scientist, or executive may not feel any need
for additional stimulation or sense of meaningfulness. For example, Roe
(1953), in her study of eminent scientists, found that these scientists
preferred working on research to virtually any other form of activity, social
or otherwise.
The empirical data contrast strongly with the view that impersonal
sources of satisfaction can never be enough in themselves. For instance,
Sermat (1978) argued that people who appear to be successfully self-reliant
still require some form of meaningful relationship with another. There are
limits to self-sufficiency and inner direction. Although Sermat may be
correct that no person can live utterly alone, it does seem possible that a
person could exist successfully for extended periods of time without
interpersonal interaction if other resources were available.
One reason that people may prefer nonsocial sources of stimulation and
meaningfulness to social ones has to do with the relative costs. One cost of
interacting with other people may be the loss of control. Davis (1973) argued
that becoming close to someone involves an ever-increasing loss of
independent action and control. Even at the most superficial stages of
friendship, a person must accommodate to others in terms of timing and
choice of activities. A person who values control above sociability might
well prefer another form of activity to social interaction or may desire to
engage in activities (such as shopping and working) alone. Some people
who choose solitary activities to maintain control may regret the loss of
companionship, but it is conceivable that a certain percentage of the people
who report having no friends, prefer this status.
Another cost of interacting with others is that the frame of reference that
they provide may be inhibiting. Moustakas's (1961) common theme in his
extensive writings about the solitary artist is that people who wish to be
creative may find it necessary to withdraw from the company of others. This
point is given support by Rubenstein et al. (1979), who found that 24% of her
subjects reported that they used time alone to be creative. As FrommReichmann (1959) pointed out, the withdrawal to be creative is usually selfchosen and temporary. A short-term period of avoiding friends and
friendship, then, may bear no relationship to loneliness.
The cost of time and effort is another aspect of interacting with people. A
common complaint of people surveyed in rural and urban California was
that friends involve too many time demands (Fischer, 1982). This can be a
problem particularly for people who already have constraints on their
241
time. For example, Fischer noted that mothers with young children
complained about this problem. In the mass media, popular columnists
often receive letters from people complaining about friends dropping by
unannounced and ask advice on how to limit these visits. Thus, friends can
occupy time that a person would rather spend otherwise.
In an earlier section I mentioned the material advantages of having
friends. However, friends are also costly in regards to goods and services.
Receiving aid from an impersonal institution simply means having to pay
back the debt. Receiving help from a friend means having to pay back the
debt and being available to provide similar help at some future time
(Walster, Walster, & Berscheid, 1978). Thus, aid from friends carries
additional costs.
Finally, research on social comparison indicates that using friends as a
frame of reference has its costs as well as its rewards. Tesser and Campbell
(1980) reported that self-esteem suffers more from comparing oneself to a
friend who excels on a personally important dimension than from
comparing oneself to a stranger.
No attempt is being made here to suggest that the rewards of human
social interaction are not substantial. However, these rewards are associated
with costs, some of which have been described previously in this chapter.
Thus, friendship should not be considered as an unmixed blessing, as it
sometimes appears in the psychological literature, but rather as a complex
and sometimes problematic relationship (cf. Rook, 1984).
Therapeutic Interventions
In the past sections I have presented the position that having alternative
social and nonsocial activities might help reduce the association between
lacking friends and feeling lonely. This same position is presented, though
only implicitly, in some of the clinical literature on loneliness. A brief review
of the therapy literature, therefore, should add support to this point.
One area of the research on loneliness as an emotional problem is
specifically concerned with coping mechanisms available to lay people.
Rubenstein el al. (1979), cited earlier, were particularly concerned with the
approaches taken by people not in therapy to deal with feeling lonely. Some
of the methods that they listed were quite active and would be available even
to socially unskilled individuals: listening to music, exercising, shopping, or
working on a hobby. The popularity of these nonsocial means of coping
with loneliness is implicit support for their ability to replace some of the
functions of friendship. However, the success of these methods may be
hindered by the extrinsic motivation to engage in these activities rather than
by intrinsic motivation; that is, an activity engaged in simply to replace a
friend may be less satisfying than that same activity chosen freely (Weiss,
1973).
In addition to these active methods, Rubenstein et al. (1979) identified a
wide variety of passive activities used to cope with loneliness. These
242
Cecilia H. Solano
methods are probably less successful because they fulfill fewer of the
functions that friends provide. These "coping" behaviors include crying,
sleeping, watching television, taking drugs, drinking, overeating, sitting,
thinking, and doing nothing. Taking drugs has been specifically discussed
by Hartog (1980). He suggested that the specific type of drug a person takes
reflects a particular coping style. For example, tranquilizers allow people to
become disinhibited and therefore socialize more easily, whereas psychedelics remove the need to cope with the social world entirely.
For people who are unable to cope on their own with their social
difficulties, therapy may be necessary. As therapy for both children and
adults, some clinicians have emphasized behavioral training and feedback
(Goswick & Jones, 1981; Oden & Asher, 1977). Although these skillsoriented programs have been successful, Young (1982) recommended a
broader based approach. He observed that most people who request therapy
for loneliness state that their problem is a lack of romantic relationships.
Young, however, stated that lonely people should not be oriented directly
toward finding a successful romantic relationship. Rather, they must first
concentrate on the development of a full social network. Young has
constructed a hierarchy of social abilities whereby clients are first taught to
feel comfortable alone, then meet acquaintances, then develop friendships,
and, only as a last stage, establish a romantic relationship. Young's program
for working with lonely individuals is particularly interesting because he
sees friendship, acquaintanceship, and romantic partners as forming an
integrated social world that must be handled as a whole.
Summary
In this chapter I have examined why people who lack friends might or
might not be lonely. First, I indicated that there are definitional issues about
loneliness and aloneness that need to be resolved in order to clarify the
question. Second, I considered the current research on loneliness and found
it useful in helping to understand the ways in which people achieve or fail to
achieve friendships. Finally, I suggested that future research on loneliness
and friendship needs to be balanced by the perspective that friends are only
one source of interpersonal satisfaction. A broader approach includes an
analysis of the role that other relationships and nonsocial sources of
satisfaction play. I also considered the implications for this broader
approach to therapy techniques. In the developing field of close relationships, the interrelatedness of social life needs to be emphasized by
researchers who should not concentrate on anyone social relationship in
isolation, even such an important one as friendship.
References
Allan, G. (1979). A sociology of friendship and kinship. London: George Allen &
Unwin.
Altman, I. (1975). The environment and social behavior. Monterey, CA: Brooks Cole.
243
Asher, S. (1978). Children's peer relationships. In M.E. Lamb (Ed.), Social and
personality development (pp. 91-1l3). New York: Holt, Rinehart, Winston.
Bell, R (1981). Worlds offriendship. Beverly Hills, CA: Sage.
Berger, P., & Luckmann, T. (1966). The social construction of reality. New York:
Doubleday.
Bott, E. (1957). Family and social network. New York: Free Press.
Bradburn, N. (1969). The structure of psychological well-being. Chicago: Aldine.
Brain, R (1976). Friends and lovers. New York: Basic Books.
Byrne, D. (1971). The attraction paradigm. New York: Academic.
Cheek, 1., & Busch, C. (1981). The influence of shyness on loneliness in a new
situation. Personality and Social Psychology Bulletin, 7, 572-577.
Cutrona, C. (1982). Transition to college: Loneliness and the process of social
development. In L. Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of current
theory, research, and therapy (pp. 291-309). New York: Wi1ey-Interscience.
Davis, K, & Todd, M. (1982). Friendship and love relationships. Advances in
Descriptive Psychology, 2, 79-122.
Davis, M. (1973). Intimate relations. New York: Free Press.
Dweck, c., & Goetz, T. (1978). Attributions and learned helplessness. In 1. Harvey,
W. Ickes, & R Kidd (Eds.), New directions in attribution research (Vol. 2, pp. 158191). Hillsdale, NJ: Erlbaum.
Festinger, L. (1954). A theory of social comparison processes. Human Relations, 7,
117-140.
Fischer, C. (1982). To dwell among friends. Chicago: University of Chicago Press.
Foa, E., & Foa, U. (1974). Social structure of the mind. Springfield, IL: Thomas.
Frank, R, & Greenberg, M. (1979). Zooming in on TV audiences. Psychology Today,
13,91-103,114.
Fromm-Reichmann, F. (1959). Loneliness. Psychiatry, 22, 1-15.
Gans, H. (1962). The urban villagers. New York: Free Press.
Gerson, A, & Perlman, D. (1979). Loneliness and expressive communication.lournal
of Abnormal Psychology, 88,258-261.
Gordon, S. (1976). Lonely in America. New York: Simon & Schuster.
Goswick, R, & Jones, W. (1981). Loneliness, self-concept, and adjustment. Journal of
Gottlieb, B. (1981). Social networks and social support. Beverly Hills, CA: Sage.
Gough, H. (1964). California Psychological Inventory: Manual. Palo Alto, CA:
Consulting Psychologists.
Gough, H., & Heilbrun, A (1964). Manual for the Adjective Checklist. Palo Alto, CA:
Consulting Psychologists.
Hansson, R, & Jones, W. (1981). Loneliness, cooperation, and conformity among
American undergraduates. Journal of Social Psychology, 115, 103-108.
Harre, R (1977). Friendship as an accomplishment: An ethogenic approach to social
relationships. In S. Duck (Ed.), Theory and practice in interpersonal attraction (pp.
339-354). London: Academic.
Hartog, J. (1980). The anatomization. In J. Hartog, J. Audy, & Y. Cohen (Eds.), The
anatomy of loneliness (pp. 1-12). New York: International Universities Press.
Horowitz, L., & French, R de S. (1979). Interpersonal problems of people who
describe themselves as lonely. Journal ofConsulting and Clinical Psychology, 47,762764.
Jacobs, L., Berscheid, E., & Walster, E. (1971). Self-esteem and attraction. Journal of
Johnson, M., & Leslie, L. (1982). Couple involvement and network structure: A test of
the dyadic withdrawal hypothesis. Social Psychology Quarterly, 45, 34-43.
Jones, W. (1982). Loneliness and social behavior. In L. Peplau & D. Perlman (Eds.),
Loneliness: A sourcebook of current theory, research, and therapy (pp. 238-252). New
York: Wiley-Interscience.
244
Cecilia H. Solano
Jones, W., Freemon, J., & Goswick, R (1981). The persistence ofloneliness: Self and
other determinants. Journal of Personality, 49,27-48.
Jones, W., Sansone, c., & Helm, B. (1983). Loneliness and interpersonal judgements.
Personality and Social Psychology Bulletin, 9,437-442.
Jong-Gierveld, J. de. (1971). Social isolation and the image of the unmarried.
Sociologica Neerlandia, 7, 1-14.
Kiefer, C. (1980). Loneliness and Japanese social structure. In J. Hartog, J. Audy, & Y.
Cohen (Eds.), The anatomy of loneliness (pp. 425-450). New York: International
Universities Press.
Komarovsky, M. (1967). Blue collar marriage. New York: Random House.
Kurth, S. (1970). Friendship and friendly relations. In G. McCall (Ed.), Social
relationships (pp. 136-170). Chicago: A1dine.
LaGaipa, J. (1977). Testing a multidimensional approach to friendship. In S. Duck
(Ed.), Theory and practice in interpersonal attraction (pp. 249-270). London:
Academic.
LaGaipa, J., & Wood, H. (1981). Friendship in disturbed adolescents. In S. Duck & R
Gilmour (Eds.), Personal relationships III Personal relationships in disorder (pp. 169190). London: Academic.
Lopata, H. (1979). Women as widows: Support systems. New York: Elsevier.
Lorenz, K (1966). On aggression. New York: Harcourt, Brace, & World.
Loucks, S. (1980). Loneliness, affect, and self-concept: Construct validity of the
Bradley Loneliness Scale. Journal of Personality Assessment, 44, 142-147.
Lowenthal, M. (1964). Social isolation and mental illness in old age. American
Sociological Review, 29, 54-70.
Lowenthal, M., Thurnher, M., & Chiriboga, D. (1975). Four stages of life. San
Francisco: Jossey-Bass.
Lynch, J. (1977). The broken heart: The medical consequences of loneliness. New York:
Basic Books.
Maroldo, G. (1981). Shyness and loneliness among college men and women.
Psychological Reports, 48, 885-886.
Marshall, N. (1972). Privacy and environment. Human Ecology, 2,93-110.

McCormick, S., & Kahn, A (1980). Behavioral characteristics of lonely and nonlonely
college students. Paper presented at the annual meeting of the Midwestern Psychological Association, St. Louis.
Mead, G.H. (1934). Mind, self, and society. Chicago: University of Chicago Press.
Milardo, R, Johnson, M., & Huston, T. (1983). Developing close relationships:
Journal of Personality and Social Psychology, 44,964-976.
Moore, D., & Schultz, N. (1983). Loneliness at adolescence: Correlates, attributions,

and coping. Journal of Youth and Adolescence, 12, 95-108.
Moustakas, C. (1961). Loneliness. Englewood Cliffs, NJ: Prentice-Hall.
Myers, J., & Roberts, B. (1959). Family and class dynamics in mental illness. New York:
Wiley.
Oden, S., & Asher, S. (1977). Coaching children in social skills for friendship making.
Child Development, 48, 495-506.
Paloutzian, R, & Ellison, C. (1982). Loneliness, spiritual well-being, and the quality
of life. In L. Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of current theory,
research, and therapy (pp. 224-237). New York: Wiley-Interscience.
Peplau, L., Bikson, T., Rook, K, & Goodchilds, J. (1982). Being old and living alone.
In L. Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of current theory,
research, and therapy. New York: Wiley-Interscience.
Peplau, L., & Perlman, D. (1982). Loneliness: A sourcebook of current theory, research,
and therapy. New York: Wiley-Interscience.
Peplau, L., Russell, D., & Heim, M. (1979). The experience ofloneliness. In I. Frieze,
245
D. Bar-tal, & J. Carroll (Eds.), New approaches to social problems: Applications of

attribution theory (pp. 53-78). New York: Jossey-Bass.
Perlman, D., Gerson, A, & Spinner, B. (1978). Loneliness among senior citizens: An
empirical report. Essence, 2,239-248.
Ponzetti, J., & Cate, R (1981). Sex differences in the relationship between loneliness
and academic performance. Psychological Reports, 48, 758.
Putallaz, M., & Gottman, J. (1981). Social skills and group acceptance. In S. Asher &
J. Gottman (Eds.), The development of children's friendships (pp. 116-149). Cam-
bridge, England: Cambridge University Press.

Roe, A (1953). A psychological study of eminent psychologists and anthropologists,
and a comparison with biological and physical scientists. Psychological Monographs, 67, I-55.
Rook, K (1984). The negative side of social interaction: Impact on psychological

well-being. Journal of Personality and Social Psychology, 46, 1097-1108.
Rubenstein, e., & Shaver, P. (1980). Loneliness in two northeastern cities. In A
Hartog, J. Audy, & Y. Cohen (Eds.), The anatomy of loneliness (pp. 319-337).
New York: International Universities Press.
Rubenstein, e., Shaver, P., & Peplau, L. (1979). Loneliness. Human Nature, 2,
58-65.
Russell, D., Peplau, L., & Cutrona, e. (1980). The revised UCLA Loneliness Scale:
Concurrent and discriminant validity evidence. Journal of Personality and Social
Schachter, S., & Singer, J. (1962). Cognitive, social, and physiological determinants
of emotional state. Psychological Review, 69,379-399.
Schmidt, N., & Sermat, V. (1983). Measuring loneliness in different relationships.
Schwartz, J.e. (1972). Effects of peer familiarity on the behavior of preschoolers in

a novel situation. Journal of Personality and Social Psychology, 24, 276-284.
Sermat, V. (1978). Sources oflone1iness. Essence, 2, 271-276.
Slater, P. (1970). The pursuit of loneliness. Boston: Beacon.
Sloan, W., & Solano, e. (1984). The conversational styles of lonely and nonlonely
males with strangers and roommates. Personality and Social Psychology Bulletin,
10, 293-301.
Solano, e., Batten, P., & Parish, E. (1982). Loneliness and patterns of self-disclosure.
Solano, e., & Koester, N. (1984). Loneliness and communication apprehension. Unpublished manuscript, Wake Forest University.
Spitz, R, & Wolf, K (1946). Anaclitic depression: An inquiry into the genesis of
psychiatric conditions in early childhood. The Psychoanalytic Study of the Child, 2,
313-342.
Suedfeld, P. (1982). Aloneness as a healing experience. In L. Peplau & D. Perlman

(Eds.), Loneliness: A sourcebook of cu"ent theory, research, and therapy (pp. 54-67).
New York: Wiley-Interscience.
Suomi, S., & Harlow, H. (1978). Early experience and social development in rhesus
monkeys. In M. Lamb (Ed.), Social and personality development (pp. 252-271).
New York: Holt, Rinehart, Winston.
Tesser, A, & Campbell, J. (1980). Self definition: The impact of relative performance
and similarity of others. Social Psychology Quarterly, 43,341-347.
Thibaut, J., & Kelley, H. (1959). The social psychology of groups. New York: Wiley.
Trivers, R (1971). The evolution of reciprocal altruism. Quarterly Review of Biology,
46,35-57.
Walster, E., Walster, G., & Berscheid, E. (1978). Equity: Theory and research. Boston:
Allyn Bacon.
Weiss, R (1973). Loneliness: The experience of emotional and social isolation. Cam-
246
Cecilia H. Solano
bridge, MA: Massachusetts Institute of Technology Press.

Weiss, R. (1975). Marital separation. New York: Basic.
Williams, 1., & Solano, C. (1983). The social reality of feeling lonely: Friendship
and reciprocation. Personality and Social Psychology Bulletin, 9,237-242.
Willmott, P., & Young, M. (1960). Family and class in a London suburb. London:
Routledge & Kegan Paul.
Young, 1. (1982). Loneliness, depression and cognitive therapy: Theory and application. In L. Peplau & D. Perlman (Eds.), Loneliness: A sourcebook of cu"ent
theory, research, and therapy (pp. 379-405). New York: Wiley-Interscience.
Chapter 13
A Cognitive-Behavioral Approach to
Friendship Disorders
Jeffrey E. Young
In this chapter I outline a cognitive-behavioral framework for understanding disorders of friendship. I explain how specific patterns of thinking
and behavior create barriers to friendship satisfaction. I also offer a
developmental perspective, detailing the progression of cognitive organization that influences later friendship patterns. Finally, I describe interventions targeted at cognitive and behavioral blocks to friendship. The
emphasis in the chapter is on understanding long-term, chronic difficulties
with friendship.
The hypotheses discussed later in this chapter are based on extensive
clinical experience with patients suffering from friendship disorders. For
example, the concept of early negative schema was inferred from the
phenomenological experience of many patients in individual and group
therapy. It would be interesting for researchers in the future to compare
hypotheses derived clinically from disturbed individuals with those developed through other methods.
Cognitive-Developmental View of Friendships

Early Schemas
The term schema throughout this chapter refers to stable and enduring
patterns of thinking that develop in early childhood and are elaborated
throughout an individual's lifetime. These schemas serve as templates for
the processing of later experience. For example, children with very critical
parents may develop the schema, "Nothing I do is ever good enough for
other people." This schema is tacitly accepted by the children and forms
the basis for future expectations in relationships. They may come to expect
rejection as a result of this schema and, in fact, will distort later experiences
in friendships to confirm this schema. For instance, children who believe
248
Jeffrey E. Young
that nothing they do is good enough may be excessively cautious in

friendships, always trying to avoid situations that might expose their
inadequacy. Furthermore, they may become exquisitely sensitive to any
reactions from friends that even remotely suggest rejection, and they may
discount and disqualify positive feedback because it is inconsistent with
their view of themselves as inadequate. Schemas, therefore, become selfperpetuating; experience is processed through a mental filter that almost
guarantees that schemas will be reinforced.
Schemas are especially important in understanding chronic friendship
disorders. It has been my experience that the majority of clients with
friendship problems developed early schemas that interfere with their
capacity for satisfying friendships. Three types of schemas seem especially
important in friendship development: (a) self-concept schemas, (b) perception-of-others schemas, and (c) relationship schemas.
Self-concept schemas represent the view of self, as reflected by the
reactions of significant others. Children's primary means of knowing about
themselves is initially through interactions with parents and siblings and
later with peers. Some of the most frequent negative self-concept schemas
observed in clients revolve around the following themes: being a failure,
incompetent; being unlovable; being different; and being unworthy of
respect or attention.
Perception-of-others schemas represent children's views of other people.
These early perceptions are based on generalizations from parents and
siblings to the world at large. Some of the most frequently observed negative
schemas in this category involve a view of other people as: selfish, critical,
uncaring and unloving, controlling and manipulative, lacking in empathy,
superior to oneself, different from oneself, dangerous or malicious, favoring
other people to oneself, and deceptive, dishonest, not meaning what they
say.
Relationship schemas represent children's view of how relationships
operate. Children's early expectations about relationships reflect the quality
of their connections with parents and siblings. These relationship schemas
are closely connected to self-concept and perception-of-others schemas. For
example, children with a self-concept schema of being inferior and a
perception-of-others schema of being critical will typically have a relationship schema centering around the likelihood of rejection. Included among
the commonly observed dysfunctional relationship schemas are those
revolving around the themes of: rejection and failure, loss, entrapment
(being controlled by other's needs), emotional distance and coldness,
alienation and exclusion, competition or constant striving for attention and
approval, and overdependence.
As children move from their families into circles of peers when they begin
school, these early schemas become elaborated. For some children, the early
family environment will severely bias and distort peer interactions, so that
their school experience confirms their family life.
Cognitive-Behavioral Approach to Friendship Disorders
249
Continuing the earlier example, children who are constantly criticized at

home may enter school with a cautious, withdrawn manner. This social
withdrawal may lead to ridicule or exclusion by peers. Thus, their
experience at school will validate their earlier self-concept schema as
undesirable, their perception-of-others as critical, and their relationship
schema of rejection.
For other children, their peer experiences may be highly discrepant with
their family life. The home environment does not seem to have a profound
influence on their school life. To illustrate, a boy who is never nurtured at
home may approach peers with a markedly independent style (perhaps
because he was never allowed to be dependent). This independence may
help him in becoming a leader at school. The leadership role carries with it
a high level of social acceptance and support. The peer experience of
acceptance for this leader will seem discrepant from his family experience
of emotional distance and coldness. The child may resolve this discrepancy
by developing a second set of schemas specifically for social relationships.
These may be quite distinct from his earlier schemas regarding close (that is,
family) relationships. The following diagram illustrates the discrepancy:
Type oj Schema
Close Relationships
Social Relationships
Self-Concept
Perception-oj-Others
Relationship
Unlovable
Cold, unemotional
Nonnurturing
A valued leader
Accepting
Inclusion
It is my contention that the close schemas developed in the family lay the
groundwork for future intimate relationships and, in some cases, have a
profound effect on later friendships (Guida no & Liotti, 1983). The social
schemas developed through childhood and teenage peer groups generally
have the greatest impact on later patterns of friendship. For some people,
close schemas and social schemas overlap considerably; for others, the two
sets of schemas are quite different.
Whereas schemas are inferred from individuals' attitudes toward relationships and from their early family and peer experiences, behaviors are the
observable component in friendships. Early schemas lead to expectations of
how others will respond, which, in turn, lead to specific behaviors. The
schema thus becomes self-fulfilling. Table 13-1 illustrates this point.
To summarize, I have postulated that early experiences with family
members and peers lead to social and close schemas. These schemas bias an
individual's expectations in regard to friendship. These distorted expectations lead to inappropriate behaviors, which subsequently affect the
quality of an individual's friendships.
Levels of Friendship
In this section I discuss levels of friendship. Although these levels are not
specific or unique to a cognitive approach, I believe it is essential for any
250
Jeffrey E. Young
Table 13-1. Illustration of Relationship Between Early Schemas and Friendship

Behaviors
Early Schemas
Expected Responses
From Others
Friendship Behaviors
In social situations
I'm a lovable person
Withdrawal and shyness
others will reject
when people get to
in social situations.
me because I'm boring
Difficulty maintaining
know me well (close
conversations.
schema), but I'm boring and ugly. In close
friendships, people
and ugly on the surface
Avoidance of eye contact.
Open and self-disclosing
will like and accept me.
(social schema).
with good friends.
In social situations,
I'm basically unlovable
Socially skilled and confident in groups. Avoids
others will be drawn
(close schema), but I
self-disclosure with
to me. In close friendlook good on the
ships, people will soon
friends. Prefers
surface (social schema).
see how bad I am and
acquaintances to good
abandon me.
friends.
Other people will try to
Rebels against almost any
Other people are
get me to do things
demands or expectations
basically manipulative
from others. Must always
their way, if I let them.
and controlling (close
be in control with friends.
and social schemas).
Selects passive friends.
Avoids close friendships
to avoid feeling trapped.
Avoids becoming too close
I will eventually be
Close people will
abandoned again.
or dependent on anyone
eventually leave you
friend. Has many Level B
(close schema).
and C friends.
I will be humiliated
Becomes aggressive when
Other people are
competence is questioned
unless I stand up for
demeaning and want
in any way. Emphasizes
to hurt me (close and
myself at all times.
only positive qualities
social schema).
about self with other
people. Eventually
alienates most potential
friends.
theorist to have some operating belief about what it is that individuals need
or desire from friendships.
In this regard, my clinical experience confirms the theories of Sullivan
(1953) and Weiss (1973). I have found that most people suffer if they lack
close friendships to provide understanding, support, and stability or if they
lack social friendships to provide a sense of community and belonging.
In working with clients, I distinguish four levels of friends, which I term
Levels AA, A, B, and C. (In addition, I attach special importance to
relationships with parents, siblings, spouses, lovers, and children; however,
251
these will not be discussed here because they are beyond the scope of this
volume.) These four levels of friends can be differentiated along five
dimensions: history and expected duration of friendship; frequency of
contact; degree of mutual disclosure, caring, and trust; and shared
interests.
The first two dimensions refer to relatively objective characteristics of a
friendship. The history and expected duration of friendship refers to the length
of time the friends have known each other and to their expectation that the
friendship will continue even if they separate geographically. Levels AA and
A friends have already had a substantial history with each other and
anticipate that the friendship will continue in some form, regardless of
distance. For Levels Band C friends, history is not especially important, and
there is generally little concern with maintaining the friendship in case of
geographic distance.
The second variable is frequency of contact. Level AA friends have a daily
or almost daily "check-in." Level A friends may be equivalent to AA friends
along the other dimensions, except that the frequency of contact is not as
high. (When a Level AA friend moves away, the friendship often becomes a
Level A because of the less frequent contact.) Level B friends have
considerably more frequent contact than Level C friends, but not as high as
a Level AA.
The remaining two dimensions refer to subjective components of the
friendship. The degree of mutual disclosure, caring, and trust is especially
critical. Levels AA and A friends experience a high degree of concern,
caring, and trust. Furthermore, they look forward to sharing private
concerns, fears, and problems. These friends open up to each other.
Concern, caring, and trust are not as important with Level Band C friends,
and self-disclosure is much more limited.
The fourth dimension, shared interests, is often the initial basis for
developing all levels of friendship. Thus, friends at all levels are likely to
share some common interests. Level B friendships are usually based
primarily on shared interests. Level B friends are usually good companions,
buddies, or colleagues who work or play together and between whom there
exists a one-to-one relationship. The degree of caring is limited, however, in
comparison to Levels AA and A friends.
Level C friends usually belong to the same social group (for example,
church, athletics) or work at the same company; but they rarely associate
with each other one-to-one outside of the shared group or community. Table
13-2 summarizes these distinctions.
People vary considerably in their desire for friends at these four levels.
Nevertheless, it is helpful in working with friendship problems to have
clients map their social networks in terms of these levels. It is also important
to ask clients whether they are disturbed by a deficit at any of these levels. A
priori assumptions cannot be made about a client's friendship needs.
However, having noted these caveats, I will make a few observations. First,
most people seem to need one AA friend (often a lover or spouse). Without
Jeffrey E. Young
252
Table 13-2. Levels of Friendship
Characteristics
of Friendships
Level A
LevelAA
Brief
designation
Check-in;
best friend
History and
expected
duration
Frequency of
contact
High
Disclosure,
caring, and
trust
Shared interests
Very high
Daily, almost
daily
Moderate to
high
Close friend,
when
available
High
Level B
Companion,
good buddy
Acquaintance
Moderate
Low to moderate
Variable,
Moderate (e.g.,
once every 2
depends on
weeks)
distance and
time available
High to very
Moderate
high
Moderate to
high
Level C
Moderate to
high
Low,
occasional
Low
Moderate
such a daily, "check-in" relationship, loneliness is frequently a problem

(Morgan, 1985).
Second, most people with normal friendship patterns maintain one or
more Level A friends. The presence of Level A or AA friends provides
security, even when a love relationship breaks up. Furthermore, the
presence of Level A friends demonstrates the client's capacity to maintain
stable friendships over time.
Third, most people without Level B or C friends are limiting their lives
unnecessarily. These friends provide us with a connection to the community
at large and also supply a degree of novelty and stimulation. It is easy to
become bored and insular with only AA and A friends. Furthermore, Level
Band C friends provide access to a network of people that can be tapped
later for additional Level AA and A friends (and lovers).
To summarize, most people need a few friends to provide closeness and
stability (Levels AA and A), and a broader circle of friends to provide
novelty and a social community (Levels B and C). In the next section, I
examine the role that specific schemas play in the lives of clients who have
difficulty establishing friends at these different levels.
Classification of Friendship Disorders

Problems Initiating Friendships
I divide the disorders of friendship into two primary stages: problems
initiating friendships and problems deepening friendships. (Some general
253
problems with friendship are common to both stages.) Problems initiating

friendship refer to all difficulties involved in the process of meeting
potential new friends, whether in a group setting or one-to-one. The
initiating stage corresponds most closely to the social schemas described
earlier. In this stage people are most concerned with how they come across
to others at a superficial level. They are especially concerned with physical
attractiveness, social confidence, and conversational skills.
Social anxiety. One of the most common symptoms in this stage is social
anxiety. Many patients are so worried about how they appear to others in a
new situation that they become extremely anxious. Many of them develop
panic attack symptoms, such as heart pounding, hyperventilating, blushing,
sweating, and trembling in social situations. The development of anxiety
symptoms creates a vicious cycle. They become so focused on the symptoms
themselves that they are unable to concentrate on the conversation. They
become convinced that others will be able to see their anxiety and will view
them as crazy, insecure, or weak. These patients then try to control the panic
symptoms, but, paradoxically, the more they attempt to control them, the
worse they become.
The social anxiety disorder can be viewed as a cognitive dysfunction. At
one level of analysis, these patients' initial fears that they will be socially
inept are often the result of early family and peer experiences. As discussed
previously, these early experiences lead to strongly held schemas about
being socially undesirable in some way. Because schemas operate to
perpetuate themselves (cognitive consistency), socially anxious people are
hypervigilant to any cues that they are making a bad impression on others.
They are unable to accept information contrary to their schema, even when
they receive a favorable response from others, except perhaps when the
positive feedback is so overwhelmingly enthusiastic that they cannot distort
it. This type of unconditional feedback is quite rare in initial meetings. At
one level, then, social anxiety can be viewed as the logical consequence of
early negative social schemas playing themselves out in the initial stages of
contacting other people. Feedback from others is biased during information
processing to support the schema.
At a second level of analysis, social anxiety can be seen as an unsuccessful
attempt to control one's own behavior. Socially anxious patients become
aware of physiological symptoms, which opens up a second channel of
information. Not only are they trying to listen to the conversation of other
persons, but they are also attempting to monitor their physical state. In a
sense they are trying to listen simultaneously to two channels of communication. I call this process "spectatoring," by which I mean that they are
tuning into their own behaviors and physiological state as if they were
outside observers or spectators. This spectatoring acts as a diversion from
their primary task: participating actively in the social encounter. Instead,
they are distracted by their own thinking, which makes it almost impossible
for them to give sufficient attention to other people.
254
Jeffrey E. Young
This process has been termed "metacognition" (Hartman, 1983). Metacognition refers to the fact that the patients' problem is not only with the
content of their thoughts, but more importantly with the focusing of
attention. Socially anxious patients are engaging in a maladaptive process
of thinking; they are attempting to monitor two channels of information
simultaneously and intensively. Even if the content of the thinking were
entirely logical, the metacognitive problem of direction of attention would
be crippling.
At yet another level of analysis, the thinking of socially anxious patients is
disturbed. In addition to maladaptive focusing of attention, the patients
usually exaggerate the deleterious effects of social anxiety in the eyes of
other people. They often believe that everyone else can see their shaking
hands, unnatural smile, or sweating and that others will certainly devalue
and reject them at the first opportunity. This distorted view of how obvious
and distasteful their anxiety must be to others reinforces their negative
social schema and intensifies the anxiety symptoms.
To review, the cognitions of socially anxious patients are distorted in three
respects. First, the cognitions typically stem from negative social schemas
that bias their interpretation of feedback in a consistently self-degrading
way. Thus, they see themselves as making a far worse impression than they
are actually making. Second, at a metacognitive level, the patients are not
giving their undivided attention to the conversation because they are so
worried about the quality of their "performance" and by their anxiety
symptoms. This misfocusing of attention may actually impede the quality of
their social interaction. And, in a third respect, the patients are exaggerating
the degree to which their social anxiety will lead others to lose respect for
them.
Poor body image. Another common problem for patients with difficulty
initiating friendships is negative body image. Many patients are convinced
that they are so unattractive that no one would want to be friends with them.
This ugliness schema often arises during childhood or, especially, teenage
years. The schema can develop in several ways. First, the patients may have
been raised in a family where one or more siblings were considered more
attractive. Although the patients may in fact have been quite appealing in
appearance, they may have felt ugly relative to others in the family. Because
these schemas can develop very early, the children may not have had the
benefit of reinforcement for their physical appearance from people outside
the family. Because schemas are self-perpetuating, they might still be
convinced of their ugliness, despite considerable reactions to the contrary at
a later point in their life.
Another frequent route by which patients develop an ugliness schema is
that they may in fact have been homely or fat as children or as adolescents.
The repeated reminders of their unattractiveness, such as exclusion from
peer groups, can create schemas that are nearly impervious to change, even
255
when the same patients often blossom into extremely attractive adults.
Often, patients who were fat as children still think of themselves as fat
adults, even though they may be slimmer than average.
A third manner in which the ugliness schema can develop is through an
environment that encourages perfectionism or an excessive preoccupation
with physical appearance. In some families, children are taught that
nothing is ever good enough. These patients always focus on the flaws in
whatever they are evaluating. These perfectionists view themselves as ugly
because they do not feel they meet up to their own, or their family's ideal
physical specimen. They may feel they are ugly if they have one pimple, or
are 5 pounds overweight, or if their cheekbones are not quite high enough. A
related problem in families like this is an exaggerated emphasis on the
importance of physical appearance relative to other personal qualities.
Some children are taught, through parents' subtle and not-so-subtle
messages, that physical appearance is the sine qua non of social relationships. As patients, they cannot conceive that anyone would want to be
friends with them unless they are handsome or beautiful. They are
unconvinced that personality, loyalty, perceptiveness, and humor enter into
other's selection of friends.
Poor social skills. Another problem in initiating friendships is poor social

skills. As discussed later in this chapter, much attention has been paid to the
behavioral component of social relationships and to social skills training.
My own belief is that lack of social skills is one of many problems
encountered with these patients, but by no means the most important or the
most common.
Some patients, perhaps because they had inadequate role models as
children, lack social sensitivity. They seem unable to see how their own
behavior is affecting other people. In my experience, these patients often
had grossly impaired relationships with one or both parents in terms of
intimacy and communication. Because the parents provided so little
accurate feedback about the child's behavior or because the parents'
expectations were so vastly deviant from the norm, the patients never
learned what specific behaviors would trigger what specific reactions from
other people. As children they never learned to appreciate nuances of
feeling that would predictably be set off by their own social behavior.
For other patients with social skills deficits, their problems again stem
from early negative schemas. In these instances, the social behavior is an
adaptation to the schema. For example, in Table 13-1, I charted the example
of a patient with the schema, "Other people are demeaning and want to hurt
me." This patient engaged in a variety of self-defeating behaviors. He
bragged about himself in the hope that it would boost his image in the eyes
of others. He became belligerent and defensive whenever other people
criticized or joked with him, and he would become involved in extended,
public feuds when he believed his reputation might be damaged. As a result
256
Jeffrey E. Young
of these behaviors, which stemmed directly from his perception-of-others

schema as demeaning, the patient was practically a social isolate. He had
managed to alienate almost all potential friends.
Low conversational esteem. Many patients with friendship disorders have
very little confidence in their ability to carry on conversations. They often

complain that they never know what to say next, that they are boring, or that
they always make a bad impression on others. Unlike individuals with poor
social skills, these patients only believe that they cannot carry on a
conversation. In reality, their social skills are at least satisfactory. Patients
with social anxiety almost always have low conversational esteem, but the
reverse is not true. Many individuals with low conversational esteem do not
suffer from social anxiety. The primary differentiating factor seems to be
that unlike people with social anxiety, these individuals are not constantly
monitoring themselves while they are engaged in conversations. Patients
with this problem may avoid social situations because they expect failure,
and they are frequently self-critical after a social encounter has ended, but
they are not morbidly obsessed with their performance while it is ongoing.
More important, perhaps, they are not focused on controlling anxiety
symptoms; therefore, most of their attention is available for the conversation
itself, unlike the social anxiety cases. In short, then, patients in this category
are not as preoccupied with "spectatoring" and do not suffer from the
anxiety that characterizes the socially anxious individual.
Patients with low conversational esteem usually developed the problem
when they were children, either through f<tmily or peer experiences. An early
self-concept schema developed centering around the belief that, "I never say
the right thing." These individuals may in fact have been criticized, rejected,
ignored, overlooked, overshadowed, or excluded by parents, siblings, or
peers. However, they wrongfully attributed these responses from others to
deficits in conversational skills. The evidence rarely supports their belief
that they were ignored because of their verbal skills. However, this problem
can become self-perpetuating: Because they believe they will fail in
conversations, they become shy and withdrawn, and others often do
overlook them.
General Problems With Friendship
Low close esteem. The problem of low close esteem, and the two problems
that follow, present obstacles both to initiating and deepening friendships.

Patients with low close esteem believe that they are inherently unworthy,
unlovable, and inadequate. The closer they get to others, the more certain
they are that they will be found out. This conviction is based on an early
self-concept schema that, "Once people get to know the 'real me,' they'll
reject me." This schema almost invariably develops through the child's
interactions with parents. As children these people were usually given the
257
message that nothing they did was good enough, that they were not valued,
or that they were unloved.
Patients with low close esteem cope with their schema by keeping as
distant from others as possible. Some of them remain isolated because, they
reason, why even initiate relationships if the eventual outcome will be
rejection. Others in this category develop a polished exterior, keeping
friends at arm's length. They may have many Level B or Level C friends, but
no Levels A or AA friends. This latter group frequently develops what I call
the "fraudulence schema": "I can try to give the impression that I'm an
acceptable human being; but if they knew the truth, they'd abandon me."
This schema is unfortunately self-perpetuating. The more successful these
people are at creating a poised exterior, the more convinced they become
that the interior is inadequate. They avoid those close situations that might
disprove their schema. Low close esteem may be the most serious and most
resistant to change of all the disorders of friendship.
Alienation. Some individuals with friendship problems feel alienated from
others around them. They have the schema: "I am basically different from
other people." This schema can develop either through feeling left out in
one's family or through being excluded by peers. Unlike patients with low
close esteem, these individuals usually do not feel inherently inadequate or
inferior. Rather, they feel different, usually because of their interests or
values. Common examples include the "bookworm," who prefers reading to
playing outside or the "dreamer," who prefers movies and fantasies of
excitement to the routine of everyday life. The problem of alienation is
usually intensified during adolescence, when peer pressure for conformity is
highest. In high school, teenagers who are different may have to live in
isolation because they often find it difficult to meet similar-minded
friends.
As adults, alienated people are reluctant to initiate new friendships or
deepen superficial ones because of the expectation that there will be little in
common. These patients often exaggerate the discrepancy between themselves and others.
Mistrust. Some people with friendship disorders have a basic lack of trust in
other people. This mistrust often arises from traumatic experiences with
family members or peers. As children, they may have been psychologically,
sexually, or physically abused by parents, siblings, or peers. The perceptionof-others schema that develops is predicated on the notion that "Other
people are dangerous, and you can never predict when they will hurt you."
Naturally, this schema leads to extreme avoidance of others, both in terms
of initiating and deepening friendships. If the abuse was primarily in the
peer situation, the patient may feel comfortable with a few close friends but
will be reluctant to develop Level B or C friends. If the abuse was in the
family, the pattern of isolation is likely to be more extreme, and the
avoidance may restrict the possibility of friends at any level.
258
Jeffrey E. Young
Problems Deepening Friendships

Fear of self-disclosure. As I mentioned earlier in this chapter in discussing
levels of friendship, one of the primary factors in differentiating Levels AA
and A friends from Levels Band C friends is the degree of self-disclosure. I
define this as the degree to which friends share private problems, thoughts,
and feelings. Problems with self-disclosure are probably the most frequent
cause of difficulty in deepening friendships. The schema most closely linked
to "constriction" (a term I use to convey the problem of having private
thoughts and feelings bottled up inside) has already been discussed under
the section on low close esteem: "Once people get to know the 'real me,'
they'll reject me." As I mentioned earlier, the origins of this schema usually
involve extremely critical or rejecting parents. Patients with self-disclosure
problems go to considerable lengths to prevent others from finding out too
much about themselves. They may only be comfortable discussing a narrow
range of impersonal topics and usually recoil at any mention of private
feelings. Some of these patients will terminate Level B friendships if they
believe the friend is trying to get "too close." They sometimes report feeling
smothered or trapped by closeness.
Pacing problems. Another difficulty in deepening friendships is not knowing
how fast to pace the relationship. The greatest risk here is of pressing too
quickly to be close. Some of these patients come across as "needy" or
"clinging" to others, leading potential friends to back away. These people
are not satisfied with Level B or C friends; they often want everyone to
become a Level A or AA. Individuals with pacing problems do not realize
that developing trust and closeness takes time. Close friendships require
some history of experience; otherwise, the relationship may be based on
"pseudo-closeness."
The predisposing schema for pacing problems is often one of dependency: "I cannot bear to be alone; I need others around me at all times to
support me." This self-concept schema is generally the result of overprotective parents who give the child the message: "You need us very much.
Without us, you may not be able to handle all the problems that will arise."
As adults, these patients constantly work to recreate the parental closeness
they felt as children in their friendships. Sometimes they are successful, but
in other cases they drive potential friends away.
.
Unrealistic expectations. The problem of unrealistic expectations is similar to
pacing difficulties in that both types of patients often want more from
friendships than other people are usually willing to provide. Patients with
unrealistic expectations often have rigid expectations about what friends
should and should not do. For example, one patient of mine believed that
friends should always be on time, never cancel plans, always be available to
listen to problems, know what she expected of them without her having to
spell out her expectations, and place her higher in importance than they
259
placed any other friends. If her friends let her down, she became angry and
would express annoyance with them. Like other patients in this category,
she lacked flexibility. Her thinking did not allow for the possibility that
others might disagree with her standards of friendship. Furthermore, she
did not make distinctions among different levels of friends. A friend was
either a AA or not a friend at all.
A related problem in this category, although usually less serious, is
difficulty dealing with flux in relationships. Some patients have difficulty
tolerating the experience of change in relationships. Frequently, a friend at
one level drops to a lower level because oflife changes. Friends move away,
develop other friends, get married, become more involved in their work, etc.
Any of these factors may lead a friend to become less available. For some
people, these changes in friendship level are extremely distressing and often
lead them to back away from deepening friendships in the future.
It is difficult to generalize about the origins of this problem. In many
cases, these patients grew up with parents who were very rigid about the
"right" and "wrong" way to behave with others. Their thinking style is
generally rigid and is applied to friendships as well as most other aspects of
life. The relationship schema might be expressed as follows: "Friends
should follow my definition of friendship at all times and forever.
Otherwise, they are bad and unfair."
. For other patients in this category, the schema revolves more around the
possibility of loss and the need for stability. They may have grown up
without a stable base of support and companionship, and thus they want to
insure that they will not be alone again. The relationship schema might be
stated as, "I need to have a predictable core of friends or else I'll be alone
and isolated again."
Low-assertiveness. Some patients have difficulty asking to have their needs
met in friendships. In some ways, this group is at the other extreme of those
with unrealistic expectations. These people frequently feel frustrated and
angry because they do not get what they want and feel mistreated by friends.
The problem, though, is that they do not feel comfortable asking for what
they want. Unlike individuals with unrealistic expectations, their desires are
usually perfectly reasonable. Because they do not assert themselves, these
patients rarely feel satisfied with their friendships.
This problem seems to be based on two interlocking schemas. First, many
of these individuals have a self-concept schema that "I am not entitled to
ask for more than what I have." Second, they frequently have a perceptionof-others schema that "Others will reject me if I ask for what I want."
Generally, these schemas develop when parents either actively discourage
assertiveness in their children to minimize the demands on themselves or
when parents model passivity in their relationships with other people.
Restriction of affect. One subgroup of patients with difficulty deepening
friendships is characterized by a restriction of affect. They have difficulty
260
Jeffrey E. Young
expressing emotions, especially warmth and affection. They come across as

cold and impersonal to others, and their friendships are rarely satisfying,
even to themselves. People with restricted affect often experience friendships
as obligations. The schema is: "I do what I'm supposed to do for friends."
They do not expect affection or feelings, nor are they usually aware why they
are dissatisfied. As might be expected, these patients typically are products
of emotionally cold parents who did not express feelings to their children.
Such parents characteristically function according to duty, not love or
spontaneous involvement. In the diagnostic nomenclature, these patients
are described as having the features of an obsessive-compulsive personality
style.
Problems in friendship selection. Some patients have difficulty deepening
friendships because of their choice of friends. They seem to select people

who are difficult to get close to. For example, they may gravitate toward
aloof, critical, cold, inaccessible, or conceited people-frequently toward
other people with friendship problems. Often patients in this category try to
prove their own value by winning over people who seem hard to get. They
are typically more concerned with status and prestige than with closeness.
The underlying self-concept schema is usually based on social inferiority:
"I am not accepted by the people who are popular and successful." They
then devote much of their lives to the pursuit of those trappings (for
example, money, cars, titles, degrees) that will eventually earn them the
status they so strongly desire. Unfortunately, the quality of their friendships
usually suffers in the process.
Another possible explanation for the tendency of these patients to select
inaccessible partners is that they choose partners who confirm their
expectations. By selecting rejecting people, they get feedback that reinforces
their self-concept schemas of being unlovable, unacceptable, etc. Thus,
problems in partner selection may be the outgrowth of patients' desires for
"cognitive consistency."
Fear of entrapment. The last problem in deepening friendships is the fear of
being entrapped by the demands of other people. These patients view close
friendships as the potential loss of individuality, self-determination, and
privacy. Because of this schema, they are usually ambivalent about close
friendships and may devote considerable energy to protecting their privacy.
They often overreact to the reasonable requests of friends and set up
protective barriers when none are needed.
These patients are typically the product of enmeshed families and
intrusive parents. As children, their parents may have been involved in
every decision they made, however trivial. The family members may have
kept no secrets from one another, and time spent alone may have been
discouraged. As adults, these patients fight desperately to avoid feeling
smothered again. The relationship schema can be expressed as follows:
"People will not respect my boundaries. Therefore, I must fight hard to
maintain them."
261
Cognitive Therapy for Friendship Disorders

General Principles of Cognitive Therapy
The most basic principle of cognitive therapy is that our thoughts and
beliefs (cognitions) playa major role in determining how we feel and
behave. Cognitive therapists actively work with patients to help them
change beliefs about themselves and about others, with the expectation that
these cognitive changes will lead in tum to improved emotional and
behavioral functioning.
Levels of cognition. Cognitive therapy postulates several levels of cognition,
differing in terms of stability and duration. I discuss three types of

cognitions: early schemas, automatic thoughts, and underlying assumptions. The most stable and enduring cognitions (that is, those beliefs that
originate earliest and are most resistant to change) I call early schemas.
Patients are rarely aware of their early schemas. I have already discussed
this concept in considerable detail earlier in the chapter.
The least stable cognition is termed automatic thought (Beck, 1976). An
automatic thought is a subliminal thought (or image) that arises in a
particular situation. Patients mayor may not be aware of their automatic
thoughts but can be trained to identify them with practice. Consider a young
woman who has just called a friend to arrange plans for a movie. The friend
tells her she is too busy to make plans that week. The woman might have
several automatic thoughts, including:
1. She's no friend of mine anymore.

2. She doesn't care about me anymore.
3. I have no friends now.
4. Next time she calls, I'll say I'm too busy.
5. I'm frightened of being alone.
Depending on the nature of her automatic thoughts, the woman could have
a wide range of emotional reactions (ranging from panic to fury) and of
behavioral responses (ranging from terminating the friendship to planning
revenge).
The third type of cognition is an underlying assumption (Beck, 1976). The
underlying assumption is intermediate between an early schema and an
automatic thought in terms of its duration and stability. Patients are rarely
aware of their underlying assumptions. These assumptions comprise the
rules, values, and standards that people use to judge their own and others'
behavior. (The term corresponds approximately to the psychoanalytic
notion of superego). In relation to friendship, some common underlying
assumptions include:
1. Friends should be there all the time for you or else they're not really
friends.
2. Friends should not criticize one another.
262
Jeffrey E. Young
3. Friends should always include each other in whatever they do.

4. If I'm nice to someone, he/she should be my friend.
These underlying assumptions differ from early schemas in three important
respects:
1. Assumptions are conditional; schemas are unconditional. By this, I mean
that an assumption can usually be phrased as an "if ... then" statement: "If
he calls me when I ask, then he is a good friend." With such assumptions,
there is the possibility of carrying on normal friendships, so long as the rules
and expectations are met. Schemas are unconditional: "No one could love
me; people will reject me; I can't talk appropriately to other people." So long
as the schema is present, normal relationships are nearly impossible.
2. Assumptions generally apply equally to everyone; schemas (particularly selfconcept schemas) are more unique to the individual patient. For example,
patients who hold the assumption that "Two close friends should know what
each other is thinking" will apply the rule equally to themselves and to
others. Patients with the schema "I am unlovable" do not extend the belief to
include the notion that "Everyone is unlovable."
3. Schemas have a more profound impact on feelings and behaviors than do
assumptions. Schemas, because they are at the core of our view of ourselves
and other people, and because they originate so early, generalize across all
aspects of our lives and dramatically affect our emotional reactions to
various situations. A typical underlying assumption, because it is often
limited to a particular class of situation (for example, making plans) and
because it is conditional, is not so pervasive in its impact.
Differentiation from other therapeutic approaches. Beyond these theoretical
constructs, cognitive therapy can be differentiated from other approaches in

several respects (Beck, Rush, Shaw, & Emery, 1979). First, cognitive
therapists are very active and structured in therapy sessions. They do not
assume a detached or nondirective role. They work with patients to set a
mutually acceptable agenda for each session. They conceptualize patient
problems in cognitive-behavioral terms and then plan and implement
interventions to bring about cognitive, emotional, and behavioral change. In
this sense, cognitive therapy is more problem-oriented than insightoriented. The therapists' task is to identify a central problem; pinpoint the
early schemas, assumptions, and automatic thoughts leading to the central
problem; explain to the patient the relevance of these cognitions to the
central problem (including emotions such as loneliness and behaviors such
as avoidance of social contact); outline a plan of change for the patients;
implement specific techniques to change cognitions and behaviors; confront the patients' avoidance of change; and persist until the central
problem has remitted.
Another important difference between cognitive therapy and more
traditional approaches is the use of homework assignments (Young & Beck,
1982). At the end of each session, the patient and therapist agree on a self-
263
help assignment designed to assist the patient in translating new insights

learned during the session into life outside of therapy. Homework thus
facilitates transfer of learning outside the session. Examples of homework
assignments related to friendship disorders include:
1. Approach a colleague at work and suggest going out to lunch.
2. Write down any automatic thoughts you had that gave you difficulty in
asking the colleague to lunch.
3. List all evidence in your life supporting and contradicting your schema
that "No one could ever like me."
4. Write down the advantages and disadvantages of insisting that a friend
meet your very strict underlying assumptions regarding friendship.
5. Carry a flash card around with you, reminding you of all the disadvantages of avoiding social contact. Read it at least three times a
day.
It has been my experience that systematic homework assignments are one
of the most important advances of cognitive therapy. Patients who complete
these self-help assignments are much more likely to improve than are those
who neglect them.
Collaborative empiricism. In cognitive therapy, the patient and therapist

work as an investigative team (Beck & Young, 1985). Each of the patient's
key schemas and assumptions is approached just as a scientific team would

approach a question. The cognition is posed as a hypothesis to be tested (for
example, "Is it true that I am basically unlovable?"). The patient and
therapist collect all evidence that either supports or refutes the hypothesis,
including experiences from the past, circumstances in the present, and
unexplored possibilities for the future. Next, the team subjects all the
evidence to a logical analysis and then reaches one of several conclusions:
1. The hypothesis is true; I have to work around it or accept it.
2. The hypothesis is obviously false and is based on my selective biasing of
information to perpetuate an early schema.
3. There is insufficient evidence to evaluate the hypothesis; I must
undertake "experiments" to gather enough information to test my belief
fairly.
4. The evidence proves that I have a problem, but my problem may not be
what I thought it was (for example, I thought I was unlovable, but
actually I push people away). I need to change my behavior in the
future.
In summary, the patient and therapist collaborate to test the patient's
schemas and assumptions in an empirical fashion. The conclusion then

leads the patient to alter, systematically, feelings and behaviors, primarily
through homework assignments.
264
Jeffrey E. Young
Friendship Assessment
The first step with patients presenting friendship disorders is to diagnose the
problem. This involves a careful assessment of the patient's past experiences
with friends. The initial task is to determine whether the individual's
problems are with initiating friendships, deepening friendships, or both.
The following general questions (with appropriate probing after each one)
often yield information about initiating friendships:
Do you find it difficult to meet new people?
Are you anxious or uncomfortable around people you don't know?
Do you expect people to like you when you first meet them?
How often do you engage in social activities involving other people (for
example, clubs, parties, dinners)?
5. How many friends do you currently have?
6. Do you find that, no matter how hard you try, people don't want to be
your friend?
7. How do you feel you come across to people when you first meet them?
1.
2.
3.
4.
In my practice, after asking these general questions, I then explain to new

patients the four levels of friends described earlier in the chapter (Levels
AA, A, B, and C). I ask them to classify their friends during the past 2 years
under these categories and to indicate how long each friendship lasted. If
there are no lasting Level AA or A friends, I then ask a second set of
questions:
1. Why do you think you have so few AA or A friends?
2. Let's discuss some of your Level B friends. Why haven't you gotten
closer to
(name of friend)?
3. Do you find it difficult to talk to friends about private problems or
feelings? Why?
4. Do you trust other people enough to get close to them?
5. If people really got to know you, would they like you? (Why not?)
6. Have you had bad experiences in the past with friends? Describe
them.
7. I notice that most of your Level AA and A friendships end pretty
quickly. Let's go through each of them and discuss why you think the
friendship ended.
8. In what ways do you feel your friends disappoint you?
9. You tell me that you don't really feel close to your friends. Do you have
ideas about what's missing? What would you need to feel close?
10. Do you find that, when you try to get closer to friends, they back off or
end the friendship?
After assessing whether the problem is initiating friendships, deepening

them, or both, I start to examine the patient's early negative schemas and
underlying assumptions. Some of the questions I explore include:
265
1. Tell me about your early childhood. Did your parents, brothers, and
sisters value you? Respect you? Include you? Criticize you? Ignore
you?
2. Did you feel close to your family? (Why not?)
3. Were your parents loving and affectionate or more cold and aloof with
you?
4. Did you have enough privacy? Did your parents intrude in everything?
Did they leave you alone too much? Were they uninvolved?
5. Were you the favorite child? The least favored?
Then, I investigate early peer experiences:
6. As a child, how did you get along with the other kids in your
neighborhood? In your school?
7. Did they ignore you? Make fun of you? Include you? Pick on you? Were
you a leader? An outcast?
8. Did you feel lonely as a child? Did you isolate yourself? Why?
9. Did you have any close friends? Did they disappoint you?
10. As a teenager, did you have a group you felt a part of? Did you value that
group? Did you wish you could be part of some other group?
11. As a teenager, did you have close friends? Did they disappoint you?
12. Did you feel you had much in common with other people?
Finally, I select at least two recent friendships and try to get a detailed
description of how the patient relates to friends. I ask for specific examples
of situations that upset the patient and for events that upset their friends. I
sometimes role-play specific events to get an even better sense of how the
individual interacts with friends.
After piecing all this information together, I arrive at a tentative diagnosis
of the friendship disorder. The diagnosis for each patient includes a
classification of friendship disorders drawn from the detailed list provided
earlier in the chapter, early schemas and assumptions related to each
disorder, possible origins, and a list of dysfunctional behaviors related to the
disorder. A structured form of providing this information appears in Figure
13-1.
As an example of a case assessment, I have provided in Table 13-3 a
cognitive-behavioral assessment of a patient named Clyde. (This analysis
corresponds to Part II of the assessment from the structured form.)
In the next section, I discuss the treatment strategies that follow from
cognitive-behavioral assessments such as these.
Initiating Treatment
After the assessment process is completed, the therapist begins the process
of intervention.
266
Jeffrey E. Young
Stages in friendships. I have already mentioned the important distinctions

between initiating and deepening friendships. These steps can be further
subdivided into six smaller steps:
1. To enter a situation where there is a possibility of meeting a new

friend.
ASSESSMENT OF FRIENDSHIP DISORDERS

Patient's Name _ _ _ _ _ _ _ _ _ _ _ _ __
Part I. Classification
A Problems initiating friendships
1. Social anxiety
2. Poor body image
3. Poor social skills
4. Low conversational esteem
B. General problems in friendship
5. Low close esteem
6. Alienation
7. Mistrust
C. Problems deepening friendships
8. Fear of self-disclosure
9. Pacing problems
_ _ _ 10. Unrealistic expectations
_ _ _ 11. Low assertiveness
_ _ _ 12. Restriction of affect
_ _ _ 13. Problems in friendship selection
_ _ _ 14. Fear of entrapment
Part II. Cognitive-Behavioral Assessment
For each of the problems checked above, answer the following questions:
A Early schemas
1. Self-concept
2. Perception-of-others
3. Relationship
B. Possible origins of problem
1. Early family experiences
2. Childhood peer experiences
3. Teenage peer experiences
4. Later experiences
C. Underlying assumptions about friendship
D. Dysfunctional behavior patterns in friendships
Figure 13-1. A structured form for assessment of friendship disorders.
1. I'm a coward. (self)

2. I'm clumsy and
awkward. (self)
3. People will laugh at
me. (perception-ofothers)
1. I'm a weak person.
Problems initiating:
social anxiety
Problems deepening:
fear of self-disclosure
(self)
2. If people get to know
me well, they'll see
I'm weak and reject
me. (perception-ofothers)
3. I'm a phony. (self)
Early Schemas
Classification
It's bad to be weak and

vulnerable.
Clyde was the youngest

child. He was overshadowed by his
brothers, and viewed
himself as weak. His
parents overprotected
him. He tried to
compensate by acting
tough outside the
home.
Underlying
Assumptions
If people act tough,
they get more respect.
Possible Origins
Clyde was small for his
age and developed
motor skills late.
Other kids picked on
him.
Table 13-3. Cognitive-Behavioral Assessment of Clyde
2. Will not show any

weakness or
vulnerability.
3. Devalues other people
he sees as weak.
1. Avoids self-disclosure.
situations
2. Acts "macho."
1. Avoids most social
Dysfunctional
Behaviors
2:l
;;l
.s.:::r'
....t:I
::s
'T1
::2.
(1)
g-
'0
eL
$.
o
Il:>
g.
tel
~.
~.
268
Jeffrey E. Young
2. To approach a new person in an appropriate setting and initiate a

conversation.
3. To initiate a second meeting with this new person, in which you engage in
an activity of common interest.
4. To have contact with the new friend on a regular basis in which
"nonthreatening" topics are discussed.
5. To trust the friend enough to engage in mutual self-disclosure. To solve
problems, weaknesses, etc. To be yourself.
6. To maintain the friendship at this level for a sustained period of time. To
make a commitment and expect a commitment in return.
The premise of this approach is that patients must be comfortable with
earlier stages in the development of friendships before they can progress to
higher levels (young, 1982). Thus, the treatment begins at the earliest of
these six stages that the individual has not yet mastered. Thus, when both
stages are problems, initiating generally takes precedence over deepening
friendships.
Explanation of rationale. With the target stages established, the therapist then
outlines for the patient his or her assessment of why the person is blocked at
this stage. The assessment is, of course, based on the information gathered
during the assessment process. For example, the therapist might explain to
Clyde (see Table 13-3) that his problem in initiating conversations is based
on social anxiety, that the anxiety seems to be based on three particular
schemas (for example, ''I'm a coward"), that the origin seems to be in his
early experience of being picked on by other children, that he believes that
tough people get more respect, and that these beliefs lead to specific
dysfunctional behaviors. The therapist enlists Clyde's assistance in revising
or elaborating on this tentative conceptualization of the social anxiety
problems. The therapist also asks Clyde whether he is willing to work on
these problems.
Changing Schemas
Once the patient has agreed with the conceptualization, the therapist begins
the intervention.
1. The therapist asks Clyde to describe memories of incidents in the past
that trigger feelings of being weak, clumsy, and awkward or the object of
ridicule. These memories "activate the schema," which in tum usually
triggers a strong emotional reaction in the session, such as sadness, shame,
or anger. For homework, Clyde is asked to list any other situations during
the past few months in which this schema was activated.
2. In the next session, the therapist asks Clyde for all the evidence from his
past experience that he believes proves the validity of the schema that he is
basically weak, clumsy, and awkward. The therapist reviews each piece of
evidence with Clyde to determine whether he has accurately evaluated the
269
meaning of the experience. For example, if Clyde gives an example of being

picked on by a gang of older boys in which he ran away instead of fighting,
the therapist might try to show Clyde that his running away was a wise
decision rather than proof of cowardice.
3. In the following session, the therapist might ask Clyde for all the
evidence from his past experience that he believes is contrary to his schema
of cowardice and weakness. Any examples where he stood up for himself
(either verbally or physically) would be relevant to this phase of treatment.
The therapist sometimes has to "poke around" for a while before the patient
will concede that he has, on occasion, been strong. The therapist demonstrates how his schema works to magnify incidents that support the schema
and to discount or minimize events that contradict it. Clyde could then be
assigned homework to continue listing any events, past or present, that are
relevant to testing the validity of the schema.
4. By the end of this process, Clyde and the therapist attempt to draw some
conclusions about the validity of each schema. For example, they might
conclude: "Because I was small as a child, I often was unable to defend
myself. Since childhood, however, I have appropriately stood up for my
rights and am no longer weak. I keep viewing myself as weak and cowardly,
even though evidence no longer supports it." This is called the "rational
response" to the schema.
Clyde would be asked to make up a flash card for this schema. The flash
card is simply an index card, with the schema on one side and the rational
response on the other. Clyde would be instructed to read the flash card
several times a day, with special emphasis on situations that might
previously have triggered the schema. (If the schema had proved true rather
than exaggerated, the therapist would have worked with Clyde on a
behavioral program to assert himself more often and more effectively).
5. Next, the therapist would help Clyde see how his schema of weakness
has led him to engage in several self-defeating behavioral patterns: to avoid
situations, to act inappropriately macho and tum people off, to fear selfdisclosure, and to devalue other people when they show vulnerability. The
therapist would assign Clyde behavioral tasks to change, gradually, these
self-defeating behaviors. In each situation, Clyde would write a new flash
card explaining why his previous behavior has been based on an inaccurate
schema of weakness and stressing the importance of experimenting with
new behaviors. As Clyde changes his behavior and as his schema gradually
becomes neutralized, his friendships will improve, both in number and in
quality.
Treating Specific Problems Initiating Friendships
Social anxiety. As described earlier in this chapter, social anxiety usually has
three levels of cognitive dysfunction; each level must be addressed in
treatment. At the first level, social anxiety usually stems from one or more
270
Jeffrey E. Young
schemas about being socially undesirable. To intervene at this level,

therapists follow the procedures described previously in the section on
treating schemas: they help socially anxious patients gather evidence
supporting and contradicting the schema, draw conclusions regarding its
accuracy, and combine behavioral experiments with flash cards to fight the
schema.
At the second level, social anxiety involves "spectatoring," maladaptive
focusing on one's own performance and anxiety symptoms. To counteract
this problem, therapists train the patients to refocus on the content of the
conversation at hand, rather than their own performance. Each time
patients start evaluating their performance, they are instructed to picture a
traffic stop sign. The stop sign is a cue to refocus on the conversation itself
and to repeat verbatim to themselves what the other person in the
conversation is saying until they are able to join in again. The rational
response is: "There is no value in trying to control or monitor my anxiety or
my performance. It just makes me more nervous. So I'll refocus on the
conversation and practice being a good listener."
At the third level of cognitive analysis, social anxiety derives from the
belief that others will see the patient's anxiety and will be extremely critical.
The typical rational response to this cognition is: "Even though I feel very
anxious, my symptoms are not that obvious to other people. And, even if
others do notice my anxiety, they will probably be sympathetic rather than
critical or demeaning."
Once therapists have completed the interventions just described, it is
essential that the patients take every possible opportunity to practice
overcoming the social anxiety. This practice can be gained through roleplaying during the therapy session, often with videotape; participating in a
group therapy experience; and engaging in behavioral tasks outside the
therapy to initiate social contact in various social situations. It is essential
that therapists confront actively the patients' tendencies to avoid these social
contacts. This avoidance can be overcome by asking patients to keep
records of social encounters each week and by asking them to agree to a
contract to engage in a certain number of encounters each week. If patients
still resist social contact, therapists can try to insist on a group therapy
experience as a precondition of continuing individual therapy. In my
experience, little or no progress will be made with social anxiety so long as
patients avoid initiating contact.
Poor body image. Again, therapists must address the core schema, usually
revolving around ugliness, height, weight, or sexiness. Once evidence is
gathered, patients can usually see that their self-concept schema is greatly
exaggerated and that they are more attractive than they think. Furthermore,
therapists can show patients that they are placing undue emphasis on
physical attractiveness, that people choose friends on the basis of many
271
characteristics (including, for example, personality, common interests, and

loyalty), and that physical appearance is only one quality others consider.
Finally, some patients can be taught techniques for improving their
appearance, such as good grooming, dressing in a more becoming fashion,
and dieting.
Poor social skills. Some patients do not have appropriate social skills for
dealing effectively with initiating friendships. Most often, the problem is
poor empathy. Patients may talk too much about themselves and not really
listen to the other person. Therapists can use role-playing to teach these
people listening skills. One technique for doing this is the Johnny Carson
technique. The therapist and patient take turns playing the roles of Johnny
Carson and of a guest being interviewed. The patient can be shown that
Johnny carries on a conversation primarily by asking the guest questions
and by sharing personal experiences that confirm the opinions the guest
was offering. (For more detail on this technique, see Young, 1981.) In many
of these role-plays, therapists may have to explain to patients how certain
remarks they make trigger undesired responses from others. Through this
feedback process, patients can gradually improve their empathy.
With some people who have poor social skills, the problem can be traced
to an early schema, as mentioned previously in this chapter. For example,
the therapist can help a patient who has a perception-of-others schema of
demeaning, by encouraging the patient to test the schema and achieve a
more realistic view of others. With this new view, the patient is much less
likely to engage in social behaviors that tum other people off, such as
bragging or acting defensive.
Low conversational esteem. These patients only believe that they cannot carry
on conversations. Therapists must examine early schemas about never
being able to say the right thing. Once the schema is challenged, as
explained in the section on changing schemas, the patients gradually stop
expecting failure in social situations. This is, again, accomplished through a
combination of flash cards and specific behavioral experiments designed to
help the patients gain confidence in social situations. These patients must
also be helped to talk more spontaneously, usually through role-playing in
the session. The patients are instructed to carry on a social conversation
with their therapist without trying to monitor what they are about to say
next. Once these people stop censoring what they say, they usually get
increasingly positive feedback from others.
Treating General Problems With Friendships
Low close esteem and fear of self-disclosure. As discussed earlier, these patients
usually believe that they are inherently unworthy, unlovable, and inadequate. The primary solution to this problem, as with most of the others
272
Jeffrey E. Young
described here, is to change the early schema using the techniques outlined
previously. In general, these individuals come to see that the reason their
family members were critical or rejecting of them was not because of any
inherent inadequacy that they had, but because of psychological problems
of the parents (such as unrealistically high expectations or general
frustration with life) or family structure (for example, the older child and
younger child may have gotten a lot of attention whereas the middle child
was neglected).
This new interpretation of the schema must be combined with specific
assignments to share more private feelings with friends. As patients expose
more of themselves, they see that others value them more, and their
friendships become more satisfying. These experiments in self-disclosure
counteract patients' beliefs that no one can understand them, that they are
different from other people, and that other people would reject them if they
disclosed their weaknesses or "shameful" thoughts.
Alienation. These patients believe that they are basically different from other
people. Again, therapists work to change the schema by examining and
reinterpreting evidence. These individuals can be shown that they exaggerate the discrepancy between themselves and others and can be guided to
discover ways in which they are similar to specific friends. Therapists teach
these patients to look for people who share common interests and to stop
focusing selectively on the ways in which they differ from others.
Mistrust. These patients need help challenging their schema that other
people are dangerous and will hurt them. In addition to changing the
schema using the standard approach described, therapists must gently push
these patients to act in ways that give other people an opportunity to show
that they can be trusted. For example, a homework assignment might be to
ask a friend for help or to share a secret. Group therapy experience can be
extremely valuable for people with mistrust schemas.
Treating Specific Problems With Deepening Friendships
Pacing problems. Some patients press too quickly to deepen a friendship,
usually because of schemas involving dependency. First, therapists must
attempt to alter the schema that they need other people around all the time
to support them. This can be best accomplished by asking the patients to do
as many activities as possible independently, to prove that they can take
care of themselves. These individuals should also be encouraged to spend
periods of time by themselves so that they can see that they can deal with
aloneness.
Concurrently, therapists can list specific behaviors that other people view
as "clinging" or "too needy." These include expecting to spend a lot of time
together at the beginning of the friendship, acting possessive toward a new
273
friend by showing jealousy of time that they spend with other people, and
confiding information that is too personal in the early stages of a friendship.
Patients are instructed to hold off or moderate these behaviors until the
friendship has had more time to develop.
Unrealistic expectations. Patients with this problem need information from

therapists about what is realistic to expect from friends. My approach is to
explain to these individuals that they will not usually be able to control the
behavior of their friends by criticizing them or demanding better treatment.
Rather, they should learn what each friend is willing to offer them in a
friendship and to adjust expectations accordingly. I explain that people
have different ideas of what a friend should do, and there are few absolute
standards of right and wrong. For example, people have different ideas
about how important it is to be on time, in what circumstances it is
acceptable to cancel plans, and how much friends should sacrifice for each
other. Given these differences, patients must accept the idea that there are
friends of different levels and that they cannot criticize a Level B friend for
not acting like a Level AA friend. If they complain too much, friends will
back away from them even more.
A related problem is difficulty in dealing with flux in relationships. I
usually explain to people who have this problem that friendships are
continually changing: people move, make new friends, marry, and take on
other responsibilities. These patients must come to accept these changes as
natural and inevitable. This flux does not indicate that friends cannot be
trusted or that friends no longer value them. The only long-term stability in
friendship comes from periodically developing and nurturing new friends.
Level C friends can move up to Level B, Level B to Level A, and so on, to
compensate for friendships that drop in closeness. Unfortunately, individual friends do not usually provide as much long-term stability as do marital
partners and family members. This is because most people seem to place
higher priority on spouses, lovers, and close family members than on
friends when they are forced to make choices.
Low-assertiveness. These patients have trouble asking for what they want.
Cognitive therapists must first address the core schemas involved. These
usually include a lack of sense of entitlement with other people and the fear
of being rejected if they ask too much. Therapists use the techniques
described earlier to question the validity of these schemas. Unassertive
people are helped to see that they are as worthwhile as other people, that
they have a legitimate right to ask to have their needs met, and that others
will not reject them for making reasonable requests. Role-playing and
homework assignments are especially useful in guiding these people to ask
for what they want.
Restriction of affect. These patients have difficulty experiencing and express-
274
Jeffrey E. Young
ing warmth and closeness. The solution here first requires educating the
patients about people's needs for affection. They must be helped to
recognize their own unmet needs for closeness. Sometimes this can be
accomplished by asking them to describe the emotional climate in their
home and how they felt as children in this environment. These patients
need help seeing the link between their feeling that "something is missing"
from their friendships and their desire for closeness.
Once these patients identify their needs for intimacy, they are encouraged
to act in ways that will bring them closer to friends: to express liking and
caring, to compliment their friends, to touch their friends when appropriate,
to confide more, and to buy gifts or send cards on special occasions. Group
therapy is extremely useful in providing opportunities for these patients to
express their emotions in a safe environment.
Problems in friendship selection. These patients often have the schema, "I am
not accepted by people who are popular and successful." They then try to
prove their own social desirability by selecting "friends" who are extremely
successful, popular, or hard-to-get.
The approach that 1 use with these individuals is, first, to explore the
origins of the schema. After doing this, 1 contrast the advantages and
disadvantages of choosing most friends on the basis of their social status.
These patients usually conclude that the inability to be comfortable and to
"be myself' around such friends outweighs the advantages of feeling more
desirable if one can win them over. 1 guide them to see that the primary
advantages of a close friend are to provide companionship, to know each
other very well, to serve as a "safe haven" from the pressures of life, and to
trust and depend on another person. High status has no connection to
meeting these important needs.
Other patients in this category select "friends" who confirm their schemas
of undesirability. They choose others who are critical of them or who have
minimal interest in being friends. These patients then try in vain to win
these uninterested people over. The first step in working on this problem is
for therapists to point out the pattern to the patients: they are selecting
rejecting "friends" and then finding confirmation of their early negative
schemas when they are mistreated. Next, therapists can help these patients
identify clues early on in the selection process so that they can recognize
and avoid contact with potentially rejecting acquaintances. Perhaps most
important, therapists can work on modifying the negative self-concept
schemas that almost always underly this pattern. These schemas can be
evaluated and changed using the techniques already outlined.
Fear of entrapment. These patients generally have schemas equating friendship with the loss of freedom and individuality. Therapists assist these
patients in exploring the origins of the schema. Next, therapists try to break
down the all-or-nothing thinking: "Either 1 am completely free of obliga-
275
tions to other people or else I am trapped and cannot be myself." These

individuals can be shown that most other people will respect their
boundaries and, furthermore, that it is possible to assert oneself when other
people are invading one's privacy or freedom to an inordinate degree. Most
importantly, though, patients with the fear of entrapment must come to see
that they can give up a few degrees of freedom in order to get close to other
people while still maintaining a sufficient amount of privacy and selfdetermination. Therapists can review individual friendships with patients to
help them see where and how to draw the line between reasonable
obligations and commitments to friends, on the one hand, and unreasonable expectations and intrusiveness on the other.
Summary
In this chapter, I started by outlining a cognitive-developmental view of
friendships. This view is based on the primacy of three types of schemas:
self-concept, perception-of-others, and relationship schemas.
I also emphasized the importance of different levels of friendship: Level
AA, A, B, and C friends. These levels differ in terms of history and expected
duration; frequency of contact; degree of mutual disclosure, caring, and
trust; and shared interests.
I next suggested a classification of friendship disorders. This classification involved three types of friendship problems: general problems,
problems initiating friendships, and problems deepening friendships.
In the last sections, I described a treatment approach to these friendship
disorders. This approach is based on the general principles of cognitive
therapy. I explained how three levels of cognition could be systematically
changed through a structured, active, problem-oriented treatment. The
treatment is based on collaborative empiricism and relies heavily on selfhelp homework assignments. I outlined procedures for cognitive assessment
and then illustrated in detail the process of changing schemas. Finally, I
went through each of the friendship problems and discussed briefly the
procedures for intervention.
I believe that this cognitive approach offers a systematic way of
understanding and treating friendship problems, and, though it is still in the
process of development, my own clinical experience suggests that the
approach is a promising one.
References
Beck, AT. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press.
Beck, AT., Rush, Al, Shaw, B.F., & Emery, G. (1979). Cognitive therapy of depression.
New York: Guilford Press.
276
Jeffrey E. Young
Beck, A T. & Young, IE. (1985). Depression. In D. Barlow (Ed.), Clinical handbook of
psychological disorders (pp. 206-244). New York: Guilford Press.
Guidano, V.F., & Liotti, G. (1983). Cognitive processes and emotional disorders. New
York: Guilford Press.
Hartman, L.M. (1983). A metacognitive model of social anxiety: Implications for
treatment. Clinical Psychology Review, 3, 435-456.
Lauter-Morgan, H. L. (1985). Loneliness in college students as a function of a social
network map. Unpublished doctoral dissertation, University of Pennsylvania,
Graduate School of Education, 1985.
Sullivan, H.S. (1953). The interpersonal theory of psychiatry. New York: Norton.
Weiss, R. (1973). Loneliness: The experience of emotional and social isolation. Cambridge: MIT Press.
Young, IE. (1981). Cognitive therapy and loneliness. In G. Emery, R. Bedrosian, & S.
Hollon (Eds.), New directions in cognitive therapy (pp. 139-159), New York: Guilford
Press.
Young, IE. (1982). Loneliness, depression and cognitive therapy: Theory and
application. In L.A Peplau & D. Perlman (Eds.), Loneliness: A sourcebook oJ current
theory, research and therapy (pp. 379-405). New York: Wiley.
Young, IE., & Beck, AT., (1982). Cognitive therapy: Clinical applications. In Al
Rush (Ed.), Short-term psychotherapies for depression (pp. 182-214). New York:
Guilford Press.
Author Index
Abelson, R.A., 117

Adelson, J., 86
Adler, T., 57
Ainsworth, R., 46
Alagna, S.W., 215, 216
Allport, G.W., 82, 168-169, 174, 194
Altman, I., 4,5,20,84, 102, 103, 105,
106, 122, 130, 131,230
Ames, c., 179
Ames, R., 179
Andrews, G., 221
Applegate, J.L., 129
Aran, G., 192
Archer, R.L., 93, 120
Aries, E.J., 82, 87
Aristotle, 63-64
Armor, D.J., 172
Aronson, E., 33
Asher, S., 1,52,227,242
Avery, A.W., 147, 148
Back, K., 118, 195,211,223
Bakeman, R., 120
Banikotes, P.G., 97
Barbarin, O.A., 83, 94
Barnes, J.A., 149
Barnes, R.D., 73
Bartrop, R.W., 218
Batten, P., 232
Baum, A., 210, 211, 215, 216, 221, 222,
223
Baxter, L.A., 141
Beatty, A.E., 162

Beck, A.T., 261, 262, 263
Becker, H.S., 191, 193
Behling, M., 170
Bell, R., 227
Bell-Nathaniel, A., 171
Bellack, A.S., 56
Belle, D.E., 149, 150
Bender, V.L., 84, 89
Benton, A.A., 86, 92
Berg, J.H., 2, 3, 4, 5, 93, 102, 103, 104,
109,111, 114, 116, 119, 120, 121,
122, 123
Berger, C.R., 131, 132, 137, 138
Berger J., 169
Berger, P., 235
Berkman, B., 218
Berlo, O.K., 131
Berlyne, D.E., 141
Bernard, H.R., 150
Berndt, T.J., 48, 51, 53, 54
Berscheid, E., I, 13, 29, 31, 34, 35, 36,
37, 102, 104, 111, 114, 116, 121,
122, 156,235-236,241
Bierman, K.L., 58
Bigelow, B.J., 53
Bikson, T., 228
Bjorn, L., 199
Blau, P., 200
Blau, R.M., 104
Blaylock, T., 114
Blodgett, T.B., 199
Bogdan, R., 186
278
Boissevain, J., 147, 148, 153
Bolton, C.D., 107, 108
Boorman, S., 147
Booth, A., 84, 87, 94, 160
Boss, D., 83
Bott, E., 237
Boyce, W.T., 220
Bradac, J.J., 132, 137, 138
Bradburn, N., 5, 217, 228
Braiker, H.B., 107, 108, 109, 122, 124,
162
Brain, R., 236, 237
Breiger, R., 147
Brook, R.H., 214
Brophy, J.E., 85, 92
Brown, B.B., 106
Brown, G.W., 220
Bucher, R., 188
Buhrmester, D., 2, 3, 45, 52, 54, 55, 57,
86
Bultena, G.L., 87
Burgess, R.L., 160
Burleson, B., 132
Burns, G.L., 150
Burowoy, M., 188
Busch, C., 233
Byrne, D., I, 11-12,30,34,36,235
Calabrese, R.J., 131
Caldwell, M.A., 82
Camarillo, J., 114
Campbell, B.H., 71
Campbell, J., 241
Cannon, W.B., 209, 210
Caplowitz, B., 217
Cassel, J.c., 211, 214, 218, 219
Cate, R., 233
Cerreto, M.C., 101, 160
Chaikin, A.L., 88, 103
Chapman, AJ., 1,54
Charlesworth, R., 56
Cheek, J., 233
Chelune, G.J., 131
Chesler, M.A., 83, 94
Chester, N.L., 91
Childs, L., 84, 90
Childs, M.K., 54
Chiriboga, D., 227, 237
Chodorow, N., 86
Author Index
Christensen, H., 31, 35, 102, 122, 156
Cicero, 64
Clark, M.S., 2, 3, 4, 5, 34, 102, 103,
104, 111-112, 113, 114, 115, 119,
121
Clore, G.L., 11-12, 30, 36
Cobb, S., 210, 211, 214, 219
Cohen, D.K., 172
Cohen, E.G., 169, 170, 177
Cohen, S., 214, 215, 216, 219, 220, 221,
222
Coleman, J., 191,201
Collins, B.E., 177
Collins, E.G.C., 197, 198
Collins, N., 202
Constantian, C.A., 4, 90, 93
Constanzo, P.R., 67
Cook, M., 150
Cook, T.D., 34
Coppotelli, H., 56
Corwin, R., 200
Costanza, R., 90
Cox, B., 195
Cox, V., 223
Coyne, J.C., 215
Crawford, A.C., 82
Cunningham, M., 202
Cutrona, C., 228, 231, 233
Dabbs, J.H., 120
Daly, J.A., 130
Damico, S.B., 171
Dandridge, T.C., 195
Darley, J.M., 111, 114
Darrow, C.N., 201
Davidson, L.R., 82, 84
Davies-Avery, A., 214
Davis, K.E., 2, 10, 20, 105, 106, 116,
119, 149, 152, 194,231,235,236
Davis, M., 138, 240
Davis, Y., 84, 89
de la Paz, D., 209
Deal, T.E., 192
Dean, A., 216, 219
DeAraujo, G., 219
Deaux, K., 91, 92
Delia, J., 141
Derlega, V.J., 4, 5, 83, 84, 88, 89, 93,
103, 132
279
Author Index
DeVault, R.C., 211
DeVellis, B.M., 215, 216
DeVellis, R.F., 215, 216
DeVries, D.L., 179
Diaz, R.M., 53
Dickens, W.J., 147, 150
Diogenes Laertius, 64
Dion, K.K., 121
Dodge, K.A., 56
Doise, W., 178
Domhoff, G.W., 192
Donald, C.A., 214
Douvan, E., 86
Duberman, L., 82, 84
Duck, S.W., 1,2,3, 10,30,36,130,
131, 138, 141
Dudley, D.L., 219
Duncan, B., 170
Dutton, D.G., 24
Dweck, C., 232
Eaton, W.W., 220
Eckenrode, J., 219
Eder, D., 85-86, 92, 93
Edwards, J.N., 160
Edwards, K.J., 179
Ehrlich, D., 188
Ehrlich, H.J., 174
Ellison, c., 239
Emery, G., 262
Ensel, W.M., 215, 220
Erdwins, C., 83, 88
Evans, M., 120
Farina, A., 150
Feger, H., 152
Fehr, 8.,3,9-40, 148
Felker, D., 179
Feshbach, N., 86
Festinger, L., 118,207,208,211,222,
223,235
Fincham, F., 59
Fine, G.A., 3, 191, 194
Fischer, C.S., 2,44, 147, 149, 150, 151,
229, 237, 238, 240, 241
Fischer, J.L., 84, 88
Fiske, S.T., 162
Fitt, L. W., 202
Fleming, R., 216, 221, 222, 223

Foa, E.B., 103, 234, 235
Foa, U.G., 103,234,235
Foot, H.C., 1,54
Foote, N.N., 199
Frank, R., 239
Frankenhaeuser, M., 210
Fredholm, E.M., 103
Freemon, J., 233
French, J.P.R., Jr., 223
French, R. de S., 232
Fried, N., 218
Fromm-Reichmann, F., 240
Frydman, M.L, 221
Fuehrer, A., 84, 90
Furman, W., 2, 3, 45, 52, 54, 55, 56,
57,58,86
Gangestad, S., 69, 70
Gannon, T.M., 191
Gans, H., 218, 238
Gatchel, R.J., 216, 221, 222, 223
Geer, 8., 193
Gerard, H.B., 117, 168
Gershoni, R., 53, 54, 86
Gerson, A., 228, 232, 239
Gerson, K., 147, 148, 149
Gessner, T., 83, 88
Gilligan, C., 87
Gilmour, R., 1
Gisriel, M.M., 216, 221, 222, 223
Giuliano, T., 104
Glass, D.C., 210, 211, 215
Glazer, J.A., 56
Glover, 0., 84, 89
Gluckman, M., 194
Goetz, T., 232
Goffman, E., 191,200
Goldbert, U., 214, 222
Golightly, c., 30
Gonzo, J., 56
Goodchilds, J., 228
Goode, W.J., 200
Goodstein, L.D., 161
Gordon, S., 235, 237
Gore, S., 214, 219, 223
Goswick, R., 233, 242
Gottlieb, B., 236
Gottman, J.M., 1,52,56,232
Author Index
280
Gough, H., 230

Granovetter, M., 149, 158, 192
Greenberg, M., 239
Gregory, K.L., 186
Griffitt, W., 25, 121
Gross, R., 83, 88
Gruber, J.E., 199
Grunberg, C., 218
Grunberg, N.E., 210, 215
Gruner, C., 193
Grzelak, J., 132
Guidano, V.F., 249
Guay, P., 25
Haas, J., 189, 190,200

Habif, V. L., 222
Hacker, H.M., 88
Hallinan, M.T., 3, 85-86, 92, 93, 171,
173, 175, 176
Halverson, C.F., 85, 92
Handelman, D., 193
Hansson, R., 232
Harburg, E., 18
Harlow, H., 234
Harris, T., 220
Hartman, L.M., 254
Hartog, I., 242
Hartup, W.W., 45, 56, 58
Harvey, J.H., 31, 35, 102, 122, 156
Harvey, 0.1., 178, 179
Hatfield, E., 13, 14, 104
Hauserman, N., 170
Haviland, J.B., 195
Hay, J., 19
Hays, R.B., 4, 5, 83, 89, 102, 103, 110,
Ill, 116, 118, 122, 124
Hay thorn, W.W., 106
Healy, S., 2, 90
Hebden, J.E., 186
Heider, F., 17
Heilbrun, A., 230
Heim, M., 232
Heller, K., 222
Helm, B., 233
Hershey, R., 194
Herson, M., 56
Heslin, R., 83
Hess, E., 94
Hewson, D.M., 221
Higgins, E.T., 131

Hill, C.T., 1,82
Hinde, R.A., 3, 30, 32, 91, 92
Hoberman, H., 221
Hoffman, J.E., 53, 54, 86
Holmes, J.G., 103
Holmes, T., 218, 219
Homans, G.C., 172, 180
Honeycutt, J.M., 130, 131
Hood, W.F., 178, 179
Hook, J.G., 34
Hopper, C.H., 120
Hops, H., 57
Horowitz, L., 232
House, J.S., 223
Hoyt, M.F., 177
Huckfeldt, R.R., 157
Hughes, E.C., 189
Hunter, F.T., 53
Hunter, S., 93
Huston, T.L., 10,25,31,35, 102, 122,
147, 150, 152, 156, 160,237
Hyman, H.H., 174
Ickes, W., 65, 66, 67, 73

Ingham, J.G., 214, 222
Jackson, R.M., 147, 148, 149

Jacobs, L., 235-236
Jamieson, D.W., 68
Johnson, D., 160
Johnson, F.L., 82, 87
Johnson, M.P., 25, 150, 152, 160, 237
Jones, L.M., 147, 148, 149, 151
Jones, M.O., 188
Jones, W., 229, 230, 232, 233, 242
Jong-Gierveld, J. de, 230
Jonsson, E., 211
Jorgenson, B.W., 107
Kahn, A., 33, 228

Kandel, D.B., 152
Kantor, R.M., 202
Kaplan, B.H., 214, 219
Kasl, S.V., 210
Katz, 1., 174
Author Index
Keiser, GJ., 106
Kelley, H.H., 3, 5,13,24,31,35,44,
102, 103, 107, 108, 109, 120, 122,
124, 156, 162, 239
Kennedy, A.A., 192
Kerckhoff, A.C., 20, 105, 106,
195
Kernis, M., 150
Kiefer, c., 228, 238
Killworth, P.O., 150
Ki1oh, L.G., 218
Klassen, 90
Klaus, R.A., 56
Klein, E.B., 201
Kleinman, S., 191
Klinger, E., 117
Koester, N., 231
Komarovsky, M., 238
Kram, K.E., 201
Kraus, A., 218
Krause, S., 2, 90
Kreyburg, P.C., 57
Krile, D., 56
Krulewitz, 33
Kuo, W., 220
Kurdek, L.A., 56
Kurth, S.B., 187, 197,236
Ladd, G.W., 56-57
LaGaipa, J.J., 53, 54, 57, 230, 233, 235,
236
Lahey, B.B., 222
Lamm, H., 34
Langer, E.J., 211
Langer, T., 218
Laosa, L.M., 85,92
LaRocco, J.M., 223
Larson, R.W., 152
Lavine, L.O., 88
Lazarus, L., 218
Lazarus, R.S., 210, 215
Leary, J.P., 193
Leavy, R.L.,148
Leemon, T.A., 192
Leighton, A., 218
Lembright, M.F., 201
Lepkowsky, c., 91
Lerner, M.J., 103
Leslie, L., 150, 237
281
Lever, J., 86, 92, 96
Levin, K., 145
Levinger, G., 4, 20-23, 24, 30, 31, 35,
37,82,83,89-90, 101, 102, 104,
105,106, 107, 122, 147, 156
Levinson, D., 201
Levinson, M., 201
Levitt, E., 90
LeVoi, M.E., 129
Lewis, R.A., 20, 89
Lewis, R.G., 170, 174
Light, J.M., 199,200
LilienfeId, A., 218
Lin, N., 215, 216, 219, 220
Lincoln, J.R., 200
Lindenthal, J.J., 220
Linder, D., 33
Lindzey, G., 82
Liotti, G., 249
Lippitt, R., 56
Lipset, S.M., 191,201
Little, R., 191
Lombardo, J.P., 88
Lopata, H., 227
Lorenz, K., 234
Losoff, M., 86,90
Lott, A.J., 11, 12
Lott, B.E., 11, 12
Loucks, S., 231
Lounsbury, J.W., 211
Lowenthal, M.F., 81, 84, 227, 230,
237
Luckhurst, E., 218
Luckmann, T., 235
Lydon, J.E., 68
Lynch, J., 228
MacDonald, M.G., 101
Madden, N., 180
Mandler, G., 141
Mannarino, A.P., 57
Margulis, S.T., 5
Maroldo, G., 232
Marshall, N., 230
Martin, J., 186, 192
Mason, J.W., 210
Masters, J.C., 56
Masuda, M., 218
McAdams, D.P., 2, 4, 86, 90, 93
Author Index
282
McCain, G., 223
McCarl, R.S., Jr., 192
McCarty, C., 150
McClintock, E., 31, 35, 102, 122, 156
McCormick, S., 228
McGuire, K.D., 57
McKay, G., 214, 215, 216, 219, 220,
221, 222
McKee, B., 201
Meade, M., 235
Mechanic, D., 200
Mechling, E., 194
Mechling, J., 194
Medalie, J.H., 214, 222
Michael, S., 218
Miell, 2, 3, 129, 138, 141, 142
Milardo, R.M., 3, 25, 147, 149, 150,
151, 152, 155, 159, 160,237
Milberg, S., 103
Miller, G., 139
Miller, J., 200
Miller, L.c., 93, 103, 104, 116, 119
Miller, N., 168
Mills, J., 4, 34, 103, 104, 111-112, 113
Mitchell, J.C., 147, 148
Moore, D., 234, 239
Moore, S.G., 56
Moore, S.G., 56
Morgan, B.S., 84, 87
Morgan, H.L., 252
Morton, T.L., 103, 106, 130, 131
Moustakas, C., 240
Murray, R.B., 26, 35
Murstein, B.I., 10,20,30, 101, 105,
106
Myers, H., 197
Myers, J.K., 220, 238
Narus, L.R., 84, 88

Neimeyer, G.J., 91
Nelson, J.I., 173
Newcomb, T.M., I, 17-19,35,36,81,
121, 145
Newman, 0., 211, 212
Newton, D.A., 202
Nezlek, J., 4, 55, 82, 83, 89, 150, 152
Nichols, J., 89
Niemeijer, R., 147
Nisbett, R.E., 92
Nuckolls, K.B., 219

Nusbaum, P., 193
O'Leary, V.E., 33
Oden, S., 52, 56-57, 242
Oliker, S.J., 149
Ouellette, R., 103
Packard, V., 200
Paine, R., 195
Paloutzian, R., 239
Parish, E., 232
Parsons, J.E., 47
Patchen, M., 169, 171
Paulus, P., 223
Peele, E., 211
Penny, R., 218
Peplau, L.A., 31, 35, 102, 122, 156,
228,229,231,239
Pepper, M.P., 220
Perkowitz, W.T., 119, 120
Perlman, D., 3, 30, 36, 147, 148, 150,
228, 229, 231, 232, 239
Perri, M., 150
Peters, TJ., 186, 192, 195
Peterson, D.R., 31, 35, 102, 122, 156
Pettigrew, T.F., 169, 170, 172
Pfeffer, J., 192
Phillips, S.L., 150
Planalp, D., 130, 131
Plato, 63-64
Ponzetti, J., 233
Powers, E.A., 87
Powers, M.E., 192
Preno, B.E., 103, 122
Price, K.O., 18
Purvis, J.A., 120
Putallaz, M., 232
Quinn, R.E., 197, 198
Rahe, D.F., 58
Rands, M., 82, 83, 89
Rasmussen, B., 56
Read, S.H., 117
Reimer, J.W., 201
Author Index
Reis,H.T., 55, 83, 84, 85, 89, 93, 150,
152
Rennert, K., 211
Richey, H.W., 86
Richey, M.H., 86
Ridley, e.A., 147, 148
Riemer, J., 191
Riley, R.T., 172
Riordan, C., 169
Robbins, P., 55
Roberts, B., 238
Rodin, J., 211
Roe, A., 240
Rohlen, T.P., 196
Rokeach, M., 82
Rook, K., 228, 241
Ross, J., 47
Roy, D., 189, 193
Rubenstein, C., 234, 239, 240, 241
Rubin, K.H., 56
Rubin, Z., 1, 30, 83, 85, 101, 107
Ruble, D.N., 47
Ruggiero, J., 169
Rusbult, e.E., 13-14, 15, 36, 121
Rush, A.J., 262
Russell, D., 228, 231, 232
Russell, S.W., 161
Sabshin, M., 188
Sagar, H.A., 169, 171
Salzinger, L.L., 147, 153, 154, 158,
159, 161, 162
Sansone, C., 233
Saulnier, K., 148
Scanzoni, J., 24
Schachter, S., 118,208,211,213,223,
235
Schaefer, e., 215
Schatzman, N., 188
Schill, T., 84
Schmidt, N., 231
Schofield, J.W., 169,171, 176
Schultz, N., 234, 239
Schutte, J.G., 199,200
Schwartz, J.e., 234
Segal, B., 218
Selman, R.L., 47
Selye, H., 209
Semyonov, E., 169
283
Semyonov, M., 169
Senchak, M., 84, 85
Senn, D.Y., 107
Sermat, V., 231, 240
Sharabany, R., 53, 54, 86
Sharan, S., 171, 177
Shaver, P., 55, 234, 239, 240, 241
Shaw, B.F., 262
Shaw, M.E., 67,168
Sherif, e.W., 178, 179
Sherif, M., 178, 179
Shibutani, T., 194
Sholar, W., 114
Shulman, N., 149, 152, 159
Siehl, C., 186
Simone, R.S., 215, 220
Simpson, J.A., 69, 70, 72
Singer, J.E., 210, 211, 215, 235
Slater, P., 228
Slavin, R.E., 171, 177, 178, 179, 180
Sloan, W., 232
Small, A., 83, 88
Smith, D., 2, 67, 72
Smith, D.L., 191, 193
Smith, G.H., 56
Smith, J., 1,2,54
Smith, M.S., 172
Smith, S.S., 173
Snoek, J .E., 4, 20-22, 32, 101, 105, 122
Snyder, M., 2, 65-72
Snyder, N., 117
Sokol, R., 218
Solano, e.H., 2, 5, 84, 228, 229, 230,
231,232
Solomon, B., 84, 85
Solomon, S.K., 211
Sones, G., 86
Sorensen, A.B., 175, 176
Sorenson, S., 21 I
Sorrentino, R., 106
Sparks, P., 195
Spiegel, N., 150
Spinner, B., 228, 239
Spitz, R., 234
Sprecher, S., 19, 104
St. John, N.H., 170, 172, 174
Stapp, J., 89
Steck, L., 114
Steinberg, M., 139
Stelling, J., 188
284
Steuve, C.A., 147, 148, 149
Stewart, A.J., 91
Stiles, W.B., 103, 122
Stodolsky, S., 177
Stokes, J.P., 84, 90, 103
Strauss, A., 188, 193
Stull, D.E., 82
SuedfeId, P., 228, 233
Sullivan, H.S., 41-62, 250
Sundby, H.S., 57
Sundstrom, E., 211
Suomi, S., 234
Surra, c., 3
Suttles, G., 186
Swann, W.B., 117
Swanson, c., 188, 193
Tajfel, H., 170

Tanke, E.D., 117
Taves, M.J., 200
Taylor, D.A., 4, 5, 20, 84, 102, 103,
105,106, 122, 130, 131
Taylor, S., 186
Tennant, c., 221
Tesch, S.A., 150
Tesser, A., 241
Thibaut, J.W., 5,13,24,44,239
Thoits, P.A., 148, 216
Thurnher, M., 227, 237
Timms, D.W.G., 199
Todd, M.J., 2, 10, 116,231,235,236
Tolstedt, B.E., 103
Tornblom, K.Y., 103
Traupmann, J., 13, 15
Trivers, R., 234
Trow, M., 191,201
Turner, J.c., 170, 179
University of Minnesota, 199

Utne, M., 13
Uwanna, R., 120
Vaillant, G.E., 221

Valins, S., 211, 223
Van Arsdel, P.P., 219
Van Hasselt, V.B., 56
Vaught, C., 191, 193
Vaux, A., 84
Author Index
Verbrugge, L.M., 152, 191
Vernon, P.E., 82
Vinsel, A., 106
Waldrop, M.F., 85, 92

Wallston, B.S., 215, 216
Walster, E., I, 13, 14, 15,34, 102, 104,
121, 235-236, 241
Walster, G.W., 13, 14, 15,34, 102, 104,
121, 241
Ware, J.E., Jr., 214
Waterman, R.H., Jr., 186, 192, 195
Wegner, D.M., 104
Weinstein, E., 44
Weiss, E., 218
Weiss, L., 81, 84
Weiss, R., 46, 160,236,237,241,250
Weisz, J.R., 57
Wellman, B., 147, 152, 154, 158
West, S.G., 30, 36
Whalen, S.R., 170
Wheeler, L., 4, 55, 82, 83, 89,150,152
White, B.J., 178, 179
Wicklund, R.A., 30, 36
Williams, J., 228, 229, 230
Wilmot, W.W., 141
Wilmott, P., 238
Wilson, M., 103
Wilson, T., 92
Winstead, B.A., 4, 5, 83, 84, 89, 93
Wolf, K., 234
Won-Doornick, M., 103
Wong, P.T.P., 4, 84, 89, 93
Wood, H., 230
Worchel, S., 178
Wright, H.F., 52
Wright, P.H., 2, 81, 82, 192
Yancey, W.L., 211, 212, 223

Young, J.E., 242, 262, 263, 268, 271
Young, M., 238
Youniss, J., 45, 47, 48, 53
Yuchtman-Yaar, E., 184
Zadny, J., 117

Zanna, M.P., 68
Zelditch, M., 169
Zuckerman, H., 201
Subject Index
ABCDE model, 22-23

Ability groups, 175-178, 180-181
Abusive relationships, 24-25
Acceptance, need for, 47-48
Accessibility, 118
Adolescence, 51
Adult participation, need for, 46
Affect, restriction of, 259-260
Affiliation, 208, 213
Aftereffect, 210
Alienation, 257, 272
Amancita, 64
Androgyny, 88-89
Anxiety
induction, 4-5
maternal, 46
Apathia (detachment), 64
Appraisal, 210, 221
Architectural design, 211-212
Assumptions, 261-262
Athletics, 179-180
Attitude, 17
intensity of, 17
similarity, 30
Attraction, 4, 17
antecedents to, 121
evaluation of theories, 29-37
intensity of, 17
theories, 24-29
Attractiveness, 254-255, 270-271
Autarky (self-sufficiency), 64
Automatic thought, 261
Autonomy, 189, 190-192
Availability, 118
Balance, 16-19,26,31-32,34
Belonging, 221
Bereavement, 218
Berle Psychosocial Assets Index, 219220
Body image, 254, 270-271
Boredom, 239
Buffering hypothesis, 216-217, 219-222
Careers, 239
Ceremonies, 192, 195-196
Chumships, 48-51
Classical conditioning, 11-12,25, 30
Close relationships, 102
Closeness, 102
Cluster, definition of, 154
Cognitive consistency, 16-19,253
need for, 34
Cognitive needs, 234-235
Cognitive organization, 17
Cognitive therapy for friendship
disorders, 261-275
levels of cognition, 261
Collaborations, 48-51
Commitment, 191
Communal relationships, 4, 33, 111116, 123
Communication, purposive, 130-131
Comparison level for alternatives, 13,
240
Compatibility, 68
Competition, 178-180
Complementarity, 73
286
Concordia, 64
Consensual validation, 49
Constriction, 258
Contact theory, 168-169, 181
Content of work, 190-192
Control, 189,211
Conversational esteem, 256, 271
Conversational responsiveness, 120
Cooperation, 179
Coping, 210, 241-242
nonsocial, 241
Countercultures, 186
Courtship, 107
content dimensions, 107
development, 20
Creativity, 240
Crime rates, 211-212
Cultural stereotypes, 190
Decision making, 107-116
Dependency, 13
Desegregation, 167-184
Detachment, 64
Developmental arrests, 45
Developmental theories, 32
Differential Loneliness Scale,
231-232
Differentiation, 107-116
Direct-action coping, 210
Disliking, 18
Disparaging child, 50
Dissonance reduction, 222
Distributive justice norms, 33
Drugs, 242
Egalitarian exchange relationship,

47
Emotional needs, 235-236
Emotional support, 213
Entrapment, fear of, 260, 274
Epochs, 42-43
Equity theory, 14-16, 19,31,33
Exchange relationships, 4, 33, 111-116,
123
Exchange theory, 12-16,31, 102
Exclusivity, 3
Expectations state theory, 169
Extra-curricular activities,
179-180
Subject Index
Family, 237-238
Fear induction, 4-5
Fear of entrapment, 260, 274
Femininity, 89
Fight or flight response, 209
Filter theories, 106-107
Focal emotions, 46
Friendlessness, 5, 227-246
Friends
alternatives to, 236-241
cognitive needs, 234-235
cost of, 240-241
emotional needs, 235-236
functions of, 233-236
material needs, 233-234
as need fulfillers, 51-55
nonsocial alternatives, 238-241
as source of stimulation, 234
therapeutic interventions, 241-242
Friendship
activity-oriented, 63-80
affect-oriented, 63-80
applications of theory and research, 5
assessment, 264-268
balance theory, 16-19
basis of, 67
behavioral manifestations, 69-71
characteristics of, 2
childhood, 41-62
classical concepts, 63-64
classification of disorders, 252-260
cognitive-developmental approach,
247-252
conceptions of, 67-69
and culture, 185-187
degree of mutual disclosure, 251
deterioration, 22-23
developmental theories, 19-23
disorders, 247-276
dissolution, 76, 77-78
expectations for, 13
formation, 211-213
frequency of contact, 251
and growth of social competence,
55-57
history and expected duration, 251
homogenity,69-71
impact of third parties on, 25-26
initiation, 75-76
interracial, 167-184
investment model, 13-14
Subject Index
levels of, 249-253
longitudinal research, 4
maintenance, 76-77
methodological issues, 3-5
nature, of, 73-78
organizational, 192-196
and organizational hierarchy,
199-202
orientations, 67-69
paradox of, 78-79
and personality, 63-80
platonic, 94-95
platonic cross-sex, 197-199
for the sake of pleasure, 64
primary, 63-64
problems deepening, 258-260
problems initiating, 252-256
same-sex, 81-99
scientific approach, to, 65-67
segmentation, 69-71
selection, 5, 260, 274
sex differences in, 81-99
shared interests, 251
stages, 266-268
strategies in developing, 129-143
theoretical issues, 2-3, 9-40
therapeutic benefits, 57-58
true, 63
for the sake of utility, 64
work, 191-192
worlds, 71-73
Gain/loss effect, 33
Gender stereotypes, 93
General adaptation syndrome, 209-210
Gossip, 194-195
Group membership, 215
Gullibility, 193
287
Information gathering, 131-132
about partner, 137-139
Informational support, 214
Initiation rites, 193-194
Institutionalization, 218
Instructional grouping, 175-178, 180181
Interaction disability treatment, 169170
Interdependence, 5,13,22,31,105-106
structural, 156-162
Intermittent positive reinforcement, 24
Interpersonal attraction
analysis of, 9-40
classical conditioning, II
reinforcement-affect model, 11-12
reinforcement theories, 11-12
Interpersonal competencies, 44
Interpersonal situation, 41-42
Interracial friendships, 167-184
Intimacy, 51, 84
cross-sex, 54-55
degrees of, 23
growth, 131
level of, 139
motivation, 90-91
Intimate exchange, need for, 48
Intragroup cooperation, 179
Inventory of Socially Supportive
Behaviors (lSSB), 221
Investment model, 13-14
Isolated child, 50
ISSB, see Inventory of Socially
Supportive Behaviors
Joint relevance, 34
Joking, 193-194
Kin, 237-238
Harassment, 194, 198-199
Hatred, 187-188
Helping, 233-234
Helplessness, 46
Heterosexual relationships, 32
Homophobia, 89-90, 94, 95
Impersonal resources, 239

Information exchange, 132-141
Level of awareness, 141-142

Life cycle issues, 36-37
Life events, 219-220
Likability, 69-71
Loneliness, 5, 227-246
definition, 229-230
preadolescent, 46
scales, 231-232
288
Loneliness (cont.)
self-report studies, 232
sex differences, 93
verbal behaviors, 232
Loss of control, 240
Love relationships, 2-3
in the workplace, 187-188
Low close esteem, 256-257, 271-272
Low-assertiveness, 259, 273
Malevolent child, 50
Malevolent transformation, 45, 47
Marriage, 237-238
Mate selection, 106-107
Material needs, 233-234
Mechanical solidarity, 193
Mentoring, 200-202
Metacognition, 254
Mistrust, 257, 272
Mutual investment, 105
Mutuality, 22, 32, 105
Need fulfillment, 42, 51-55

Needs, 42
for acceptance, 47-48
for adult participation, 46
for intimate exchange, 48
for tenderness, 46
Network analysis, 145-166
Nonbalanced systems, 18
Nonspecific response to injury, 209
Nonverbal behavior, 20, 83
Norms, adherence to, 120-121
Nuturance, 68
Organizational culture, 192-196

Organizational romance, 197-202
Organizational subcultures, 186
Ostracized child, 50
Pacing, 258, 272

Pair relatedness, 20-22
Palliative coping, 210
Panic attack symptoms, 253
Passive contact, 212-213
Peer orientation, 85-88
Subject Index
Peer workgroups, 177
Peers, 47-48
popularity, 56-57
therapeutic benefits, 57-58
Personal choice, 145-166
Personality, 63-80
Perspective-taking skills, 56
Physical appearance, 254-255, 270-271
Physical proximity, 118, 197
Preadolescence, 48
Purposive communication, 130-131
Racial composition of classroom, 170171, 172-175, 180

Reinforcement, II-12, 26, 30-31, 33
intermittent positive, 24
negative, 25
Relatedness, 20-22
Relationship development
ABCDE model, 22-23
stages of, 20
theories, 19-23
Relationships
breadth of, 29, 35-37
casual, 101-128
chronically conflicted, 24-25
close, 101-128
decisions about closeness of,
117-121
decisions about future of, 137
decisions about nature of, 107-116
differentiation, 107-116
distressed, 24-25
egalitarian exchange, 47
predictions, 109-110
Religion, 239
Responsiveness, 119-120
Restriction of affect, 259-260, 273-274
Reverse small world procedure, 150
Reward structure, 178-180
Role playing, 177-178
Romantic partners, 237-238
Romantic triangle, 34
Rumor, 194-195
Safety, 234
Same-sex friends, 81-99
intrasex differences, 88-91
289
Subject Index
Schemas, 247-249, 261
close, 249
social,249
School organization and interracial
friendships, 167-184
Scripts, 117
for appropriate behavior, 133-141
Selective resources, 103
Self-disclosure, 4, 20, 22, 48, 87, 88-89,
93
fear of, 258, 271-272
flexibility in, 131
instrumental analysis of, 132
levels of, 101-128
patterns, 161-162
Self-esteem, 57, 217, 221, 235-236
Self-monitoring, 63-80
Self-Monitoring Scale, 66--67
Self-presentation, 92, 93
Self-sufficiency, 64
Self-worth, 48
Selfishness, 64
Sex differences in same-sex friends,
81-99
developmental perspective, 85-88
Sex roles, 95
Sexual harassment, 194, 198-199
Sexuality, 51
Similarity, 73, 81-82
Social anxiety, 253-254, 256, 269-270
Social comparison, 44, 48, 208, 217,
222
Social competence, 44, 222
growth of, 55-57
Social constraint, 145-166
Social exchange, 101-128
casual relationships, 102-104
closeness, 102-103
communication, 103
norms, 103
quality of resources, 103
transition from casual to close
relationship, 105-107
Social interaction, 229-230
Social interdependence theory, 5, 13
Social loneliness, 236--237
Social needs, 42
Social networks, 145-166
clique structure, 152-155
collective influence of, 157
density, 152-155
interactional attributes, 146
interconnectedness, 152-155
of married individuals, 159
overlap, 155
size, 149-152
structural attributes, 146, 149-163
Social penetration, 5, 20, 106
Social skills, 255-256, 271
Social status, 235-236
of students, 174, 177
Social support, 207-226
assets-benefits hypothesis, 217-218
buffering hypothesis, 216--217, 219222
emotional support, 213, 215
and friendship, 213-222
function of, 214-215
informational support, 214
instrumental,215
measurement, 215-217
sources of, 214-215
tangible support, 213-214
Social-personality development,
41-51
Socialization, 190
Space, 211-213
Spectatoring, 270
Sports participation, 196
Spousal abuse, 24
Spouses, 159, 237-238
Status characteristics, \99-200
Stereotypes, 93, 95
cultural, 190
Stories, 192
Strategic planning, 141-142
Stress attitudes about, 211
Stress, 83, 207-226
attitudes about, 211
physiological response to, 209-211
Structural interdependence, 156--162
Sullivan, H.S., 41-62
Superego, 261
Surface contact, 22, 105
Symptom reporting, 220-221
Taboo topics, 140

Tangible support, 213-214
Television, 239
290
Subject Index
Unilateral awareness, 21-22, 105
Unrealistic expectations, 258-259, 273
Tenderness, need for, 46

Three Mile Island, 221, 222
Topic changing, 139-140
Trust, 189
Tutoring, 177-178
Vandalism, 211-212
UCLA Loneliness Scale, 231-232

Unattractiveness, 254-255, 270-271
Unbalanced systems, 18
Underlying assumption, 261
Work relations, 187-188

Workplace, 185-206
informal organization of, 188
as a setting, 188-190
Springer Series in Social Psychology

Attention and Self-Regulation: A Control-Theory Approach to Human Behavior
Charles S. Carver/Michael F. Scheier
Gender and Nonverbal Behavior
Clara Mayo/Nancy M. Henley (Editors)
Personality, Roles, and Social Behavior
William Ickes/Eric S. Knowles (Editors)
Toward Transformation in Social Knowledge
Kenneth J. Gergen
The Ethics of Social Research: Surveys and Experiments
Joan E. Sieber (Editor)
The Ethics of Social Research: Fieldwork, Regulation, and Publication
Joan E. Sieber (Editor)
Anger and Aggression: An Essay on Emotion
James R. Averill
The Social Psychology of Creativity
Teresa M. Amabile
Sports Violence
Jeffrey H. Goldstein (Editor)
Nonverbal Behavior: A Functional Perspective
Miles L. Patterson
Basic Group Processes
Paul B. Paulus (Editor)
Attitudinal Judgment
J. Richard Eiser (Editor)
Social Psychology of Aggression: From Individual Behavior to Social Interaction
Amelie Mummendey (Editor)
Directions in Soviet Social Psychology
Lloyd H. Strickland (Editor)
Sociophysiology
William M. Waid (Editor)
Compatible and Incompatible Relationships
William Ickes (Editor)
Facet Theory: Approaches to Social Research
David Canter (Editor)
Action Control: From Cognition to Behavior
Julius Kuhl! Jiirgen Beckmann (Editors)
Springer Series in Social Psychology
The Social Construction of the Person

Kenneth J. Gergen/Keith E. Davis (Editors)
Entrapment in Escalating Conflicts: A Social Psychological Analysis
Joel Brocknerl1effrey Z. Rubin
The Attribution of Blame: Causality, Responsibility, and Blameworthiness
Kelly G. Shaver
Language and Social Situations
Joseph P. Forgas (Editor)
Power, Dominance, and Nonverbal Behavior
Steve L. Ellysonl1ohn F. Dovidio (Editors)
Changing Conceptions of Crowd Mind and Behavior
Carl F. Graumann/Serge Moscovici (Editors)
Changing Conceptions of Leadership
Carl F. Graumann/Serge Moscovici (Editors)
Friendship and Social Interaction
Valerian J. Derlega/Barbara A. Winstead (Editors)
An Attributional Theory of Motivation and Emotion
Bernard Weiner
Public Self and Private Self
Roy F. Baumeister (Editor)

(Springer Series in Social Psychology) Barbara a. Winstead, Valerian J. Derlega (Auth.), Valerian J. Derlega, Barbara a. Winstead (Eds.)-Friendship and Social Interaction-Springer-Verlag New York (198 (1)

Hochgeladen von

Dokumentinformationen

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

(Springer Series in Social Psychology) Barbara a. Winstead, Valerian J. Derlega (Auth.), Valerian J. Derlega, Barbara a. Winstead (Eds.)-Friendship and Social Interaction-Springer-Verlag New York (198 (1)

Hochgeladen von

Copyright:

Verfügbare Formate

SSSP

To Wally and Stella Derlega, Bob and Lois Winstead,

studying personal relationships owe appreciation to Steve Duck and Robin

1. Friendship and Social Interaction: An Introduction .............

Theoretical Issues in the Study of Friendship ... . . . . . . . . . . . . . .

2. Theories of Friendship: The Analysis of Interpersonal Attraction. . . .

Reinforcement Theories ....................................

3. The Changing Functions of Friends in Childhood:

Sullivan's Theory of Social-Personality Development. . . . . . . . . .

4. Personality and Friendship: The Friendship Worlds of

A Scientific Approach to the Study of Friendship .............

Behavioral Manifestations of Friendship .....................

5. Sex Differences in Same-Sex Friendships .....................

Sex Differences in Adult Same-Sex Friendships. . . . . . . . . . . . . . .

6. Differences in Social Exchange Between Intimate and Other

Some Differences in Exchange Between Close and Casual

7. Strategies in Developing Friendships ..........................

Dorothy Miell and Steve Duck

8. Personal Choice and Social Constraint in Close Relationships:

Conceptualizing Social Networks. . . . . . . . . . . . . . . . . . . . . . . . . . ..

9. School Organization and Interracial Friendships ...............

Theoretical and Empirical Research on Interracial

Organizational Characteristics of Schools and Interracial

Culture and Friendship. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 185

12. People Without Friends: Loneliness and Its Alternatives .........

13. A Cognitive-Behavioral Approach to Friendship Disorders .......

Cognitive-Developmental View of Friendships. . . . . . . . . . . . . . ..

Author Index ................................................... 277

Andrew Baum, Department of Medical Psychology, School of Medicine,

Dorothy Miell, Psychology Discipline, Faculty of Social Sciences, The Open

University, Walton Hall, Milton Keynes MK7 6AA, England

Robert M Milardo, Department of Child Development and Family Relation-

ships, School of Human Development, University of Maine at Orono,

Daniel Perlman, Division of Family Sciences, University of British

Columbia, Vancouver, British Columbia, Canada V6T lW5

Dave Smith, Department of Psychology, University of Minnesota, Min-

neapolis, Minnesota 55455, U.S.A.

Mark Snyder, Department of Psychology, University of Minnesota, Min-

neapolis, Minnesota 55455, U.s.A.

Cecilia H Solano, Department of Psychology, Wake Forest University,

Winston-Salem, North Carolina 27109, U.S.A.

Barbara A. Winstead, Department of Psychology, Old Dominion University,

Norfolk, Virginia 23508, U.S.A.

Jeffrey E. Young, Department of Psychiatry, College of Physicians and

Surgeons, Columbia University, New York, New York 10027, U.S.A.

Friendship and Social

Barbara A Winstead and Valerian J. Dedega

In the introduction we address ourselves to three aspects of the study of

Theoretical Issues in the Study of Friendship

Friendship and Social Interaction: An Introduction

involvement. Friendships are free of the expectation of exclusivity, but also

Methodological Issues in the Study of Friendship

Barbara A Winstead and Valerian J. Derlega

Friendship and Social Interaction: An Introduction

Applications of Friendship Theory and Research

Barbara A Winstead and Valerian I Derlega

Friendship and Social Interaction: An Introduction

Theories of Friendship: The Analysis

Daniel Perlman and Beverley Fehr

The purpose of this chapter is to present an overview of theories of

Daniel Perlman and Beverley Fehr

Thus, again through the process of classical conditioning, we come to like

Exchange and Equity

major exchange theory is Thibaut and Kelley's interdependence theory