Birchmeier - 2011 - Strategic Uses of Social Technology An Interactive Perspective of Social Psychology

Strategic Uses of Social Technology
An Interactive Perspective of Social Psychology
On an everyday basis, we communicate with one another using var-

ious technological media, such as text messaging, social networking
tools, and electronic mail, in work, educational, and personal settings.
As a consequence of the increasing frequency of use and importance
of computer-supported interaction, social scientists in particular have
heeded the call to understand the social processes involved in such inter-
actions. In this volume, the editors explore how aspects of a situation
interact with characteristics of a person to help explain our technolog-
ically supported social interactions. The person-by-situation interac-
tion perspective recognizes the powerful role of the situation and social
forces on behavior, thought, and emotion, but also acknowledges the
importance of person variables in explaining social interaction, includ-
ing power and gender, social influence, truth and deception, ostracism,
and leadership. This important study is of great relevance to modern
readers, who are more and more frequently using technology to com-
municate with one another.
Z a c h a r y B i r c h m e i e r has a Ph.D. in Psychology from Miami Uni-

versity. He has published scientific studies about online social relation-
ships, including how students who are shy can build confidence online.
B e t h D i e t z - U h l e r is Professor of Psychology at Miami Univer-

sity. Her research interests include intragroup and intergroup behavior,
reactions to threats to social identity, sport fans and spectators, and
computer-mediated communication.
G a r o l d S t a s s e r is Professor of Psychology at Miami University. His

published work has appeared in Psychological Review, Review of Personal-
ity and Social Psychology, Psychological Inquiry, Journal of Personality and
Social Psychology, Journal of Experimental Social Psychology, and Organi-
zational Behavior and Human Decision Processes. His research interests
are communication in decision-making groups, coordination of social
behavior, collective choice and problem solving, and computational
models of social interaction.
Strategic Uses of Social
Technology
An Interactive Perspective of Social Psychology
Edited by
Zachary Birchmeier, Beth Dietz-Uhler, and
Garold Stasser
c a m b r i d g e u n i ve r s i t y p r e s s
Cambridge, New York, Melbourne, Madrid, Cape Town,
Singapore, São Paulo, Delhi, Tokyo, Mexico City
Cambridge University Press
The Edinburgh Building, Cambridge CB2 8RU, UK
Published in the United States of America by Cambridge University Press,
New York
www.cambridge.org
Information on this title: www.cambridge.org/9780521899260

C Cambridge University Press 2011
This publication is in copyright. Subject to statutory exception

and to the provisions of relevant collective licensing agreements,
no reproduction of any part may take place without the written
permission of Cambridge University Press.
First published 2011
Printed in the United Kingdom at the University Press, Cambridge
A catalogue record for this publication is available from the British Library
Library of Congress Cataloging in Publication data

Strategic uses of social technology : an interactive perspective of social
psychology / edited by Zachary Birchmeier, Beth Dietz-Uhler, and
Garold Stasser.
p. cm.
Includes bibliographical references and index.
ISBN 978-0-521-89926-0 (hardback)
1. Social networks – Psychological aspects. 2. Social interaction.
I. Birchmeier, Zachary. II. Dietz-Uhler, Beth. III. Stasser, Garold.
HM741.S78 2011
302.30285 – dc23 2011023025
ISBN 978-0-521-89926-0 Hardback
Cambridge University Press has no responsibility for the persistence or

accuracy of URLs for external or third-party internet websites referred to
in this publication, and does not guarantee that any content on such
websites is, or will remain, accurate or appropriate.
Contents
List of figures page vii

List of tables viii
Notes on the contributors ix
Acknowledgments xii
1 Introduction: A social psychological analysis of

computer-supported social interaction 1
zachar y birchmeier, beth dietz-uhler, and
garold stasser
2 A SIDE look at computer-mediated interaction: Power
and the gender divide 16
russell spears, mar tin lea, tom postmes,
a n d a n k a wo l b e r t
3 Trust, deception, and identity on the internet 40
melanie c. g reen and jordan m. car penter
4 An interactional approach to social influence in
computer-mediated communication 63
kai sassenberg
5 Social interaction in cyberspace: Social construction
with few constraints 84
susanne abele
6 Dynamics of leader emergence in online groups 108
a n d r e a b. h o l l i n g s h e a d
7 Ostracism in cyberspace: Being ignored and excluded in
electronic-based interactions 127
e r i c d . we s s e l m a n n a n d k i p l i n g d . w i l l i a m s
v
vi Contents
8 Opinion-based groups: (Racist) talk and (collective)

action on the internet 145
c r a i g m c g a r t y, g i r i s h l a l a ,
and karen m. douglas
9 A juxtaposition of social influences: Web 2.0 and the
interaction of mass, interpersonal, and peer sources
online 172
j o s e p h b. w a l t h e r , s t e p h a n i e t o m t o n g ,
d av i d c . d e a n d r e a , c a l e b t . c a r r , a n d
b r a n d o n va n d e r h e i d e
10 The virtual social world: The continually changing
landscape of social interaction 195
garold stasser, beth dietz-uhler, and
zachar y birchmeier
Index 211
Figures
5.1 The Prisoner’s Dilemma Game page 90

5.2 The Battle of the Sexes Game 91
6.1 Mean leader consensus scores by media and time period in
ten FTF and ten CMC groups 118
6.2 Number of leaders by group size in twenty-four internet
support groups 119
6.3 Leader consistency in twenty-four internet support groups
by consecutive time periods 121
9.1 A typology of online influence sources 179
vii
Tables
4.1 Examples for moderators of the impact of media

characteristics on social influence sorted by the mechanism
underlying the moderation page 78
viii
Notes on the contributors
susanne abele
Department of Psychology
Miami University
zachar y birchmeier
Graduate & Continuing Studies
Stephens College
jordan m. carpenter
University of North Carolina at Chapel Hill
caleb t. carr
Department of Communication
University of Oklahoma
d av i d c . d e a n d r e a
Department of Communication and Department of Epidemiology
Michigan State University
beth dietz-uhler
Miami University
karen m. douglas
School of Psychology
University of Kent
melanie c. g reen
University of North Carolina at Chapel Hill
a n d r e a b. h o l l i n g s h e a d
Annenberg School for Communication and Journalism
Marshall School of Business
ix
x Notes on the contributors
University of Southern California
girish lala
The Australian National University
mar tin lea
The School of Psychological Sciences
The University of Manchester
craig mcgar ty
Murdoch University
tom postmes
Department of Social and Organizational Psychology
University of Groningen
kai sassenberg
Knowledge Media Research Center
Tübingen, Germany
russell spears
Cardiff University/University of Groningen
garold stasser
Miami University
stephanie tom tong
Department of Communication Studies
University of Puget Sound
b r a n d o n va n d e r h e i d e
School of Communication
The Ohio State University
j o s e p h b. w a l t h e r
Department of Communication and Department of
Telecommunication,
Information Studies & Media
Michigan State University
Notes on the contributors xi
e r i c d . we s s e l m a n n
Department of Psychological Sciences
Purdue University
kipling d. williams
Department of Psychological Sciences
Purdue University
a n k a wo l b e r t
Department of Social Psychology
University of Amsterdam
Acknowledgments
This book grew out of the Conference on Group Processes in Computer-

Supported Interaction: Technological and Social Determinism, held in
April 2005 at the Marcum Center, Miami University, Oxford, Ohio. This
conference would not have been possible without the generous support of
the National Science Foundation (BCS 0434200) and the Department
of Psychology, Miami University. We also thank the many speakers and
participants for their contributions to the conference and their thoughtful
work on group processes in computer-supported interaction. This work
laid the foundation for the work reported in this book.
xii
1 Introduction
A social psychological analysis of

computer-supported social interaction
Zachary Birchmeier, Beth Dietz-Uhler,
and Garold Stasser
The use of computer-supported social interaction (CSI) has become

a primary feature of communication among individuals, due in part
to its structural features (e.g., freedom of time and geographical con-
straints) and psychological features (e.g., anonymity). As a conse-
quence, many social scientists have investigated the social processes in
computer-supported interactions, including online impression forma-
tion, relationship development, and group dynamics. Because individuals
communicate via the use of computers in many personal, educational,
and professional settings, it is important to continue and encourage the
study of social processes in such environments. These studies have iden-
tified a number of influences on the behaviors (i.e., conformity, economic
choices, etc.), thoughts (i.e., attitude change, impression formation, etc.),
and, to a lesser degree, physiological/emotional states of people involved
in computer-supported interaction.
The goal of this volume is to impose the global theoretical framework
of the person-by-situation interaction (Snyder and Ickes, 1985) onto the
study of computer-supported social interaction. This perspective recog-
nizes that people are affected by the expectations and limitations of social
situations, but to varying degrees. The extent that a person will respond
or react to social forces has been found to depend on the levels of inter-
nal characteristics (e.g., self-esteem) that a person brings with them to
the social situation. For example, social psychologists have observed that
those individuals who express higher levels of self-esteem are more resis-
tant to social pressure and as a result are less likely to conform to the
influence of others (Baumeister, 1982), and are more likely to persist in
the face of obstacles to their goals (Gist and Mitchell, 1992). Originally
posited by Lewin (1935), reinvigorated by Mischel (1968), and then
tested and validated in decades of social psychological research across
a broad range of social phenomena, the explanatory rubric of social
1
2 Zachary Birchmeier, Beth Dietz-Uhler, and Garold Stasser
psychology can now be applied to the communication forums that have

emerged in the relatively brief existence of internet technologies. Before
discussing the details and benefits of the person-by-situation interaction
perspective, we will review the existing theoretical frameworks that have
been applied to computer-supported interactions.
Contemporary literature reviews

In reviewing the available literature on computer-mediated communica-
tion, a number of patterns can be observed. To date, the largest amount
of research and theory on this topic has focused on the applied contexts
of collaboration and team performance, as summarized in Bailey and
Kurland (2002), Baltes et al. (2002), Coovert and Burke (2004), Gib-
son and Cohen (2003), Hertel et al. (2005), Hinds and Kiesler (2002),
Kreijns et al. (2003), and other treatises. These works present theo-
ries from industrial and organizational psychology, as well as in human
resource management. Broader psychological issues of online impres-
sion and relationship formation, as well as general patterns of group
dynamics, have received less attention from psychologists. The existing
psychological theories and research on these more general social issues
have been summarized formally in a handful of review papers (McKenna
and Bargh, 2000; Riva, 2002), including an entire volume of the Journal
of Social Issues (2002, v. 58), as well as in a few edited books (Amichai-
Hamburger, 2004a; Avgerou et al., 2004; Gackenbach, 1998; Joinson,
2003; Riva and Galimberti, 2001; Thurlow et al., 2004).
In Gackenbach’s (1998) seminal text, the authors described the
potential for internet technologies to both enhance and transform
(Haythornthwaite et al., 1998) a broad range of psychological and social
phenomena as they emerged in the first few years following the inception
of the world-wide web. Following the description of those possibilities
for enhanced business, social, and educational practices, Riva and Gal-
imberti’s (2001) book contributed an impressive depth of detail on the
status of virtual reality use in psychological research and practice, as
well as ventures in telemedicine and e-therapy. The text also included
some analysis of computer-mediated social interaction from psycholin-
guistic and communications perspectives. Galimberti and Riva (2001)
expand on the idea that computer-mediated communication involves a
negotiation of shared meaning with strong hermeneutical overtones:
Context may be co-constructed by social actors, but they use communication

to exchange meanings, not pieces of information. More precisely, the content of
communication is interpretations of the situations which actors are involved in.
Introduction 3
In this sense, the most effective way of clarifying the meaning of messages is to
relate them to a shared context of meaning. (p. 14)
Joinson’s (2003) text cast a wide net in documenting the pervasive

prosocial outcomes (i.e., internet dating and friendship formation) and
deviancy (i.e., deception) that had been observed to date, but also
inferred a theoretical framework that would characterize the global effects
of internet technology for social relations, as noted below. Shortly there-
after, Thurlow et al. (2004) published a textbook for undergraduate audi-
ences that also organized the existing phenomena in a coherent structure
involving learning and critiquing of theories, application of social internet
technologies, and exploration of newer uses of the tools (e.g., in legal
and health communications). The authors in Avgerou et al.’s (2004)
text blended both sociological and information systems perspectives to
provide an interesting account of “innovations” and “actors” that are
embedded in computer-mediated social contexts with functional inter-
dependence.
Bottom-up approaches
Each treatise has summarized the existing theory and research on the
implications of the internet for social contexts. Each has organized the
available theories by categories of group dynamics, online relationships,
and social influence, among others. Each has also drawn general conclu-
sions about computer-mediated communication in a bottom-up fashion
by noting general effects of technology on social communication and
relationships.
Technological and social determinism

Sherman’s (2001) chapter in Riva and Galimberti (2001) summarized
the classic theories (e.g., media richness, Daft et al., 1987; social pres-
ence, Short et al., 1976) that differentiate computer-mediated modes of
communication modes from more traditional ones. Spears et al.’s (2002)
article in the Journal of Social Issues parsed out global assumptions made
by classic theorists of CMC into technological deterministic models and
social deterministic models. In technological deterministic accounts, tech-
nology has universal effects on social processes, which can lead to a
widening of communication options on the one hand and a limit to the
number of available social context cues on the other (Kiesler and Sproull,
1992). Walther (1996; Tidwell and Walther, 2002) has offered to explain
how users can alter their information-seeking and behavioral strategies
to ask deeper questions and adhere more strongly to norms of disclosure

reciprocity in the service of impression and relationship formation when
communicating with others via relatively impoverished media.
In social deterministic accounts, individual users of technology con-
struct their own meaning, which can lead to a decrease in self-regulation,
and an enhanced sensitivity to social information and expectations. Social
influence becomes very strong when online, but taking cues from oth-
ers can have prosocial or antisocial influences on a person’s thoughts,
feelings, and behaviors, depending on the tone of the social climate that
emerges in a particular interaction setting (Spears et al., 2002).
Interactionist response
Spears et al. (2002) conclude that an interactionist model of social and
technological influence may better explain variability in both the use
of technology for social purposes and the tendency for technology to
enhance social interaction. In sum, observations of social behavior in
computer-mediated settings often parallel those in face-to-face contexts,
but the social qualities of communication and cognition are also com-
monly enhanced by specific communication modes. Personal influences
on online cognitions and behaviors are noted where the authors dis-
cuss individuals’ strategic uses of social technologies (i.e., connecting to
anonymous, online forums populated with like-minded others) as hav-
ing cognitive effects on those users (i.e., enhanced social influence and
attitude polarization).
Similarly, Joinson (2003) presented the Strategic and Motivated user
model that includes Expected and Emergent effects (SMEE). This model
accounts for users’ selections of communication media and forums as a
determinant of the effects of those media (i.e., cognitive changes in self-
awareness and social identity salience; behavioral changes in language use
and self-disclosure, etc.). These changes in users’ cognitive and behav-
ioral states are then construed to influence future choices of communica-
tion media. Elements of these ideas were also addressed in Mantovani’s
(2001) chapter in discussing the uses of and reactions to the computer-
mediated social context. Attending to the ability for individuals to choose
communication media for various purposes, Mantovani denotes that
user goals of “information seeking” and “social gathering” both involve
social interdependence, but with differing amounts of synchronicity and
interactivity between communicators. These ideas are congruent with
the intended uses for the world-wide web from its creators: to share
Introduction 5
information, to develop a shared knowledge base, and to connect like

minds (King et al., 1997).
Top-down theories
Some of the available reviews and books on computer-mediated commu-
nication have sought to structure the available theories and research in
a top-down fashion. For example, Galimberti and Riva’s (2001) chapter
applied theories of psycholinguistics and communication, in ephemeral
terms, to characterize online social contexts. Sherman’s (2001) chap-
ter in the same text also applied social cognition theories of bias in
social information processing to characterize impression formation and
categorization processes among online interactants (see also Rafaeli
et al., 2004). Authors in Amichai-Hamburger’s (2004a) text applied
topic-specific social psychological principles and theories to specific
domains of online communication, including persuasion (Guadagno and
Cialdini, 2004), prosocial behavior (Sproull et al., 2004), and prejudice
(Glaser and Kahn, 2004).
Similarly, Spears et al. (2002) provided a notable theoretical framework
for conceptualizing online group dynamics in light of a Social Identity
model of Deindividuation Effects. Their SIDE model characterized the
effects of subjective group norms on spontaneous social identity salience
that had been inferred and tested previously from observations of offline
groups (Hogg and Abrams, 1988; Reichert et al., 1995; Tajfel and Turner,
1986). The existing social identity framework of group dynamics was then
applied against the potential for the internet to involve distanced users in
meaningful social categorizations and to provide enhanced group identity
salience via anonymous meeting forums. Thus, the online context was
illustrated as a ripe field for potent social influence.
Amichai-Hamburger’s (2004b) chapter discusses the existing research
that involves strategic choices of social internet technologies as a function
of personality traits, and calls for more research on the interaction of
technologies and individuals. The author notes Kraut et al.’s (2002)
observations that more extraverted individuals are more likely to use
internet technologies in order to interact directly with others. Amichai-
Hamburger’s research has also shown that, when more introverted or
neurotic individuals seek direct interaction with others online, they are
more likely to express aspects of their self-concept that are not expressed
during typical face-to-face interactions (i.e., “Real-Me” characteristics;
Bargh et al., 2002; see also Maldonado et al., 2001). The author also
speculates on how an individual’s levels of need for closure, need for
cognition, attachment style, and locus of control may affect their goals,
intentions, and strategies when spending time connected to the internet.
Interactionist framework
To date, a handful of chapters in separate texts have posited an inter-
actionist framework in analyzing computer-mediated communication.
These inferences have been drawn from topic-specific research find-
ings that personality characteristics, identity content, and personal biases
can moderate the choice and impact of communication media. What is
needed at this point is to apply the rubric of social psychology more sys-
tematically onto the fundamental processes of impression, relationship,
and group processes online with the goal of identifying new directions of
research in the growing modes of social connectivity.
This text systematically applies the person-by-situation rubric of social
psychology onto computer-mediated communication by including the
technological source of influence. By beginning with a broad, domain-
level framework and then imposing this structure onto the existing works,
the potential for more comprehensive analyses and applications of those
works is increased. The social psychological orientation is unique in that it
provides comprehensive, pluralistic explanations for socio-behavioral and
socio-affective levels of analysis in addition to the socio-cognitive level.
Thus, online behaviors such as questioning or disclosing information
can also be explained, and programs can be proposed for improving the
efficacy of online social support networks as well as electronic group
decision making.
In this text, prominent psychologists summarize the existing research
in their respective areas of expertise. They then organize these findings
within the established rubric of social psychology. Each author also spec-
ulates as to how other sources of influence that have not yet been tested
may affect individuals who communicate online. By identifying the influ-
ences that shape individuals’ thoughts, feelings, and behaviors online,
and evaluating the relative strength and interactive properties of those
sources of influence, other scientists may make use of this broad theo-
retical framework to make recommendations for improving social and
collaborative relations that are supported by computers. In addition, the
potential sources of influence that have not yet been explored are out-
lined in more detail, thus identifying future directions for research and
intervention.
Many who study computer-mediated communication are social
psychologists. In general, social psychologists have been concerned
historically with evaluating how strong the power of the situation can
Introduction 7
be in influencing how people relate to, think, and feel about others.
The preference of a situational explanation of social phenomena is often
portrayed to be at odds with longstanding, classic psychological explana-
tions of individual behavior that have focused on the validation of stable
traits that characterize individuals (e.g., self-monitoring) and correspond
to behaviors (e.g., conformity). Kurt Lewin’s seminal propositions, that
those inner, personal, and external social factors would be shown to com-
bine in meaningful ways to influence people’s behaviors, have been tested
and elaborated upon by social psychologists in the last seventy years.
Person-by-situation interaction defined

Theory and research on social thought, feeling, and behavior that involves
the person-by-situation interaction have become integrated into con-
temporary instructional texts for college students on social psychology
courses (Breckler et al., 2006; Myers, 2005). Myers (2005) presented a
metaphor about the complexity of influences that direct human thought,
feeling, and behavior:
This great truth about the power of external pressures would sufficiently explain
our behavior if we were passive, like tumbleweeds. But unlike tumbleweeds, we
are not just blown here and there by the environment. We act; we react. We
respond, and we get responses. We can resist the social situation and sometimes
even change it. (p. 203)
In a set of papers, social psychologists Mark Snyder and William Ickes

have qualified these basic assumptions to allow for fundamental influ-
ences to interact with one another in meaningful ways (Ickes et al., 1997;
Snyder, 1983; Snyder and Ickes, 1985). In their analyses, influences on
behaviors, thoughts, and feelings in social contexts interact when the
effect of one influence (i.e., the power of the situation) depends on the
state of another factor. In a contemporary example of distanced work,
members of an ad-hoc committee of scientists from several nations that
have been assembled by the World Health Organization are charged
with the task of preventing a global pandemic of the bird flu. Members’
individual decisions to share or withhold a risky, potentially controversial
solution with the rest of the group will depend on each member’s personal
goals (i.e., to achieve the optimal solution to the problem at hand, or to
avoid being ostracized from the think tank), and whether those scientists
will be identified or left anonymous in meeting minutes or public reports.
One type of interaction among personal and social influences involves
the multiplication of the two. Since both influences form the product
of the expression, extreme levels of one factor can essentially negate the
other factor (as when the extreme factor is near zero) or increase the
weight of the other factor geometrically. Snyder and Ickes (1985) differ-
entiated cases when personal variables moderate the impact of the situ-
ation, as well as the converse. They illustrated their points with observa-
tions that individuals who express relatively low levels of self-monitoring
are more likely to resist social pressures to express particular attitudes,
or otherwise behave differently from situation to situation. Conversely,
social situations with particularly weak or unclear expectations for behav-
ior (e.g., a Yahoo! American Idol chat room) can allow for individuals
to express aspects of their self-concepts that are central to their iden-
tities but that also possess some risk in being a target of prejudice by
others. The multiplicative interaction of personal and social influences
has been shown to account for significant amounts of variability in peo-
ple’s behaviors, thoughts, and feelings, above and beyond the effects
of each influence alone, in research on many social issues (Snyder and
Ickes, 1985), including the effects of prejudice (Steele, 1997), aggression
(Baumeister et al., 2000), and persuasion (Haugtvedt et al., 1992).
In contemporary social contexts involving internet connectivity, the
additional factors of media richness and social presence must also be
considered. In a technological x social interaction perspective, the con-
straints imposed by less rich distanced technologies can either increase or
decrease the impact of the social influence when trading emails or posting
messages on a blog. For example, research by Wesselmann and Williams
(this volume) has shown that reactions to social ostracism are character-
ized by quiet withdrawal in a typical offline setting, but that ostracized
participants display a virtual bravado in chatroom settings by continuing
to send messages to the group. The interaction of technological and social
influences is also illustrated in scientific evidence reported by Green and
Carpenter (this volume). They find that the willingness to trust another
person when online depends on the perceived similarity between parties,
as well as the ability to verify their identity through social ties.
The features of the communication media that are chosen for group
interactions can also make or break the quality of any decisions or judg-
ments that are reached. For example, important research on electronic
teamwork has found that the perception of anonymity that social tools
create may help those teams make better decisions by increasing the likeli-
hood of participation from minority factions (McLeod et al., 1997). Den-
nis et al. (1999) also found that, in group tasks requiring greater interde-
pendence (e.g., decision-making contexts), only groups composed of all
women were able to benefit from using richer media for communication.
We can also observe triple interactions among the technological, social,
and personal sources of influence: when online, which individuals are
Introduction 9
impacted greatest by which situations? Thus, it can be observed that

high self-monitors may be particularly likely to be influenced by partici-
pation in online groups that have clear, strong expectations for members.
Conversely, low self-monitors may express their political beliefs with solid
consistency, regardless of whether they are asked when they are online
or offline, or are in the presence of judgmental others. The observed
evidence for the SIDE model offers strong support for the technology x
social x personal interaction. Spears et al. (this volume) have found that
the degree to which individual members of online groups will shift their
attitudes in response to other members depends on each member’s level
of identification with the group ideals, but that the salience of the group
context depends on the degree of anonymity allowed by the mode of
communication. Also, Sassenberg (this volume) reports that the strength
of social influence online depends on the target’s level of private self-
awareness.
Research on electronic teamwork has also found support for the
technology x social x personal interaction perspective. For example, Hollings-
head (this volume) shows how the amount of information shared
during computer-mediated group discussions depends on the idiosyn-
cratic goals that each group member may enter with. Competitive mem-
ber goals (such as to be liked, or to be viewed as the most competent) can
lead those members to focus their contributions on advocating for their
initially preferred option, without necessarily attending to new informa-
tion that might disconfirm their belief. Moreover, Abele (this volume)
discusses how the nature of dependencies affects interpersonal impres-
sions. These initial impressions affect subsequent behavior in a way that
often confirms the impressions.
Dynamic interactionism
Using Snyder and Ickes’ theoretical framework, explanations of social
processes can better fit reality by accounting for the tendency in indi-
viduals to respond to situations differently depending on our unique
attributes as individuals. The influences of social and personal factors
can also be correlated. Snyder and Ickes (1985) elaborated on how tech-
nology mode, social context, and personal characteristics can be dynamic
in their influences:
In each case individuals appear to gravitate actively toward social situations that
will foster and encourage the behavioral expression of their own characteristic
dispositions and interpersonal orientations. To the extent that they succeed in
regularly and consistently spending time in these situations, and to the extent
that these situations promote the regular and consistent display of behavioral
manifestations of their characteristic dispositions, these individuals will come to

display the cross-situational consistency and temporal stability that we regard as
personality. (p. 918)
Thus, they acknowledged how a person’s choice of situations to experi-

ence, or groups to belong to, is dependent upon their idiosyncratic values
or capabilities (e.g., a person with liberal political affiliations might seek
out groups of other liberals online), but that continued exposure to those
environments can amplify the person’s predisposed interest (e.g., that
person begins contributing to activist campaigns of liberal orientation).
Moreover, Snyder and Ickes also acknowledged the potential for
individuals to be active participants in social contexts by influencing
other group members and being instrumental in negotiating norms. The
authors illustrated this principle with evidence on individual differences
in perceived locus of control (Rotter, 1966). For example, individuals
who attribute their personal outcomes as being internally controlled are
more likely to expend effort in overcoming obstacles to personal goals,
and are also more successful in influencing others when motivated to do
so (Lefcourt, 1982). Thus, intentional influence among communicators
in online social forums may be a function of the amount of perceived
“congruency” between the individuals’ dispositional preferences for the
forum, and what the forum actually provides (Secord and Backman,
1965; Snyder and Ickes, 1985).
Arguably, the greatest potential value of applying the person-by-situation
framework to CSI is in the illumination of dynamic interactionism at
work. By identifying the ways that individuals can become empowered
by social technologies (especially those individuals who are disadvantaged
in offline contexts), we can then make recommendations for programs
to benefit those individuals. For example, people who are stigmatized by
others in the physical world can achieve satisfying, enjoyable relationships
online. McKenna and Bargh (1998) reported that people who are appre-
hensive about being evaluated on the basis of a potentially stigmatizing
identity feature (e.g., a homosexual orientation, speech impediment, etc.)
can reduce their apprehension via online group participation, and gain
confidence in disclosing that feature to friends and family offline. In gen-
eral, people around the world who have been attributed with lower status
can exert power by connecting with like-minded others using networked
computers. However, McGarty, Lala, and Douglas’ observations (this
volume) of online social action groups note that the road from opinion-
ated group discussion to concerted effort is often peppered with discon-
tent, and that consensus among those ‘like-minded’ individuals is seldom
reached. Also, Sassenberg (this volume) reports that individuals’ choices
Introduction 11
of groups to join are limited by what is available and permeable. These

observations communicate a need for continued research in this area to
translate these researchers’ findings into evidence-based social programs,
policies, and practices for healthy living.
Conclusion
Interpersonal relationships and group dynamics that use internet tools
to exert power and achieve personal and group goals can be effectively
explained using the rubric of social psychology. The interaction of per-
sonal and social influences has been shown to account for significant vari-
ability in people’s behaviors, thoughts, and feelings, above and beyond
the effects of each influence alone. Using this theoretical framework,
explanations of social processes can better fit reality by accounting for
the tendency in individuals to:
r choose situations to experience, or groups to belong to, that are con-
gruent with their values and capabilities
r respond to situations differently depending on their unique attributes
as individuals
r be active participants in social contexts by influencing other group
members and negotiating norms.
Including the impact of individual predispositions, choices, and
motivations on social uses of internet tools allows for more effective
management of those processes by identifying the factors that can be
manipulated to facilitate the expression of power for minority inter-
ests, prevent ostracism, or facilitate healthy, productive interdependence
among people using networked computers. The potential for social tech-
nologies to connect remote minds, to communicate emergent ideas, and
to empower individuals should establish the relevancy of the CSI per-
spective for interested readers in all communities and cultures. Theo-
retical analyses of online communication and collaboration have yet to
apply systematically the perspective of social psychology that Snyder and
Ickes summarized. The most important contributions of this volume are
the authors’ plots of trajectories for future research. Continued research
will enhance the prosocial, instrumental, and therapeutic uses of social
technologies.
References
Amichai-Hamburger, Y. (2004a). The social net: Human behavior in cyberspace.
New York: Oxford University Press.
(2004b). Personality and the internet. In Y. Amichai-Hamburger (Ed.), The

social net: Human behavior in cyberspace (pp. 27–56). New York: Oxford
University Press.
Amichai-Hamburger, Y., Wainapel, G., and Fox, S. (2002). “On the internet,
no one knows I’m an introvert”: Extroversion, neuroticism and internet
interaction. CyberPsychology & Behavior, 5(2), 125–128.
Avgerou, C., Ciborra, C., and Land, F. (Eds.) (2004). The social study of informa-
tion and communication technology: Innovation, actors, and contexts. New York:
Oxford University Press.
Bailey, D. E., and Kurland, N. B. (2002). A review of telework research: Find-
ings, new directions, and lessons for the study of modern work. Journal of
Organizational Behavior, 23, 383–400.
Baltes, B. B., Dickson, M. W., Sherman, M. P., Bauer, C. C., and LaGanke, J.
S. (2002). Computer-mediated communication and group decision making:
A meta-analysis. Organizational Behavior and Human Decision Processes, 87,
156–179.
Bargh, J. A., McKenna, K. Y. A., and Fitzsimons, G. M. (2002). Can you see the
real me? Activation and expression of the “true self” on the internet. Journal
of Social Issues, 58(1), 33–48.
Baumeister, R. F. (1982). Self-esteem, self-presentation, and future interaction:
A dilemma of reputation. Journal of Personality, 50, 29–45.
Baumeister, R. F., Bushman, B. J., and Campbell, W. K. (2000). Self-esteem,
narcissism, and aggression: Does violence result from low self-esteem or
from threatened egotism? Current Directions in Psychological Science, 9(1),
26–29.
Breckler, S. J., Olson, J., and Wiggins, E. (2006). Social psychology alive (1st
edn.). Belmonk, CA: Thompson-Wadsworth.
Coovert, M., and Burke, J. (2004). Leadership and decision making. In
Y. Amichai-Hamburger (Ed.), The social net: Human behavior in cyberspace
(pp. 219–246). New York: Oxford University Press.
Daft, R. L., Lengel, R. H., and Trevino, L. K. (1987). Message equivocality,
media selection and manager performance: Implications for information
systems. MIS Quarterly, 11(3), 355–366.
Dennis, A. R., Kinney, S. T., and Hung, Y. C. (1999). Gender differ-
ences in the effects of media richness. Small Group Research, 30(4), 405–
437.
Gackenbach, J. (Ed.) (1998). Psychology and the internet: Intrapersonal, interper-
sonal, and transpersonal implications. San Diego, CA: Academic Press.
Galimberti, C., and Riva, G. (2001). Actors, artifacts and inter-actions: Outline
for a social psychology of cyberspace. In G. Riva and C. Galimberti (Eds.),
Towards cyberpsychology: Mind, cognition, and society in the internet age (pp.
3–18). Washington, DC: IOS Press.
Gibson, C. B., and Cohen, S. G. (Eds.) (2003). Virtual teams that work: Creating
conditions for virtual team effectiveness. San Francisco, CA: Jossey-Bass.
Gist, M. E., and Mitchell, T. R. (1992). Self-efficacy: A theoretical analysis of
its determinants and malleability. Academy of Management Review, 17, 183–
211.
Introduction 13
Glaser, J., and Kahn, K. (2004). Prejudice, discrimination, and the internet. In
Y. Amichai-Hamburger (Ed.), The social net: Human behavior in cyberspace
Guadagno, R., and Cialdini, R. (2004). Online persuasion and compliance:
Social influence on the internet and beyond. In Y. Amichai-Hamburger
(Ed.), The social net: Human behavior in cyberspace (pp. 91–114). New York:
Haugtvedt, C. P., Petty, R. E., and Cacioppo, J. T. (1992). Need for cognition
and advertising: Understanding the role of personality variables in consumer
behavior. Journal of Consumer Psychology, 1(3), 239–260.
Haythornthwaite, C., Wellman, B., and Garton, L. (1998). Work and community
via computer-mediated communication. In J. Gackenbach (Ed.), Psychology
and the internet (pp. 199–226). San Diego, CA: Academic Press.
Hertel, G., Geister, S., and Konradt, U. (2005). Managing virtual teams: A
review of current empirical research. Human Resource Management Review,
15, 69–95.
Hinds, P., and Kiesler, S. (Eds.) (2002). Distributed work. Cambridge, MA: MIT
Press.
Hogg, M. A., and Abrams, D. (1988). Social identifications: A social psychology of
intergroup relations and group processes. London: Routledge.
Ickes, W., Snyder, M., and Garcia, S. (1997). Personality influences on the choice
of situations. In R. Hogan, J. Johnson, and S. Briggs (Eds.), Handbook of
personality and social psychology. San Diego, CA: Academic Press.
Joinson, A. N. (Ed.) (2003). Understanding the psychology of internet behaviour:
Virtual worlds, real lives. New York: Palgrave Macmillan.
Kiesler, S., and Sproull, L. (1992). Group decision making and communication
technology. Organizational Behavior and Human Decision Processes, 52(1),
96–123.
King, J. L., Grinter, R. E., and Pickering, J. M. (1997). The rise and fall of
netville: The saga of a cyberspace construction boomtown in the great divide.
In S. Kiesler (Ed.), Culture of the internet (pp. 3–33). Mahwah, NJ: Lawrence
Erlbaum Associates.
Kraut, R., Kiesler, S., Boneva, B., Cummings, J., Helgeson, V., and Crawford,
A. (2002). The internet paradox revisited. Journal of Social Issues, 58, 49–74.
Kreijns, K., Kirschner, P. A., and Jochems, W. (2003). Identifying the pitfalls
for social interaction in computer-supported collaborative learning environ-
ments: A review of the research. Computers in Human Behavior, 19, 335–353.
Lefcourt, H. M. (1982). Locus of control: Current trends in theory and research (2nd
edn.). Hillsdale, NJ: Erlbaum.
Lewin, K. (1935). A dynamic theory of personality. New York: McGraw-Hill.
Maldonado, G. J., Mora, M., Garcia, S., and Edipo, P. (2001). Personality, sex
and computer communication through the internet. Anuario de Psicologia,
32, 51–62.
Mantovani, G. (2001). Beyond the “impact” metaphor: The mutual shaping
of psychological theory and internet development. In G. Riva and C. Gal-
imberti (Eds.), Towards cyberpsychology: Mind, cognition, and society in the
internet age (pp. 39–51). Washington, DC: IOS Press.
McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the inter-
net: Identity demarginalization through virtual group participation. Journal
of Personality and Social Psychology, 75, 681–694.
(2000). Plan 9 from cyberspace: The implications of the internet for personality
and social psychology. Personality and Social Psychology Review, 4, 57–75.
McLeod, P. L., Baron, R. S., Marti, M. W., and Yoon, K. (1997). The eyes
have it: Minority influence in face-to-face and computer-mediated group
discussion. Journal of Applied Psychology, 82(5), 706–718.
Mischel, W. (1968). Personality and assessment (pp. 136, 235). New York: Wiley.
Myers, D. (2005). Social psychology (8th edn.). New York: McGraw-Hill.
Rafaeli, S., Raban, D., and Kalman, Y. (2004). Social cognition online. In Y.
Amichai-Hamburger (Ed.), The social net: Human behavior in cyberspace (pp.
57–90). New York: Oxford University Press.
Reichert, S. D., Spears, R., and Postmes, T. (1995). A social identity model of
deindividuation. European Review of Social Psychology, 6, 161–198.
Riva, G. (2002). The sociocognitive psychology of computer-mediated commu-
nication: The present and future of technology-based interactions. CyberPsy-
chology & Behavior, 5, 581–598.
Riva, G., and Galimberti, C. (Eds.) (2001). Towards cyberpsychology: Mind, cog-
nition, and society in the internet age. Washington, DC: IOS Press.
Rotter, J. B. (1966). Generalized expectancies for internal versus external control
of reinforcement. Psychological Monographs, 80, 1–28.
Secord, P. F., and Backman, C. W. (1965). An interpersonal approach to per-
sonality. In B. Maher (Ed.), Progress in experimental personality research, Vol.
2 (pp. 91–125). New York: Academic Press.
Sherman, R. C. (2001). The mind’s eye in cyberspace: Online perceptions of self
and others. In G. Riva and C. Galimberti (Eds.), Towards cyberpsychology:
Mind, cognition, and society in the internet age (pp. 53–72). Washington, DC:
IOS Press.
Short, J., Williams, E., and Christie, B. (1976). The social psychology of telecom-
munications. London: Wiley.
Snyder, M. (1983). The influence of individuals on situations: Implications for
understanding the links between personality and social behavior. Journal of
Personality, 51, 497–516.
Snyder, M., and Ickes, W. (1985). Personality and social behavior. In G. Lindzey
and E. Aronson (Eds.), Handbook of social psychology (3rd edn.). New York:
Random House.
Spears, R., Postmes, T., Lea, M., and Wolbert, A. (2002). When are net
effects gross products? The power of influence and the influence of power
in computer-mediated communication. Journal of Social Issues, 58, 91–
107.
Sproull, L., Conley, C., and Moon, J. Y. (2004). Prosocial behavior on the net.
In Y. Amichai-Hamburger (Ed.), The social net: Human behavior in cyberspace
Stasser, G. (1992). Pooling of unshared information during group discussion.
In S. Worschel, W. Wood, and J. A. Simpson (Eds.), Group processes and
productivity (pp. 48–67). Newbury Park, CA: Sage.
Introduction 15
Steele, C. (1997). A threat in the air: How stereotypes shape intellectual identity
and performance. American Psychologist, 52(6), 613–629.
Tajfel, H., and Turner, J. C. (1986). The social identity theory of intergroup
behavior. In S. Worschel and W. G. Austin (Eds.), Psychology of intergroup
relations (pp. 7–24). Chicago, IL: Nelson-Hall.
Thurlow, C., Lengel, L., and Tomic, A. (2004). Computer-mediated communica-
tion. Thousand Oaks, CA: Sage Publications.
Tidwell, L. C., and Walther, J. B. (2002). Computer-mediated communication
effects on disclosure, impressions, and interpersonal evaluations: Getting to
know one another a bit at a time. Human Communication Research, 28(3),
317–348.
Walther, J. B. (1996). Computer-mediated communication: Impersonal, inter-
personal, and hyperpersonal interaction. Communication Research, 23, 3–43.
2 A SIDE look at computer-mediated
interaction
Power and the gender divide

Russell Spears, Martin Lea, Tom Postmes,
and Anka Wolbert
Introduction
The new communications technologies are developing at such a fast pace
that it is difficult for research and theorizing to keep up. Although explor-
ing the range of applications and instantiations of the latest forms of
computer-mediated communication (CMC), texting, and video-based
phone systems provides many useful insights, research and theorizing
that lag behind the technological developments will run the risk of being
phenomenon- and even technology-driven, making it difficult to antici-
pate new uses and consequences. In this chapter we therefore adopt a
theory-focused approach to make some sense of the effects of the new
technologies (as Kurt Lewin said, there is nothing so practical as a good
theory), and a primarily experimental methodology to test this. We focus
on a theoretical framework that we have developed over a number of years
to gain insights into the effects of CMC in social and organizational set-
tings: the SIDE model. We have found this model useful in helping to
correct a tendency, in the literature on CMC in particular, to underesti-
mate the role of social influences on and within these technologies, and
an equal (and perhaps opposite) tendency to overestimate their capacity
to counteract the impact of status and power.
In particular, we think this theoretical model has been useful in helping
us to understand (and predict) some of the more counterintuitive find-
ings of behavior found using computer-mediated communication. The
idea that people actually conform when isolated from and anonymous to
their group is a good example of such an effect explained by the SIDE
model. Gender, which is a key focus in the present chapter, also forms
an interesting case study in this respect. Much theorizing and research
has proposed that women might become more assertive and less submis-
sive when liberated by the anonymity of CMC. Our research suggests
that this is not necessarily the case. The SIDE model helps to explicate
16
A SIDE look at computer-mediated interaction 17
when and why the technology helps disempowered groups to transcend

inequalities of status and power, and when it leaves them more vulnerable
to the power divide. Of course, people are not just passively exposed to
the effects of technologies such as CMC – a key argument is that they
provide strategic opportunities for people to “manage” their identities
contra “less mediated” face-to-face communication. So, for example,
when women are given the chance to conceal or deceive their gender
identity, do they do this, and if so, with what effect? And are these strate-
gies and effects similar for other groups and categories? After outlining
our model and providing some evidence of empirical support, we con-
centrate on the “gender divide” as one important test case, in which the
social and power dimensions of these communications technologies in
particular can be examined in some detail. Finally, we consider some of
the implications of these findings for gender, as well as relations between
groups divided by power and status in general.
The SIDE model: theoretical foundations

The social identity model of deindividuation effects (SIDE model) devel-
oped from the social identity theory of intergroup relations (Tajfel
and Turner, 1986) and its sister theory, the self-categorization theory
(Turner, 1982; Turner et al., 1987), together referred to as the social
identity approach. A key principle of this approach is that the self or self-
concept is not simply (exclusively) defined at the individual or personal
level, as the sum of our traits and attributes, but can also be defined
by our membership of (valued) groups. Indeed, these group forms of
self-definition (social identities) can often be more important than our
personal identity, especially in intergroup contexts where they become
more salient. Although we may have stronger commitment to some group
identities than others, the social context in which we find ourselves is
likely to define which group identities are relevant, through the social
comparisons that they evoke or imply with other groups.
As well as encompassing the salience of different identities, however
(referred to below as the “cognitive component” of SIDE), there is an
important strategic dimension to behavior derived from these salient
identities. Unlike face-to-face communication, the new communication
technologies may allow us scope to present our identities as we would
prefer to be seen, or to escape them altogether. This strategic dimension
of the SIDE model becomes particularly interesting when considering
communications technologies that make the manipulation and manage-
ment of identity possible in various ways.
18 Russell Spears, Martin Lea, Tom Postmes, and Anka Wolbert
Before getting into this we provide some brief background on the theo-
retical roots of SIDE. The social identity approach provided the basis for
a critique of classical deindividuation theory that was the launch-pad for
the SIDE model (Reicher et al., 1995; Spears and Lea, 1992, 1994).
Deindividuation theory had been used to explain the antisocial and
aggressive products of crowd behavior (Diener, 1980; Festinger et al.,
1952; Prentice-Dunn and Rogers, 1989; Zimbardo, 1969). More rel-
evant here, deindividuation principles have also been invoked by the
first researchers trying to understand negative aspects of behavior in
computer-mediated communication such as “flaming” (Kiesler et al.,
1984). The basic argument of deindividuation theory is that immersion in
the group, and anonymity associated with this, can produce a “deindivid-
uated” state of reduced self-awareness (or even a “loss of self ”), resulting
in deregulated and disinhibited behavior (Diener, 1980; Prentice-Dunn
and Rogers, 1989; Zimbardo, 1969).
Glossing over the fact that the stereotypically cool and rational context
of CMC (Lea, 1991) seems very different to the chaos of the crowd,
the (visual) anonymity in CMC does provide one starting point for the
analyses of its effect compared with face-to-face (FtF) interaction. How-
ever, we have questioned the applicability of deindividuation theory to
the context of CMC and developed a social identity critique of deindi-
viduation as applied to the explanation of collective behavior in crowds
(e.g., Reicher, 1987). Because we have described this critique in detail
elsewhere (e.g., Spears et al., 2001, 2007), we reproduce this only briefly
here. Basically, our argument is that there is very little evidence for the
operation of a deindividuated state, and a meta-analysis of the deindi-
viduation literature suggests that, far from producing antinormative or
antisocial behavior, the classical deindividuation conditions of anonymity
in the group actually enhance conformity to (more specific) group norms
(Postmes and Spears, 1998). Whether such group norms are antisocial
or prosocial is highly contingent on the nature of the groups. The theo-
retical point is that the evidence that supposedly deindividuating settings
prompt individuals to search for the meaning and guidance provided by
norms suggests a much more socially regulated and less mindless process
than implied in deindividuation theory.
Such findings, however, are quite consistent with the predictions of the
self-categorization theory, and more specifically the SIDE model. The
SIDE model does not dispute that there are real psychological effects of
group immersion and anonymity that need to be explained. Instead of
seeing these as a product of reduced self-awareness or loss of self associ-
ated with a deindividuated state, it proposes that anonymity can enhance
the salience of group identities and further “depersonalize” social
perception of others and the self. Depersonalization (Turner, 1982), a

concept central to self-categorization theory, is not the same as deindi-
viduation; rather than implying a loss of self or reduced self-awareness,
depersonalization refers to a switch to a group level of self-categorization
in which people are seen in terms of their group identities, and behavior
is consequently regulated at this level of self. In these terms behavior is
no less rational and regulated than when individual or personal identity
is salient. Because (visual) anonymity has the effect of obscuring indi-
vidual features and interpersonal differences, it will fuel this process of
depersonalization and thereby increase the salience of group identities.
The net result is that people will tend to perceive group members in
terms of stereotypical group features, or ascribe these to group members
in the absence of visible individuating information (the “deductive aspect
of stereotyping”; Turner, 1982).
This process helps to explain why behavior in anonymous groups
becomes more normative. Following self-categorization theory, the SIDE
model predicts that anonymity in the group will increase group salience
and conformity to group norms. However, while immersion in the group
may seem straightforward in the case of the crowd, this may seem strange,
if not paradoxical, in the case of computer-mediated communication
characterized by social isolation. Indeed, this very feature of isolation
has been a major factor in promoting the idea that the impact of the
social norms and standards will be weakened in CMC (e.g., Kiesler
et al., 1984). This is not a problem for the SIDE model. In line with
self-categorization principles, this approach regards the group as a psy-
chological entity (part of self; “the group within the individual”) as much
as a physical entity embodied in co-presence of its members. In short, our
group identity can be just as strong when isolated from others, and some-
times, paradoxically, more so. This key element of the SIDE model helps
to explain our claim that the isolation and anonymity of CMC can make
people more social (more responsive to group norms and standards),
not less so.
To summarize, SIDE forms a model of deindividuation effects, not
because it endorses the deindividuation concept – quite the contrary –
but because deindividuation research has uncovered a range of important
effects associated with anonymity in the group that requires explanation.
Thus, the theory helps to shed light on the psychology of groups char-
acterized by such features (as in CMC). We refrained from calling this
model a model of depersonalization effects because depersonalization is
itself contingent on group identity being salient in context. As we shall
see, if personal identity is salient (even if people appear to be part of an
interacting group), anonymity in isolated CMC groups can actually lead
members to differentiate themselves from others in the group and any

group norms that may be operating (Spears et al., 1990).
So far we have described the effects of anonymity in increasing the
salience of group identities by means of depersonalization. However, in
line with self-categorization theory and the social identity approach, it
would be wrong to essentialize anonymity as having fixed effects that
are insensitive to other contextual features, or indeed features of the
group. As we shall see below, gender forms an interesting case that
qualifies the prediction that visual anonymity will necessarily result in
increased depersonalization, group salience, and conformity. Because
gender is a social categorization that is highly visually cued, it turns out
that visibility can also increase social category salience in this case. Our
analysis of such apparent “exceptions,” however, is explicable in self-
categorization principles of salience (Oakes, 1987) and demonstrates the
flexibility of the SIDE model in being able to account for a range of differ-
ent contextual features. This is important given the developments of the
new information communications technologies (ICTs) to transcend the
visual anonymity of text-based communication in CMC, and to embrace
streaming video through our computers (e.g., SkypeTM ) and in the next
generation of mobile phones.
A useful way of thinking about the SIDE model that may clarify our
argument that situational conditions such as anonymity vs. visibility will
have variable (but explicable) effects depending on the nature of the
group is to think of the relation between group identity and contex-
tual conditions as comparable to the person x situation (PxS) inter-
action. A psychology-savvy audience will know that Mischel used the
PxS interaction to critique the view that personality traits would have
simple or “generic” main effects irrespective of situation. Indeed, self-
categorization theorists have used this metaphor to apply to the sensitivity
and power of the self-categorization theory (SCT) more generally in tak-
ing into account person and situation, and their interaction (e.g., Turner
and Oakes, 1986). In our case, of course, as with SCT, the “person” in
this equation is replaced by whatever identity is salient (our main focus
here being group identity). The key point is that the outcome of any
behavior is literally the product of this interaction term, which is why
we cannot talk of generic effects of the communication medium (or its
properties, such as anonymity or visibility, etc.), still less of (group) iden-
tity; this will also depend on how the properties of the medium uniquely
affect this identity (and vice versa, hence the interaction term). So in
the case of gender, the fact that this identity is visibly cued means that
a medium characterized by visibility will affect the salience of this iden-
tity differently to an identity that is not visibly cued (e.g., nationality).
Moreover, differences in identity content (for men and women) also help
to explain different behavioral outcomes for these groups, once identity
is made salient, consistent with this PxS metaphor.
Before we get on to these effects, however, it is important to draw a
theoretical distinction between the effects that anonymity vs. visibility
might have on the salience of group identity, and a different class of
effects influenced by such features. Another effect of anonymity already
documented within the deindividuation tradition is the idea that this
reduces our accountability to others compared with the case where we
are visible and identifiable. Deindividuation theorists had long noted
that one effect of being anonymous or disguised is that it gives people
license to indulge in antisocial behavior by escaping identification and
punishment (Festinger et al., 1952; Zimbardo, 1969). However, such a
process clearly reflects a conscious strategy rather than the deregulation
of behavior implied by the unconscious deindividuated state (Diener,
1980; Prentice-Dunn and Rogers, 1989).
The SIDE model also incorporates the role for strategic effects, and
examines how anonymity can help to release behavior that might other-
wise be punished by powerful others or out-groups, and therefore con-
strained. By strategic we simply mean behavior that is designed to address
identity-consistent needs or goals and that takes into account the ability
of other agents (e.g., powerful out-group audiences) to block or punish
such behavior when it conflicts with their own identity and interests.
Identity-consistent behavior that is suppressed or delayed under such
conditions could therefore be said to be “strategic.” There are some
important differences here with the early deindividuation formulations
of accountability, however. In the SIDE model, strategic self-presentation
is not just about protecting and promoting personal self and self-interest
but may flow from group-level self-definition. Moreover, although strate-
gic behavior here may take into account the potential disapproval of the
audience, it does not have to be negative or antisocial from the perspec-
tive of the in-group. This will depend on group norms and agendas; the
behavior expressed might be quite prosocial and in line with the interests
of the one group but happen to conflict with the interests of a powerful
out-group.
Framed in these terms, anonymity can provide a vehicle for resistance.
However, in CMC the situation is complicated by the fact that, as well as
providing a degree of anonymity, the text-based and recorded nature of
CMC, and ICTs generally, can also introduce other sources of account-
ability and even the possibility of surveillance and control (Gandy,
1993). Second, an important feature of CMC, compared with classic
in vivo group immersion, is that people are typically (but not always)
geographically isolated from each other. Although anonymity can

enhance the salience of group identity, social isolation can also reduce
the sense of (physical, embodied) support from fellow group members
that might help to resist or contest a powerful out-group.
In sum, as well as providing a framework to analyze the salience and
impact of identities as affected by anonymity and isolation, the SIDE
model provides a framework to analyze strategic behavior that is sensi-
tive to how those factors impact on power relations and accountability
to audiences. In order to distinguish these strategic effects from those
based on the salience of identity we refer to the strategic and cognitive
dimensions of SIDE respectively.
Empirical evidence for the SIDE model

So far this remains so much abstract theorizing. What is the evidence? To
provide a bit of flesh on the bones of the model we give a brief review of
evidence for the cognitive and strategic effects outlined above as applied
to the CMC context. (See Spears, Lea, and Postmes, 2001, 2007, and
Spears, Postmes et al., 2001, for more detailed reviews.) We then focus on
the case of gender, as analyzed in terms of both the cognitive and strategic
dimensions, and the implications of this analysis for gender-based power
relations in CMC and cyberspace more generally.
Our first evidence for the cognitive effects of anonymity outlined in
the SIDE model was obtained in a variant of the group polarization
paradigm (Spears et al., 1990; see also Lea and Spears, 1991). In this
paradigm, group discussion generally results in more polarized views
on an opinion topic than the average of the individual opinions before
discussion. In other words, if there is a general preference, pro or con, on
a particular attitude issue, group discussion can result in this becoming
more extreme.
In this first study, three-person groups discussed political topics
pretested to show some consensus among the psychology students par-
ticipating, and these norms were reinforced by feedback of opinions
from their study area on each topic. We manipulated two key factors in
this study. Although everyone communicated via email (CMC), half the
groups were visible to each other and in the same room (FtF), and half
were isolated and in separate rooms (representing the standard visually
anonymous CMC case). In addition to this, however, we orthogonally
manipulated the salience of group identity – that is, for half of the groups
we stated that we were primarily interested in the participants as psy-
chology students, not as individuals (group identity salient), whereas for
the other half we stated that our focus was on individual personality
differences in email use (individual identity salient). Our prediction, in

line with the cognitive dimension of SIDE, was that when group identity
was salient, and participants visually anonymous, this would enhance the
salience of group identity (depersonalization) and conformity to group
norms, resulting in greater group polarization in the direction of the
norm. However, we did not expect this effect in the case of salient indi-
vidual identity, where participants may be expected to differentiate their
opinions from the group, asserting the uniqueness of the personal iden-
tity. This is what we found: anonymity accentuated group polarization
when group identity was salient, but led to contrast away from the group
norm when individual identity was salient.
Our analysis and explanation is based on the self-categorization expla-
nation of group polarization and social influence (Turner, 1991). How-
ever, because there are competing explanations of group polarization, it
was important to replicate this effect in other social influence paradigms
where the group normative explanation is less contested. In further
research (Postmes et al., 2001) we created small CMC groups which
interacted on a collaborative task in order to render group identity salient.
In this case anonymity was manipulated by means of posting photos on
the computers in the visible condition, so that physical isolation was kept
constant in all conditions. (All participants posed for photos beforehand
in case this affected identity salience, but these were not used in the
anonymous condition.)
In a second factor of our design we surreptitiously manipulated the
salience of two different norms before group formation by means of a
priming procedure in which words associated with these norms were
embedded in scrambled sentences that participants were required to
form into complete sentences (Srull and Wyer, 1979). These sentences
primed either a prosocial or an efficiency norm. After the collabora-
tive group task participants then engaged in group discussions of a sce-
nario. They were presented with a dilemma in which a hospital was faced
with the problem of how to provide adequate care to all patients; the
administration favored solutions focusing on efficiency and cost-cutting
whereas doctors and nurses focused more on the quality of patient care.
As predicted, the primed norms reliably influenced the solutions that the
groups preferred in their discussions. In line with the cognitive dimen-
sion of SIDE, groups favored more (primed) normative solutions when
anonymous than when visible (the depersonalization process). Indeed,
norm preference was even slightly reversed in the visible conditions. In a
follow-up study we showed that this effect was the product of a genuine
social influence process occurring within the group and was not due to
some more individual cognitive priming effect alone. Results showed that
participants primed with an efficiency norm transferred this norm to two

neutrally primed group members, demonstrating a genuine transfer of
social influence occurring within the group.
Further work has moved beyond the consideration of social influence
to show that the depersonalization process associated with isolation and
anonymity in the group can, paradoxically perhaps, increase social attrac-
tion to the group (Lea et al., 2001). Focusing on nationality as the
salient group identity, this study provided insights into the underlying
self-categorization process by showing that visual anonymity increased
self-categorization in the group and the stereotyping of these in terms of
the group norms (the depersonalization process) in mediating the effect
of anonymity on social attraction to the group. Much research has now
produced converging evidence for the SIDE model’s prediction that the
visual anonymity characteristic of CMC can increase a range of group-
related effects such as conformity to group norms, stereotyping, and
group attraction (Spears, Lea, and Postmes, 2001, 2007).
Most research to date on the strategic aspect of the SIDE model has
emerged outside of the realm of CMC, and has focused on collective
action by disadvantaged groups (see, e.g., Reicher and Levine, 1994a,
1994b; Reicher et al., 1995). This is not to say that power and status
differences are irrelevant to the realm of computer-mediated commu-
nication and cyberspace. Indeed, from early on many influential theo-
rists claimed that CMC offers an environment where status and power
differences are minimized or even equalized (e.g., Siegel et al., 1986),
and where disempowered groups use the potential to manage identity in
cyberspace to liberate themselves from the strictures and stigma of low
status (e.g., Haraway, 1990; Turkle, 1995). However, one of the theoret-
ical contributions of the SIDE approach has been to question the degree
to which power differences are eliminated or transcended here. We will
explore this issue in some detail in the next section where we consider
the case of gender, explicitly in relation to power-based gender relations.
Before we do this, however, it is useful to address some theoretical
issues about power and the strategic aspect of SIDE in relation to the
context of CMC. First, a clear difference between CMC and the context
of collective action in the crowd is that the crowd is empowered by the
physical co-presence of co-acting others, whereas the social isolation of
CMC means that people may often feel disempowered and vulnerable
because of the absence of others. Moreover, the sense of surveillance that
technologies such as CMC and the internet impart (the “panopticon”;
Spears and Lea, 1994) may add to this feeling of vulnerability.
On the plus side of the equation, the communication potential of CMC
puts people in contact with others and so allows them to communicate
both support for other views (which we call opinion support) and the
willingness to act on these views (which we call action support), which
can foster resistance to the power of dominant groups. As such, email and
the internet clearly are also powerful mechanisms for coordinating and
mobilizing collective action for groups that are geographically isolated
(e.g., Brunsting and Postmes, 2002; Postmes and Brunsting, 2002). So
despite being isolated at the terminal, access to the network as a source
of power should not be underestimated.
Clearly these factors (the vulnerability of isolation and surveillance,
the power of communication and connectedness) can work in opposite
directions. Within the context of the SIDE model the picture is further
complicated by the fact that cognitive and strategic effects of CMC need
to be taken into account, and will work in different ways, having different
effects. This is illustrated in two studies by Spears, Lea et al. (2002),
which were aimed at investigating the strategic aspect of SIDE, but were
also designed to disentangle the strategic and cognitive effects associated
with CMC. In this research we adapted a paradigm developed by Reicher
and colleagues designed to investigate the effects of empowerment in
the staff–student intergroup relationship (Reicher and Levine, 1994a,
1994b).
Reicher and Levine (1994a) established different classes of behavior
relevant to testing for the effects of empowerment of a low-power group
(students) in relation to a powerful out-group (staff). Specifically, they
distinguished behaviors that were normative for the in-group and toler-
ated by the out-group (e.g., partying in one’s free time) from behavior
that was normative for the in-group but proscribed by the out-group
(e.g., missing lectures and tutorials). The crucial point is that the disem-
powered group is more likely to constrain behavior that could be seen
as punishable by the out-group when the in-group feels vulnerable to
that group’s power of sanction (e.g., when they are accountable, and
isolated). This prediction was confirmed: students were more likely to
endorse behaviors proscribed by the out-group when they were anony-
mous and therefore not accountable (Reicher and Levine, 1994a), and
when they had the implied social support of other in-group members
present, rather than being isolated (Reicher and Levine, 1994b).
We adapted this paradigm and applied it to the context of CMC in
order to assess the effect of this technology as a channel to communicate
opinion and action support that could foster resistance against a powerful
out-group. In the first study we manipulated the availability of CMC
while keeping constant co-presence of other members of the in-group
(the group of students were all located in the same room). At the same
time we manipulated the visual anonymity associated with the cognitive
aspect of SIDE by separating group members by means of screens or

having them visible to each other (FtF). The students were not able to
communicate verbally, only via computer (in the CMC conditions) or
not at all.
We found that the anonymity manipulation produced greater endorse-
ment of behavior that was normative for students and also acceptable to
the staff – further evidence for the cognitive dimension of SIDE increas-
ing conformity to group norms. In contrast, only the CMC manipulation
affected the endorsement of group normative items that were proscribed
by staff. This supported our prediction that CMC forms a channel that
helps communication and coordination of resistance against the powerful
out-group (e.g., by means of action support). In a follow-up study we
manipulated the communication of action support via email (using false
feedback) in relation to a proposal by the staff that was clearly unpopu-
lar among the students (namely to increase substantially a course credit
requirement for research participation by students). Results confirmed
that the knowledge that other students would actively support them (over
and above opinion support, which was kept constant) was crucial in pro-
ducing open resistance to the proposal.
In sum, although CMC can isolate individuals and potentially disem-
power them as a group, it also provides a channel for communication
that helps to mobilize resistance to the out-group, forming a means of
empowerment. We now consider how CMC and cyberspace can affect
power and status relations in greater detail, and how both cognitive and
strategic SIDE effects contribute to this, by considering the case of gen-
der and the “gender divide.”
The gender divide: SIDE effects affect power relations

in cyberspace
In the remainder of this chapter we use the case of gender, an all-pervasive
social categorization that is characterized by power and status differences
(Ridgeway and Smith-Lovin, 1999), and that has been of central concern
to many researchers on CMC and the internet, in order to “road test”
some of the claims and assumptions about how cyberspace affects these
differences. One reason for this is that, as we have already noted, there has
been some optimism from researchers, including feminist theorists, that
status differences associated with gender are reduced or equalized (e.g.,
Siegel et al., 1986), and that cyberspace provides the potential to escape or
liberate disempowered identities (e.g., Haraway, 1990; Turkle, 1995). As
the preceding section reveals, however, from the perspective of the SIDE
model we are rather more sanguine about the equalizing and escapist
potential of these communications technologies and see cyberspace as

much a “panopticon” as “panacea” in this regard (Postmes et al., 1998;
Spears and Lea, 1994). This may be true for both cognitive and strategic
reasons relating to the SIDE model and we now consider these in relation
to the case of gender.
Cognitive effects: gender salience

In two studies, Postmes and Spears (2002) directly addressed the impact
of the cognitive dimension of SIDE on gender relations and examined
what effect visually anonymous CMC would have when gender is salient.
In the first study, four-person discussion groups were comprised of two
men and two women, and gender was rendered salient by means of
the name label of the participants (e.g., Man2, Woman14), who dis-
cussed four topics via email. Results revealed evidence of clear gender-
based power differences emerging in this context: men initiated more
autonomous comments than did women, and women asked more ques-
tions than men, indicating that men were more dominant in the dis-
cussions. Clearly, then, anonymity did not eliminate any gender power
differences. However, without a control or comparison condition it is also
difficult to conclude that anonymity actually increased gender salience
and stereotypical power differentials, as the cognitive SIDE would pre-
dict. The nature of the discussion topic was also not systematically varied,
so these could have favored male expertise and interest.
In a second study we therefore manipulated the salience of identity and
the gender relevance of the discussion topics. First we manipulated the
salience of gender identity by making clear the gender identity of the par-
ticipants in the name label (as before), or by not indicating gender in this
way. In a second (orthogonal) factor we also manipulated the provision of
individuating personal information by giving participants personal pro-
files of the participants in half of the conditions vs. providing no such
information (all participants remained visually anonymous throughout).
The rationale here was that such information should individuate par-
ticipants and make personal identity salient, thereby detracting from
the salience of gender identity. When no individuating information is
exchanged, however, provided that gender identity is salient through
the name labels, this identity should be depersonalized and participants
should be more likely to perceive and react to each other in terms of their
gender identity in line with the cognitive dimension of SIDE. Finally,
we also manipulated the nature of the discussion topic such that it was
more gender stereotypical either of women (personal relationships and
emotions) or for men (computers).
Based on the cognitive dimension of SIDE we predicted that the men

would be particularly dominant when gender identity was salient and
depersonalized, and discussion was on a masculine topic. However, this
dominance might reverse under conditions of high gender salience for a
feminine discussion topic. The same indicator of dominance as before
(autonomously initiated comments) was used to compute an index of
gender dominance (favoring men or women). Results confirmed the
predicted pattern (a three-way interaction on this index). Men were
also more dominant overall; the depersonalized gender salient/feminine
topic condition was the only condition in which women dominated
the men.
These studies provide further support for the cognitive dimension of
SIDE, but they also provide less cause for optimism that the typical visu-
ally anonymous conditions of CMC will eliminate power differences in
gender relations, at least if gender becomes salient via names. Because
names are typically gender-identifying, women may be advised to with-
hold such gender-identifying information if they can (we consider this
under the rubric of strategic behavior further below). However, this may
be unlikely or difficult in practice. Moreover, research suggests that gen-
der can be identified well above chance simply by the linguistic style of
communicators in anonymous CMC (Thomson and Murachver, 2001),
so with more prolonged communication this may become salient in any
case.
So, if gender salience combined with the depersonalizing effect of
anonymity increases gender power differences, is individuation the
answer? In the previous study we provided individuating information
by means of personal profiles, but a more instant and accessible form
of individuation is, of course, provided when participants are visible to
each other. To see or not to see, that is the question! This possibility is
now more of a reality as the new ICTs employ streaming video webcams
(e.g., Skype), and the new generation of mobile phones also renders visual
communication increasingly likely. In such cases we might assume from
the SIDE model that the individuation implied in FtF contact will inter-
rupt the depersonalization process that can reinforce gender stereotypes
and power relations.
Alas, this is not necessarily the case because gender is one type of
social category where visible cues convey perhaps as much about the
social category as they do about the individual. Whereas many groups
and social categories are not reliably distinguished by their visual features
(e.g., study major, many nationalities), others may be reliably associated
with diagnostic visual cues (the “big three” of ethnicity, age, and gender
are prime examples). In such cases, visibility might actually increase the
salience of these social categories, as well as providing the opportunity

for individuation.
This requires us to refine the predictions for the cognitive dimen-
sions of SIDE, which hitherto have proposed that anonymity increases
the salience of group identity. More generally, however, this refinement
is very much in keeping with self-categorization principles on which
the SIDE model is based. Self-categorization theory proposes that any
group properties that reflect differences between groups will increase the
salience of that social categorization (a process referred to as “compara-
tive fit” or “meta-contrast”; see Oakes, 1987). So where there are clear
visual cues to category differences in appearance, as there are with gen-
der, visibility may actually enhance the salience of that group identity,
especially when these differences also correlate with other underlying
stereotypic differences (i.e., whether the discussion topic is relevant to
gender differences, reinforcing meta-contrast and introducing normative
as well as comparative fit; Oakes, 1987).
We tested this idea in a CMC study in which two different bases
of social categorization were available, namely nationality and gender
(Lea et al., 2007). In the case of the nationality manipulation (Dutch vs.
British) we did not expect this distinction to be visibly cued, so the classic
SIDE prediction that anonymity should enhance salience and deperson-
alization should prevail. In the case of gender, by contrast, we predicted
that visibility would actually enhance the salience of this categorization.
We were interested in explaining how social attraction and group cohe-
siveness would be affected by anonymity vs. visibility within the group.
Each discussion group was comprised of four people, with two located in
Britain and two in the Netherlands (the nationality categorization), with
a man and woman in each location (the gender categorization). In other
words, each group could be categorized in terms of nationality and/or
gender.
Anonymity vs. visibility was manipulated by means of a streaming
video-conferencing system so that in the visible condition participants
could see each other in real time. Finally, the salience of these identities
was manipulated by two discussion topics that were relevant to gender
(“Because women are more in touch with the feelings of others they
make better leaders than men” and “There is too much sport on TV”)
and two that were relevant to nationality (“The failure of the UK to
join the Economic Union straight away is short-sighted and selfish” and
“The bad reputation of British food is justified”). Pretesting of these
discussion topics confirmed that the relevant groups respectively differed
in their agreement with the statements, and were on opposite sides of
the issue, producing a high comparative and normative fit associated
with high salience. The order in which these topics were discussed was
counterbalanced.
For the discussion topic where nationality was salient we found
the classic cognitive SIDE effect whereby group cohesiveness (of own
nationality) was enhanced by anonymity, mediated by prototypical-
ity and self-categorization (the depersonalization process) and social
attraction within the nationality subgroup. However, in the case where
the discussion topics made gender salient, the visible condition led to
increased group cohesiveness within the gender subgroup mediated by
self-categorization and social attraction to their subgroup. In short, the
social attractiveness and cohesion with the group followed a different
route depending on whether the social category was strongly visibly cued
(as in the case of gender) or not (nationality). Although we did not
measure dominance and power in this study, the implication is that mak-
ing gender visible may also strengthen stereotypical group differences in
power relations. So, in the case of gender, removing visual anonymity
does not seem to represent the best solution to breaking down the gen-
der divide any more than anonymity does. In sum, the evidence from the
cognitive side of SIDE is that, whether visible or visually anonymous,
CMC does not seem to eliminate the power and stereotypical differ-
ences that could work against women in gender-based communication
contexts.
Gender and the strategic SIDE

Although gender differences can be accentuated by CMC, there is more
encouraging evidence from the previous studies that, where gender was
hidden or not salient, these differences were reduced or eliminated. This
suggests that women may derive some strategic benefits from using
the potential for gender anonymity within CMC and cyberspace by
not disclosing their gender or adopting neutral and even male identi-
ties (“gender-bending”). This may be all the more important because,
despite the claims for greater gender equality in cyberspace, the internet
and many chat forums can be quite male dominated, both numerically
and in terms of local norms (e.g., Fallows, 2005; Sherman et al., 2000).
This was the experience of one of our team (Anka Wolbert), who found
that, when contributing to chat forums on computers, she was more
likely to be ignored or not taken seriously when others realized she was a
woman(!). Her experiences inspired the current program of research in
which we examined whether women might be more likely to adopt neu-
tral or male identities on-line as a way of avoiding the status disadvantage
they might suffer in cyberspace (Spears et al., in prep; Wolbert, 2000). In
other words, do women use the gender-bending potential of cyberspace

to circumvent male power?
As in our earlier work on the cognitive SIDE, we considered whether
this possibility might also be moderated by the nature of the discussion
topic involved and whether this stereotypically favors men or women.
Again, it may also be that power and the strategic manipulation of gen-
der identity is domain-specific. Thus we varied the discussion topics that
were pretested to suit men or women. Participants (men and women)
were seated at computers in cubicles and were presented with a (bogus
but realistic) chat room where they would be expected to discuss one of
various topics to be chosen by the computer with other people in the chat
room. For this purpose they had to choose an “avatar” to represent them
as their on-line identity in the chat room. They were required to select
a different avatar for each possible discussion topic that could come up.
They also saw the avatars representing the other participants already in
the chat room. The twelve avatars from which they could choose were
pretested for the gender identity such that they varied from very mas-
culine (e.g., superman) through neutral to very feminine. Although the
discussion never took place, we were able to assess the gender identity of
the avatars that men and women chose to represent themselves, depend-
ing also on the nature of each discussion topic.
Results revealed that women did indeed vary the choice of avatar to
match the discussion topics, choosing more masculine or neutral avatars
when the discussion topic was in the “male” domain and more feminine
avatars for “female” topics. Interestingly, however, men did not show any
such sensitivity in matching identity to topic and preferred a masculine
identity irrespective of the topic. This asymmetry arguably reflects the
power asymmetry between men and women. Because of the higher status
of men, they may suffer a loss of status and even stigma by switching to
a female identity, whereas women are more able to benefit from that
switch.
One positive consequence of this gender-bending strategy is that
women may be taken more seriously by men on “male” discussion topics.
Such dissimulation might thus have some strategic benefits for women
trying to get round the gender divide. However, it is important to note
that one consequence of this strategy is that the presence of women as
a group will become even more marginal and perhaps even invisible on
the internet, which may serve to accentuate male dominance and male
norms and behavior. Indeed, the very fact that men are not prepared to
move from the sanctuary of their gender, whereas women are, could itself
be seen as further confirmation of women as the “second sex” who have
to adapt themselves to men.
In a second study (De Wolf, 2002; Spears et al., in prep), we were

interested in following up this work and seeing whether the tendency of
women to be more flexible in managing their gender identity could itself
be open to further manipulation and moderation. This study formed a
replication of the first study in most respects, although it was conducted
with a larger sample and on-line via a website, rather than by means
of simulation in the lab. The key additional factor manipulated in this
study was a subtle manipulation of the salience of gender identity. Half of
the respondents to the on-line study answered some general questions,
including reporting their sex before going into the chat room part of the
study. For the other half these details were recorded after the chat room
and avatar choices had already been made. The idea here was that being
made aware of one’s gender identity by answering this question might
further influence the strategic choice of avatars.
Results confirmed this to be the case. First, as before, in all condi-
tions the male respondents consistently preferred the masculine avatars
irrespective of the nature of the discussion topics (stereotypically male
or female). When gender identity was not made salient by the gender
question, women once again preferred the feminine avatars to represent
them on female topics and more neutral and masculine avatars for male
topics. However, when gender identity was made salient for women,
their preferences reversed and now they preferred slightly more neutral
and feminine avatars for discussion of male topics and more masculine
avatars for discussion of female topics. It is as if, when made aware of their
gender identity, women do not want to be seen as expert on “women’s
issues” and stereotyped accordingly. Moreover, they may also be wary
of adopting a masculine identity for male topics if this raises expecta-
tions about what they can contribute or their interests in these issues.
In other words, making gender identity salient simply by asking people
to denote their sex before the chat room phase may introduce another
level of strategic identity management for women; they seem to become
aware of the down-side of gender-bending for them, and of being seen in
gender-stereotypical terms.
To summarize, we have considered the possibilities that CMC and the
internet provide for managing gender identity: how this can be used to
resist power relations, but also how it can backfire and reinforce gender
relations. So far, however, most of our research has taken place in the
lab, and even when we have modeled more regular real-life contexts,
these have been quite recreational (e.g., the chat room). It therefore
seems important to assess some of these themes in other settings where
gender power relations are perhaps even more consequential, namely in
the workplace and in the organization. This was the aim of the final study
we describe here.
In this study we conducted a large survey of a multi-national company
as part of a European-funded project investigating email use in organiza-
tions (COMMORG). As well as providing insights into the role of how
men and women manage their gender identity using email in the orga-
nization, this study allowed for interesting cross-cultural comparisons,
given that the survey was conducted within the same company across
different European countries (specifically the UK, the Netherlands, Italy,
and Greece, although we dropped the Greek comparison from the fol-
lowing analyses as the sample was too small). One of the questions we
were interested in here was whether women were more likely than men
to conceal their gender identity when using email, for many of the rea-
sons we have already addressed above (Spears et al., unpublished data).
Given their vulnerable power and status position, especially within the
typically male-dominated setting of the company, we thought women
would have more to lose by revealing their gender identity in email com-
munication with others to whom they were not known, and more to gain
by concealing this where possible.
We also thought this strategy might be particularly likely in cultures
that score higher on gender differentiation and status/power differences.
Research shows that this gender divide (as evidenced by employment
rates and the gender wage gap, for example) is stronger in southern Euro-
pean countries (e.g., Italy) than in northern European countries (e.g., the
UK, the Netherlands; see, e.g., Olivetti and Petrongolo, 2008). In short,
we expected that gender concealment for women might be more impor-
tant in Italy than our northern European comparison countries. As well
as this cross-cultural moderator variable we reasoned that gender con-
cealment is particularly likely to be an issue for company employees, and
especially women, if they are more conscious of status differences while
using email. We therefore used an item measuring this status awareness
as an additional moderator variable.
The results revealed interesting and interpretable cultural differences
as a function of gender. In the Netherlands, the pattern was fairly flat
in that both men and women employees reported being very unlikely to
conceal their gender with an unknown (male or female) other, irrespective
of status consciousness. However, Italian women who were aware of
status differences were more than twice as likely to conceal their gender
identity from unknown others compared with the Dutch case, with the
UK in between. Moreover, men in these two countries were less likely
to conceal their gender compared with women when aware of status
differences in this context. Italian women also reported being reliably

more conscious of their gender identity in FtF interaction compared with
CMC, whereas this difference was much smaller in all other comparisons,
supporting the view that there may be good reasons why Italian women
might benefit from gender concealment.
To summarize, we have evidence that women in particular might use
the potential for anonymity in email and on the internet, in social, recre-
ational, and organizational contexts, to strategically manage their gender
identity in ways that protect them from the status and power differen-
tials that characterize the gender divide. However, as we have seen, there
are also reasons to question whether such strategies will always be so
successful at the intergroup level, assuming that the gender agenda is to
equalize these status and power differentials at the group level. While
concealing gender or taking on a more neutral or masculine persona may
protect individual women personally and in the short term, one downside
is that they become less visible and perhaps less able to take credit for
contributions in cyberspace as women. Within the terms of social iden-
tity theory (Tajfel and Turner, 1986), this can be seen as an individual
mobility or social creativity strategy, rather than a group-level strategy
that involves taking on the powerful out-group directly through open
challenge (“social competition”). So long as many contexts of email use
and the internet remain dominated by men or male norms, such strate-
gies and their “SIDE effects” are perhaps to be expected.
Final comments, conclusions, and some

strategic implications
In this chapter we have reviewed the many ways in which characteristic
features of computer-mediated communication technologies can influ-
ence the salience of identities, and the outcomes associated with them,
but can also be used strategically to further the aims and agendas of
users. We have viewed this through the prism of our theoretical frame-
work (the SIDE model), focusing on our own associated program of
research. However, the value of this narrow and rather theory-driven
focus is that, hopefully, it provides at least one perspective on the wide
range of effects in the literature, and perhaps makes at least some sense
of the more puzzling patterns and paradoxes that emerge.
Simply because we associate CMC and cyberspace with (new) technol-
ogy, there is a danger that we focus primarily on the technology and less
on the social factors that shape its effects. However, a recurring theme
throughout our work is that, paradoxically, the isolation and anonymity of
CMC can lead to more powerful social influences: the wire (and wireless!)
extends the reach of identities, and anonymity can sometimes intensify

their effects. The same seems to be true for the reach of power differen-
tials associated with these identities – these do not necessarily disappear
or get diluted in cyberspace, as the example of gender shows. More-
over, the attempts to use the technology to manage and contain power in
creative ways can also backfire, reproducing these power differentials in
different ways or at another level.
We deliberately chose gender as our case study here partly because it
has been held up in previous research and theorizing as the paradigm case
of a social categorization associated with status and power differentials
that can be transcended within cyberspace. Our research shows that the
case for this is far from convincing or clear-cut, but rendered all the more
interesting by some key features associated with it, namely that it is heavily
visually cued, but also cued by name and even linguistic style (Thomson
and Murachver, 2001). This leads to more complex interactions with
media features such as anonymity. For this reason we cannot generalize
effects from this case to other social categories and intergroup relations.
Yet the SIDE model does provide us some tools to analyze when and
why certain media will have certain effects for certain groups, depending
on features of the group, the contents of group identity, and the power
relations between groups.
One criticism is that an analysis focused on anonymity is rendered
redundant by the new generation of ICTs, with enhanced bandwidth
making anonymity less and less likely (Joinson, 2005). There are two
responses to this. First, anonymity is not a unitary, all-or-nothing prop-
erty. Moreover, a key insight of the SIDE model is that anonymity needs
to be analyzed with respect to different levels of identity; different media
affect the flow of personal and group-level cues differently (see also Tanis
and Postmes, 2008), and also for different kinds of groups. The point
here is that social effects of anonymity depend on the balance of cues
and the factors that filter them, and as we have shown, the SIDE model
is up to the task of analyzing the effects of visibility as well as anonymity.
The second point, central to this volume, is that people have strategic
choice about not only which media they choose but what they choose
to reveal about their identities (personal or group level) because of the
properties of these media. The fact that text-based CMC remains for
many the medium of choice suggests that a degree of anonymity will be
with us for some time yet.
While our insights offer possible recipes for strategic use of these tools
and technologies, they also provide some cautionary lessons showing the
unintended or unanalyzed consequences that can sometimes “bite back.”
Our approach suggests there is scope to explain the patterns that emerge
in terms of the distinctive properties of technology, such as anonymity,

but two caveats are worth stating. First, these effects are not generic
properties of the medium but are highly sensitive to the content and
character of identities, and the context in which they are expressed.
The variable effect of anonymity is a good example of this. Whereas
in many cases anonymity accentuates the salience of group identity, the
example of gender again shows this is not always the case. Second, the
very fact that the effects of technological features depend on the salience
of particular identities, and on the content of the norms and stereotypes
that accompany them, shows that the analysis of the social agents who
use the technology always deserves priority over how the technology may
structure or constrain them. Having said this, we think we have provided
a powerful theoretical framework that should help policy makers and
practitioners to anticipate the effects of these technologies and tune them
to the needs of users.
References
Brunsting, S., and Postmes, T. (2002). Social movement participation in the
digital age – Predicting offline and online collective action. Small Group
Research, 33, 525–554.
De Wolf, M. (2002). Dissimulatie, zelfselectie en sekse (Dissimulation, self-
selection, and sex). Unpublished Masters thesis, University of Amsterdam.
Diener, E. (1980). Deindividuation: The absence of self-awareness and self-
regulation in group members. In P. Paulus (Ed.), The psychology of group
influence (pp. 209–242). Hillsdale, NJ: Erlbaum.
Dubrovsky, V. J., Kiesler, S., and Sethna, B. N. (1991). The equalization phe-
nomenon: Status effects in computer-mediated and face-to-face decision-
making groups. Human Computer Interaction, 6, 119–146.
Fallows, D. (2005). How women and men use the internet (Pew Internet
and American Life Project Rep.). www.pewinternet.org/Reports/2005/
How-Women-and-Men-Use-the-Internet/05-Patterns-of-Internet-Use.
aspx?view=all
Festinger, L., Pepitone, A., and Newcomb, T. (1952). Some consequences of
de-individuation in a group. Journal of Abnormal and Social Psychology, 47,
382–389.
Gandy, O. H., Jr. (1993). The panoptic sort: A political economy of personal infor-
mation. Oxford: Westview.
Haraway, D. (1990). A manifesto for cyborgs: Science technology, and social-
ist feminism in the 1980s. In L. Nicholson (Ed.), Feminism/postmodernism
(pp. 190–233). New York: Routledge.
Joinson, A. N. (2005). Who’s watching you? Power, personalization and on-
line compliance. Paper presented at Conference on Group Processes in
Computer Supported Interaction: Technological and Social Determinism.
Miami University, Oxford, OH.
Kiesler, S., Siegel, J., and McGuire, T. (1984). Social psychological aspects
of computer-mediated communications. American Psychologist, 39, 1123–
1134.
Lea, M. (1991). Rationalist assumptions in cross-media comparisons of
computer-mediated communication. Behaviour & Information Technology,
10, 153–172.
Lea, M., and Spears, R. (1991). Computer mediated communication, de-
individuation, and group decision-making. International Journal of Man–
Machine Studies, 34, 283–301.
Lea, M., Spears, R., and De Groot, D. (2001). Knowing me, knowing you: Effects
of visual anonymity on self-categorization, stereotyping and attraction in
computer-mediated groups. Personality and Social Psychology Bulletin, 27,
526–537.
Lea, M., Spears, R., and Watt, S. E. (2007). Visibility and anonymity effects on
attraction and group cohesiveness. European Journal of Social Psychology, 37,
761–773.
Le Bon, G. (1995). The crowd: A study of the popular mind. London: Transaction
Publishers. (Original work published in 1895.)
Oakes, P. J. (1987). The salience of social categories. In J. C. Turner, M. A.
Hogg, P. J. Oakes, S. D. Reicher, and M. S. Wetherell (Eds.), Rediscover-
ing the social group: A self-categorization theory (pp. 117–141). Oxford: Basil
Blackwell.
Olivetti, C., and Petrongolo, B. (2008). Unequal pay or unequal employment?
A cross-country analysis of gender gaps. Journal of Labor Economics, 26,
621–654.
Postmes, T., and Brunsting, S. (2002). Collective action in the age of inter-
net: Mass communication and online mobilization. Social Science Computer
Review, 20, 290–301.
Postmes, T., and Spears, R. (1998). Deindividuation and anti-normative behav-
ior: A meta-analysis. Psychological Bulletin, 123, 238–259.
(2002). Behavior online: Does anonymous computer communication reduce
gender inequality? Personality and Social Psychology Bulletin, 28, 1073–1083.
Postmes, T., Spears, R., and Lea, M. (1998). Breaching or building social bound-
aries? SIDE-effects of computer mediated communication. Communication
Research, 25, 689–715.
(2002). Intergroup differentiation in computer-mediated communication:
Effects of depersonalization. Group Dynamics: Theory, Research and Practice,
6, 3–16.
Postmes, T., Spears, R., Lee, A. T., and Novak, R. J. (2005). Individuality and
social influence in groups: Inductive and deductive routes to group identity.
Journal of Personality and Social Psychology, 89, 747–763.
Postmes, T., Spears, R., Sakhel, K., and De Groot, D. (2001). Social influence
in computer-mediated groups: The effects of anonymity on social behavior.
Personality and Social Psychology Bulletin, 27, 1243–1254.
Prentice-Dunn, S., and Rogers, R. W. (1989). Deindividuation and the self-
regulation of behavior. In P. B. Paulus (Ed.), The Psychology of Group Influence
(2nd ed., pp. 86–109). Hillsdale, NJ: Lawrence Erlbaum.
Reicher, S. D. (1987). Crowd behaviour as social action. In J. C. Turner, M. A.

Hogg, P. J. Oakes, S. D. Reicher, and M. S. Wetherell (Eds.), Rediscover-
ing the social group: A self-categorization theory (pp. 171–202). Oxford: Basil
Blackwell.
Reicher, S. D., and Levine, M. (1994a). Deindividuation, power relations
between groups and the expression of social identity: The effects of visi-
bility to the out-group. British Journal of Social Psychology, 33, 145–163.
(1994b). On the consequences of deindividuation manipulations for the strate-
gic considerations of self: Identifiability and the presentation of social iden-
tity. European Journal of Social Psychology, 24, 511–524.
Reicher, S. D., Spears, R., and Postmes, T. (1995). A social identity model of
deindividuation phenomena. European Review of Social Psychology, 6, 161–
198.
Ridgeway, C. L., and Smith-Lovin, L. (1999). The gender system and interac-
tion. Annual Review of Sociology, 25, 191–216.
Sherman, R. C., End, C., Kraan, E., Cole, A., Campbell, J., Birchmeier, Z.,
and Klausner, J. (2000). The internet gender gap among college students:
Forgotten but not gone? CyberPsychology & Behavior, 3(5), 885–894.
Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. (1986). Group processes in
computer-mediated communication. Organizational Behaviour and Human
Decision Processes, 37, 157–187.
Spears, R., and Lea, M. (1992). Social influence and the influence of the
“social” in computer-mediated communication. In M. Lea (Ed.), Con-
texts of computer-mediated communication (pp. 30–65). Hemel Hempstead:
Harvester-Wheatsheaf.
(1994). Panacea or panopticon? The hidden power in computer-mediated com-
munication. Communication Research, 21, 427–459.
Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., and Ter Haar, W. (2002).
Computer-mediated communication as a channel for social resistance: The
strategic side of SIDE. Small Group Research, 33, 555–574.
Spears, R., Lea, M., and Lee, S. (1990). De-individuation and group polarization
in computer-mediated communication. British Journal of Social Psychology,
29, 121–134.
Spears, R., Lea, M., and Postmes, T. (2001). Computer-mediated communica-
tion: Social pain or social gain? In P. Robinson and H. Giles (Eds.), The
handbook of language and social psychology (pp. 601–623). Chichester: Wiley.
(2007). Computer-mediated communication and social identity. In A. N. Join-
son, K. Y. A. McKenna, T. Postmes, and J. E. Katz (Eds.), The Oxford
handbook of internet psychology (pp. 253–269). Oxford: Oxford University
Press.
Spears, R., Lea, M., Rogers, P., Van Den Besselaar, P., and Ekkel, L. Gender
concealment in the organization: An international comparison. Unpublished
data, COMMORG, University of Amsterdam.
Spears, R., Postmes, T., Lea, M., and Watt, S. E. (2001). A SIDE view of social
influence. In J. Forgas and K. Williams (Eds.), Social influence: Direct and
indirect processes (pp. 331–350). Philadelphia, PA: Psychology Press.
Spears, R., Postmes, T., Lea, M., and Wolbert, A. (2002). When are net effects
gross products? The power of influence and the influence of power in
computer-mediated communication. The Journal of Social Issues, 58, 91–
107.
Spears, R., Wolbert, A., and De Wolf, M. (in prep). Girls will be boys? Gender
dissimulation and the Internet. Unpublished ms, Cardiff University.
Srull, T. K., and Wyer, R. S., Jr. (1979). The role of category accessibility in
the interpretation of information about persons: Some determinants and
implications. Journal of Personality and Social Psychology, 37, 1660–1672.
Tajfel, H., and Turner, J. C. (1986). The social identity theory of intergroup
behavior. In S. Worchel and W. G. Austin (Eds.), Psychology of intergroup
relations (pp. 7–24). Chicago, IL: Nelson Hall.
Tanis, M., and Postmes, T. (2008). Cues to identity in online dyads: Effects of
interpersonal versus intragroup perceptions on performance. Group Dynam-
ics: Theory, Research, and Practice, 12, 96–111.
Thomson, R., and Murachver, T. (2001). Predicting gender from electronic
discourse. British Journal of Social Psychology, 40, 193–208.
Turkle, S. (1995). Life on the screen: Identity in the age of the internet. New York:
Simon and Schuster.
Turner, J. C. (1982). Towards a cognitive redefinition of the group. In
H. Tajfel (Ed.), Social identity and intergroup relations (pp. 15–40).
Cambridge: Cambridge University Press.
(1991). Social influence. Milton Keynes: Open University Press.
Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., and Wetherell, M.
S. (1987). Rediscovering the social group: A self-categorization theory. Oxford:
Basil Blackwell.
Turner, J. C., and Oakes, P. J. (1986). The significance of the social identity
concept for social psychology with reference to individualism, interactionism
and social influence. British Journal of Social Psychology, 25, 237–252.
Wolbert, A. (2000). Gender in CMC: Anonymity, power and dissimulation.
Unpublished Masters thesis, University of Amsterdam.
Zimbardo, P. G. (1969). The human choice: Individuation, reason, and order
vs. deindividuation, impulse, and chaos. In W. J. Arnold and D. Levine
(Eds.), Nebraska symposium on motivation (Vol. 17, pp. 237–307). Lincoln,
NE: University of Nebraska Press.
3 Trust, deception, and identity on the internet
Melanie C. Green and Jordan M. Carpenter
The internet is a defining feature of modern life. Individuals increas-

ingly conduct social interactions online or use other forms of medi-
ated communication, such as text messaging. Understanding how this
communication medium affects interpersonal interactions and relation-
ships is a key task for psychologists. Specifically, researchers have been
interested in the extent to which individuals are willing to trust oth-
ers that they encounter online, and the degree to which this trust is
justified.
Developing trusting relationships is important if connections forged
online are to move to deeper levels. Trusting relationships may be more
likely to provide social support and other beneficial outcomes. Of course,
just as in offline relationships, appropriate calibration of trust levels is
also important. If individuals are too trusting in their online interactions,
others may take advantage of them. An optimal level of trust is one in
which the person is open to gaining deep and fulfilling relationships, but
is not gullible enough to be cheated or taken advantage of (Rotter, 1980;
Yamagishi, 2001).
Trust stems from both dispositional and situational factors. For exam-
ple, some individuals are dispositionally more trusting, due to factors
such as attachment styles established in childhood (e.g., Fraley and
Shaver, 2000). Situational factors such as amount of risk in a particular
situation, the presence or absence of institutional safeguards, and actions
of the interaction partner (e.g., commitment and pro-relationship behav-
iors; Wieselquist et al., 1999) may also influence willingness to extend
trust. Furthermore, these factors may interact: individual propensity to
trust may have a stronger influence in situations where information about
trustworthiness is ambiguous, rather than when an interaction partner is
clearly trustworthy or untrustworthy (Gill et al., 2005). One important
situational factor in online relationships is the presence and characteris-
tics of the technology itself. This chapter will highlight the influence of
situational and dispositional factors on the way in which trust develops
or fails to develop in online contacts.
40
Trust, deception, and identity on the internet 41
Our focus here is on social interactions: for example, conversing with

other individuals on a website or message board, sending (and receiving)
email, or communicating with friends or strangers over Instant Messen-
ger (IM) or social networking sites such as Facebook. These interactions
are in the service of relationship building, entertainment, or sharing infor-
mation about common interests (e.g., hobbies, sports teams, political
issues). Although trust in business or financial contexts (e.g., decisions
to use a credit card online, ways of deciding to trust a particular online
merchant) is an important topic, it is beyond the scope of this chapter,
as is the development of trust in work teams (e.g., Wilson et al., 2006).
Definition of trust
The concept of trust has been explored across disciplines (including phi-
losophy, economics, marketing, and psychology), and researchers have
not always shared a common understanding of the term. From a psy-
chological perspective, trust can be broadly defined as “confident expec-
tations of positive outcomes from an intimate partner” (Holmes and
Rempel, 1989, p. 188). Trusting individuals expect that a relationship
partner will act with their best interests at heart. More specifically, trust
can be conceptualized as a three-part relation, involving two individuals
and an action: a person trusts another person to do (or not do) a specific
action (Hardin, 2001). This concept inherently captures a person by sit-
uation interaction; a person’s trust in another may derive in part from
a general optimism about the intentions of others, but is also strongly
determined by the situation (characteristics of the partner, degree of risk
involved in trusting the other, and so on).
Global versus specific trust

Global trust refers to trusting a particular partner across a variety of
domains. In contrast, specific trust means that a person may be trusted
only for a circumscribed set of activities (e.g., a person might trust their
auto mechanic to fix their car properly, but might not trust that person
with personal or private information). Interestingly, research in our lab
suggests that trust in internet friendships or romantic relationships is a
global feeling towards the internet partner (Green, 2005). In a set of
survey studies, participants who had formed a relationship online were
asked how much they trusted their partner. They also reported how
much they trusted their partner in specific domains: emotional support,
honesty in interactions, keeping a secret, physical safety, lending money,
expectation of repayment of money, and inviting the person to one’s
42 Melanie C. Green and Jordan M. Carpenter
home. Responses to these items were highly intercorrelated and formed

a reliable trust scale. If a partner was trusted for emotional support,
that partner also tended to be trusted to not harm the person physically
and to repay money lent to him/her. The relationships examined in this
study were relatively enduring ones, lasting on average a year or more.
However, similar results emerged when people in an experimental study
were asked to imagine meeting someone online.
Of course, these findings carry the caveat that individuals’ self-reported
trust or responses to hypothetical situations may differ from their actual
behavior. Furthermore, global trust may be most likely when some kind of
relationship (continued contact over time) has been established. In other
contexts where contact is more fleeting (for example, a travel advice board
with many short-term members who may simply be seeking advice or
planning a specific trip), individuals may be more discriminating in how
much trust they wish to extend; they might trust another member’s report
about a hotel or restaurant, but would not think of lending that person
money. This distinction, in which longer-term relationships lead to global
trust but more transitory relationships may lead to specific or domain-
limited trust, remains to be tested empirically. In other words, future
research may explore the effect of familiarity or length of acquaintance
on the specific versus global nature of trust.
Development of trust
Although trust in continuing relationships may be a relatively global ori-
entation toward another person (the judgment that the person will act
with one’s best interests at heart), trust is not an all-or-nothing deci-
sion. Rather, trust develops in stages as relationships progress (Holmes
and Rempel, 1989). Although trust may grow over time in both online
and offline relationships, the unique aspects of internet communica-
tion change some aspects of this process. For example, revealing per-
sonal appearance and other identity details typically takes place in stages
in relationships that begin on the internet. Those stages can include
adding modes of communication, such as the telephone (e.g., McKenna
et al., 2002). As relationships progress, individuals are willing to take
greater levels of risk, such as revealing more intimate information,
with the idea that the benefits gained, such as relationship closeness
or social support, will also be greater (e.g., Altman and Taylor, 1973,
Chapter 2).
There is evidence that the online context can either diminish or accel-
erate the progression of trust. For example, one study found that in-
person friendships are more trusting and generally higher in quality than
online friendships in the early stages, but that this disparity is reduced
in more extended relationships (one year or more; Chan and Cheng,
2004). However, other work, such as Walther’s “hyperpersonal” per-
spective, suggests that internet relationships can progress to intimate
levels more rapidly than in-person relationships (McKenna et al., 2002;
Walther, 1996). Because online partners can be separate from an indi-
vidual’s in-person social circle, there is less potential cost to sharing
intimate or sensitive information with online partners. This information
is less likely to be passed on to one’s offline friends or family members
(cf. Derlega and Chaikin, 1977). (For example, if an individual dis-
closes being homosexual to conservative friends, they may risk rejection
and exclusion from their social groups; if individuals complain about an
annoying relative to other family members, that relative might find out
about the insult.) Therefore, the person may be less concerned about
and expect less impact from consequences of disclosure, such as embar-
rassment, ostracism, or other social costs. The reduced social presence
created by the online context (e.g., Short et al., 1976) may also reduce
the emotional impact of any negative interactions. Of course, recent
work on ostracism suggests that even cyberostracism can be quite painful
(e.g., Williams et al., 2000), though not always as damaging as in-person
ostracism (Williams et al., 2002). Disclosure is associated with trust, so
conditions that foster high disclosure, such as low levels of shyness, high
levels of perceived security, and interacting with others who are them-
selves self-disclosing (e.g., Joinson, 2001), may also be those that evoke
greater trust.
However, this previously identified progression from relatively impov-
erished forms of information (pseudonyms, text) to richer and more per-
sonal forms of contact is likely to change with new technological devel-
opments and usage patterns, at least in some online contexts. Dating and
social networking sites, for example, typically include photographs along
with personal profiles, and provide information about hobbies, musical
tastes, and even (for some matchmaking sites) income range. Of course,
with dating sites in particular, the expectation is that compatible partners
will quickly move to in-person meetings.
Personal websites also often include a wealth of detail. For example,
people may reveal intimate details on their blogs or online journals, with-
out developing a personal or trusting relationship with each individual
reader. Indeed, bloggers may not be able to identify who or even how
many people read the intimate details that are posted. Some individ-
uals may hope to create trust through their disclosures as a means of
building relationships with others, whereas others may be seeking vali-
dation, attention, or other goals that do not require trust. The role of
communicator motivations in evoking trust from others is a potentially

fruitful area for future research, especially in the online context.
Effects of anonymity
Many researchers have commented on the “paradox” of computer-
mediated communication: that its relative anonymity allows easy decep-
tion and lowered moral standards, and yet in practice, that same
anonymity allows users to express themselves more honestly (Caspi and
Gorsky, 2006). The invisibility and anonymity characterizing much inter-
net communication creates a sense of privacy and autonomy which results
in higher levels of self-disclosure, an important factor in trust formation
(Christopherson, 2007).
For example, McKenna and Bargh (1998) discovered that many mem-
bers in an online gay and lesbian community joined the group long before
coming out to families and friends, and that membership of such groups
often resulted in higher levels of acceptance regarding their homosexual
identities, implying that the internet allows people to “practice” being
open about certain aspects of themselves before they feel comfortable
doing so in other contexts. However, this anonymity still makes many
users uneasy: as personal, one-on-one relationships develop out of online
interaction, users still appreciate that their partners gradually strip away
their anonymity by providing more and more information about them-
selves and eventually provide other means of contact, such as a telephone
number or address, as a sign of trust (Whitty and Gavin, 2001).
The relationship between anonymity and self-disclosure can be some-
what complicated, however. For one thing, not only is the potential dis-
closer of personal information anonymous, the social partner may be as
well. While being anonymous oneself reduces social anxiety, having an
interaction with an anonymous partner hampers the intimacy of personal
disclosures and reduces relationship closeness (Sato and Yoshida, 2008).
However, interacting with anonymous others does not always carry a
pall of distrust and anxiety: for example, one study found that on health
websites, anonymous sources were as trusted as cited ones (Rains, 2007).
Online interactions within existing relationships

When internet use was relatively less common, individuals were likely
to encounter mostly strangers on the internet. While the internet still
provides the opportunity to connect with individuals across the globe,
using the internet to communicate with existing friends and family mem-
bers is becoming increasingly common, especially among young people.
Research has drawn mixed conclusions about whether the addition of

online contact to existing relationships is beneficial or harmful. Some
research suggests that supplementing in-person relationships with online
contact can improve the quality of friendships (Gross et al., 2002; Valken-
burg and Peter, 2007). More broadly, some studies have shown that
informational uses of the internet can be a positive influence on social
capital (e.g., Shah, Kwak, and Holbert, 2001; Shah, McLeod, and Yoon,
2001; Shah et al., 2002). Other research suggests that internet use can
have some costs. In one study of college students, frequent users of
Instant Messenger also reported lower well-being (Green et al., 2005);
internet and television use also does not appear to develop social capital-
relevant skills such as public speaking and negotiation as well as in-person
interaction (Green and Brock, 2008). A longitudinal study of internet
use suggested that in-person contact fostered additional communication
between individuals, but email did not (Shklovski et al., 2004).
One potential explanation for these mixed conclusions is the influence
of individual differences and people’s idiosyncratic goals while using the
internet. For instance, research has indicated that explicitly social behav-
iors on the website Facebook (i.e., messaging a friend to schedule dinner)
emerged out of a curiosity about people’s perspectives, while more isolat-
ing behaviors on Facebook (i.e., searching strangers’ profiles) emerged
out of a defensiveness against other people’s perspectives (Carpenter et al.,
2011).
Additional influences on trust
Dispositional tendencies
All else equal, individuals may be relatively more or less trusting due to
differing dispositional levels of trust (chronic optimistic or pessimistic
outlooks established by temperament or early experience). Dispositional
trust may make individuals more open to interpersonal risks online just as
they would be in traditional social settings, but there is no evidence thus
far that dispositional trust interacts in a particular way with the internet
context. However, other individual differences such as self-efficacy with
computers or knowledge about the internet may also influence whether
individuals are willing to extend trust to online interaction partners.
Individuals who do not feel confident in their ability to manage issues
such as maintaining appropriate levels of privacy or resisting the appeals
of con artists may be more likely to adopt a blanket attitude of distrust
as a means of self-protection.
Generalized trust of others “on the internet”

The context of the internet itself is a powerful situational factor influenc-
ing individuals’ generalized trust in others. Perhaps because of the diffi-
culty in verifying some forms of information online (or perhaps because
of frequent warnings about online dangers, such as theft of personal
information or the possibility of being deceived), many people appear to
have a general wariness toward online others. In response to standard
trust questions such as “Generally speaking, would you say that most
people [on the internet] can be trusted, or can’t you be too careful?,”
people reported significantly less trust in people on the internet than
they did about people more generally (Green, 2005). Studies of bargain-
ing show similar effects. Online negotiators have lower pre-negotiation
levels of trust relative to in-person negotiators (Naquin and Paulson,
2003). Thus, the situational variation between online and in-person con-
texts appears to create different expectations for the trustworthiness of
potential interaction partners.
Deception
One reason why individuals may be suspicious of online others is because
of well-publicized examples of dangerous internet deception. News out-
lets feature stories about the capture of pedophiles who solicit young
potential victims on chat sites, for example. At the time of writing,
headlines were trumpeting the case of a teenager who was driven to
suicide after the mother of an acquaintance allegedly pretended to be a
teenage boy online, and cruelly rejected the girl after gaining her affection
(Taylor, 2007). The salience of these reports may make the internet seem
like a hotbed of deception, even though such dramatic cases are relatively
rare.
Deception is traditionally considered a shameful or at least unseemly
activity, and the mere word calls up images of sinister and Machiavellian
manipulators. Indeed, Satan himself is referred to as “the prince of lies.”
However, the truth is that people generally have a nuanced and sophisti-
cated view of lying. Fabrications can emerge from a variety of motivations,
and the perception of these motivations can often determine the accept-
ability of the deceptive behavior. Lying to benefit the self at the expense
of others carries a high stigma, whereas lying to help others is more
acceptable. Lies can even be self-serving and still socially acceptable as
long as they do not directly harm others, since such behavior is often seen
as a fact of everyday life (Goffman, 1974). The new technology of the
internet provides fresh challenges to the study of deception, since it may
remain fundamentally the same activity as face-to-face lying but carries

with it new contexts, motives, and opportunities to deceive.
Deception in everyday life

DePaulo et al. (1996) discovered that, despite the stigma surrounding
deception, it remains a mundane activity that most commonly embodies
the popular idea of the “white lie.” Most people lie once or twice daily
in a full 20–30 percent of their (non-internet) interactions. Though the
majority of lies are specifically self-serving, this rarely implies monetary
or physical gain but rather intangible benefits such as self-esteem or a
boosted social status. Almost as common are lies specifically intended
to benefit the person being deceived, usually to avoid hurting their
feelings. In both cases, most people lie “up,” that is, they lie to exag-
gerate their positive emotions and avoid expressing negative ones. Lies
require little cognitive preparation, and liars generally feel minimal anxi-
ety about their lies being exposed, despite the fact that most lies are bla-
tant untruths rather than subtle, misleading implications (DePaulo et al.,
1996).
Still, face-to-face deception is not as harmless as many believe: it is
usually accompanied by feelings of discomfort, and interactions featur-
ing lies are less intimate and meaningful than those without falsehood.
Furthermore, both in online and face-to-face contexts, individuals who
lie to a partner end up trusting that partner less (Green, 2005; Sagarin
et al., 1998).
Deception online
Attempts to apply deception research to online interactions suggest that
there are important differences between online and in-person decep-
tion. One important situational factor is that cyberspace communication
allows for deception that would be easily caught in face-to-face interac-
tions, both due to the lack of nonverbal cues and the fact that attributes
such as gender, location, and appearance are not readily apparent. Not
only does the online situation affect the potential for detecting decep-
tion, it also affects the motives people have for lying to others. Like
face-to-face deception, the perceived reasons for lying determine the
lie’s acceptability, but the reasons for online deception differ from those
offline (Utz, 2005). Similarly, certain individual traits may take on dif-
ferent importance in predicting online versus face-to-face deception. For
instance, Lu (2008) found that individuals high in attention-seeking and
internet dependency were more likely to deceive in a specifically online

context.
Reactions to internet deception

Some cyberspace liars have been involved in highly dramatic and bizarre
circumstances, such as the case of people inventing and maintaining
fictional illnesses, tragedies, and entire people (Joinson and Dietz-Uhler,
2002; Van Gelder, 1985). Because such elaborate lies require time to
build and maintain, they can be perpetrated by individuals who are active,
high-status members of online communities, and their peers often feel
intensely betrayed and angry if the lies are exposed, derogating the liar as
mentally ill or cruel even if that is not the case (Joinson and Dietz-Uhler,
2002).
Motivations for deception

Self-protection and identity play. In offline communication, lies are mostly
pragmatic attempts to smooth over mundane interaction, and liars gen-
erally have practical, immediate reasons for telling untruths. However,
lying on the internet tends to come from an entirely different set of
motivations (Caspi and Gorsky, 2006). Deception is more commonly
self-serving than in face-to-face interactions, but not for self-esteem or
monetary reasons. A survey by Caspi and Gorsky (2006) found that the
most frequent motivation for lying online was self-protection, usually
reported by women who lie about their location or marital status to deter
threats to their privacy or safety. In fact, lies about location were almost
twice as common as other types of lies for this reason.
The next most common reason for lying was participation in “identity
play,” which motivates users to present themselves as being of a different
gender, age, sexual orientation, or occupation than is actually the case,
allowing them to create an idealized identity with which to address the
world or simply to experiment and play (Turkle, 1995). Some of these
experiments carry potential rewards: men are more likely than women
to lie about their jobs or level of income, perhaps in an effort to appear
attractive to potential partners (Whitty, 2002). Even outlandish lies such
as these tend to carry few emotional repercussions on the internet, and
in fact many deceivers admit feeling good about their lies (Caspi and
Gorsky, 2006). Still, in most cases, these lies do not carry malicious
intent, and individuals can generally accurately guess the motivations for
online deception. Users forgive even lies about such basic information as
gender if they attribute it to safety concerns or identity play, rather than

more sinister motives (Utz, 2005).
Deception versus self-presentation

Whitty and Gavin (2001) theorized that the anonymity and disinhibi-
tion on the internet allow users to express a different “self ” than in the
everyday world. Bargh et al. (2002) took this a step further by suggesting
that self-presentation on the internet actually provides users with more
opportunities to express their “true selves,” which may differ drastically
from the identity they present in face-to-face interactions. The resulting
behavior may seem deceptive or at least misleading, but it is nonetheless
genuine and in fact provides the backbone for much social behavior on
the internet, including healthy relationship formation.
In an interesting study on this phenomenon, Vazire and Gosling (2004)
examined the veracity of self-presentation on personal websites. Such
examination is useful in this area since the primary function of these sites
is to communicate deliberate and coherent messages about their owners.
They found that websites do communicate stable impressions of their
creators to viewers; perhaps more surprisingly, these impressions are gen-
erally accurate, though they tend to give heightened impressions of some
traits, usually extraversion and agreeableness. These results imply again
that, despite the anonymity and the high levels of deliberate as opposed
to accidental self-expression on the internet, users tend to present stable
images of themselves that are a mixture of their true and idealized selves,
though they are also generally objectively accurate.
Beliefs about internet deception

Despite these findings, which reveal that internet deception is no more
common than the face-to-face kind, there is still a common perception
of the internet as a kind of wilderness where deception is rampant and
no one can be trusted. Indeed, even common users of the internet dra-
matically overestimated the amount of lying, though they themselves
lie or encounter lies relatively infrequently (Caspi and Gorsky, 2006;
Green, 2005). Caspi and Gorsky theorized that this is the result of the
confirmation bias. As noted at the start of this section, publicized and
dramatic instances of online deception may have resulted in mundane
lying becoming more salient to the average user.
There is not a direct relationship between general trust toward online
communication and the actual amount of deception, and many individ-
uals feel suspicion very easily toward information that cannot easily be
confirmed, even if this information is provided in an environment where

most people are honest. The social organization of an online community
can limit this effect, but members are still aware of the potential danger:
because of the openness, users in online communities assume others to
be honest, but because of the potential for deception, those who violate
that trust may be violently denounced (Joinson and Dietz-Uhler, 2002).
Of course, just as in in-person situations, deception can also sometimes
be forgiven (e.g., Birchmeier, Joinson, and Dietz-Uhler, 2005).
Lie detection
Part of the anxiety about internet communication involves the limited
amount of information present in many online interactions, which exac-
erbates people’s already poor ability to spot lies. Obviously, successful
deception depends on the listener having a limited amount of informa-
tion about both the liar and the subject of the lie. Hearing a battered car
described as pristine is perfectly reasonable, unless the car is visible or the
listener knows the speaker is, for instance, a used car salesman. Beyond
that, most people, even trained investigators, have difficulty spotting liars
(DePaulo, 1994). There are some cues that often appear when any given
individual lies, such as dilated pupils, an increased rate of blinking, a
higher pitched voice, and a shorter, more hesitant way of speaking and
phrasing sentences. However, these cues become much more difficult
to notice when the liar is highly motivated to deceive, and they are not
present in many lies. These cues often vary drastically from person to
person: even after being trained to spot lies in one person, individuals are
poor at generalizing that skill in response to others. Still, liars’ uncon-
scious behavioral cues can inspire reactions of doubt and discomfort in
listeners, even if it does not make them good at determining, in a vacuum,
who is lying and who is not (DePaulo, 1994). Without these cues, and
without the clues that come from context, most people would truly be
blind at noticing deception. This has important implications for online
communication, which frequently eliminates non-verbal cues such as rate
of blinking. The lowered generalized trust in “people on the internet”
(as described above) may be a reaction to this absence of deception cues;
people may feel that they are less able to assess truth-telling accuracy
online, and so develop a self-protective suspicion of others.
Despite this handicap, some researchers have found promising pat-
terns which may start to uncover clues to typical online deception. Zhou
et al. (2004) discovered that there may be predictable patterns in the use
of language dominance, which is a state of attempting to overtly influ-
ence the behavior of others through communication, in online deception.
Deceivers tend to begin interactions by using submissive language and

then steadily increasing dominance in their use of language over time.
Also, users tend to more naturally and easily detect deception in certain
modes of online communication. People had a more difficult time recog-
nizing deception in the typing modality, in which messages are displayed
to the receiver as they are being typed, than in the messaging or chatting
modalities, in which the messages appear fully formed (Zhou and Zhang,
2007).
Choice of communication modes

An additional complexity arises when individuals have the option to
communicate online, in person, or by telephone. People may use the
internet strategically for deception in order to avoid sending nonverbal
signals of deceit. Individuals may generally trust an interaction partner,
but may become suspicious of a particular communication (e.g., “why
did he email to cancel our meeting rather than calling me?”). Future
research could explore the influence of suspicion about discrete events
or statements on overall trust.
Recent research in our laboratory suggests that individuals may view
online communication as less appropriate for communications requiring
high trust (Crittenden and Green, 2007). Individuals were provided with
information that a friend might wish to communicate to them. Some of
this information was important (e.g., a mutual friend being in the hos-
pital), while some was trivial (e.g., that the friend saw a famous actor
at the mall). Similarly, some of this information required high trust –
that is, could not be verified otherwise (e.g., the friend saw the person’s
boyfriend/girlfriend kissing someone else) – while other information did
not require this trust (e.g., that the dean of your department is retir-
ing). Participants were randomly assigned to one of four communication
methods – in person, email, Instant Messenger, or telephone – and
rated how appropriate they thought it would be for their friend to use
this method, and how much they would like receiving the information
via this method. For important information or information which
required trust, individuals preferred in-person communication. For less
important or low-trust information, individuals preferred Instant Mes-
senger or the telephone. Individuals appeared to be trading off the conve-
nience of mediated communication (telephone or IM) with the additional
information provided by in-person communication. Recent research by
Whitty and Carville (2008) may also suggest that the nature of the social
relationship plays a role regarding deception and choice of medium. Indi-
viduals are more likely to choose email over the telephone or face-to-face
contact when telling self-serving lies, but when telling other-oriented lies
(for instance, sparing another’s feelings) to a friend, they do not prefer
any medium to another.
Deception summary
Online deception is a complex phenomenon emerging not only from
people’s motivation to tell untruths but also from the particular oppor-
tunities in computer-mediated communication and specific personality
traits that may lead to deception. Among the situational factors, the invis-
ibility of the internet gives people the chance to lie about a wider variety
of things than would be possible otherwise, while the distance of text-
based communication allows for more types of self-expression, which
may involve untruths. Among the person factors, lower levels of trust,
a motivation to deceive, and attention-seeking may all lead to online
dishonesty. Together, these factors could help to explain why people
lie online. The difficulty in reading people’s intentions online may
decrease users’ inclinations to trust others, perhaps leading to more self-
protective lies. Similarly, a specific desire to engage identity play may be
especially easy to sate given the relative anonymity of the internet, and
likewise that anonymity may inspire people to play with their identities.
Levels of information
Early research on internet interaction began with the premise that online
communication inherently involves less information than face-to-face
or telephone interactions, since cues such as body language and tone
of voice are invisible or absent (e.g., Kiesler et al., 1984; Sproull and
Kiesler, 1986). This assumption is largely correct, but it glosses over one
of the most basic and alluring features of the internet: its protean nature
as technology changes. Whereas in the past communication largely took
the form of purely text-based exchanges such as email and chatrooms,
the internet now offers many means of communication, each of which
carries its own limits on and opportunities for the sharing of information.
Forms of self-presentation
As technology improves, there are more and more ways for internet
users to present information about themselves, some more subtle than
others. Even in highly anonymous settings, such as chatrooms, users
can choose a personal nickname, font, and profile. Some media allow
users to provide visual information such as an icon, and many allow
for the creation of full-fledged avatars, which are some of the most stable
means of online self-disclosure, since they are present in every interaction.
Interestingly, the features of these avatars differ according to the type of
communication. IRC, a form of real-time chatting, is characterized by
high levels of anonymity and freedom, and it inspires users to create
fantastic, highly imaginative avatars. Instant Messenger, which is usually
used between real-life friends and family, inspires more concrete and
realistic avatars. Since the construction of an avatar is an expression
of identity, their features carry clues to self-expression and self-image.
Thus, avatars, and other forms of self-disclosure, are closely tied to an
individual’s objectives, which are in turn tied to the medium they use to
communicate (Kang and Yang, 2006).
Furthermore, even in the absence of avatars, individuals often use
many of the features of online communication, such as editing, sentence
structure, and response time, to deliberately manage the impressions
they make on people and attempt to form specific relationships with
desired partners (Walther, 2007). By matching their behavior with the
expectations tied to the particular method of communication, and by
utilizing self-presentation that corresponds with these expectations (e.g.,
the use of a realistic versus imaginative avatar; responding quickly rather
than slowly to a message), users can appear more trustworthy to their
online partners and acquaintances.
In sum, individuals may have particular impression-management goals
(e.g., appearing competent, appearing attractive), but the way they pur-
sue those goals is influenced by the norms and capabilities of a given
online environment. Thus, the person factors (goals to create a particu-
lar impression) interact with situational affordances to create behavior.
Understanding the role of avatars in creating interpersonal trust is
becoming increasingly important with the growing popularity of virtual
worlds such as Second Life – 3D virtual worlds in which avatars can be
highly customized. Research suggests that, with the growing realism of
these representations, individuals come to expect the nonverbal signals
conveyed by the avatars to mirror face-to-face social signals (Donath,
2007). For example, individuals consciously or unconsciously pay atten-
tion to the gaze of an avatar (Yee et al., 2007). Furthermore, avatars have
an effect on the behavior of the user, not just on how others relate to
him or her socially. Taller avatars can inspire more confident behavior,
while attractive avatars can result in more open self-disclosure than less
attractive avatars (Yee and Bailenson, 2007). This behavior potentially
sets the stage for self-fulfilling prophecies (behavioral confirmation), as
others may act differently toward users with more versus less attractive
avatars, or who engage in more versus less self-disclosure. Understanding
social interactions in these enriched online contexts is a fertile area for

future research.
Providing more “channels”

Interactions characterized by higher levels of information-sharing tend to
involve more trust and a lower opportunity for deception, but there are
limits to that effect online. Webcams, which are becoming increasingly
common, allow users to communicate online while providing a real-time
image of their face (or whatever they choose to place in front of their
camera). This feature has the added benefit of allowing people to pick
up on body language or subtextual cues, perhaps giving them confidence
in their ability to recognize any lies and thus improving trust. However,
in one study, shy and socially phobic individuals displayed less trust and
made significantly fewer self-revelations in webcam interactions than in
purely text-based ones. Individuals who were not shy did not reduce their
self-disclosure behavior, implying that shyness may become “activated”
by the presence of another person’s face (Brunet and Schmidt, 2007;
see also Sheeks and Birchmeier, 2007). Therefore, despite not having
access to potential cues of deception, some people actually behave in a
more trusting manner when receiving less information about their social
partners, and the very traits that inspire these tendencies, such as social
anxiety and shyness, may be part of what causes these people to seek
online interactions and relationships in the first place. This is a clear
example of personality factors (discomfort with social situations) inter-
acting with the situation of being online. The online situation – at least
when it is text-only – removes elements such as the physical presence of
others that might normally evoke negative reactions from shy or socially
anxious individuals.
Social networking sites

Due to the rapid pace of culture and technology, there is an unfortu-
nate dearth of psychological research on some of the newer ways of
communicating online, most notably social networking websites such as
Facebook and MySpace. Interestingly, these popular sites seem to work
against some of the features of previous internet communication, such as
anonymity (most users readily use their real names, and the sites them-
selves are organized by physical location) and invisibility (most users
also provide pictures of themselves, which are prominently displayed).
Thus, the average Facebook page provides dramatically more informa-
tion about its owner than an everyday face-to-face interaction could,
though it is all carefully selected and controlled by both the individual

and the site designers.
We suggest that one fruitful area for future research would be the verac-
ity of social networking profiles and the specific ways in which people use
them to present themselves. For instance, both Facebook and MySpace
prompt users to provide quite specific information, such as their favorite
musicians and movies. Therefore, it is possible that those preferences
are gaining an inflated degree of importance in the social interactions of
internet users, since that is the obvious information through which an
individual is judged. It is possible that lying about one’s favorite band
could become a common form of deception on the internet, though that
is unlikely to damage the foundations of trust in most relationships. Still,
it may not be unreasonable for users to expect that small details can be
important on social networking sites. Perceptions of individuals in such
contexts can be heavily affected by seemingly peripheral information:
for instance, simply receiving public messages from physically attrac-
tive people can make an individual seem more attractive (Walther et al.,
2008).
Interestingly, the rising popularity of Facebook may be reducing one
of its original benefits. When Facebook began, access was limited to
individuals on one’s own campus, or to fellow college students. This
restricted nature of the community may have made individuals feel more
comfortable about interacting with strangers online. Now that Facebook
and other social networking sites do not carry these restrictions, users may
approach Facebook interactions with more caution or distrust (although
that remains to be empirically tested).
Reputation systems and identity verification

In physical communities, one way of determining another individual’s
trustworthiness is by asking others about them. A person’s reputation
provides information about their past behavior, and thus gives a basis for
predicting future trustworthy or untrustworthy behavior. Online reputa-
tion systems, such as eBay’s feedback system, attempt to provide the same
benefits in the online world. Although these systems are most commonly
used for financial websites, they are also used on political blogs, technical
websites, and other discussion arenas. Individuals might be more willing
to trust information from longtime or highly rated fellow members.
Furthermore, identity verification is an important aspect of online
trust. Individuals trust others more when they know another person is
who they say they are, through a mutual acquaintance, seeing a pro-
file on an employer’s website, or other independent verification (Green,
2005). This knowledge gives people additional confidence in any other

information the person has provided.
Applications and strategic implications

Examining online trust within a person-by-situation framework leads to
several important applications and implications. The same online factors
(presence or absence of video images, for instance) may have different
effects depending on the personality traits of the individuals using the
technology.
Training for trust

As individuals gain experience with online interactions, they may become
more knowledgeable about cues to trustworthiness. Much like learning
to identify email scams, individuals might be able to determine whether a
particular online interaction partner is likely to be telling the truth about
important information. Social networking and other websites might pro-
vide guidance to make users more savvy interaction partners. However,
most of the current advice seems geared toward preventing individuals
(especially children and teens) from being too trusting. For example,
young internet users are cautioned not to give too much information
about where they live or go to school, whereas older users are cautioned
against giving information such as social security numbers or financial
account information.
Designs for trust

Certain types of online systems take some of the risk out of an inter-
action. Reputation systems, as mentioned above, provide information
about a user’s past behavior. Websites that provide for some kind of
identity verification may have a similar value in increasing trust. Sites
with consequences for bad behavior (e.g., banning individuals who act
abusively toward others) can provide some safeguards. If encouraging
trust is an important goal for a particular online community or busi-
ness, taking advantage of these existing mechanisms may be a helpful
approach. Finally, the rise of Second Life and other rich virtual reality
systems means that research on subtle or nonverbal factors that influence
interpersonal trust can inform the development of systems to encourage
online trust (such as the studies on gaze discussed above).
Word-of-mouth marketing
Although this chapter has focused primarily on interpersonal relation-
ships, one area for future attention is the rise of viral or word-of-mouth
marketing online. These practices may blur the lines between interper-
sonal sharing and a business pitch. That is, when a fellow website user
recommends a product or service, that person may be doing so sim-
ply because of their own satisfaction with the product, or they may be
receiving some compensation from the company for doing so. Or, in
a more extreme case, the entire online persona may be the creation of
a marketing department. As these practices become more widespread,
policies may need to be developed to regulate them. From a research
perspective, it is an open question how individuals’ knowledge of these
marketing practices affects their willingness to trust others in the realm
of consumer products specifically, but also their general trust in online
others.
Summary and conclusion

The internet is a valuable tool for communicating, building knowledge,
and forging social connections, but along with those virtues come new
opportunities and reasons to mistrust, distrust, and deceive. Computer-
mediated communication often comes with distinct levels of informa-
tion, which can allow for more extravagant lies, as well as making
lies more difficult to detect. These situational factors might seem to
encourage deception. However, as we have discussed, this very free-
dom also inspires honesty and self-disclosure, and despite highly pub-
licized instances of predatory and dangerous deception, most untruths
are intended as self-protective behavior, harmless play, or even a healthy
expression of alternate selves. Thus, personality factors, such as indi-
viduals’ innate honesty or desire for true connections with others (need
for belongingness), also help determine whether individuals are likely
to use the internet to take advantage of others versus connection with
others.
Thus, far from being a dangerous wilderness of duplicity, the internet
may allow individuals to express traits, desires, and facets of themselves
that would be difficult to convey in face-to-face interactions. Likewise,
trust online, while likely to be related to individuals’ general tendencies to
believe in others and self-disclose, is almost certainly filtered through the
idiosyncratic levels of information characterizing the methods of commu-
nication offered by the internet. The situational forces that are present
online – the extent of information available (e.g., text versus video), the
amount of control users have over self-presentation through avatars and

personalized content, and the degree of anonymity, to name just a few –
can interact with dispositional levels of trust to create more or less trusting
behavior. Though researchers face the challenge of racing technological
advancement, the need to understand trust and deception in the increas-
ingly ubiquitous online world is ever more important. As we shed more
light on the complex, dynamic interplay between individual differences
and the internet’s freedom, we can perhaps begin to protect ourselves
from hurtful lies while still enjoying the freedom of expression and open-
ness that have made the internet such a popular means of interaction in
the first place.
References
Altman, I., and Taylor, D. A. (1973). Social penetration: The development of inter-
personal relationships. New York: Holt, Reinhart and Winston.
Bargh, J. A., and McKenna, K. Y. A. (2004). The internet and social life. Annual
Review of Psychology, 55, 573–590.
Bargh, J. A., McKenna, K. Y. A., and Fitzsimons, G. M. (2002). Can you see the
real me? Activation and expression of the “true self” on the internet. Journal
of Social Issues, 58, 33–48.
Birchmeier, Z., Joinson, A. N., and Dietz-Uhler, B. (2005). Storming and form-
ing a normative response to a deception revealed online. Social Science Com-
puter Review, 23(1), 108–121.
Brunet, P. M., and Schmidt, L. A. (2010). Sex differences in the expression and
use of computer-mediated affective language: Does context matter? Social
Science Computer Review, 28(2), 194–205.
Carpenter, J. M., Green, M. C., and LaFlam, J. (2011). People or profiles: Indi-
vidual differences in online social networking use. Personality and Individual
Differences, 50, 538–541.
Caspi, A., and Gorsky, P. (2006). Online deception: Prevalence, motivation, and
emotion. CyberPsychology & Behavior, 9, 54–59.
Chan, D. K. S., and Cheng, G. H. L. (2004). A comparison of offline and online
friendship qualities at different stages of relationship development. Journal
of Social and Personal Relationships, 21(3), 305–320.
Christopherson, K. M. (2007). The positive and negative implications of
anonymity in internet social interactions: “On the internet, nobody knows
you’re a dog.” Computers in Human Behavior, 23(6), 3038–3056.
Crittenden, S., and Green, M. C. (2007). It’s not just what you say: Prefer-
ences for interpersonal communication medium depend on topic. Society
for Personality and Social Psychology meeting, Memphis, January.
DePaulo, B. M. (1994). Spotting lies: Can humans learn to do better? Current
Directions in Psychological Science, 3(3), 83–86.
DePaulo, B. M., Kashy, D. A., Kirkendol, S. E., Wyer, M. M., and Epstein, J.
A. (1996). Lying in everyday life. Journal of Personality and Social Psychology,
70, 979–995.
Derlega, V. J., and Chaikin, A. L. (1977). Privacy and self-disclosure in social
relationships. Journal of Social Issues, 33(3), 102–115.
Donath, J. (2007). Virtually trustworthy. Science, 317(5834), 53–54.
Fraley, R. C., and Shaver, P. R. (2000). Adult romantic attachment: Theoretical
developments, emerging controversies, and unanswered questions. Review
of General Psychology, 4, 132–154.
Gill, H., Boies, K., Finegan, J., and McNally, J. (2005). Antecedents of trust:
Establishing a boundary condition for the relation between propensity to
trust and intention to trust. Journal of Business and Psychology, 19, 287–
302.
Goffman, E. (1974). Frame analysis: An essay on the organization of experience.
Cambridge, MA: Harvard University Press.
Green, M. C. (2005). Trust in internet relationships. Unpublished manuscript.
(2007). Trust and on-line social interaction. In A. Joinson, K. Y. A. McKenna,
T. Postmes, and U. Reips (Eds.), Oxford handbook of internet psychology
(pp. 43–52). Oxford: Oxford University Press.
Green, M. C., and Brock, T. C. (2008). Antecedents and civic consequences of
choosing real versus ersatz social activities. Media Psychology, 11(4), 566–
592.
Green, M. C., Hilken, J., Friedman, H., Grossman, K., Gasiewski, J., Adler, R.,
and Sabini, J. P. (2005). Communication via Instant Messenger: Short and
long-term effects. Journal of Applied Social Psychology, 35(3), 445–462.
Gross, E. F., Juvonen, J., and Gable, S. L. (2002). Internet use and well-being
in adolescence. Journal of Social Issues, 58, 75–90.
Hardin, R. (2001). Conceptions and explanations of trust. In K. S. Cook (Ed.),
Trust in society (pp. 3–39). New York: Russell Sage Foundation.
Holmes, J. G., and Rempel, J. K. (1989). Trust in close relationships. In C.
Hendrick (Ed.), Close relationships (pp. 187–220). Thousand Oaks, CA:
Sage.
Joinson, A. N. (2001). Knowing me, knowing you: Reciprocal self-disclosure in
internet-based surveys. CyberPsychology & Behavior, 4(5), 587–591.
Joinson, A. N., and Dietz-Uhler, B. (2002). Explanations for the perpetration of
and reactions to deception in a virtual community. Social Science Computer
Review, 20(3), 275–289.
Kang, H., and Yang, H. (2006). The visual characteristics of avatars in computer-
mediated communication: Comparison of Internet Relay Chat and Instant
Messenger as of 2003. International Journal of Human–Computer Studies,
64(12), 1173–1183.
Kiesler, S., Siegel, J., and McGuire, T. W. (1984). Social psychological aspects
of computer-mediated communication. American Psychologist, 39(10), 1123–
1134.
Lu, H. (2008). Sensation-seeking, internet dependency, and online interpersonal
deception. CyberPsychology & Behavior, 11, 227–231.
McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the
internet: Identity “de-marginalization” through virtual group participation.
McKenna, K. Y. A., Green, A. S., and Gleason, M. J. (2002). Relationship
formation on the internet: What’s the big attraction? Journal of Social Issues,
58(1), 9–31.
Naquin, C. E., and Paulson, G. D. (2003). Online bargaining and interpersonal
trust. Journal of Applied Psychology, 88(1), 113–120.
Rains, S. (2007). The anonymity effect: The influence of anonymity on per-
ceptions of sources and information on health websites. Journal of Applied
Communication Research, 35(2), 197–214.
Rotter, J. B. (1980). Interpersonal trust, trustworthiness, and gullibility. American
Psychologist, 35(1), 1–7.
Sagarin, B. J., Rhoads, K., and Cialdini, R. B. (1998). Deceiver’s distrust: Den-
igration as a consequence of undiscovered deception. Personality and Social
Psychology Bulletin, 24, 1167–1176.
Sato, H., and Yoshida, F. (2008). Self-disclosure on the internet: The effects of
anonymity on the self and the user. Japanese Journal of Psychology, 78(6),
559–566.
Shah, D. V., Kwak, N., and Holbert, R. L. (2001). “Connecting” and “discon-
necting” with civic life: Patterns of internet use and the production of social
capital. Political Communication, 18(2), 141–162.
Shah, D. V., McLeod, J. M., and Yoon, S. (2001). Communication, context,
and community: An exploration of print, broadcast, and internet influences.
Communication Research, 28(4), 464–506.
Shah, D. V., Schmierbach, M., Hawkins, J., Espino, R., and Donavan, J. (2002).
Nonrecursive models of internet use and community engagement: Ques-
tioning whether time spent online erodes social capital. Journalism & Mass
Communication Quarterly, 79(4), 964–987.
Sheeks, M. S., and Birchmeier, Z. P. (2007). Shyness, sociability, and the use of
computer-mediated communication in relationship development. CyberPsy-
chology & Behavior, 10(1), 64–70.
Shklovski, I., Kraut, R., and Rainie, L. (2004). The internet and social par-
ticipation: Contrasting cross-sectional and longitudinal analyses. Journal of
Computer-Mediated Communication, 10(1).
Short, J., Williams, E., and Christie, B. (1976). The social psychology of telecom-
munications. London: John Wiley.
Sproull, L., and Kiesler, S. (1986). Reducing social context cues: Electronic
mail in organizational communication. Management Science, 32, 1492–
1512.
Taylor, B. (2007). Authorities investigate internet postings posing as
the woman linked to an online hoax as teen. Associated Press.
www.breitbart.com/article.php?id=D8T1D73G0&show_article=1
Turkle, S. (1995). Life on the screen: Identity in the age of the internet. New York:
Touchstone.
Utz, S. (2005). Types of deception and underlying motivation: What people
think. Social Science Computer Review, 23, 49–56.
Valkenburg, P. M., and Peter, J. (2007). Online communication and ado-

lescent well-being: Testing the stimulation versus the displacement
hypothesis. Journal of Computer-Mediated Communication, 12(4), article 2.
http://jcmc.indiana.edu/vol12/issue4/valkenburg.html
Van Gelder, L. (1985). The strange case of the electronic lover: A real-life story
of deception, seduction, and technology. Ms, 14(4), 94, 99, 101–104, 117,
123, 124.
Vazire, S., and Gosling, S. D. (2004). e-Perceptions: Personality impressions
based on personal websites. Journal of Personality and Social Psychology, 87,
123–132.
personal, and hyperpersonal interaction. Communication Research, 23(1), 3–
43.
Walther, J. B. (2007). Selective self-representation in computer-mediated com-
munication: Hyperpersonal dimensions of technology, language, and cogni-
tion. Computers and Human Behavior, 23, 2538–2557.
Walther, J. B., Van Der Heide, B., Sang-Yeon, K., Westerman, D., and Tong, S.
T. (2008). The role of friends’ appearance and behavior on Facebook: Are
we known by the company we keep? Human Communication Research, 34,
28–49.
Whitty, M. T. (2002). Liar, liar! An examination of how open, supportive and
honest people are in chat rooms. Computers in Human Behavior, 18, 343–352.
Whitty, M. T., and Carville, S. E. (2008). Would I lie to you? Self-serving lies
and other-oriented lies told across different media. Computers in Human
Behavior, 24, 1021–1031.
Whitty, M. T., and Gavin, J. (2001). Age/sex/location: Uncovering the social cues
in the development of online relationships. CyberPsychology & Behavior, 4(5),
623–630.
Wieselquist, J., Rusbult, C. E., Agnew, C. R., and Foster, C. A. (1999). Com-
mitment, pro-relationship behavior, and trust in close relationships. Journal
Williams, K. D., Cheung, C. K. T., and Choi, W. (2000). Cyberostracism: Effects
of being ignored over the internet. Journal of Personality and Social Psychology,
79, 748–762.
Williams, K. D., Govan, C. L., Croker, V., Tynan, D., Cruickshank, M., and
Lam, A. (2002). Investigations into differences between social- and cyber-
ostracism. Group Dynamics: Theory, Research and Practice, 6, 65–77.
Wilson, J. M., Straus, S. G., and McEvily, B. (2006). All in due time: The
development of trust in computer-mediated and face-to-face teams. Organi-
zational Behavior and Human Decision Processes, 99, 16–33.
Yamagishi, T. (2001). Trust as a form of social intelligence. In K. S. Cook (Ed.),
Trust in society (pp. 121–147). New York: Russell Sage Foundation.
Yee, B., and Bailenson, J. N. (2007). The Proteus effect: The effect of transformed
self-representation on behavior. Human Communication Research, 33, 271–
290.
Yee, N., Bailenson, J. N., Urbanek, M., Chang, F., and Merget, D. (2007). The
unbearable likeness of being digital: The persistence of nonverbal social
norms in online virtual environments. CyberPsychology & Behavior, 10(1),

115–121.
Zhou, L., Burgoon, J. K., Zhang, D., and Nunamaker, J. F. (2004). Language
dominance in interpersonal deception in computer-mediated communica-
tion. Computers and Human Behavior, 20, 381–402.
Zhou, L., and Zhang, D. (2007). Typing or messaging? Modality effect on decep-
tion detection in computer-mediated communication. Decision Support Sys-
tems, 44, 188–201.
4 An interactional approach to social influence
in computer-mediated communication
Kai Sassenberg
Social interaction leads – online as much as offline – to mutual, more or

less equal social influence of the interacting parties on each other. Inter-
net users send out information, summaries of subjective experiences,
and opinions using communication tools. Recipients acquire uncon-
sciously new insights or strive to learn or form an opinion about a
topic. Research studying the facilitating and hindering factors for such
media-based social influence has a long-standing history (for a sum-
mary of the early work see Williams, 1977). Nonetheless, the interplay
between personality characteristics and media attributes as preconditions
of social influence has hardly received any attention yet. Therefore, the
current chapter will apply interactionism (Endler and Magnusson, 1976)
to online social influence. The combination of personal and situational
variables is relevant in the context of new communication media as in any
other context, because individuals may respond differently to the specific
situational conditions provided by computer-mediated communication
(CMC). (For initial evidence see Amichai-Hamburger et al., 2002; Cress,
2005; Guadagno and Cialdini, 2002; Hertel et al., 2008; Sassenberg
et al., 2005.) Hence, considering individual differences for the prediction
of media effects will allow more precise predictions about the impact of
CMC on social influence.
In this chapter, different types of social influence will be introduced
and the features of CMC that are relevant for social influence will be
discussed. Afterwards, the research on these types of social influence in
CMC will be reviewed. This review serves two purposes. First, it will
show that almost all existing research in this domain has focused on the
impact of situational characteristics that either (a) are inherent in the
technology (e.g., few social context cues) or (b) result from the social
setting in which CMC takes place (e.g., the impact of group norms
on a person’s attitudes). Second, it will outline how the moderation of
The author would like to thank Ulrike Cress, Joachim Kimmerle, and Katrin Wodzicki for
their extraordinarily fruitful comments on this chapter.
63
64 Kai Sassenberg
social influence processes by communication media characteristics is best

explained by the interactionist approach that includes inter-individual
differences (e.g., personality traits). Before this approach is described,
the few studies that take into account the moderating impact of inter-
individual differences between users on the effects of situational charac-
teristics (i.e., personality traits or cognitive styles) will be summarized.
(For a framework model putting forward a similar argument for CMC
research in general see Joinson, 2003.)
Types of social influence

An in-depth understanding of social influence in CMC requires a detailed
picture of the different forms of social influence and the processes
underlying them. Two types of social influence that were originally dis-
tinguished by Deutsch and Gerard (1955) and have been discussed
subsequently by many others (for a summary, see Wood, 1999) are nor-
mative influence (the conformity to others’ expectations in order to reap
social rewards and avoid social punishments) versus informational influ-
ence (conformity to others that results from the assumed validity of pro-
vided information). It has often been suggested that what Deutsch and
Gerard (1955) called normative influence actually refers to compliance
(e.g., Allen, 1965).
Informational influence should be further differentiated in influence
based on group membership and interpersonal influence (see also Sassen-
berg and Jonas, 2007; Sassenberg et al., 2005), based on the assumptions
of the social identity approach (Tajfel and Turner, 1979; Turner et al.,
1987). This approach suggests that the self-concept consists of a personal
identity (i.e., defining idiosyncratic individuals’ aspects) and many social
identities (i.e., derived from membership of social groups). As social
identities are parts of the self, they make individuals internalize group
norms, which in turn affect their thinking and behavior. Hence, this
so-called norm-based influence is not a sign of conformity to social pres-
sure stemming from the group but a self-determined action as a group
member (e.g., members of a sports team strive for its success because
they want to and not because they are forced to). Norm-based influence
should therefore be distinguished from normative influence. The more
that information communicated by other ingroup members is in line with
ingroup norms, the more they will be seen as valid and thus be influential
(Turner, 1991).
Certainly, social influence can also result from information that is
not associated with group norms. This is most likely the case when
Social influence in computer-mediated communication 65
personal identity is salient. The so-called interpersonal influence is fos-

tered by interpersonal bonds (Postmes and Spears, 2000; Sassenberg
and Boos, 2003). Interpersonal bonds can be merely affective or based on
trust (e.g., perceived expertise). Without interpersonal bonds, a salient
personal identity makes individuals strive for distinctiveness from the
communication partners and the fulfillment of other personal needs,
reducing social influence (Spears et al., 1990). All in all, there are three
forms of social influence: compliance, norm-based influence, and inter-
personal influence. The following introduces the features of CMC rele-
vant to social influence.
Characteristics of CMC
CMC and face-to-face (FTF) communication differ in three situational
features that are relevant for social influence (Spears and Lea, 1994):
anonymity of others, anonymity of the self, and physical isolation.
Anonymity of others
In CMC, senders can be more anonymous to a target than they are in
FTF communication, depending on the degree that the communication
medium transmits names, e-mail addresses, nicknames, avatars, paralin-
guistic cues, physical appearance, non-verbal communication (language
use, facial expressions, posture, gesture, etc.), information about identity,
etc.
The social identity model of deindividuation effects (SIDE; Postmes
et al., 1998; Spears and Lea, 1992, 1994; see also Spears et al., this vol-
ume) suggests that the scarce accessibility of information under condi-
tions of anonymity has a strong impact on norm-based and interpersonal
influence (see also Walther, 1996). According to the SIDE model, the
social categorization of the self and other group members as different
individuals (salient personal identity) or as members of the same groups
(salient social identity) is more pronounced in anonymous communica-
tion settings. When social identity is initially salient, group norms will
become even more important because anonymity increases the salience
of this social identity even further. As a result, a person’s attitudes tend to
shift towards group norms. However, when personal identity is initially
salient, communicators’ personal attitudes are more important and resis-
tant to influence in CMC than in FTF communication. Taken together,
SIDE assumes that when social identity is salient, anonymity of oth-
ers increases social influence (i.e., norm-based influence), and that when
66 Kai Sassenberg
personal identity is salient, anonymity reduces social influence (i.e., inter-

personal influence).
Anonymity of the self

The limited access to person-related information is not restricted to
sender-related information for the recipient (i.e., anonymity of others)
but also applies to recipient-related information for the sender. Thus,
individuals are less identifiable and often also feel less identifiable than in
FTF communication. Both types of anonymity co-occur in most forms
of CMC. At times, however, senders are not aware of their audience
but the audience is well aware of the sender’s identity (e.g., postings to
LISTSERV) and the other way round (e.g., e-mails in response to a
personal web page). More importantly, the psychological consequences
of both types of anonymity differ and therefore they are distinguished
here. Other than anonymity of others, anonymity of the self reduces the
likelihood that individuals’ behavior, attitudes, etc. are judged by others.
Therefore, higher anonymity of the self to a recipient that has the power
to sanction one’s behaviors leads to less compliance with expectations of
this/these other(s) (Reicher et al., 1995; Spears and Lea, 1994). There-
fore, higher anonymity of the self in CMC should reduce compliance.
Physical isolation
A joint feature of CMC and other media is the physical isolation from
interaction partners (if not from any other individual). Physical isolation
often co-occurs with both types of anonymity, but it has an impact on
the psychological state that goes beyond the impact of both types of
anonymity. In physical isolation, others who could distract attention from
the self are not present. Therefore, physical isolation in CMC leads to
heightened private self-awareness (i.e., the increased tendency to monitor
one’s behavior in the light of one’s own standards; Franke, 1997; Joinson,
2001; Matheson and Zanna, 1988, 1989; Sassenberg et al., 2005). This
suggests that physical isolation could be relevant for interpersonal and
norm-based influence because both the personal and the social self can
be affected by private self-awareness.
Three types of social influence in CMC – results

from situationalist research
The evidence about the impact of CMC and the respective relevant
media characteristics on the three forms of social influence is summarized
below. (For a more extensive summary, see Sassenberg and Jonas, 2007,
as well as Spears and Lea, 1992, for the historical development of these
arguments.)
Compliance
Anonymity of the self in CMC should reduce compliance. Early studies
on CMC demonstrated that the use of the media indeed reduced com-
pliance to others (e.g., Smilowitz et al., 1988). Recently, research found
evidence for the impact of anonymity of self to different targets. Partici-
pants working on choice dilemmas more often followed the unanimous
vote given by others when they responded in public compared with a
private response format (Lee and Nass, 2002, Study 1). In intergroup
contexts, group members use more stereotype-consistent language while
describing an outgroup (i.e., abstract words for stereotype-consistent
content and concrete words for stereotype-inconsistent content) when
they are identifiable to their own group than when they are not (Douglas
and McGarty, 2001, 2002, Study 1). Identifiability of the self to an
outgroup manipulated via contact information that had to be provided
elicits object evaluations in line with the expectations of this group
(Sassenberg and Kreutz, 2002), although this compliance to an out-
group is reduced if support from the own group is available (Spears et al.,
2002). The identifiably based media difference between FTF and CMC
is even stronger in collectivist cultures, because compliance resulting
from FTF communication is stronger in collectivist cultures, whereas
this culture difference disappears in CMC (Cinnirella and Green,
2007).
Norm-based influence
Norm-based influence is the type of influence for which CMC research
has yielded the most detailed insights. Spears et al. (1990) conducted
a study in which groups of psychology students discussed four topics
via CMC and were informed about the norms of psychology students
concerning these topics beforehand. Half of the groups were located in
one room but not allowed to speak (identifiable) and the other half were
located in separate rooms (anonymous). Social identity was made salient
by approaching participants as psychology students. The study demon-
strated that anonymity of others in CMC leads to more norm-based
influence when a social identity is salient and a topic that is relevant to
this group is discussed. This finding has been replicated several times (for
a summary, see Spears et al., 2001). Anonymity of others also increases
68 Kai Sassenberg
the likelihood that a dominant norm is internalized by group members

who initially did not share the norm in groups with a salient shared social
identity (Postmes et al., 2001). Norm-based influence is not restricted to
attitudes but can also be found for communication behavior (e.g., abbre-
viations, autonomous language, etc.; Lea and Spears, 1991; Postmes and
Spears, 2002; Sassenberg, 2002).
A norm has to be known collectively by (most of) the group members to
guide any norm-based influence (Sassenberg and Boos, 2003; Waldzus
and Schubert, 2000). Hence, norm-based influence can occur only in
groups that form norms. Some groups do not form any collective norms.
Prentice et al. (1994) distinguished these “common bond” groups that are
based on interpersonal ties holding the group together (e.g., a group of
friends). In other “common identity” groups, members’ ties to the group
as a whole let them stick to the group (e.g., a sports team). For members
of common identity groups, norms are much more important than for
members of common bond groups, because the group itself attracts them
and not the members (Sassenberg, 2002). Indeed, Postmes and Spears
(2000) found that social influence in anonymous CMC (compared with
non-anonymous CMC) was higher in common identity groups than in
common bond groups.
In line with the assumptions of SIDE, the impact of anonymity of
others on norm-based influence is mediated by the salience of the social
categorization (Lea et al., 2001; Postmes et al., 2001). However, sev-
eral studies failed to demonstrate that private self-awareness mediates
the impact of anonymity of others or physical isolation on norm-based
influence (Matheson and Zanna, 1989; Postmes et al., 2001).
Whereas increased salience of the social identity is the intrapersonal
mediator of norm-based influence, findings by Sassenberg and Postmes
(2002) suggest that statements indicating agreement are its intragroup
mediator during communication. Norm-based influence is not only
accompanied by intragroup agreement, it also increases the risk of con-
flict between groups. Postmes et al. (2002) have shown that disagreement
and attitude change away from outgroup opinions is much more likely in
anonymous than in non-anonymous CMC.
Interpersonal influence
When personal identity is salient to communicators, anonymity of others
in CMC will result in less social influence between interaction part-
ners for two reasons. First, group norms are less relevant when personal
identity is salient. Second, anonymity of others during CMC increases
awareness of one’s personal identity in the absence of any salient social
categorization. Under these conditions, the communicators tend to fol-

low their individual needs and values more (Spears and Lea, 1992).
Spears et al. (1990) provided empirical evidence for the reduced
interpersonal influence among individuals with salient personal iden-
tity caused by anonymity of others (for similar findings, see Sassenberg
and Boos, 2003). Other studies in which participants had to reach a joint
decision on judgmental tasks show that agreement within groups not
sharing a social identity is much harder to reach in CMC than in FTF
discussions (Hiltz et al., 1986; Sproull and Kiesler, 1986). Analyses of
the communication in these studies found effects that support the pro-
cess assumptions put forward by SIDE – namely, more striving to fulfill
personal needs. Anonymous interaction partners (compared with non-
anonymous interaction partners) insisted more on their personal views
(Lea and Spears, 1991).
The importance of interpersonal relations for interpersonal influence
in CMC is underlined by studies showing that the number of people con-
forming to the request to fill in a long questionnaire was higher among
targets that were contacted by the researcher beforehand (Guéguen,
2002). Similarly, Postmes and Spears (2000) found more attitude change
in newly formed common bond groups (i.e., groups sticking together
because of interpersonal bonds) when anonymity was low compared with
when anonymity was high.
The impact of status differences is another important factor of inter-
personal influence that is moderated by CMC. Generally in communi-
cation, social influence has more impact when the target believes the
communicator possesses a higher status (e.g., a graduate student versus
an undergraduate student). In line with the prediction that interpersonal
influence is lower in CMC (because of its anonymity), research often
found that the proportion of communication is more evenly distributed
in CMC, thereby muting status differences (compared with FTF interac-
tions; e.g., Dubrovsky et al., 1991; Zigurs et al., 1988), especially when a
group support system is used (i.e., software supporting the group decision
process by providing tools for different steps in the decision process such
as visualization of group members’ attitudes; Rains, 2005). Although sta-
tus differences are leveled out during discussion in terms of participation,
prior expectations about an interaction partner have a stronger impact
with increasing levels of anonymity of others (Epley and Kruger, 2005).
Therefore, task-relevant status differences are more influential in CMC.
Research has shown information provided about the task-relevant exper-
tise of (otherwise anonymous) group members increases their impact on
group discussion and decision in the respective domain of expertise (i.e.,
they shared more information they uniquely held and task performance
70 Kai Sassenberg
was increased; Sassenberg et al., 2001). In sum, task-irrelevant status

differences are equated in anonymous CMC (as evident in the lack of
impact of status on participation in group discussions). However, the
effect of task-relevant status differences (e.g., expertise) is amplified in
anonymous CMC.
Interpersonal influence is moderated not only by anonymity but also
by physical isolation. Even though the two media characteristics are
hard to disentangle, research suggests that physical isolation has an
impact on the psychological state independent of anonymity: private
self-awareness is higher in CMC than in FTF communication (Joinson,
2001; Matheson and Zanna, 1988, 1989; Sassenberg et al., 2005), but
anonymity does not foster private self-awareness (Postmes and Spears,
1998; Postmes et al., 2001). Therefore, the effect of the communica-
tion media on private self-awareness will most likely result from physical
isolation. At the same time, private self-awareness hinders social influ-
ence (Froming and Carver, 1981; Scheier, 1980). Therefore, the height-
ened private self-awareness in CMC mediates the media differences in
interpersonal influence (Sassenberg et al., 2005). To conclude, physi-
cal isolation in CMC increases private self-awareness and reduces the
likelihood of being convinced by communication partners during group
discussions.
Summary of insights and limitations

The summarized research gives a detailed picture of the situational char-
acteristics and mechanisms that show how CMC can lead to increased
or decreased social influence compared with FTF communication. Com-
pliance, which is based on the accountability toward powerful others, is
reduced by the increased anonymity of the self usually afforded in CMC
as well as by support from ingroup members and a low risk of being
sanctioned. That means, for example, that critical contributions to a
newsgroup are more likely for a person using a nickname rather than the
real name and if the person has the impression that others sharing their
opinion are also involved in the discussion.
The source of norm-based influence is the ingroup which becomes espe-
cially salient when the communication partners are anonymous. Precon-
ditions for this type of influence are a shared norm, the initial salience of
the group membership, and the perception of a common identity among
group members. In the case of the newsgroup mentioned above, the fact
that others are using nicknames rather than real names will increase the
likelihood that the contribution follows the norms of the contributor’s
own group and is stronger opposing the norms of outgroups.
Finally, interpersonal influence is most likely to occur among people

who are not nested within a common identity group but who still have
a good relationship (e.g., liking, trust). Psychologically, interpersonal
influence occurs when salience of one’s personal identity is low (which
occurs when people are identifiable) and when private self-awareness
is low (which occurs when we are in the physical presence of others).
Hence, in the newsgroup, individuals knowing each other well (from
real or virtual interactions) are more likely to influence each other in
their opinions, especially when context conditions reduce private self-
awareness (e.g., features of the newsgroups supporting awareness of the
presence of others, such as a counter of the people logged on).
This summary indicates that there is substantial knowledge about the
mechanisms underlying social influence in CMC. Nonetheless, the cur-
rent state of research has limitations – namely, interpersonal influence
and the effect of physical isolation as well as the interplay of the three
forms of social influence are understudied. Moreover, social influence (in
CMC and beyond) involving participants who have known each other for
a while – which certainly makes up most of the actual communication –
has rarely been studied. (For a discussion of further limitations see
Sassenberg and Jonas, 2007.) These foregoing limitations can be
addressed by additional research using the traditional paradigms in
CMC research. However, understanding the role of individual differ-
ences requires new approaches to research. To be more precise, most
research on the impact of media use on social influence has looked at the
effects of contextual variables and intrapersonal processes. The moderat-
ing impact of individual differences has rarely been taken into account.1
The following section summarizes the few existing studies using such
a person x situation interaction approach, and then introduces a more
general interactional model to media-based social influence.
Interactionism and social influence in CMC

The ability of individual differences to further moderate media-based
social influence processes has previously been addressed sporadically
1 One might object that the SIDE model looks at person x situation interactions, when
predicting that the salient self-category and anonymity interact in their impact on norm-
based as well as on interpersonal influence. However, the salient self-category is rather a
situationally varying state than an individual difference variable. Therefore, it has been
manipulated rather than assessed in the relevant research (see Spears et al., this volume).
The chronic social identification might moderate the predicted pattern. However, this
prediction to the best of my knowledge still awaits empirical testing.
72 Kai Sassenberg
rather than systematically. The following sections summarize these iso-

lated approaches stemming from different lines of research.
Overview of existing evidence
Media use
A domain in which individual difference variables have been taken into
account is media choice and media usage (for an overview, see Amichai-
Hamburger, 2005; Renner et al., 2005). McKenna and Bargh (2000)
report, for example, that individuals high in social anxiety are more likely
to form relationships online (see also Amichai-Hamburger et al., 2002).
Hertel et al. (2008) demonstrated a person x situation interaction con-
cerning media choice: introverted as well as emotionally unstable indi-
viduals prefer e-mail for conflict resolutions, whereas extraverted and
emotionally stable individuals prefer face-to-face interactions (compared
with individuals with the respective other personality characteristics).
(For an approach to person x situation interaction research concerning
media use and beyond, see Joinson, 2003.)
Information exchange
Cress (2005) demonstrated that the impact of anonymity on informa-
tion exchange differs depending on individuals’ social value orientations
(i.e., she classified individuals in prosocials maximizing joint outcomes or
equality and proselfs maximizing absolute or personal outcomes or per-
sonal outcomes in comparison with others’ outcomes; van Lange, 1999).
Anonymity of others reduces the amount of information exchanged
by proselfs and increases the information exchanged among prosocials.
These results can be explained based on the relation between social value
orientations and social identification: prosocials identify higher with the
interacting group than proselfs. Therefore, it seems likely that the former
experience norm-based influence (which is stronger) when anonymity of
others is high and the latter experience interpersonal influence (which is
weaker) when anonymity of others is high.
Kimmerle et al. (2007) studied the interaction between individual
differences in trait-based trust in others, sensation seeking, and self-
monitoring with the situational variable of group awareness. Group
awareness was manipulated by giving or not giving information about
the highly cooperative behavior of the fellow group members. Hence,
it can also be seen as a manipulation of the presence of a cooperative
group norm. Group awareness increased cooperation among individuals
low in trust, low in sensation seeking, and high in self-monitoring. At the

same time it reduced cooperation intentions (but not actual cooperation)
among individuals high in sensation seeking and low in self-monitoring.
These findings demonstrate that self-focused individuals (low trust, low
sensation seeking, and high self-monitoring) cooperate more if they get
information about their group, whereas other-focused individuals coop-
erate less when getting this information (see also Kimmerle and Cress,
2008). In sum, individual differences moderate the impact of character-
istics of CMC (i.e., anonymity) as well as characteristics of the specific
communication (i.e., group awareness) on information exchange.
Attitude change and the person–communication

match idea
There is also research on the impact of interindividual differences on
attitude change in CMC which points to the relevance of several spe-
cific communication characteristics (rather than media characteristics).
Guadagno and Cialdini (2002, Study 1) examined the moderating impact
of gender on the effect of a single persuasive message on attitudes. Their
female participants showed less attitude change from e-mails than from
FTF communication with the same content, whereas no such differ-
ences occurred for male participants. Guadagno and Cialdini (2002)
argue that not the individual differences per se (i.e., gender) but the
match between an individual’s characteristics and the communication
characteristics (i.e., the medium, the style, etc.) lead to elevated atti-
tude change. The fit between the communication style of a message and
the recipient’s gender (i.e., “cooperative” for females and “independent”
for males) rules out media differences in social influence. The stronger
influence of CMC compared with FTF communication on males resulted
from the fact that communication via CMC generally has an independent
style (Guadagno and Cialdini, 2002, Study 2).
In a somewhat related finding, websites with simple messages have
more impact on viewers who speak a different primary language than
the website, whereas websites with complex messages have more impact
on viewers who speak the same primary language as the website (Luna
et al., 2003). Hence, the fit between the recipients (i.e., their skills or
their characteristics) on the one hand and the communication content or
the perceived characteristics of the communication medium on the other
fosters social influence.
However, a good match between person and medium can not only fos-
ter social influence but also lead to particularly strong resistance against
74 Kai Sassenberg
social influence. My own work on the effect of chronic private self-

awareness and the communication medium supports this idea (CMC
versus FTF communication; Sassenberg et al., 2005). Because CMC
increases state private self-awareness, participants with high trait pri-
vate self-awareness match very well with this medium and thus show
increased resistance against interpersonal influence (an amplification of
their natural characteristic). Participants with low private self-awareness
are equally persuadable via both CMC and FTF communication media.
Summary
On the one hand, the evidence summarized in this section clearly demon-
strates that individual differences matter in the context of media effects
on social influence. There is evidence concerning a wide range of phe-
nomena (media choice, online behavior, and communication outcomes)
as well as a wide range of personality characteristics. On the other hand,
there is no overarching framework predicting person x media interaction
effects on social influence (but see Joinson, 2003, for a framework model
not addressing social influence). The following section will introduce
such a framework (based on the synergistic person x situation interaction
approach; e.g., Endler and Magnusson, 1976) and apply it to media-
based social influence.
The person x situation interaction approach

For a long time, it has been widely accepted in psychology that charac-
teristics of individuals, characteristics of the environment, and the inter-
action between the two determine human cognition and behavior (e.g.,
Cronbach, 1957; Lewin, 1936). Nonetheless, theoretical as well as empir-
ical approaches regarding person x situation interactions are rare, even
though some applications in specific domains led to interesting insights
(e.g., anxiety, Egloff and Hock, 2001, Spielberger, 1972; aggression in
the context of media violence, Bushman, 1995; justice behavior, Schmitt
et al., 2003; for theoretical frameworks see Endler and Magnusson, 1976;
Mischel and Shoda, 1995).
Most of the existing research found a specific form of a person x situ-
ation interaction; that means person and situation characteristics which
are functionally equivalent do not just add up in their impact on men-
tal processes and behavior but amplify their effects. In statistical terms
this implies an ordinal interaction. Interaction effects can be explained
based on different cognitive structures resulting from personality dif-
ferences which are activated by situational characteristics, similar to a
priming effect (e.g., Bushman, 1995). This is referred to as cognitive

process henceforth. Alternatively, an interaction of effects can also result
from the motivational effects specified in an expectancy x value model
within which the personality characteristic determines the value and the
situation determines the expectation (e.g., Schmitt et al., 2003). This is
referred to as motivational process henceforth. The stronger preference of
shy individuals for e-mail communication in case of conflicts (compared
with individuals who are not shy; Hertel et al., 2008) results most likely
from the higher value they attribute to control during conflict resolu-
tion and the greater expected control they expect to have during e-mail
exchange than in FTF communication.
Some person x situation interactions do not fit the ordinal pattern
but rather show a disordinal interaction. For example, Robins and John
(1997, Study 2) found that narcissists’ positive image of themselves
becomes even further exaggerated when they can watch themselves on
a videotape, whereas non-narcissists evaluate themselves worse under-
going this treatment (compared with usual self-judgment). The authors
apply two different metaphors to explain the divergent responses: narcis-
sists behave like an “egoist distorting information to enhance self-worth,”
whereas non-narcissists behave like “a scientist seeking out information
in a dispassionate search for truth about the self” (Robins and John,
1997, p. 41) and thereby recognize all their mistakes. Hence, two dif-
ferent processes (search for truth and search for self-enhancement) are
necessary to understand disordinal interactions.
Taken together, person x situation interactions can result from cogni-
tive as well as from motivational processes. They can occur in the form
of ordinal interactions based on one underlying process or in the form
of disordinal interactions, which can be considered as a combination of
two ordinal interaction effects.
Application of the interactionist approach to

social influence in CMC
The three situational characteristics of CMC relevant to social influence
are the prime candidates for situational features within person x situation
interactions: anonymity of others, anonymity of the self, and physical
isolation.
Anonymity of others
Anonymity of others has an impact on two social influence processes:
interpersonal influence and norm-based influence. Hence, individual
76 Kai Sassenberg
difference variables that are relevant for both types of social influence
will lead to a disordinal interaction, whereas moderators that are relevant
to only one type of social influence will lead to an ordinal interaction.
Cress (2005) provided an example for a moderator relevant for interper-
sonal as well as for norm-based influence: social value orientations. In
line with the assumption given above, she found a disordinal interaction.
Proselfs perceived the individuals they were interacting with as a collec-
tion of individuals whereas prosocials perceived them as their ingroup.
Therefore, proselfs were more likely to experience interpersonal influ-
ence, and consistent with this assumption they showed less social influ-
ence (i.e., cooperation) when anonymity of others was high (compared
with when anonymity was low). Contrarily, prosocials perceived their
interaction partners as ingroup and showed more social influence when
anonymity of others was high (i.e., norm-based influence). Both inter-
actions underlying the disordinal interaction found by Cress (2005) are
based on cognitive processes, because the impact of social value orien-
tation on the salience of the personal or social identity depends on the
different perceptions of social situations resulting from the social value
orientations.
To the best of my knowledge, there are limited examples in the lit-
erature of individual difference variables leading to ordinal interactions
with anonymity on social influence. However, a potential example would
be trait-based trust in others (as studied by Kimmerle et al., 2007, in
the context of group awareness). Norm-based influence should not be
affected by trait-based trust, because the salience of a social identity
which is required for norm-based influence leads to high levels of trust
independent of the level of trait-based trust (i.e., a strong situation over-
rules the effects of a trait). The salience of the personal identity, however,
does not induce trust situationally. Thus, interpersonal influence, which
requires strong interpersonal bonds, will profit from chronic trust. This is
especially the case when anonymity of others is high and chronic trust can
affect behavior because the situational information about the trustwor-
thiness is low. When anonymity is low, information about trustworthiness
is available and will most likely reduce the trust of individuals high in trait
trust, because they start to consider the visible features of the interaction
partners. Taken together, interpersonal influence is high when trait trust
and anonymity co-occur. This is because in this situation individuals are
especially likely to expect that they can trust the communication part-
ners and therefore evaluate the incoming information more. Hence, the
expected trust x anonymity of others interaction would be driven by a
motivational process.
Anonymity of the self

Moderations of the impact of anonymity of the self on social influence
should exclusively result in ordinal interactions, because anonymity of
the self impacts on only one social influence mechanism: conformity.
Candidates for moderators are all inter-individual differences related to
conformity. Even though research has not yet directly addressed per-
son x situation interactions involving the moderation of the impact of
anonymity of the self on social influence by shyness and social anxiety,
there is some evidence that suggests those characteristics will most likely
increase compliance (Hertel et al., 2008; McKenna and Bargh, 2000).
The motivation to conform should be increased by shyness and anxiety,
when anonymity of the self is low, but in case of high anonymity of the
self the behavioral difference between socially anxious and non-anxious
individuals should be reduced. Similarly, the findings that the difference
in compliance between independent and interdependent cultures does
not occur in CMC can be attributed to the high anonymity of the self
in CMC (Cinnirella and Green, 2007). In sum, compliance should be
extraordinarily high when shyness, anxiety, or an interdependent orien-
tation co-occur with low anonymity of the self. In the case of all three
moderators, a motivational process is at work, because anonymity of the
self reduces the expectation to be evaluated and the three individual dif-
ference variables increase the importance (i.e., the value) of evaluations.
An individual difference variable that might interact with anonymity
of the self based on a cognitive process is public self-awareness. Because
of its more cognitive nature, Fenigstein et al. (1975) differentiated social
anxiety from public self-awareness. The slightest signal of identifiability of
the self might automatically elicit conformity to the public’s expectations
among individuals who are chronically high in public self-awareness. In
the case of anonymity of the self, the impact of public self-awareness on
compliance should be virtually non-existent. Unfortunately, this mecha-
nism has not been studied yet in the context of social influence in CMC.
Physical isolation
For physical isolation, ordinal rather than disordinal interactions with
interpersonal difference variables in the impact on social influence are to
be expected, because isolation is related to only one type of social influ-
ence: interpersonal influence. The interaction between physical isolation
and private self-awareness reported above (Sassenberg et al., 2005) sug-
gests a cognitive effect. The tendency to resist interpersonal influence
78 Kai Sassenberg
Table 4.1 Examples for moderators of the impact of media characteristics on

social influence sorted by the mechanism underlying the moderation
Moderation processes and relevant moderators

Media characteristics Cognitive process Motivational process
Anonymity of others Social value orientation Chronic trust
Anonymity of the self Public self-awareness Shyness, anxiety,

interdependence
Physical isolation Private self-awareness,
social comparison orientation
among those high in private self-awareness when they are physically iso-
lated results from higher levels of accessibility of personal values rather
than from different evaluations of these values (Froming and Carver,
1981). A similar effect might result for individuals with high levels of
social comparison orientation (Gibbons and Buunk, 1999), as social
comparisons occur automatically when others are present (Huguet et al.,
1999). Hence, individuals high in social comparison orientation should
be more prone to interpersonal influence only if others are present. Inter-
actions of effects based on a motivational process are somewhat less likely
for physical isolation, because the effects of physical isolation are mostly
based on the distraction caused by the presence of others and the acces-
sibility of self and other related information, rather than by changes in
expectations or values.
Summary and conclusion

The impacts of media characteristics on the three types of social influ-
ence (i.e., compliance, interpersonal influence, and norm-based influ-
ence) are moderated by individual difference variables that affect either
message processing (i.e., cognitive process) or the expectancies and val-
ues in the situation that are relevant for action (i.e., motivational process).
Anonymity of the self and physical isolation are more likely to show ordi-
nal interactions with individual difference variables, because they impact
on only one social influence process. Differing from this, anonymity of
others can result in ordinal as well as disordinal interactions with indi-
vidual difference variables, because it is relevant for two social influence
processes. This type of interaction depends on the relevance of the indi-
vidual difference variable for one or both social influence processes.
The list of potential moderators in Table 4.1 is certainly not exhaus-
tive. All person characteristics related to the chronic accessibility of the
actual self and standards related to the self or brought up by others might
increase the resistance to interpersonal influence and compliance as well
as increase norm-based influence in interaction with the media charac-
teristic relevant for the specific type of social influence. In addition, all
person characteristics increasing or reducing the value of information,
behavior, or attitude change might also moderate the impact of media
characteristics on all three forms of social influence. Finally, all person
characteristics related to the relations between interaction partners might
impact in different ways on the three types of influence.
Taken together, research on social influence in CMC has rarely taken
into account interpersonal differences. The few existing studies summa-
rized here and the numerous potential effects discussed demonstrate the
importance of this understudied domain. Further research should not
only test the predictions derived here but also strive for a more integra-
tive theory of person x media interaction effects on social influence.
References
Allen, V. L. (1965). Situational factors in conformity. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 2, pp. 133–175). New York:
Academic Press.
Amichai-Hamburger, Y. (2005). Personality and the internet. In Y. Amichai-
Hamburger (Ed.), The social net: The social psychology of the internet
(pp. 27–56). Oxford: Oxford University Press.
Amichai-Hamburger, Y., Wainapel, G., and Fox, S. (2002). “On the internet
no one knows I’m an introvert”: Extraversion, neuroticism, and internet
interaction. CyberPsychology & Behavior, 2, 125–128.
Bushman, B. J. (1995). Moderating role of trait aggressiveness in the effects of
violent media on aggression. Journal of Personality and Social Psychology, 69,
950–960.
Cinnirella, M., and Green, B. (2007). Does “cyber-conformity” vary cross-
culturally? Exploring the effect of culture and communication medium on
social conformity. Computers in Human Behavior, 23, 2011–2025.
Cress, U. (2005). Ambivalent effect of member portraits in virtual groups. Journal
of Computer-Assisted Learning, 21, 281–291.
Cronbach, L. J. (1957). The two disciplines of scientific psychology. American
Psychologist, 12, 671–684.
Deutsch, M., and Gerard, H. B. (1955). A study of normative and informational
social influences upon individual judgement. Journal of Personality and Social
Psychology, 51, 629–636.
Douglas, K. M., and McGarty, C. (2001). Identifiability and self-presentation:
Computer-mediated communication and intergroup interaction. British
Journal of Social Psychology, 40, 399–416.
(2002). Internet identifiability and beyond: A model of the effects of identifia-
bility on communicative behavior. Group Dynamics, 6, 17–26.
80 Kai Sassenberg
making groups. Human Computer Interaction, 6, 119–146.
Egloff, B., and Hock, M. (2001). Interactive effects of state anxiety on emotional
Stroop interference. Personality and Individual Differences, 31, 875–882.
Endler, N. S., and Magnusson, D. (1976). Toward an interactional psychology
of personality. Psychological Bulletin, 83, 956–974.
Epley, N., and Kruger, J. (2005). When what you type isn’t what they read:
The perseverance of stereotypes and expectancies over e-mail. Journal of
Experimental Social Psychology, 41, 414–422.
Fenigstein, A., Scheier, M. R., and Buss, A. H. (1975). Public and private self-
consciousness: Assessment and theory. Journal of Consulting and Clinical
Franke, G. H. (1997). Über die Möglichkeiten der computerunterstützten
Darbietung beim revidierten Freiburger Persönlichkeitsinventar: Zwei
experimentelle Studien [Computer-based administration of the Freiburg
Personality Inventory: Two experimental studies]. Zeitschrift für Experi-
mentelle Psychologie, 44, 332–356.
Froming, W. J., and Carver, C. S. (1981). Divergent influences of private and
public self-consciousness in a compliance paradigm. Journal of Research in
Personality, 15, 159–171.
Gibbons, F. X., and Buunk, B. P. (1999). Individual differences in social com-
parison: Development of a scale of social comparison orientation. Journal of
Personality and Social Psychology, 76, 129–142.
Guadagno, R. E., and Cialdini, R. B. (2002). Online persuasion: An examination
of gender differences in computer-mediated interpersonal influence. Group
Dynamics, 6, 38–51.
Guéguen, N. (2002). Foot-in-the-door technique and computer-mediated com-
munication. Computers in Human Behavior, 18, 11–15.
Hertel, G., Schroer, J., Batinic, B., and Naumann, S. (2008). Do shy people
prefer to send e-mail? Personality effects on communication media prefer-
ence in threatening and non-threatening situations. Social Psychology, 39,
231–243.
Hiltz, S. R., Johnson, K., and Turoff, M. (1986). Experiments in group deci-
sion making: Communication process and outcome in face-to-face versus
computerized conferences. Human Communication Research, 13, 225–252.
Huguet, P., Galvaing, M. P., Monteil, J. M., and Dumas, F. (1999). Social
presence effects in the Stroop task: Further evidence for an attentional view
of social facilitation. Journal of Personality and Social Psychology, 77, 1011–
1025.
Joinson, A. N. (2001). Self-disclosure in computer-mediated communication:
The role of self-awareness and visual anonymity. European Journal of Social
(2003). Understanding the psychology of internet behaviour: Virtual worlds, real
lives. Basingstoke: Palgrave Macmillan.
Kimmerle, J., and Cress, U. (2008). Group awareness and self-presentation in
computer-supported information exchange. International Journal of Com-
puter Supported Collaborative Learning, 3, 85–97.
Kimmerle, J., Cress, U., and Hesse, F. W. (2007). An interactional perspective on

group awareness: A tool alleviating the information-exchange dilemma (for
everybody?). International Journal of Human–Computer Studies, 65, 899–910.
Lea, M., and Spears, R. (1991). Computer-mediated communication, de-
individuation and group decision making. International Journal of Man–
Machine Studies, 39, 283–301.
Lea, M., Spears, R., and de Groot, D. (2001). Knowing me, knowing you:
Anonymity effects on social identity processes within groups. Personality and
Social Psychology Bulletin, 27, 526–537.
Lee, E.-J., and Nass, C. (2002). Experimental tests of normative group influence
and representation effects in computer-mediated communication. Human
Communication Research, 28, 349–381.
Lewin, K. (1936). Principles of topological psychology. New York: McGraw-Hill.
Luna, D., Peracchio, L. A., and de Juan, M. D. (2003). The impact of lan-
guage and congruity on persuasion in multicultural E-marketing. Journal of
Consumer Psychology, 13, 41–50.
Matheson, K., and Zanna, M. P. (1988). The impact of computer-mediated
communication on self-awareness. Computers in Human Behavior, 4, 221–
233.
(1989). Persuasion as a function of self-awareness in computer-mediated com-
munication. Social Behaviour, 4, 99–111.
McKenna, K. Y. A., and Bargh, J. A. (2000). Plan 9 from cyberspace: The
implications of the internet for personality and social psychology. Personality
and Social Psychology Review, 4, 57–75.
Mischel, W., and Shoda, Y. (1995). A cognitive-affective system theory of person-
ality: Reconceptualizing situations, dispositions, dynamics, and invariance
in personality structure. Psychological Review, 102, 246–268.
Postmes, T., and Spears, R. (1998). Deindividuation and anti-normative
behaviour: A meta-analysis. Psychological Bulletin, 123, 238–259.
(2000). Refining the cognitive redefinition of the group: Deindividuation effects
in common bond vs. common identity groups. In T. Postmes, R. Spears, M.
Lea, and S. Reicher (Eds.), Side issues centre stage: Recent developments in
studies of de-individuation in groups (pp. 31–45). Amsterdam: KNAW.
(2002). Behavior online: Does anonymous computer communication reduce
gender inequality? Personality and Social Psychology Bulletin, 28, 1073–
1083.
Postmes, T., Spears, R., and Lea, M. (1998). Breaching or building social bound-
aries? SIDE-effects of computer mediated communication. Communication
Research, 25, 689–715.
(2002). Intergroup differentiation in computer-mediated communication:
Effects of depersonalisation. Group Dynamics, 6, 27–37.
Postmes, T., Spears, R., Sakhel, K., and de Groot, D. (2001). Social influence
in computer-mediated communication: The effects of anonymity on group
behavior. Personality and Social Psychology Bulletin, 27, 1243–1254.
Prentice, D. A., Miller, D. T., and Lightdale, J. R. (1994). Asymmetries in attach-
ments to groups and to their members: Distinguishing between common-
identity and common-bond groups. Personality and Social Psychology Bulletin,
20, 484–493.
82 Kai Sassenberg
Rains, S. A. (2005). Leveling the organizational playing field – virtually. Commu-

nication Research, 32, 193–234.
198.
Renner, K.-H., Schütz, A., and Machilek, F. (Eds.) (2005). Internet und
Persönlichkeit [Internet and personality]. Göttingen: Hogrefe.
Robins, R. W., and John, O. P. (1997). Effects of visual perspective and narcissism
on self-perception: Is seeing believing? Psychological Science, 8, 37–42.
Sassenberg, K. (2002). Common bond and common identity groups on the
internet: Attachment and normative behavior in on-topic and off-topic chats.
Group Dynamics, 6, 27–37.
Sassenberg, K., and Boos, M. (2003). Attitude change in computer-mediated
communication: Effects of anonymity and category norms. Group Processes
& Intergroup Relations, 6, 405–422.
Sassenberg, K., Boos, M., and Klapproth, F. (2001). Wissen und Prob-
lemlösekompetenz: Der Einfluss von Expertise auf den Informationsaus-
tausch in computervermittelter Kommunikation [Knowledge and problem-
solving competence: The influence of expertise on information exchange
in computer-mediated communication]. Zeitschrift für Sozialpsychologie, 32,
45–56.
Sassenberg, K., and Jonas, K. J. (2007). Attitude change and social influence on
the net. In A. N. Joinson, K. Y. A. McKenna, T. Postmes, and U.-D. Reips
(Eds.), The Oxford handbook of internet psychology (pp. 273–288). Oxford:
Sassenberg, K., and Kreutz, S. (2002). Online research and anonymity. In B.
Batinic, U.-D. Reips, and M. Bosnjak (Eds.), Online social sciences (pp. 213–
229). Seattle: Hogrefe and Huber.
Sassenberg, K., and Postmes, T. (2002). Cognitive and strategic processes in
small groups: Effects of anonymity of the self and anonymity of the group
on social influence. British Journal of Social Psychology, 41, 463–480.
Sassenberg, K., Rabung, S., and Boos, M. (2005). Attitude change in face to face
and computer-mediated communication: Private self-awareness as mediator
and moderator. European Journal of Social Psychology, 35, 621–632.
Scheier, M. F. (1980). Effects of public and private self-consciousness on the
public expression of personal beliefs. Journal of Personality and Social Psy-
chology, 39, 514–521.
Schmitt, M., Eid, M., and Maes, J. (2003). Synergistic person x situation interac-
tion in distributed justice behavior. Personality and Social Psychology Bulletin,
29, 141–147.
Smilowitz, M., Compton, D. C., and Flint, L. (1988). The effects of computer
mediated communication on an individual’s judgement: A study based on
the methods of Asch’s social influence experiment. Computers in Human
Behavior, 4, 311–321.
Spears, R., and Lea, M. (1992). Social influence and the influence of
the “social” in computer-mediated communication. In M. Lea (Ed.),
Contexts of computer-mediated communication (pp. 30–65). Hemel Hempstead:

Harvester-Wheatsheaf.
(1994). Panacea or panopticon? The hidden power in computer-mediated com-
munication. Communication Research, 21, 427–459.
Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., and ten Haar, W. (2002).
in computer-mediated communications. British Journal of Social Psychology,
29, 121–134.
influence. In J. P. Forgas and K. D. Williams (Eds.), Social influence: Direct
and indirect processes (pp. 331–350). New York: Psychology Press.
Spielberger, C. D. (Ed.) (1972). Anxiety: Current trends in theory and research.
New York: Academic Press.
Sproull, L., and Kiesler, S. (1986). Reducing social context cues: Electronic mail
in organizational communication. Management Science, 32, 1492–1512.
Tajfel, H., and Turner, J. C. (1979). An integrative theory of intergroup conflict.
In W. G. Austin and S. Worchel (Eds.), The social psychology of intergroup
relations (pp. 33–47). Monterey, CA: Brooks/Cole.
Turner, J. C. (1991). Social influence. Milton Keynes: Open University Press.
Blackwell.
Van Lange, P. A. M. (1999). The pursuit of joint outcomes and equality in out-
comes: An integrative model of social value orientation. Journal of Personality
and Social Psychology, 77, 337–349.
Waldzus, S., and Schubert, T. (2000). Group norm and category norm in anony-
mous situations: Two sources of social influence. In T. Postmes, R. Spears,
M. Lea, and S. Reicher (Eds.), Side issues centre stage: Recent developments in
studies of de-individuation in groups (pp. 31–45). Amsterdam: KNAW.
personal and hyperpersonal interaction. Communication Research, 23, 1–43.
Webb, W. M., Marsh, K. L., Schneiderman, W., and Davis, B. (1989). Interaction
between self-monitoring and manipulated states of self-awareness. Journal
Williams, E. (1977). Experimental comparisons of face-to-face and mediated
communication: A review. Psychological Bulletin, 84, 963–976.
Wood, W. (1999). Motives and modes of processing in the social influence of
groups. In S. Chaiken and Y. Trope (Eds.), Dual-process theories in social
psychology (pp. 547–570). New York: Guilford.
Zigurs, M. I., Poole, S., and DeSanctis, G. L. (1988). A study of influence in
computer-mediated group decision making. MIS Quarterly, 12, 625–644.
5 Social interaction in cyberspace
Social construction with few constraints

Susanne Abele
No doubt, social interactions supported by modern technologies have

become an important part of our personal, professional, and political
lives. At least since the last decade social psychologists have become inter-
ested in studying interactions that are mediated by modern technologies
such as the internet. Research topics have focused on how computer-
mediated interactions differ from traditional face-to-face interactions.
One aspect that has drawn a lot of attention is that computer-mediated
interactions have greater potential for anonymity. As a consequence,
many studies have focused on the effects of whether or not the actor
is identifiable to the people she is interacting with, referencing the SIDE
model (Spears et al., 1990, 2001). SIDE stands for Social Identity De-
individuation Effects. The model proposes that social identity moderates
the effects of anonymity. It proposes that deindividuation can accentuate
social influence when one identifies strongly with an online group but
reduce social influence in the absence of strong identity.
When studying social interactions that are supported by new technolo-
gies, there is an inclination to start with the premise that social behavior is
in some or all ways fundamentally different as a result of the technology.
A notable exception is the research on the SIDE model, which started
with the position that deindividuation and social identity have been stud-
ied in social psychology at least over the last three decades (Diener, 1977,
1979), and that we could learn from applying what we know from these
research traditions to advance our understanding of the role of anonymity
in computer-mediated interactions.
I am proposing the following approach to study of computer-mediated
interactions. First, we should think carefully about what distinguishes
a computer-mediated interaction from face-to-face interaction. As has
been pointed out, one difference is the greater opportunity for anonymity.
Another difference is the scarcity of social cues. Others suggest that the
available social cues in computer-mediated interactions are often dif-
ferent in nature than the ones that are salient in face-to-face interac-
tions (e.g., Walther, 2007). However, I will claim that there is another
84
Social interaction in cyberspace 85
important difference, which has been neglected so far, namely the oppor-
tunity for deliberative action. Second, once we identify important dis-
tinctions between computer-mediated and face-to-face interactions, we
should think about what these distinctions mean conceptually. Third, we
glean as much as possible from existing theory and empirical knowledge
to give us insight into how behavior is affected by these differences. As
pointed out before, the research on the SIDE model has pursued this
avenue.
In this chapter, I will propose that one of the most important and so
far neglected differences between computer-mediated and face-to-face
interactions is that computer-mediated interactions have many more pos-
sibilities to be deliberative. Whereas in a lot of face-to-face interactions
our behavior and responses to others need to be immediate, and con-
sequently automatic processes play a big role (see, e.g., Bargh, 2007;
Dijksterhuis and Bargh, 2001), computer-supported interactions give
much more opportunity for delayed responses, and hence a deliberative
planning of how to achieve our goals in the interaction (see also Walther,
2007). When one is writing an email, responding to an email, and com-
municating in a chat room, people have time to think about whether their
message has the intended effect on the receiver. That means that people
have more room to think strategically about their actions, which in turn
means that people think more about how they want the interaction part-
ner to interpret and react to their messages. In other words, what might
contribute to a major part of the differences between face-to-face and
computer-supported interactions are the perceptions of the interaction
partner, the degree to which anticipated reactions are considered, and
how these anticipated reactions shape one’s message. From that angle,
anonymity of others is as important as one’s own anonymity. Put differ-
ently, what one knows or imagines about the others is equally relevant as
what others know about oneself.
Note that the proposition is not that face-to-face interactions never
give room for deliberation. If we are teaching or giving a presentation
to our colleagues, we plan carefully what we want to say, thereby taking
into account the knowledge and standpoint of the interaction partner or
audience. The idea is that overall opportunities for deliberation are much
more plentiful in a computer-mediated than in a face-to-face interaction,
and that is one aspect that differentiates computer-mediated and face-
to-face interactions.
Let us take the example of dating. You can meet people at a variety of
venues, including at a bar, a party, a speed-dating event, a singles’ club,
or a friend’s house. You could also meet people on the internet either
via one of the more “serious” dating sites such as matchmaker.com or
86 Susanne Abele
eharmony.com, or on a forum like sugarbabe.com or fling.com. Whether

one is searching online or not, there is a range of places to choose from
that are more or less suitable depending on whether one is seriously look-
ing for a long-term partner or just wants to spend some amusing time and
play games. However, in any of the face-to-face initial dating encounters,
you will have less time for deliberating about your actions and reactions
than on any of the computer-mediated forums. No question, the image
of the place in terms of the expected length of the relationships that are
being sought will determine other important aspects of the interaction.
But in any of the face-to-face encounters you have to act and react to your
partner’s actions immediately. On any of the internet dating encounters,
people have time and will consider what they want their counterpart to
think about them. So if it is a site with a more serious image and big-
ger prospects to find a long-term relationship, people may more possibly
refrain from communicating things that are misleading, but nevertheless
consider deeply what their counterparts might be looking for in terms of
qualities, what they want the counterpart to think about themselves, and
how to present such an image. On a site with a more playful image, peo-
ple will predominantly also consider what they want their counterparts
to think about them and how they can achieve that. For example, a male
going on sugarbabe.com assumes that the women searching on that site
look for qualities such as wealthy. Moreover, he will probably assume that
the nature of the relationship will be short-lived and may even remain
computer-mediated. Hence, he will adjust his online communication and
presentation with these considerations in mind.
In sum, communication has many more opportunities to be deliber-
ative when it is computer-mediated. When communication is delibera-
tive, people tune the message to fit the intended recipient or audience
(Walther, 2007). In order to be able to do that effectively, one needs
to know, or at least infer characteristics of, the recipient or audience.
Thus, it is important how much and what one knows about the message
recipient(s).
But in computer-mediated communication people often might know
less about their counterparts than they would ordinarily do in face-to-
face interactions. Computer-supported interactions are, compared with
traditional face-to-face interactions, deprived of a lot of social cues and
information. It has been proposed that, as a consequence, the social
information that is present might get more amplified (Spears et al., 2002;
Walther, 1992; Walther et al., 2005).
Social cues and information can refer to the (interacting) person(s) or
the situations. Cues such as facial mimicry, physical appearance, into-
nation of voice, and emphasis of words all refer to characteristics of the
person, or the way the communicators package their message. In con-

trast to social cues that are derived from characteristics of the person,
features of the situation are equally present in face-to-face and computer-
supported interaction. Compounding the problem created by the scarcity
of social cues is the fact that the context of online interactions is very
malleable. In the physical world, interactions often occur in well-defined
situations. Physical markers tell the nature of the situation and often what
to expect from others. We expect people to behave differently at home
than they do at the office. Remarks that are regarded as humorous and
playful in a bar may be viewed as tasteless and rude in church. However,
in the online world, physical surroundings tell us little about the nature
of a social interaction. It matters little whether we log on while we are in
a library or at the beach. We often do not know where others are located
and, even if we do, this knowledge tells us little about the online inter-
action. Thus, in the online world, social context depends heavily on
the actual or inferred relationships among the interacting parties. That
is, social context is defined by interdependencies. So studying contexts
from this perspective requires answering two main questions. First, how
can we distinguish interdependencies in a systematic and consistent way?
Second, how can people use information about the social context to tune
their communication to accomplish the desired outcomes?
Computer-mediated interaction presents a conundrum. On the one
hand, there is ample opportunity for deliberative action and consid-
eration of how others will react to our behavior. On the other hand,
there are fewer social and situational cues on which to base our deliber-
ations. One guiding assumption is that, when there are fewer cues, the
available cues matter more. These considerations lead to three related
questions. First, how does one form impressions of others when there
are limited social cues? Second, how does one assess social situations in
terms of interdependencies? Third, when interpersonal impressions and
situational assessments are based on limited information, how do they
interact? Implicit in this third question is the idea that limited knowledge
of others may shape our assessment of the social situation and limited
knowledge about situations may shape our perceptions of others. That
is, if the social context is judged to involve conflict, we may be inclined to
perceive others as more competitive than if the social situation is judged
to be cooperative. Similarly, if the available social cues suggest that oth-
ers are competitive, we may interpret an ambiguous social context to be
conflictual. To elaborate on this line of reasoning, we need a more precise
language to classify situations in terms of interdependencies.
Although not immediately obvious, the science of experimental eco-
nomics addresses similar questions. Indeed, I will argue that game theory
88 Susanne Abele
and behavioral game theory, in particular, provide useful ways of framing

answers to these questions. Experiments in economics are traditionally
characterized by anonymity and defining social contexts by rules of inter-
dependency, not by physical locations, while investigating the interplay
between people, institutions, and markets (see, for example, Kagel and
Roth, 1995). For example, when experimental economists study markets,
they are not referring to physical locations such as the farmer’s market or
the stock exchange; they are referring to a set of rules that defines how
the parties’ behaviors affect others.
Within the broader area of experimental economics, there is a research
area looking at behavior in specifically defined interdependencies, which
is called experimental game theory. It looks at how people behave in a
situation where they are faced with making a decision when the outcome
of their decision depends not only on what they decide but also on some-
body else’s decision. Experimental game theory is, in contrast to classical
game theory, interested in the actual behavior that people display in these
situations. Thus, experimental game theorists are interested in study-
ing situations where the independent variables clearly and unambigu-
ously define interdependencies. As a result, carefully constrained inter-
actions are studied without providing the opportunity for face-to-face
communication.
New technologies control social interactions by limiting the amount
and type of social information available. In that sense technology-
supported interactions have a lot of similarities with the controlled and
highly defined interactions studied in experimental game theory. In both
types of interactions, what the interaction partners know about each other
is constrained. But even from relatively limited information, people can
and do make inferences. When there is little information about another,
people tend to supplement this information in two ways.
First, we tend to assume that the other is like ourselves (Abele et al.,
2004; Krueger and Acevedo, 2005). We can think about what we expect
from others in the situation, how we would want them to be, and how
we would act, and then project our own attitudes, wishes, goals, and
strategies onto the interaction partner. We would then communicate,
behave, and react to our projected strategies. Indeed, classical game the-
ory invokes a similar projection assumption in the form of common
knowledge rationality. That is, game theory assumes that players are per-
fectly rational and that those players assume that the other players are
also perfectly rational. In this sense, although game theory does not state
it that way, this assumption can be taken as a projection assumption, in
the sense that each player assumes that the other players are as rational
as herself and from that perspective one can assume, predict, or project
what they will do. Since every player is assumed to be rational – and
assumes the other players are as well – predicting and projecting what
others do is equivalent, from the perspective of classical game theory.
Second, if people want to focus on their interaction partners but have
limited information about them, they can use the information about
the situation to make inferences about their interaction partners’ needs,
goals, and strategies. In order to look more closely at what inferences
people can and do make about the information they have about the situ-
ation, we need to be able to classify situations by the feature that defines
them. This feature is the nature of the interdependencies. Describing and
classifying interdependencies is the stuff of games.
Characterizing situations by interdependency

Examining the impact of situations on social interactions and on per-
ceptions of others in the situation is aided by a systematic classification
of situations. If one were to ask people what kind of different situations
they could define, they would probably come up with categories such
as a threatening situation, a romantic situation, or a hostile or friendly
situation. These categories refer to how they are affectively perceived by
the people involved. Certainly, such a categorization would be subjec-
tive. What one person might perceive and hence categorize as hostile
might not be perceived that way by another. More technical classifica-
tions, which do not refer to affective experiences and can be objectively
applied to interactive situations, are provided by game theory (von Neu-
mann and Morgenstern, 1947) and interdependence theory (Kelley and
Thibaut, 1978; Thibaut and Kelley, 1959). Both of these theoretical
perspectives provide a formal approach to classifying interactions. There-
fore, the experiments that have been conducted in these areas consist of
clearly defined situations and have high anonymity between the interac-
tion partners. Both aspects make the findings very relevant for the study
of computer-mediated interactions.
Concepts from game theory and interdependence theory

Game theory analyzes strategic decision-making situations, that is, situa-
tions where decision makers are interdependent, by describing the solu-
tions that would result if all players made rational choices. In this context,
rationality can be defined as maximizing one’s outcomes. In applying
game theory, the assumption is not that people interacting via computers
should be more rational than in face-to-face interactions. Nevertheless,
90 Susanne Abele
C D
10 20
C
10 0
A
0 5
D
20 5
PDG
Note. A = person A; B = person B; C = cooperate; D = defect
Figure 5.1 The Prisoner’s Dilemma Game
situations can be reliably classified by the solutions that rationality would

prescribe.
Two essential concepts from game theory can be used to classify situ-
ations. The first is the Nash equilibrium: a state in which no interaction
partner has an incentive to change her current position or choice, given
the choices of the other(s). If all players have made a choice from a given
set of possible alternatives, and no one can improve her outcome by
switching to a different alternative, as long as all the others stick to their
chosen alternative, then the players are in a Nash equilibrium. More pre-
cisely, the chosen alternatives and their corresponding pay-offs constitute
a Nash equilibrium. In contrast, a Pareto efficient solution to an interac-
tion is one in which no player can improve her situation without harming
another or others. A combination of behaviors that maximizes the joint
outcomes across all players (i.e., maximizes collective outcomes) is one
prominent solution to many interactions that are Pareto efficient.
Using the principles of Nash equilibrium and Pareto efficiency, social
interactions can be characterized as conflictual or cooperative. Accord-
ing to game theory, an inherently conflictual situation exists when there
is no Nash equilibrium or the available Nash equilibria are not Pareto
efficient. Many of the classic conflicts studied in behavioral economics
and psychology are ones for which Nash equlibria are not Pareto effi-
cient. Consider the classic Prisoner’s Dilemma Game (PDG) illustrated
in Figure 5.1. The Nash equilibrium is reached if both players defect. It is
F O
16 0
F
10 0
W
6 10
O
6 16
BOS
Note. W = wife; H = husband; F = football; O = opera
Figure 5.2 The Battle of the Sexes Game
costly for either player to switch to cooperate if the other player continues
to defect. However, it is Pareto efficient if both players cooperate. When
both cooperate, the only way for either player to improve her outcome
(i.e., switch to defect) harms the other player.
As pointed out earlier, being in a state of Nash equilibrium does not
necessarily mean that the involved parties get the best outcome possible.
When there is a conflict, the involved parties would be better off not to end
up in a Nash equilibrium. Imagine, for example, competing companies:
they would be better off forming a cartel, instead of each pursuing their
individual profit maximization. The fact that this often does not happen
(even without government interventions preventing it) demonstrates that
there is some validity to the Nash concept, even though nobody claims
that people are invariably rational.
Conversely, social situations are inherently cooperative when Nash
equilibria are Pareto efficient. Consider another classic game from behav-
ioral economics, the Battle of the Sexes (BOS), as depicted in Figure 5.2.
The story of the game is one where players are a husband and wife who
are going on a date and need to choose an activity for the evening. There
are two options: A (preferred by the husband, e.g., demolition derby)
or B (preferred by the wife, e.g., opera). Although each player has an
individual preference, being together is valued even more. In this game,
92 Susanne Abele
there are two Nash equilibria – both players responding A or both play-
ers responding B. That is, if both players respond A, neither benefits
by switching to B as long as the other stays with A. Similarly, if both
respond B, neither can improve his/her situation by changing to A unless
the other player also changes. Importantly, in this game, AA and BB are
also Pareto efficient. Once in AA, there is no combination of choices
that improves one player’s outcome without reducing the other player’s
outcome. The same conclusion applies to BB.
In their classic treatise of interdependence theory, Kelley and Thibaut
(1978; Thibaut and Kelley, 1959) distinguished three components of
interdependency in social relationships: behavioral control, fate control,
and reflexive control. Behavioral control refers to the degree to which
one person’s choice affects the value or pay-off associated with others’
choices. The BOS depicted in Figure 5.2 illustrates behavioral control.
Player 1’s choice of A or B changes the values associated with A and B
for Player 2. Thus, Player 1 has behavioral control over Player 2. In this
case, Player 2 also has behavioral control over Player 1. Thus, the behav-
ioral control is mutual. Fate control refers to the degree to which one’s
decision directly affects the outcomes of others, meaning, by varying
her behavior, Player 1 can affect Player 2’s outcome regardless of what
Player 2 does.
There is a conflict inherent in a situation when two or more people have
mutual fate control or mutual behavioral control, and their outcomes are
non-correspondent, meaning they do not prefer the same outcome. The
Prisoner’s Dilemma Game depicted in Figure 5.1 has a large compo-
nent of fate control and the fate control is mutual. Regardless of whether
Player 1 chooses to defect or to cooperate, Player 2’s choice of defect-
ing reduces Player 1’s outcome dramatically. The fate control is mutual
because Player 1’s choice affects Player 2’s outcome in the same way.
The third component is reflexive control, which is, in essence, a per-
son’s preferred ordering of the possible choices or actions presented in
the situation. Indeed, reflexive control is the value reflected back to the
actor by the various actions or choices and, technically, is not derived
from the interdependency. Nonetheless, reflexive control is relevant to
classifying situations. For example, the BOS depicted in Figure 5.2 is not
a game of pure coordination because, underlying the mutual behavioral
control, the two parties prefer different actions. Thus, the BOS is pre-
dominantly a coordination problem with an undertow of conflict. As a
result, it can be experienced as a conflict because Player 1 prefers AA
over BB and Player 2 has the reverse preference. Even though AA and
BB are Nash equilibria and Pareto efficient, there is no solution that is
optimal for both parties, both individually and collectively.
In summary, both game theory and interdependence theory provide a

precise language for characterizing social situations. Additionally, these
theoretical perspectives provide ways of quantifying the degree of con-
flict in a situation. For instance, in the subsequent discussion of public
goods games, I will introduce the concept of marginal rate of substitution
(MRS), which measures the degree of conflict in the situation.
Interdependencies are underlying useful frames of reference of any
social situation. In the negotiation literature, for instance, a distinction is
made between zero-sum negotiation situations and situations which pro-
vide opportunities for integrative solutions (Thompson, 2005). Nobody
would make the inference that findings from one type of negotiation sit-
uation would replicate in the other type of negotiation situation, because
of the assumption that the interdependency is a fundamental aspect of
the situation. Interdependency conceptually categorizes situations. The
following subsections illustrate two examples of fundamentally different
kinds of interdependencies: one entails conflict, the other one requires
coordination.
Conflict: public goods game

The often studied public goods game models the conflict between the
individual’s interest and the group’s interest. Involved interaction part-
ners are provided with an endowment. They can either keep this endow-
ment to themselves or put it in a public pool. Everyone’s pay-off from
the public pool is the total amount contributed multiplied by a factor b.
Endowments kept are multiplied by a factor a. Thereby a > b must always
hold, in order for the situation to be a public goods game. The ratio a / b
is called the marginal rate of substitution and represents, numerically,
what conflict a person is faced with in the decision over each unit of
endowment. It is an excellent and unambiguous indication of how much
conflict is involved in the situation. The smaller the marginal rate of sub-
stitution, the less conflict is involved. In terms of interdependence theory,
the public goods game involves a high degree of mutual fate control. That
is, each person’s decision to contribute or not affects the other’s outcome
regardless of what they decide to do.
Coordination
A situation embodies a coordination problem when both interaction
partners are interested in ending up in the same Nash equilibrium.
They might prefer different equilibria, but they still have an incentive
94 Susanne Abele
to end up in the same one. Imagine, for example, that your university
has an online vote for new members of the graduate council. From your
department, there are two faculty members as candidates for becoming
members of the council, and one is favored over the other by a nar-
row margin. However, membership in the university council requires
a two-thirds vote of approval from the candidate’s home department.
In other words, it could be that if all of your faculty members went
with their first choice of vote, your department would have no repre-
sentative in the graduate council. But if either side could convince a
certain proportion of the other side to vote for their favorite, your depart-
ment would be successful in having a representative in the council. That
is, if the faculty members of your department coordinated their votes
to be bundled on one candidate, your department would have a very
good chance of having a representative. Everybody prefers a successful
bundling of votes on one candidate over spread-out votes. There are
just different preferences on which of the two candidates to bundle the
votes.
If all interaction partners prefer the same equilibrium, then the equi-
libria that are presented by the situation differ in the risk they pose to
the interaction partners. In terms of interdependence theory, a situation
calls for coordination if the interaction partners have mutual behavioral
control and corresponding outcomes.
Dynamic reframing of interdependencies: the role of

personal experience
Even formal games often leave room for multiple interpretations, how-
ever. Kelley and Thibaut (1978; Thibaut and Kelley, 1959) argued that
people can reinterpret pay-off matrices in social relations. Repeated inter-
actions often provide opportunities to transform social situations. For
example, repeated episodes of an interaction that has the flavor of a
BOS may come to be experienced as conflictual as the two parties try
to get each other to choose their preferred option. The wife who has
gone to fifteen demolition derbies may refuse to cooperate further until
her husband agrees to go to the opera. In a similar way, a PDG con-
flict can be transformed into a cooperative venture by trading “defect”
responses: “You take the heat for me this time and I will take the heat
for you next time.” Such transformations require cognitive reframing of
a game. As we shall see, such reframing can be prompted by seemingly
subtle cues.
Person perceived and person perceiver by

situation interaction
In the 1960s it was proposed that behavior should not be conceptualized
as a function of personalities but rather as a function of the interaction
between personalities and situations (Mischel, 1968, 1999, 2004). For
instance, people differ in their competitiveness or cooperativeness. In a
situation that requires primarily coordination, persons high in cooper-
ativeness may not differ significantly in the behavior they display from
people high in competitiveness. However, in a situation of conflict, such
as a public goods game, people high in cooperativeness are more likely
to contribute to the common pool than people high in competitiveness.
Individual differences in inclination to be competitive or cooperative
are captured in the concept of social value orientation (Messick and
McClintock, 1968). This refers to the notion that people have rather
stable preferences in whether they prefer to maximize their individual
outcome, to maximize the difference between their individual outcome
to others’ outcomes, or to maximize the group’s outcome. These indi-
vidual differences have shown to be predictive of behavior in prisoner’s
dilemma and resource dilemma games (Liebrand, 1984; van Lange and
Liebrand, 1989), but would not be expected to be predictive in romantic
relationships (unless the romance has turned into a dilemma).
Moreover, even in the person–situation–interaction model, situations
should not be conceptualized as stable entities. There are individual dif-
ferences in the perception of cooperative and competitive interpersonal
situations (King and Sorrentino, 1983). The authors showed that dif-
ferent aspects of the situation seem to be salient to different kinds of
individuals. So interindividual differences in perceptions can account for
different behaviors in the same situation (Endler and Magnusson, 1976).
Liebrand et al. (1986) found an interaction between social value orien-
tations and participants’ interpretations of others’ and own behavior.
Mainly cooperators and individualists differ in the degree of importance
they attach to the evaluative and the potency dimension. Cooperators
tend to view the dimension “cooperative–competitive” as evaluative, but
individualists tended to view it as a dimension of power. Viewing a com-
petitive move as an expression of power or something “bad” and likewise
viewing cooperation as good or a sign of weakness, of course, has strong
behavioral implications for an interaction.
Besides, when studying how the greater opportunities for thought-
ful deliberation and cognitive tuning in computer-supported interaction
affect behavior in these situations, it should not be neglected that in a
96 Susanne Abele
social interaction the person perceived is also a person perceiver, and vice
versa. People perceive one another in social interactions. While person
A forms an impression about the goals, wishes, strategies, and charac-
teristics of person B, person A is aware that person B is at the same
time forming an impression of her. Hence, the interaction partners are
making judgments about each other’s traits, similarity, smartness, and
likeability (Kenny, 1994, 2004). Therefore, what we need is a naı̈ve the-
ory of person-by-situation interactions. What sorts of people do persons
expect to interact with in which kind of situation? This would be one way
to investigate the impact of greater opportunities for deliberativeness in
computer-mediated interactions on people’s behavior.
Three lessons from behavioral game theory to

computer-mediated interactions
On a more general level, how do the person and situation interaction
and the chances for deliberativeness and cognitive tuning in computer-
mediated interactions affect behavior in these situations? We should
refrain from viewing persons perceived and situations as stable entities.
Following, three lessons to be learned from behavioral game theory will
be outlined. All three lessons address how the interaction of person and
situation can provide cues that we may use in computer-mediated inter-
actions. In particular, they identify ways that one can glean information
from the scarcity of cues about persons and situations in the virtual world.
Lesson 1: Subtle cues shape the subjective definition of the situ-
ation in the absence of strong cues from the physical environ-
ment.
Lesson 2: As an extension of lesson 1, limited information about
others in the situation shapes the subjective definition of the
situation.
Lesson 3: The converse of lesson 2 is that limited informa-
tion about the situation affects perceptions of others in the
situation.
Lesson 1: Perceptions of situations: labels, sanctioning systems,

and timing
Labels. Apart from the objective properties of the situation or the person-
alities of the individuals involved, the mere label of a social interaction
can impact people’s behavior. Again we refer to research in experimental
game theory in order to demonstrate the effect of labeling a situation
in different ways on behavior and decisions in computer-mediated

situations.
Pillutla and Chen (1999) had participants play a linear public goods
game. Like the PDG, linear public goods games are conflictual in that
the Nash equilibrium is for everyone to “defect” – that is, to give noth-
ing to the public good – but the Pareto efficient solution is for everyone
to cooperate by giving the maximum amount to the public good. The
linear public goods game is also high in mutual fate control: the amount
contributed by each person benefits the others regardless of how much
they contribute. However, this mutual fate control is noncorrespondent
with everyone’s preference, which is to donate less. Thus, the structure
of this game is highly conflictual. The game Pillutla and Chen used was
identical in two experimental conditions. The only aspect that differed
was the label that the authors gave to the game. They called one an
investment game, in which the participant’s decision was described as
deciding how much to invest in a joint investment fund. This scenario
was the economic decision frame. The second condition was the non-
economic decision frame, in which participants were in the role of a
representative of a company and had to decide how much of a contribu-
tion they wanted to make to a social event. When the game was labeled
in this way as a social task, participants contributed almost 40 percent
of their endowment to the common pool, which was significantly more
than participants contributed when the game was called an investment
task. In the latter, participants contributed a third of their endowment.
Hence it can be concluded that providing a social label resulted in people
behaving more cooperatively in the face of interdependency, which was
objectively conflictual.
Sanctioning systems. In a similar vein, Tenbrunsel and Messick (1999)
showed first that the establishment of even very weak sanctioning systems
alters the perception of the situation, even though the interdependency
between parties is identical. Second, they showed that people behave
differently as a result of having a different perception of the situation.
More specifically, they showed that a 5 in 100 chance of paying a fine
for defecting in a PDG increased people’s tendency to defect. But aside
from showing that both cooperation and the expectation of cooperation
decreased when a mild sanctioning system was present, Tenbrunsel and
Messick also showed that implementing a sanctioning system changed
the perception of the situation. In a second experiment, they gave, in
two conditions, the identical scenario to a social dilemma game. They
also varied the absence or presence of a mild sanctioning system, which
was represented by a low probability of giving something up. Apart from
making a choice, participants were also asked how they would describe
98 Susanne Abele
the situation. In the presence of the sanctioning system, the majority

of participants described the situation as a business situation, and only
20 percent indicated that they viewed the decision as an ethical one.
When no sanctioning system was present, about half of the participants
described the situation as a business decision, and about half described
it as an ethical decision. Again, participants were less cooperative when
a sanctioning system was present.
The authors concluded that the presence or absence of a sanctioning
system acted as a “situational cue that triggers an assessment of the
type of decision that one is making” (p. 698). Moreover, they had some
evidence that the absence or presence of a mild sanctioning system can
impact expectations of others’ behaviors.
Decision timing. An even subtler cue, which affects how people concep-
tualize a social situation, is decision timing. Abele et al. (2004) showed
that people make different decisions when deciding simultaneously in
a coordination game, compared with deciding in a pseudo-sequential
order. Deciding pseudo-sequentially means that participants make their
choices in a sequential order, but without knowledge of earlier moves.
Hence, the information structure is identical to the situation in which
participants choose simultaneously. Moreover, in both timing conditions
participants were in an identical interdependency, a coordination game,
meaning players had a corresponding interest. Participants made signifi-
cantly more cautious decisions under simultaneous decision timing com-
pared with the pseudo-sequential timing structure. Abele et al. hypothe-
sized and found empirical support for their idea that pseudo-sequential
decisions activate concepts of social interactions, which in turn increase
individuals’ interpersonal trust and decrease cautiousness in situations
of interdependence. Simultaneous game structures are more likely to
activate concepts of games of chance, which makes the possibility of an
actual total loss more salient. Thus, a cue such as the presence of a
weak sanctioning system, but also a very subtle cue like decision timing
without knowledge of earlier moves, can affect how people conceptualize
the social situation they are in, although the objective interdependency
is identical. Note also that these data were obtained from social inter-
actions, which were solely mediated by computers, and the interaction
partners remained anonymous to each other.
Lesson 2: Perception of others affects the construal of the situation

Imagine yourself in a situation in which your role is to propose a divi-
sion of an endowment between yourself and another party, whereby that
other party remains anonymous to you, and their role consists of merely
accepting your proposal. This situation is called a Dictator Game. Your

role is referred to as the allocator, and the other party’s role is referred
to as the receiver.
Now imagine the same situation with a slight variation. Again your
role is to propose the division of an endowment; however, this time, the
other party, who again remains anonymous to you, can either accept or
reject your proposal. If they accept, the endowment is divided according
to the proposal. But if the other party does not accept your proposal,
neither you nor the other party gets anything from the endowment. This
situation is referred to as an Ultimatum Bargaining Game. Here the role
of the other person can be referred to as the receiver or the responder.
In both situations the first turn is yours. You can decide which division
you want to propose. In both cases, the other player can at most get the
amount that you allocate to her. In the second case, she also has the veto
option of punishing you if she finds your proposition unacceptable. So in
order to avoid the possibility that the other player might not accept your
proposal, you may decide to allocate a bit more to the other player due to
strategic fairness considerations (van Dijk and Vermunt, 2000). Possibly,
somebody might have other fairness concerns, such as a normative fair-
ness concern or a concern for distributive fairness, and that person might
feel obliged to offer the other player 30–40 percent of the endowment,
as an act of kindness and fairness. If this is the case, however, there is no
reason to imagine somebody would do so in a situation where the other
person has no possibility for a veto, and the same person would not do so,
when the other person has a possibility for a veto. But that is exactly what
van Dijk and Vermunt (2000) observed in their study. They had partic-
ipants play either an Ultimatum Bargaining Game or a Dictator Game.
In both types of games the units of the endowment were worth twice
as much to the allocator as they were to the receiver. The authors var-
ied whether this fact was known to the receiver or not, but it was always
clear to the allocator. They found that in the Dictator Game, participants
tended to allocate the same proportion of the endowment to the other
person, regardless of whether the receiving person knew that the points
were worth twice as much to the allocator. Indeed, in the Dictator Game,
they gave on average about twice as many points to the recipient in light
of the fact that the points were worth less to the recipient. However, in
the Ultimatum Bargaining Game, allocators gave more to the responder
only when the responder knew that the endowment was worth twice as
much to the allocator. When this fact was unknown to the responder, the
allocators gave significantly less.
These results are counterintuitive. Allocators had the power to affect
the other’s outcomes (i.e., given fate control). The role of the other person
100 Susanne Abele
was varied: in the case of the Ultimatum Bargaining Game, the responder
also had reciprocal fate control because she could reject the allocation
and, in doing so, reduce the allocator’s outcome to zero. In the case of the
Dictator Game, the recipient had no power. However, allocators seemed
to redefine the situation depending on whether or not the recipient had
power. In the face of a powerful partner, allocators acted as though the
situation was a strategic, economic exchange. In the face of a powerless
partner, they acted as though the situation was social exchange governed
by norms of fairness. The allocations that participants made imply that
they weigh the importance of distributive justice very differently as a
function of recipients’ veto power.
In a similar vein, Cooper et al. (1999) showed that labeling the other
player in a game as “manager” or “planner” evoked more strategic play
than when the players were labeled as player A and player B. Further-
more, Ben-Ner et al. (2004) showed that women, when designated as
allocator in a Dictator Game, give significantly less to other women than
they give to men or persons of unknown gender.
All of these studies demonstrate that what we know about our inter-
action partner affects how we define the situation, which in turn affects
our behavior. Moreover, it is consistent with the idea that the less social
information we have, the more the available information gets weighted
and has an influence. In a computer-mediated interaction we do have
less social information, and the currently outlined research gives another
insight into how exactly the little information that is left can have an
impact. Moreover, the games described above could often be an extract
of a longer computer-mediated interaction. The responder, or receiver,
will have different feelings and perceptions of her interaction partner as a
result of the first player’s behavior. As a consequence, she might behave
quite differently on one of her subsequent moves depending on her per-
ception of whether the allocator seemed to have deemed distributive
justice as important. The responder’s behavior again will affect subse-
quent reactions on the allocator’s side. In other words, the very subtle
difference in the role of just one interaction partner could set the tracks
for a cooperative or competitive tone of a non-face-to-face interaction.
Lesson 3: The situation affects perception of others

The fundamental attribution error states that people tend to attribute
behavior to dispositions of the actors. That means that people are more
likely to make a person attribution than a situation or an object attribution
(Ross, 1977). One reason that has been proposed for this phenomenon
is that the observer is focused on the actor when observing the behavior.
I am going one step further and propose that, even before the behavior
has occurred, the observer can make inferences from the situation about
the characteristics of others in the situation. One mechanism that might
underlie the situation-to-person attribution is the association between
certain kinds of situations and certain kinds of people. This association
may be reinforced sufficiently in social interactions that people often do
not correct for the fact that any kind of person could end up in any
situation. In a world in which computer-supported interactions prevail
more and more, we encounter a lot of social interactions in which we
do not meet the other actor in person, and before the interaction all the
information we have is the information about the kind of situation we
will be in with another person.
Imagine you are expecting to meet someone on a blog. In trying to
predict how others will react, it is natural to want to know what type
of people they are. For these predictions and assessments, people use
the information available to them. Anticipating meeting a finance expert
on an investment blog will probably evoke different expectations about
what the person is like than if you were expecting to meet someone
on a site that matches “busy professionals” for lunch dates. In both
instances, you likely do not have much information about your interaction
partner(s). Nevertheless, people have a desire to form an impression
of the person they are dealing with and will use any information that
is available to them. In the foregoing example, you would likely form
a different impression of the person on the investment blog than the
person on the lunch date site. However, these persons could be one
and the same. Hence, people’s tacit assumptions about the other person
may vary, depending on the nature of the situation. For example, it
would matter whether there is a conflict of interests, how much conflict
is present, or whether it is more a matter of coordinating behavior to
achieve mutually beneficial outcomes. Moreover, people’s assumptions
about others, stemming from the specifics of the situation, could have an
impact on how people behave in the situation.
Empirical support
In order to test the idea that features of the situation can affect perceptions
of interaction partners even before any interaction has taken place, I put
participants in a social situation. The situation was identical in the two
conditions, but varied in only one aspect, which was the degree of conflict
involved in the interdependency. Participants played a public goods game
with a continuous pay-off function. Degree of conflict was manipulated
by the marginal rate of substitution. Participants played the game with
102 Susanne Abele
either a marginal rate of substitution of 2/3, or with a marginal rate of

substitution of 4/5. The bigger the marginal rate of substitution, the less
conflict is embedded in the interaction.1
Both interactions were operationalized as dyadic interactions, and were
completely computer-mediated. Moreover, participants were aware that
their interaction partner remained anonymous to them and vice versa.
The rules of the game were explained in detail, and they were told that
they were going to play that game with one other person, and payment
was contingent on the outcome of their play. They were then told that,
before they began playing, the experimenter wanted to know the impres-
sion they had of their interaction partner. So before they played the game
participants rated their partner on a nine-point rating scale. They were
asked about their general impression of similarity. Also included were
items assessing two primary dimensions of stereotype contents: pleas-
antness (warmth) and intelligence (competence) (Fiske et al., 2002).
Moreover, items related to social motivation were included: selfishness,
cooperativeness, and competitiveness. After participants had given their
ratings, which were not communicated to each other but just to the exper-
imenter, participants played the game. They were then informed about
the decision of their interaction partner and the outcome of the game.
It turned out that the more conflictual the public goods game that
players were going to play, the more intelligent, the meaner, and the less
cooperative they judged the other person to be. Moreover, the initial
pre-interaction impressions were (statistically) predictive of decisions. As
judged intelligence increased, contributions increased. As judged mean-
ness increased, contributions decreased.
The results replicated in a series of four experiments. Another study
looked at the possible underlying processes mediating these effects. This
study was basically a replication of the study just described, with the
addition of another condition in which there was a passive observer of the
interaction. That person was assigned to one specific dyad, was informed
about the rules of the game, rated the players before the game, then
learned what they decided, and finally reevaluated the two players. The
purpose of including the observer condition was to get an insight as to
1 If a public goods game has a marginal rate of substitution of 4/5 (or 2/3 respectively) it
means that every unit of her endowment which the player puts into her private pool will
be multiplied by 4 (2 respectively). Every unit of endowment that the player allocates to
the public pool will be multiplied by 5 (3 respectively). The pay-off for each individual
player is calculated by taking the sum of her individual private pool and her share of the
public pool. The public pool is divided equally among all players. Given that the number
of players remained constant, which was always two, a game with a MRS of 4/5 is less
conflictual for the individual player than a game with a MRS of 2/3.
whether one really needs to be in the interaction as an actor in order to

show the effects on interpersonal perceptions. Hence, the question was
whether there is a motivational component to the effect, or whether it
seems to be due to a socially learned association between certain types
of situations and certain types of people. It turned out that the observers
perceived the players very much as the players perceived each other. That
is, perceptions of the players were affected by the situation, prior to the
interaction, regardless of whether the perceiver was an observer or acting
in the situation as a player. One interpretation is that we learn to associate
certain types of people with different situations. That is, inferring traits
of people from the nature of social contexts may be a socially learned
association.
So in sum, I found, first, that even this subtle difference affects the
interpersonal impressions of the interaction partners. More specifically, I
found that the more conflict there was in the anticipated interaction, the
smarter and less pleasant people judged their opponent to be. Second,
the initial interpersonal impressions were related to the decisions.
Implications for computer-mediated interactions

In the reported study of impression formation, the social context was
stripped from the interaction, and the two different situations varied
only by the contingencies among the players. With respect to computer-
mediated interactions, these results have important implications. In many
electronic interactions, such as email, telephone conversations, or even
video conferences, situations can be more enriched with social cues
than was the case in our experimental paradigm. Nevertheless, in some
computer-mediated contexts, such as online auctions, the interactions
can be as sterilized as they were in the situation I looked at. These lessons
apply to online auctions for private consumers such as on eBay, as well
as professional ones involving millions of dollars and often affecting the
future of groups of companies (for example, the auctions on the license
for blocks of bandwidth for universal mobile telecommunications sys-
tems). The results imply that the fewer the social cues that traditionally
characterize a face-to-face interaction, the more people use what little
information they have.
The effects of these minimal-cues situations are just as relevant today
as they were in the early days of the internet. The trend toward mak-
ing computer-mediated communication more information rich (e.g.,
video-conferencing) is simultaneously accompanied by the advent of
more portable and compact devices. Some of the more recent technolo-
gies developed to support interactions, for instance the BlackBerry, are
104 Susanne Abele
designed to improve convenience (allowing usage on the plane, in the

tram, on the business trip in Saudi Arabia) and speed. Users of Black-
Berries do not tend to write emails as extensive and formal as someone
would do sitting in a quiet office of a university. With the trend of improv-
ing usability of technologies which allow for distant communications, we
might also approach the trend of facing interactions with fewer and fewer
social cues involved.
Conclusion
In this chapter I have argued that a major and thus far neglected difference
between computer-mediated and face-to-face communication lies in the
greater opportunities for deliberation when the interaction is technology-
supported. However, as others have noted, computer-mediated interac-
tions often contain fewer and different social cues than do face-to-face
interactions. Both of these characteristics of computer-mediated interac-
tions highlight the importance of two issues. First, how do we use cues
stemming from the specific situation in order to form impressions of the
other interaction partner and make predictions of their reactions to our
alternative course of actions? Second, how can deliberation about one’s
interaction partner’s strategies and goals affect the interaction?
I argued that experimental game theory addresses similar questions.
Consequently, I suggested incorporating game theory as a tool to clas-
sify situations, and research findings from behavioral game theory as a
guide to study effects of deliberation processes in computer-supported
interactions.
When there is opportunity and motivation to focus on the other inter-
action partner, there are two things we can do to compensate for the
scarcity of social cues. First, we can use projection and react to our
projected strategies and behavior. Second, we can also use and draw
inferences from the limited information the situation provides. To exam-
ine these limited situations, game theory and interdependence theory can
provide a useful classification of situations.
Behavior is a function of personalities and situations. Moreover, differ-
ent personality types perceive situations and also other actors differently.
Besides, a person perceiver is also a person perceived, and vice versa, in
any interaction. Along that line, I outlined three lessons to be gleaned
from behavioral game theory with implications for computer-mediated
interaction. First, cues such as the label of an interaction, mild sanc-
tioning systems, and even decision timing can affect the perception of a
situation. Second, limited information about the other person can affect
the perception of the situation, whether the interdependency is viewed as
cooperative or competitive, and whether the interaction is viewed as an

economic or social/ethical exchange. Third, limited information about
the situation can affect how we perceive the other person interacting with
us, and hence how we behave.
I have argued that the study of computer-mediated interaction should
start with the assumption that people are the same type of social animal
whether they are interacting face-to-face or over electronic media. Thus,
the study of computer-mediated interaction should start with under-
standing how it differs from face-to-face interactions and how those dif-
ferences constrain or shape behaviors. The early emphasis on anonymity
and deindividuation in computer communication certainly took this
strategy and the SIDE model built on this early work. However, the
SIDE model suggests that it is not simply the anonymity of the actor that
is important – the lack of individuating information about others is also
important, as is whether others are viewed as like the actor (ingroup) or
not (outgroup). I am suggesting that this movement toward considering
what is known or not known about others is critical because computer-
mediated interaction affords, even promotes, a deliberative process. In
this deliberation, actors try to anticipate the reaction of others to alter-
native messages and presentations. An essential lesson from behavioral
economics is that controlling or constraining social information does not
prevent actors from engaging in this deliberative process. Rather, it makes
them more industrious in mining what they can from the scant informa-
tion available. Like the proverbial flap of a butterfly wing in the southern
hemisphere spawning a hurricane in the northern hemisphere, a morsel
of information about others or about the nature of one’s relationship with
others can change the course of computer-mediated exchanges.
References
Abele, S., Bless, H., and Ehrhart, K. M. (2004). Social information processing
in strategic decision making: Why timing matters. Organizational Behavior
and Human Decision Processes, 93, 28–46.
Bargh, J. A. (Ed.) (2007). Social psychology and the unconscious: The automaticity
of higher mental processes. Philadelphia, PA: Psychology Press.
Ben-Ner, A., Kong, F., and Putterman, L. (2004). Share and share alike? Gender-
pairing, personality, and cognitive ability as determinants of giving. Journal
of Economic Psychology, 25(5), 581–589.
Cooper, D. J., Kagel, J. H., Lo, W., and Gu, Q. L. (1999). Gaming against
managers in incentive systems: Experimental results with Chinese students
and Chinese managers. American Economics Review, 89, 781–804.
Diener, E. (1977). Deindividuation: Causes and consequences. Social Behavior
and Personality, 5, 143–155.
106 Susanne Abele
(1979). Deindividuation, self-awareness, and disinhibition. Journal of Personality

and Social Psychology, 37, 1160–1171.
Dijksterhuis, A., and Bargh, J. A. (2001). The perception-behavior expressway:
Automatic effects of social perception on social behavior. In M. P. Zanna
(Ed.), Advances in experimental social psychology (Vol. 33, pp. 1–40). San
Diego, CA: Academic Press.
Endler, N. S., and Magnusson, D. (1976). Toward an interactional psychology
of personality. Psychological Bulletin, 83, 956–974.
Fiske, S. T., Cuddy, A. J., Glick, P., and Xu, J. (2002). A model of (often mixed)
stereotype content: Competence and warmth respectively follow from per-
ceived status and competition. Journal of Personality and Social Psychology,
82, 878–902.
Kagel, J. H., and Roth, A. E. (1995). The handbook of experimental economics.
Princeton, NJ: Princeton University Press.
Kelley, H. H., and Thibaut, J. W. (1978). Interpersonal relations: A theory of
interdependence. New York: Wiley.
Kenny, D. A. (1994). Interpersonal perception: A social relations analysis. New York:
Guilford.
(2004). PERSON: A general model of interpersonal perception. Personality and
Social Psychology Review, 8, 265–280.
King, G. A., and Sorrentino, R. M. (1983). Psychological dimensions of goal-
oriented interpersonal situations. Journal of Personality and Social Psychology,
44, 140–162.
Krueger, J. I., and Acevedo, M. (2005). Social projection and the psychology of
choice. In M. D. Alicke, D. Dunning, and J. I. Krueger (Eds.), The self in
social perception (pp. 17–41). New York: Psychology Press.
Liebrand, W. B. (1984). The effect of social motives, communication and group
size on behaviour in an N-person multi-stage mixed-motive game. European
Journal of Social Psychology, 14(3), 239–264.
Liebrand, W. B., Jansen, R. W., Rijken, V. M., and Suhre, C. J. (1986). Might over
morality: Social values and the perception of other players in experimental
games. Journal of Experimental Social Psychology, 22(3), 203–215.
Messick, D. M., and McClintock, C. G. (1968). Motivational bases of choice in
experimental games. Journal of Experimental Social Psychology, 4(1), 1–25.
Mischel, W. (1968). Personality and assessment. New York: Wiley.
(1999). Implications of person–situation interaction: Getting over the field’s
borderline personality disorder. European Journal of Personality, 13, 455–461.
(2004). Toward an integrative science of the person. Annual Review of Psychol-
ogy, 55, 1–22.
Pillutla, M. M., and Chen, X.-P. (1999). Social norms and cooperation in social
dilemmas: The effects of context and feedback. Organizational Behavior and
Human Decision Processes, 78(2), 81–103.
Ross, L. D. (1977). The intuitive psychologist and his shortcomings: Distortions
in the attribution process. In L. Berkowitz (Ed.), Advances in experimental
social psychology (Vol. 10, pp. 173–220). New York: Academic Press.
Ross, L. D., Amabile, T. M., and Steinmetz, J. L. (1977). Social roles, social
control, and biases in social perception processes. Journal of Personality and
Social Psychology, 35, 485–494.
Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., and ten Haar, W. (2002).
in computer-mediated communications. British Journal of Social Psychology,
29, 121–134.
influence. In J. P. Forgas and K. D. Williams (Eds.), Social influence: Direct
and indirect processes (pp. 331–350). New York: Psychology Press.
Tenbrunsel, A. E., and Messick, D. M. (1999). Sanctioning systems, decision
frames and cooperation. Administrative Science Quarterly, 44, 684–707.
Thibaut, J. W., and Kelley, H. H. (1959). The social psychology of groups. New
York: Wiley.
Thompson, L. L. (2005). The mind and heart of the negotiator. Upper Saddle
River, NJ: Pearson Prentice Hall.
van Dijk, E., and Vermunt, R. (2000). Strategy and fairness in social decision
making: Sometimes it pays to be powerless. Journal of Experimental Social
van Lange, P. A., and Liebrand, W. B. (1989). On perceiving morality and
potency: Social values and the effects of person perception in a give-some
dilemma. European Journal of Personality, 3(3), 209–225.
von Neumann, J., and Morgenstern, O. (1947). Theory of games and economic
behavior. Princeton, NJ: Princeton University Press.
Walther, J. B. (1992). Interpersonal effects in computer-mediated interaction: A
relational perspective. Communication Research, 19, 52–90.
(2007). Selective self-presentation in computer-mediated communication:
Hyperpersonal dimensions of technology, language, and cognition. Com-
puters in Human Behavior, 23, 2538–2557.
Walther, J. B., Loh, T., and Granka, L. (2005). Let me count the ways: The
interchange of verbal and nonverbal cues in computer-mediated and face-
to-face affinity. Journal of Language and Social Psychology, 24, 36–65.
6 Dynamics of leader emergence in
online groups
Andrea B. Hollingshead
Communication technologies and the internet have created an explosion

in the number and types of groups that exist in contemporary society –
some of which could not have been imagined twenty years ago. The
ubiquitous range of information and communication technologies (from
hand-held devices to laptops to videoconference rooms to virtual reality
immersion labs) provides a variety of channels (audio, video, text, tactile,
and graphics in two and sometimes three dimensions) and functional-
ity (communication, collaboration, file sharing, social networking, etc.)
which has enabled groups to move beyond the conventional “same time,
same place” model of face-to-face groups to an “anytime, anywhere” or
even an “all the time, everywhere” model of the digital world (Hollings-
head and Contractor, 2002). Some types of online groups, which were
novel fifteen years ago, are now commonplace, such as geographically
distributed work teams and internet-based discussion forums and online
communities.
With the advent of Web 2.0 and Web 3.0 on the horizon, new group
forms are continually emerging. Facebook, MySpace, and other social
networking sites have enabled people with common acquaintances, sim-
ilar interests, or complementary goals from across the country or around
the world to connect and, in some cases, form discussion groups, friend-
ship groups, political action groups, and even business ventures. Some
multiplayer online games, such as “World of Warcraft,” are organized
around “guilds” and players must work as a team to overcome obstacles
and achieve goals. Virtual worlds, such as Second Life, are user-defined
and user-created environments where “residents” can explore, meet other
“residents,” socialize, participate in individual and group activities, and
conduct business (Wikipedia, 2008).
Many of the online groups described above would fit the traditional
definition of a small group: an entity composed of people with inter-
dependent goals, interacting with one another, and having a sense of
belonging associated with their membership (Hollingshead and Con-
tractor, 2002). However, many online groups are not subject to the same
108
Dynamics of leader emergence in online groups 109
geographical, temporal, membership, and size constraints as traditional

face-to-face groups. (See Brandon and Hollingshead, 2007, for a descrip-
tion of online groups.) In traditional face-to-face groups, group size is
likely to be relatively small and its membership is by definition a closed
set. That is certainly true for some geographically distributed work teams
in organizations. However, this is not the case in many internet-based
newsgroups and support groups, where there can be hundreds or even
thousands of participants who come and go (Alexander et al., 2003).
The similarities and differences among face-to-face and online groups
raise many questions about the dynamics of leader emergence in initially
leaderless online groups, many of which have not been fully explored in
previous research.
1. How do people rise to positions of influence in online groups?
2. How does online group leadership evolve and change over time?
3. To what extent is the process of leader emergence in online groups
similar to traditional face-to-face groups?
Of course, all online groups are not created equal,1 just as face-to-
face groups differ greatly from each other. However, one goal of social
psychological research is to discover general principles about how people
influence and are influenced by their social environment, that is, by
other people. Scholars know much more about the behavior of face-to-
face groups than online groups. It is quite possible that some general
principles regarding the dynamics of face-to-face groups may also apply
to the online groups.
This volume examines the interplay between personal attributes and
situational factors on social behavior in mediated settings. As a social psy-
chologist, my work focuses more (but not exclusively) on situation-based
rather than on person-based explanations and determinants of behav-
ior, as will be evident in this chapter. However, there is an extensive lit-
erature on leader attributes and traits, which has regained prominence
recently (see Zaccaro et al., 2004, for a review). From the late 1940s to
the early 1990s, leadership researchers did not find consistent empiri-
cal evidence that leaders differed significantly from their followers with
regard to their personality traits or other attributes, and the person-based
1 Although in this chapter I have chosen the term “online groups” to describe groups
that interact using communication technologies, the array and types of online groups
have evolved much quicker than the terminology used to describe them (Brandon and
Hollingshead, 2007). Descriptors such as “computer-mediated,” “virtual,” “online,” or
“distributed” are technically imprecise. It is increasingly common to find groups that
interact using multiple media and increasingly rare to find groups that interact using
only one medium. In addition, many of the current technologies available to groups are
portable and wireless; they enable members to collaborate without having to be “online.”
110 Andrea B. Hollingshead
or “trait” perspective of leadership went out of favor. Recent improve-

ments in conceptual, methodological, and statistical tools are enabling
contemporary leadership researchers to explore the influence of personal
attributes more comprehensively and more precisely in ways that were
not possible in earlier research. In their recent review, Zaccaro et al.
(2004) proposed that leadership emergence (and effectiveness) is best
predicted by “an amalgamation of attributes reflecting cognitive capac-
ities, personality orientation, motives and values, social appraisal skills,
problem-solving competencies and general and domain-specific exper-
tise” p. 120. However, they also acknowledged that situational and con-
textual influences define the leadership processes and behaviors required
for success.
This chapter explores leader emergence in online groups. It does not
examine leadership effectiveness, which is beyond the scope of this chap-
ter. I begin with an overview of social psychological research on leader
emergence in face-to-face groups, then I present data from two unpub-
lished studies about leaders in online groups. The first study, on which I
was a collaborator, compares the process of leader consensus in face-
to-face and computer-mediated work groups. The second study, on
which I was an advisor, examines participation patterns during the first
year in the life of twenty-four internet support groups. The strength of
both of these studies is that they examine leader emergence at multiple
points in time. I end with some conclusions and directions for future
research.
What we know about leader emergence in

face-to-face groups
Group roles
The dictionary definition of a role is “the characteristic and expected
social behavior of an individual” (American Heritage of the English Lan-
guage Dictionary, 2009). This definition suggests that a role involves: 1)
a somewhat consistent patterning of an individual’s behavior over time,
and 2) some level of consensus from others in the social setting about the
role holder’s behavior.
Group and organizational researchers have long distinguished between
formal and informal roles. Formal roles are assigned to the individual,
and identify the individual’s position within the group or organization
(e.g., secretary, CEO). Attached to each position is a set of role-based
behaviors that members occupying those positions are expected to per-
form (Biddle and Thomas, 1969; Katz and Kahn, 1966).
In contrast, informal roles are negotiated by group members and

emerge through interaction. Benne and Sheats (1948) provided a clas-
sification scheme that is still widely used today to describe three types
of informal roles: task-oriented, maintenance, and self-centered. Mem-
bers in task-oriented roles move the group toward the attainment of its
goal. Members in maintenance roles provide social support for the group.
Members in self-centered roles focus on attaining their individual goals
while, at the same time, impeding the attainment of the group’s goals.
Fisher and Ellis (1990) note that, with the exception of one (or more)
leader, no single informal role is found in all groups, which is the reason
why leader emergence and dynamics are the focus of this chapter.
Definition of leader
A leader is a person who directs, guides, or influences the thoughts, feel-
ings, actions, or direction of a group. Scholars who have studied leader
emergence have used a variety of measures for identifying group leaders.
Leaders can be identified through their assigned role in the group (for
formal leaders); through observing their behavior toward group mem-
bers (e.g., high participation, dominance, assertiveness); or through the
behavior of group members toward them (the person whom group mem-
bers address most). Leaders can also be identified by asking group mem-
bers, and examining the level of consensus among them. This chapter
reviews studies that use a range of measures to identify leaders. It is not
common to find studies that use multiple measures to identify group
leaders.
Leader emergence in face-to-face groups

One of the leading theories about how leaders emerge in groups is expec-
tation states theory (Berger et al., 1985; Ridgeway, 2001; see also social
role theory, Eagly, 1987). This is one of the few theories that examines
both personal attributes and situational factors on leadership emergence.
Expectation states theory suggests that people hold shared cultural beliefs
(group stereotypes) that confer greater social status, competence, and
other positive attributes on members of certain social categories (e.g.,
men) as compared with others (e.g., women). These shared cultural
beliefs can affect the behavior and evaluations of group members by cre-
ating performance expectations that affect the degree to which group
members participate, have influence, and are evaluated by others in their
group.
The behavior of group members often conforms to performance expec-

tations based on these shared cultural beliefs. Thus, shared cultural
beliefs can have self-fulfilling effects and can influence leader emer-
gence in groups: people from socially advantaged groups are more
likely to emerge initially as leaders than people from socially disadvan-
taged groups. For example, in a meta-analysis, Eagly and Karau (1991)
found that men were more likely to emerge as leaders than women in
short-term groups and in groups that did not require complex social
interaction.
Although much of the research on expectation states theory focuses on
members’ gender stereotyping, members may engage in other forms of
stereotyping in their initial perceptions of leaders (Ridgeway, 2001). For
instance, in an organizational context, members higher in the formal hier-
archy, or with longer tenure, may initially be perceived as leaders. Racial
and ethnic stereotyping is another factor that appears to affect leader
emergence, although little actual data exists on that topic (Morrison and
Von Glinow, 1990).
Salient characteristics other than demographically based stereotypes
(such as gender, level in the hierarchy, tenure, ethnicity) may also affect
patterns of role emergence. For example, the seating pattern of group
members has been shown to affect emergent leadership. Group members
who are most centrally seated (Leavitt, 1951) or are seated at the end of
a table (Strodbeck and Hook, 1961) have a greater chance of emerging
as leaders than those seated at other locations around a table. Leathers
(1986) suggested that having visual access to all group members was the
main reason for the seating position advantage.
Participation and leader emergence

The most consistent empirical finding in the body of research on leader
emergence is that the member who speaks most in the group is often rated
or identified by others as the leader, with a couple of qualifications (see
Bonito and Hollingshead, 1997, for a more extensive review). For exam-
ple, an early study by Bass (1949) examined perceptions of leadership
in two initially leaderless groups across multiple sessions. Independent
observers ranked each member on thirteen leadership status traits. A
record was made of the amount of time each participant talked at each
session. The correlation between time spent talking and the thirteen lead-
ership traits was .93, suggesting a strong and positive association between
participation and leader emergence. The researchers varied group com-
position across the sessions and did not measure whether the same people
emerged as leaders across sessions.
Morris and Hackman (1969) found that members perceived to be

leaders participated significantly more than other members on a variety
of tasks, including production, discussion, and problem solving in ad
hoc groups. When the quantity of each member’s participation was par-
tialled out of the analysis, no noteworthy behavioral differences between
perceived leaders and other members remained. However, Morris and
Hackman (1969) noted that high participation was neither a sufficient
nor a necessary condition for perceived leadership in their experiment.
Some high participators were not perceived as leaders and made con-
tributions that were detrimental to group creativity and performance.
This qualification should not be surprising to anyone who has worked in
groups. The relation between quantity and quality of members’ partic-
ipation in groups can vary within and across groups. Often the relation
is positive, but it can sometimes be negative. Hence the labels “Know it
all,” “Big mouth,” and “Windbag” are often used to designate members
who are high on participation quantity but low on quality.
Task complexity is another important consideration regarding the rela-
tion between participation and leadership in groups. Some group tasks
require specific expertise and knowledge; others require little, if any,
expertise and knowledge. Ginter and Lindskold (1975) investigated the
impact of task ambiguity, expertise, and participation in groups. The
researchers found that when the task did not require expertise, high par-
ticipating group members were chosen to be leaders. However, partici-
pation was not related to leader choice when the task required expertise:
both low and high participating experts were chosen as leaders.
In sum, the quantity of member participation is the most reliable pre-
dictor of leader emergence in the first encounters of face-to-face groups.
Yet it is imperfect, and is likely to be much less reliable in groups working
on a task which requires special knowledge or expertise. Little is known
about the stability of informal leadership over time in face-to-face groups.
Participation and group size

Group size is an important determinant of how much and how often
each member contributes to the group. After reviewing group partici-
pation research, McGrath (1984) made several conclusions about the
relations between group size and member participation. First, the most
active group member tends to participate approximately 40–45 percent
of the time, regardless of group size. Second, disparities in participa-
tion among members tend to be minimized in smaller groups, and tend
to increase in larger groups. Third, the amount of participation by the
most active member, relatively speaking, remains constant whereas other
members contribute proportionately less to the group discussion. For

example, in groups with three members, the most active member partici-
pates 44 percent of the time on average whereas the least active member
participates 23 percent. In groups with ten members, the most active
member participates 43 percent whereas the least active member partici-
pates 3 percent. A similar pattern was discovered in much larger groups.
Mitchell (2006) investigated the relation between meeting size and mem-
ber participation during the business meetings of Quakers, a religious
group, which highly values egalitarianism and consensus building. In
small meetings (group size = 10), approximately 90 percent of the atten-
dees spoke at least once; in large meetings (group size = 1,000), only
about 10 percent of the attendees spoke at least once.
In sum, group size affects how opportunities to speak are distributed
among members in face-to-face groups. Moreland and Levine (1992)
offer several explanations for this phenomenon. First, larger groups tend
to develop status systems, and such systems are responsible for par-
ticular distributions of speaking opportunities. Second, satisfaction of
some members decreases as groups become larger, a finding that suggests
members are more likely to become alienated from the group and, as a
consequence, make fewer contributions to discussion. Finally, members
feel more efficacious in smaller groups; they feel that their contributions
play a part in the group’s decisions. Some members in larger groups
feel less efficacious and are less likely, therefore, to offer contributions to
groups. In addition to those described by Moreland and Levine (1992),
the norm that only one member speaks at a time creates fewer oppor-
tunities and more competition for speaker time in large than in small
face-to-face groups.
What we know about leader emergence in online groups
Leader emergence in computer-mediated groups: the early years

Early research on leader emergence in online groups was conducted in
the late 1980s and early 1990s before access to the internet was widely
available. At that time, systems that enabled computer-mediated com-
munication for groups involved the exchange of text-based messages.
Some systems, known as group decision support systems, required group
members to be collocated, supported synchronous communication, and
provided tools to facilitate decision making. Other systems provided only
communication, supported distributed groups, and allowed groups to
communicate asynchronously (see McGrath and Hollingshead, 1994,
for a review).
At that time, the predominant thinking regarding the differences

between computer-mediated and face-to-face communication was that
text-based communication among dispersed group members may deper-
sonalize group interaction, so that group members concentrate more
on the messages and less on the people involved in the communication
(Sproull and Kiesler, 1991). In other words, group members lose access
to social context cues that communicate the status and individuality of
participants.
Kiesler and Sproull (1992) differentiated between dynamic and static
social context cues. Static cues result from the manifest characteristics of
the individual members, such as gender and ethnicity, or accompanying
artifacts, such as formal position in the organizational hierarchy, business
attire, or the chair situated at the head of the table. Dynamic cues stem
from people’s behavior, such as nodding, smiling, gaze, and eye con-
tact. Many static and dynamic cues are absent, or severely attenuated,
in computer-mediated communication. For example, static cues such as
gender and ethnicity of group members in computer-supported interac-
tions are not as salient as in face-to-face interactions. Culnan and Markus
(1987) noted that dynamic cues are important sources of feedback and
reinforcement to emerging leaders in small, initially leaderless groups.
Members of computer-mediated groups may lose at least some informa-
tion cues that provide feedback regarding the impact of their behavior on
other group members (e.g., nonverbal sanctioning acts: glares, interrup-
tions; nonverbal signs of approval: nods).
Consequently, group members engaged in computer-mediated com-
puter interaction tend to feel more anonymous and detect less individu-
ality in their communication partners, participate more equally (because
low-status members are less inhibited), and experience more difficulty in
attaining group consensus both because of elimination of interpersonal
feedback and because of reduced concern with social norms (Kiesler and
Sproull, 1992).
Empirical support exists for this perspective, but almost all of that
research was conducted at only one point in time on ad hoc groups.
Strickland et al. (1978) found that mediated communication reduced
group consensus about leadership and increased the groups’ focus on
the task at hand. Several researchers found that an informal leader was
more likely to emerge in face-to-face groups than in computer-supported
groups in a one-session lab experiment (e.g., Poole et al., 1993; Watson
et al., 1988). Some researchers observed less domination by one indi-
vidual in computer-supported groups (e.g., Vogel et al., 1987) and more
equal participation among members in computer-supported groups than
in face-to-face groups (e.g., Easton et al., 1992; Siegel et al., 1986; Straus
and McGrath, 1994). Wickham and Walther (2007) found that computer
groups often identified more than one member as the leader. Hiltz et al.
(1991) showed that providing computer-mediated groups with a tool to
designate a discussion leader helped groups reach consensus on deci-
sions. It is important to note, however, that several studies observed no
differences in participation patterns between face-to-face and computer-
supported groups (e.g., Poole et al., 1991; Walther and Burgoon, 1992).
More recently, research has examined characteristics of members iden-
tified by the majority of their peers as informal leaders of computer-
mediated groups. Wickram and Walther (2007) found that emergent
leadership in computer-mediated groups was associated with high levels
of participation, perceived intelligence, encouragement, and authoritari-
anism in a laboratory experiment with undergraduate students. Yoo and
Alavi (2002) found that emergent leaders sent out more emails to group
members in general, more task-relevant emails, and longer emails than
other members of project teams composed of students in an executive
development program. In addition, they found that demographic char-
acteristics such as age, job experience, and experience at their current
position were not associated with emergent leadership (gender was not
reported; there were many more men than women in the sample).
Over the last fifteen years, many scholars have moved away from the
technologically deterministic view embodied in the “cues-filtered-out”
approach, proposing instead that the uses and effects of communication
technologies are better studied from an “emergent perspective” (Con-
tractor and Eisenberg, 1990; Fulk et al., 1990; Poole and DeSanctis,
1990). The emergent perspective is based on the assumption that the
uses and effects of communication technologies depend more on the
interplay between social structures, relationships, and interactions among
users than on the technology’s capabilities and features. Groups adapt
communication technologies to fit their purposes and needs. An emer-
gent perspective leads to a greater focus on the process of role emergence
rather than its presence or absence.
The process of leader emergence and how it unfolds over time in
initially leaderless computer-mediated groups had not been addressed
in previous research. Many earlier studies about leader emergence in
computer-mediated groups were conducted in a single session, and, with
a few exceptions, there was more decentralization and lower levels of
consensus about leaders compared with face-to-face groups. Many dis-
tributed work teams, committees, and the like collaborate across many
points in time. Does leader emergence just take longer in computer-
mediated groups or is it unlikely to happen at all? Study 1 addresses this
issue.
Study 1
Hollingshead and Contractor (1994) compared the dynamics of leader
consensus in initially leaderless face-to-face and computer-mediated
work groups across multiple meetings. The study was part of a larger
longitudinal project with student project teams conducted in a large
course on the social psychology of organizations (see McGrath, 1993,
for a comprehensive description of the experimental design). The work
groups were told they were the “employees” of a large consulting orga-
nization and each week they would carry out assignments for differ-
ent “client” organizations. Group size ranged between three and five
members. The computer-mediated teams interacted through a com-
puter network, which enabled only text-based communication. After each
week’s assignments, group members completed a questionnaire individ-
ually that examined their reactions to the task, their group, and one
another.
Leader consensus was measured in ten face-to-face and ten computer-
mediated project groups across five points in time. At each time point, the
groups worked on and completed a different assignment. One question
on the weekly questionnaire completed by each member asked partic-
ipants to list the roles that each group member (including themselves)
had assumed during that week’s task.
The findings indicated that leader consensus was higher in face-
to-face work groups than in computer-mediated work groups initially
(see Figure 6.1). At Time 1 (the group’s first assignment), face-to-face
work groups had significantly higher levels of leader consensus than
computer-mediated work groups (0.55 vs. 0.21). However, leader con-
sensus declined significantly across the five sessions in face-to-face work
groups. Computer-mediated work groups showed fluctuations in leader
consensus over time, but these fluctuations were not statistically signif-
icant. Neither were the differences in leader consensus between face-
to-face and computer-mediated groups in weeks 2, 3, 4, or 5. Overall,
the mean leader consensus across the five sessions did not differ signifi-
cantly between the face-to-face and computer-mediated work groups (M
= 0.29; 0.29 respectively).
One possible explanation for the findings is consistent with expectation
states theory. Face-to-face groups may have initially inferred common
stereotyped expectations about each member’s likelihood to assume a
leader role. Those expectations may be based on salient member char-
acteristics (e.g., gender, ethnicity, wearing glasses, etc.) and as a result
groups showed high levels of role consensus among members. Group
members’ expectations may have changed as they learned more about
0.6
0.5
0.4
FTF
0.3
CMC
0.2
0.1
0
Time 1 Time 2 Time 3 Time 4 Time 5
Figure 6.1 Mean leader consensus scores by media and time period in
ten face-to-face (FTF) and ten computer-mediated (CMC) groups
Source: Hollingshead and Contractor (1994).
each other’s abilities and weaknesses through their subsequent inter-

actions. Consequently, face-to-face groups showed decreasing levels of
leader consensus over time.
The attenuated pattern of role consensus over time for computer-
mediated groups in this study may be due to the reduced salience of
members’ manifest characteristics (Kiesler and Sproull, 1992). In addi-
tion to the studies reviewed above, the results are consistent with case
study field data. Eveland and Bikson (1989) studied two task forces in an
organization: one that corresponded via electronic mail exclusively and
one that had no communication restrictions (standard). The email task
force had less centralization and less stable leadership than the standard
task force.
It is worth noting that from one session to the next there were non-
systematic differences in the relative magnitude of leader role consen-
sus scores between groups in the two media conditions. This may help
explain the contradictory findings across cross-sectional studies, which
could have been conducted at different points in the groups’ role devel-
opment process. Since the majority of studies were one-shot experimen-
tal studies, it is not surprising to find many studies reported the initial
difference in leader role consensus found in this study. Those results
would imply that those differences between face-to-face and computer-
mediated work groups would endure over time. However, the results of
the present study indicate that the difference in leader role consensus
90
80
70
60
Group size
50
40
30
20
10
0
0 2 4 6 8
Number of leaders
Figure 6.2 Number of leaders by group size in twenty-four internet

support groups
Source: Alexander (2003).
scores between the face-to-face and computer-mediated work were not

statistically significant in any of the subsequent sessions.
Leader emergence in internet groups

Time constraints and turn taking are not issues in many computer-
mediated interactions where members can participate asynchronously
or simultaneously. In addition to few geographical and temporal con-
straints, many internet groups do not have constraints on group size.
An important question is, how does the elimination of these constraints
affect leader emergence in internet-based groups? Study 2 examined this
question.
Study 2
Alexander (2003) investigated the relations between participation, group
size, leader emergence, and stability in twenty-four internet-based sup-
port groups. The support group topics included anxiety, attention deficit
disorder, breast cancer, depression, diabetes, eating disorders, fibromyal-
gia, and multiple sclerosis. Alexander observed the participation patterns
in three-month intervals in the first year in the history of each support
group (four time periods).
Figure 6.2 plots the relation between the mean group size and mean
number of high participators or “leaders” over the year for each of the
twenty-four groups.2 It is evident that the number of members accounting

for most of the messages remained very small even as the group size
approached 100 members. In twenty-three out of twenty-four support
groups, there were four or fewer “leaders.” Only one group had a single
leader. Despite fewer constraints in internet support groups, the relatively
small ratio of leaders to group size is similar to face-to-face groups.
This finding is consistent with Wickham and Walther (2007): computer-
mediated groups often had more than one leader.
Alexander (2003) also examined the degree to which the same group
members emerged as informal leaders across two consecutive time peri-
ods. All twenty-four groups showed some degree of leader change, but
most groups also showed some consistency over the year. Twenty-one
groups had at least one person who was a leader across two consecutive
time periods; three groups showed no leader consistency across any two
time periods (see Figure 6.3). There were no groups where all leaders
were the same across all four time periods. These findings suggest that
leader participation patterns are likely to vary in the first year of the
group’s history.
Conclusions and future research

Three questions were asked in the introduction:
1. How do people rise to positions of influence in online groups?
2. How does online group leadership evolve and change over time?
3. To what extent is the process of leader emergence in online groups
similar to traditional face-to-face groups?
I will provide a few very tentative answers to these questions based
on the review and studies presented in this chapter. Clearly, much more
research is needed across more types of online groups, and at multiple
points in time.
In response to the first question, high participation seems to be posi-
tively associated with emergent leaders in online groups as it is in face-
to-face groups. Even though online groups have fewer constraints, a
relatively small number of group members hold leader positions and
2 Leaders in small face-to-face groups account for between 33.3 and 50 percent of
the total participation (Bonito and Hollingshead, 1997). The percentages of messages
accounted for by the highest participators were added together until the sum was between
33.3 percent and 50 percent. For example, if the most frequent participant accounted
for 42 percent of messages, only one member was identified as the group’s leader. If the
most active participant accounted for 30 percent of all messages and the second most
active participant accounted for 15 percent, then together they accounted for 45 percent
(which would fall into the 33.3–50 percent range). In that group, two members would
be identified as leaders.
Number of groups with same leader(s) 10
0
0 1 2 3
Number of consecutive time periods
Figure 6.3 Leader consistency in twenty-four internet support groups

by consecutive time periods (max = 3)
Source: Alexander (2003).
account for most of the participation in the group. However, this is likely
not to be the case for all internet-based groups. For example, Turner
et al. (2005) identified seven different patterns of participation in a social
network analysis of Usenet meta-data: Answer Person, Questioner, Troll,
Spammer, Binary Poster, Flame Warrior, and Conversationalist.
In the case of Usenet groups and other discussion forums where peo-
ple ask and answer technical questions and seek information, the con-
tributors’ qualifications and the quality of their responses are likely to
be more important in determining “leaders” than the quantity of their
participation. Members of internet groups have more control over self-
presentation and messages can be anonymous to some extent, so mem-
bers may have somewhat more freedom to decide on the roles they will
take within the group than their face-to-face counterparts. They can iden-
tify their messages with their given names, a pseudonym, or no name to
maintain a sense of anonymity. Members also express and develop identi-
ties through signature files, which appear at the end of messages and may
consist of quotes, pictures, professional titles, or other defining informa-
tion. Members can also provide information about their background and
relevant experiences in the body of the message itself. So group members
have control over the amount and content of personal information that
others receive from them.
Members’ roles in online groups can be linked to their responsibilities

within the group and to their identities outside of the group (Putnam
and Stohl, 1990). Identities outside of a group may be influenced by a
title, prestige, or status imported from outside the group. Something as
simple as an email address provides information that can serve as a status
label and can include clues about the name, work place, occupation, gen-
der, and nationality of the sender. For example, a post from someone with
an MD in their signature file may be afforded high status by an internet
support group on adult-onset diabetes but not by an online community
on patient rights and malpractice. This example shows how the commu-
nication of personal attributes can enhance or reduce the likelihood that a
particular person may emerge as leader depending on the group context.
In response to the second question, group leadership appears to be
fluid rather than static in online groups. In the studies presented, most
online groups had more than one member with high levels of participa-
tion, and more than one member was perceived to be leader at any given
point in time. All online groups had some degree of leader change over
time – with some showing more change than others.
And finally, in response to the third question, there appear to be dif-
ferences initially in the patterns of leader consensus for face-to-face and
computer-mediated groups that dissipate over time. Face-to-face groups
had a relatively high degree of leader consensus in the first session that
declined over time. In contrast, the computer-supported groups had a
relatively low degree of consensus initially, that persisted over time. At
later points in time, patterns of leader consensus between face-to-face
and online groups did not differ significantly.
To conclude, there are many unanswered questions for researchers
interested in leader emergence in online groups. The barrier between
the online and offline worlds of groups is becoming more blurred, and
is a topic worthy of study. Multiplayer online games and virtual worlds,
such as “World of Warcraft” and “Second Life,” provide players with
opportunities to take on leader roles. Players can learn, experiment with,
and apply leadership skills gained in the virtual world to real-world tasks.
An article in the New York Times (Schiesel, 2008) described a plan by
CCP, the creators of the game Eve Online, that allowed players to elect
their own representatives to the “Council of Stellar Management.” The
nine council representatives are flown to CCP headquarters in Iceland
to express their concerns and suggestions on behalf of the players to the
company in person.
And finally, with regard to social networking, many political candi-
dates, CEOs, and other leaders have Facebook and MySpace pages. How
do leaders represent themselves on those sites, and in turn, how do those
representations affect the impressions of the people they lead? Internet

groups provide users with myriad opportunities to belong to groups and
to assume positions of leadership should they so desire. People who are
shy or otherwise uncomfortable in social situations may feel empowered
and may be more likely to emerge as leaders in online than in face-to-face
groups. Future theory and research on leadership emergence in online
groups should attend to aspects of the person and the situation.
References
Alexander, S. C. (2003). Leadership and social support on the internet: A longi-
tudinal analysis of mental and physical illness groups. Dissertation Abstracts
International Section A: Humanities and Social Sciences, 63(11-A), p. 3788.
Alexander, S. C., Peterson, J., and Hollingshead, A. B. (2003). Support at your
keyboard: A study of on-line support groups. In L. Frey (Ed.), Group com-
munication in context, Vol. 2. Hillsdale, NJ: Lawrence Erlbaum Associates.
R
The American Heritage Dictionary of the English Language, Fourth Edition
(2009). Copyright C 2000 by Houghton Mifflin Company. Boston, MA:
Houghton Mifflin Company.
Bass, B. H. (1949). An analysis of leaderless group discussion. Journal of Applied
Benne, K., and Sheats, P. (1948). Functional roles of group members. Journal of
Social Issues, 4, 41–49.
Berger, J., Wagner, D. G., and Zelditch, M. (1985). Expectation states theory:
Review and assessment. In J. Berger and M. Zelditch (Eds.), Status, rewards,
and influence (pp. 1–72). San Francisco, CA: Jossey-Bass.
Biddle, B. J., and Thomas, E. J. (Eds.) (1969). Role theory: Concepts and research.
New York: Wiley.
Bonito, J. A., and Hollingshead, A. B. (1997). Participation in small groups.
Communication Yearbook, 20, 227–261.
Brandon, D. P., and Hollingshead, A. B. (2007). Categorizing on-line groups.
In A. Joinson, K. McKenna, T. Postmes, and U. Reips (Eds.), The Oxford
handbook of internet psychology (pp. 105–120). Oxford: Oxford University
Press.
Contractor, N. S., and Eisenberg, E. M. (1990). Communication networks and
new media in organizations. In J. Fulk and C. Steinfield (Eds.), Organizations
and communication technology (pp. 145–174). Newbury Park, CA: Sage.
Culnan, M. J., and Markus, M. L. (1987). Information technologies. In F. M.
Jablin, L. L. Putnam, K. H. Roberts, and L. W. Porter (Eds.), Handbook of
organizational communication: An interdisciplinary perspective (pp. 420–444).
Newbury Park, CA: Sage.
Eagly, A. H. (1987). Sex differences in social behavior: A social-role interpretation.
Hillsdale, NJ: Erlbaum.
Eagly, A. H., and Karau, S. J. (1991). Gender and the emergence of leaders:
A meta-analysis. Journal of Personality and Social Psychology, 60(5), 685–
710.
Easton, A., Vogel, D. R., and Nunamaker, J. F. (1992). Interactive versus stand-
alone group decision support systems for stakeholder identification and
assumption surfacing in small groups. Decision Support Systems, 8, 159–168.
Eveland, J. D., and Bikson, T. K. (1989). Work group structures and computer
support: A field experiment. ACM Transactions on Office Information Systems,
6, 354–379.
Fisher, B., and Ellis, D. (1990). Small group decision making: Communication and
the group process. New York: McGraw-Hill.
Fulk, J., Schmitz, J., and Steinfield, C. (1990). A social influence model of
technology use. In J. Fulk and C. Steinfield (Eds.), Organizations and com-
munication technology (pp. 117–140). Newbury Park, CA: Sage.
Ginter, G., and Lindskold, S. (1975). Rate of participation and expertise as
factors influencing leader choice. Journal of Personality and Social Psychology,
32, 1085–1089.
Hiltz, S. R., Johnson, K., and Turoff, M. (1991). Group decision support: The
effects of designated human leaders and statistical feedback in computerized
conference. Journal of Management Information Systems, 8, 81–108.
Hollingshead, A. B., and Contractor, N. S. (1994). Dynamics of leader consen-
sus in continuing face-to-face and computer-mediated work groups. Paper
presented at the annual meeting of the Speech Communication Association,
Chicago, November.
(2002). New media and organizing at the group level. In L. A. Lievrouw and
S. Livingston (Eds.), Handbook of new media: Social shaping and consequences
of ICTs (pp. 221–235). London: Sage.
Katz, D., and Kahn, R. L. (1966). The social psychology of organizations. New
York: John Wiley.
Kiesler, S., and Sproull, L. (1992). Group decision making and technology.
Organizational Behavior and Human Decision Processes, 52, 96–123.
Leathers, D. (1986). Successful nonverbal communication: Principles and applica-
tions. New York: Macmillan.
Leavitt, H. (1951). Some effects of certain communication patterns on group
performance. Journal of Abnormal and Social Psychology, 51, 704–706.
McGrath, J. E. (1984). Groups: Interaction and performance. Englewood Cliffs,
NJ: Prentice Hall.
(1993). The JEMCO workshop – Description of a longitudinal study. Small
Group Research, 24, 285–306.
McGrath, J. E., and Hollingshead, A. B. (1994). Groups interacting with technology.
Newbury Park, CA: Sage.
Mitchell, E. (2006). Participation in unanimous decision-making: The New Eng-
land monthly meetings of Friends. Philica.com Article number 14.
Moreland, R. L., and Levine, J. M. (1992). The composition of small groups. In
E. J. Lawler, B. Markovsky, C. Ridgeway, and H. Walker (Eds.), Advances
in group processes, Vol. 9 (pp. 237–280). Greenwich, CT: JAI Press.
Morris, C. G., and Hackman, J. R. (1969). Behavioral correlates of perceived
leadership. Journal of Personality and Social Psychology, 13, 350–361.
Morrison, A., and Von Glinow, M. (1990). Women and minorities in manage-
ment. American Psychologist, 45(2), 200–208.
Poole, M. S., and DeSanctis, G. (1990). Understanding the use of group

decision support systems. In C. Steinfield and J. Fulk (Eds.), Organi-
zations and communication technology (pp. 175–195). Newbury Park, CA:
Sage.
Poole, M. S., Holmes, M., and DeSanctis, G. (1991). Conflict management
in a computer-supported meeting environment. Management Science, 37(8),
926–953.
Poole, M. S., Holmes, M., Watson, R., and DeSanctis, G. (1993). Group deci-
sion support systems and group communication: A comparison of decision
making in computer-supported and non-supported groups. Communication
Research, 20, 176–213.
Putnam, L., and Stohl, C. (1990). Bona fide groups: A reconceptualization of
groups in context. Communication Studies, 41, 248–265.
Ridgeway, C. L. (2001). Gender, status, and leadership. Journal of Social Issues,
57, 637–655.
Schiesel, S. (2008). Face to face: A council of online gamers. New York Times,
28 June.
Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. (1986). Group processes
in computer-mediated communication. Organizational Behavior and Human
Decision Processes, 37, 157–187.
Sproull, L. S., and Kiesler, S. (1991). Connections: New ways of working in the
networked organization. Cambridge, MA: MIT Press.
Straus, S., and McGrath, J. E. (1994). Does the medium matter? An investigation
of process and performance in computer-mediated and face-to-face groups.
Journal of Applied Psychology, 79(1), 87–97.
Strickland, L. H., Guild, P. D., Barefoot, J. C., and Paterson, S. A. (1978).
Teleconferencing and leadership emergence. Human Relations, 31, 583–
596.
Strodbeck, F., and Hook, L. (1961). The social dimensions of a twelve-man jury
table. Sociometry, 24, 397–415.
Turner, T. C., Smith, M. A., Fisher, D., and Welser, H. T. (2005). Pic-
turing Usenet: Mapping computer-mediated collective action. Journal of
Computer-Mediated Communication, 10(4), article 7. http://jcmc.indiana.edu/
vol10/issue4/turner.html
Vogel, D., Nunamaker, J., Applegate, L., and Konsynski, B. (1987). Group deci-
sion support systems: Determinants of success. Proceedings of the 7th Inter-
national Conference on Decision Support Systems, 8–11 June.
Walther, J. B., and Burgoon, J. K. (1992). Relational communication in
computer-mediated interaction. Human Communication Research, 19, 50–88.
Watson, R. T., DeSanctis, G., and Poole, M. S. (1988). Using a GDSS to facil-
itate group consensus: Some intended and unintended consequences. MIS
Quarterly, September, 463–478.
Wickham, K. R., and Walther, J. B. (2007). Perceived behaviors of assigned and
emergent leaders in virtual groups. International Journal of e-Collaboration, 3,
1–17.
Wikipedia (2008). http://en.wikipedia.org/wiki/Second Life.
Yoo, Y., and Alavi, M. (2002). “Electronic mail usage pattern of emergent lead-
ers in distributed teams.” Sprouts: Working Papers on Information Environ-
ments. Systems and Organizations, Vol. 2, Summer. sprouts.aisnet.org/211/1/
020309.pdf
Zaccaro, S. J., Kemp, C., and Bader, P. (2004). Leader traits and attributes. In J.
Antnakis, A. T. Cianciolo, and R. J. Sternberg (Eds.), The nature of leadership
(pp. 101–124). Thousand Oaks, CA: Sage Publications, Inc.
7 Ostracism in cyberspace
Being ignored and excluded in

electronic-based interactions
Eric D. Wesselmann and Kipling D. Williams
This chapter presents a basic overview of ostracism, describing the rea-

sons for its use, the situations in which it can occur, and the effects
it has on its targets. Ostracism can happen in both face-to-face and
electronic-based (e.g., internet, cell phones, and virtual reality) inter-
personal interactions. The authors describe a variety of experimental
research methods that have been used to study the effects of ostracism
in both of these interpersonal contexts, comparing and contrasting the
results of these different studies. Particular attention is paid to stud-
ies involving electronic-based interactions. Both immediate and long-
term effects of ostracism on targets’ behaviors are discussed, focusing
specifically on how people can be negatively impacted by being ostra-
cized in an electronic-based interaction. The authors discuss how the
nature of electronic-based interactions may affect people’s reactions to
ostracism in novel ways, providing greater understanding of the simi-
larities and differences between face-to-face and electronic-based inter-
actions. These findings are reexamined and explained in the context of
virtual bravado – uninhibited behavior perhaps due to the anonymity
provided by electronic-based communication. Finally, future directions
of research are suggested for studying ostracism in electronic-based
interactions, as well as people’s general behavior in these types of
interactions.
Introduction
There are several needs that are essential to our lives as human beings,
and the desire to satisfy these needs influences our attitudes, behaviors,
and interactions. Four of the most fundamental human social needs are
The authors would like to thank Alvin T. Law, Amanda Wesselmann, and James H.
Wirth for substantive comments on earlier versions of this chapter.
127
128 Eric D. Wesselmann and Kipling D. Williams
the needs for belonging, self-esteem, control, and meaningful existence

(Williams, 2001, 2009). Humans have a need to belong – they desire
interpersonal relationships and meaningful interactions (Baumeister and
Leary, 1995). Humans also have a desire to have reasonably high self-
esteem (Leary et al., 1995; Tesser, 1988). Humans also have more action-
focused needs; they desire some degree of control over their environment
(Friedland et al., 1992), and knowledge that their existence matters to
others (Greenberg et al., 1986; Solomon et al., 1991).
Baumeister and Leary (1995) argue that perhaps the most impor-
tant of these needs is belonging. They review considerable literature
from various academic fields, demonstrating how important social rela-
tionships are to a person’s physical, mental, and emotional well-being.
Belonging through social relationships can certainly have an influence
on the other three fundamental needs as well (Williams, 2001). Belong-
ing and the other fundamental needs are often satisfied in individuals’
everyday social interactions with other people. These interactions were
traditionally face-to-face, but with the advent of different modes of com-
munication humans can now communicate with each other regardless of
time constraints and geographic location (Bargh and McKenna, 2004;
McKenna and Bargh, 1999). Despite the debate concerning the possi-
ble harmful effects of the internet on people’s social relationships (Kraut
et al., 1998), several research studies have demonstrated that people do
form lasting and meaningful relationships on the internet. The internet
(and other forms of electronic-based communication) has unique char-
acteristics that afford people opportunities to overcome obstacles that
normally inhibit face-to-face social interactions. Several studies have
found electronic-based communication can be helpful for individuals
who are usually inhibited by social anxiety, loneliness, or lack of social
skills (McKenna and Bargh, 1999; McKenna et al., 2002; Peter and
Valkenburg, 2006; Peter et al., 2005; Reid and Reid, 2007; Sheeks and
Birchmeier, 2007; Stritzke et al., 2004). Other research has found that
electronic-based communication can lead to increased organizational
equality or influence across status and expertise (Dubrovsky et al., 1991),
and it can help people who possess stigmatized identities (e.g., non-
mainstream sexual orientation or political views) be more confident in
expressing their true selves to friends and family (McKenna and Bargh,
1998).
Even though the internet boasts many social advantages, this new
mode of communication also affords individuals more opportunity to
experience the darker side of social interactions – interpersonal rejection,
exclusion, and ostracism (Williams et al., 2002). Anecdotally, these neg-
ative treatments seem to lose none of their sting by being transferred to
Ostracism in cyberspace 129
a different communication medium. In some extreme cases, people have

reportedly committed suicide as a result of these negative treatments over
electronic-based communication (ABC News, 2007; ABC Primetime,
2006; Kowalski et al., 2008; Smith and Williams, 2004). The theoretical
similarities and differences between rejection, exclusion, and ostracism
(as well as their effects) are still being debated (Williams, 2009); even so,
it is difficult to review the literature on one without discussing the others
(Leary et al., 2006). This chapter will discuss research involving all three
concepts, but the primary focus will be on ostracism research.
Ostracism – being ignored and excluded – can be experienced at
the individual or group level, and has many harmful effects (Williams,
1997, 2001, 2009). Ostracism has immediate and robust psychologi-
cal and physical consequences. Myriad studies have demonstrated that
the experience of ostracism thwarts the four fundamental human needs
and significantly increases negative moods (Williams, 2009). Other stud-
ies that have focused on physiological measures have demonstrated that
ostracism has negative physical effects on targets – blood pressure often
increases, and fMRI data suggest the part of the brain involved in experi-
encing physical pain is also activated while experiencing ostracism (Eisen-
berger et al., 2003; Williams and Zadro, 2004). Prolonged exposure to
ostracism can have profound and various effects on behavior: people have
been known not only to suffer from psychological disorders or engage
in self-destructive behaviors (Williams, 2001), but in extreme cases may
behave violently towards others (Leary et al., 2003, 2006; Williams and
Wesselmann, 2011).
Ostracism can occur in three main modes – physical, face-to-face, and
cyber-ostracism. Physical ostracism involves being ignored and excluded
by being physically separated or isolated from the group. An illustrative
example of this is when small children are placed in “time out” for
misbehaving. Face-to-face ostracism involves being ignored and excluded
in the physical presence of others. An example of this mode is when a
child is left out of peer activities at school. Cyber-ostracism occurs through
electronic media other than face-to-face interactions, when recognition
and communication are anticipated within an acceptable time frame but
do not occur. The medium could be as basic as electronic mail or text
messaging, or could involve real-time social interactions such as internet
chat rooms or online games (Williams et al., 2000). Cyber-ostracism
is the primary focus of this chapter. Research comparing face-to-face
vs. cyber-ostracism has found little difference in the immediate effects
of ostracism (Williams et al., 2002). The authors of this chapter argue
that the most significant difference is in effects of these two modes of
ostracism on behavioral reactions, and that the difference may be due in
part to the anonymity and lack of social cues inherent in electronic forms
of communication.
Studying cyber-ostracism
Several programs of research have examined the effects of cyber-
ostracism on individuals and their subsequent behavior. The majority
of these programs have focused on experimental methods, using a vari-
ety of electronic media, constructs, and manipulations of ostracism to
gain a more thorough understanding of the intricacies of these phenom-
ena (Campbell and Fiske, 1959). These methods have been as simple as
text messages and as complex as online interactions that give a variety of
social and contextual cues that may affect perceptions of ostracism.
Perhaps the most common paradigm for studying social ostracism is
Cyberball, a virtual online ball-tossing game (Williams et al., 2000).1
This game is an electronic version of the original face-to-face paradigm
developed by Williams and Sommer (1997) to experimentally manip-
ulate ostracism in a laboratory setting. Cyberball not only allowed for
a more practical and controlled examination of ostracism but also gave
researchers an opportunity to find out how powerful ostracism is and
what are the minimal conditions necessary for ostracism to elicit its neg-
ative effects. Participants were led to believe they were participating in
a real-time group exercise in mental visualization. They were placed in
a three-person game of ball toss (in reality, the other two players were
virtual confederates pre-programmed to either include or ostracize the
participant). They were assured the point of the game was not their
ball-tossing performance but how well they mentally visualized the game
setting (e.g., the characteristics of the other players, the weather con-
ditions). In the Inclusion condition, confederates tossed the ball to the
participant approximately 33 percent of the time (equal division of tosses
between each of the three players); in the Ostracism condition, partic-
ipants received the ball once or twice from each confederate and then
were subsequently ignored and excluded for the remainder of the game.
Afterwards, participants were asked a variety of questions about their
thoughts and feelings during the game. The goal of these questions was
to index the degree to which participants recognized they were being
ostracized, as well as to obtain a measure of mood and need satisfaction
for the four basic needs (Williams et al., 2000).
1 Cyberball (for either MAC or PC) is available for free download at www1.psych.purdue.
edu/∼willia55/Announce/cyberball.htm
A second method of manipulating ostracism in an online interaction

used chat rooms (Williams et al., 2002). Participants entered a chat
room with two other confederates and began chatting about a variety
of topics. Ostracism in this paradigm has been manipulated in a variety
of ways. In some studies participants were informed via instructions of
their impending ostracism, and in other studies the confederates simply
began ostracizing the participant during the interaction. Researchers have
also manipulated the type of topic being discussed, the degree to which
confederates agree or disagree with the participant, and the in-group/out-
group status of the participant and confederates.
A third paradigm investigated whether or not ostracism is dependent
upon real-time social interactions. Smith and Williams (2004) utilized
the common phenomenon of text messaging on cell phones to determine
whether ostracism could be detected only when targets see others inter-
acting with them (either in person or with an electronic representation).
There are two main differences between this paradigm and the previous
two cyber-paradigms that allow for examining the necessity of real-time
interactions. The first difference is that in the cell phone paradigm there
was no way for participants to know whether the other people (con-
federates) were in fact communicating with each other and excluding
them, unlike in Cyberball or chat rooms where their interaction (i.e.,
ball tosses or conversation) was clearly visible to the participant. Second,
because participants do not see others interacting, nor do they know if
and when their text message was received by others, it is possible for
participants to attribute their ostracism to technological failures (either
for their phone or the phones of others). This is different to Cyberball or
chat rooms because confederates interact with ostracized participants at
the beginning of each interaction, making it clear to the participant that
the ostracism was not due to technological problems.
The fourth way of examining cyber-ostracism involves implementing
technology that will allow greater opportunity for examining real-time
interactions – virtual reality. Thus far, virtual reality has been used in
two ways: to further examine real-time interactions in settings with more
social information than Cyberball (Cocktail Party World), and to examine
the minimal conditions necessary to detect ostracism (Minimal World).
Cocktail Party World placed participants in a virtual world where they
walked through a cocktail party and interacted with virtual confederates.
Williams et al. (in preparation) manipulated whether the virtual confeder-
ates simply ostracized participants, turned and scowled at them (negative
attention), or even walked through the participant as if they did not exist.
This manipulation allowed the researchers to compare whether negative
attention was considered better than no attention (ostracism), and to
compare two different degrees of ostracism (being unacknowledged vs.

walked through as if one does not exist).
Minimal World, conversely, examined the minimal conditions neces-
sary for people to detect ostracism. The other cyber-paradigms discussed
involve social cues that imply other people are present and involved in the
interaction. Kassner et al. (2010) hypothesized that it is the knowledge
(or the implication) of a social interaction that gives ostracism its power.
Inspired by the Heider–Simmel paradigm (Heider and Simmel, 1944)
where people watched a series of shapes and were asked to construct a
story out of it, Kassner and colleagues placed participants in a virtual
environment where they saw two squares and a sphere in front of them.
This paradigm mimics the Cyberball paradigm, but replaces the player
icons with squares and the ball with a sphere. The sphere moved back and
forth between the two squares, and occasionally moved towards the par-
ticipants’ point of view and disappeared. Participants were instructed to
press one of two buttons when the sphere disappeared – one button sent
the sphere back to the left square and the other button sent the sphere
to the right square. The main manipulation in this research was whether
or not participants were instructed to mentally visualize a “coherent
story” about the movement of the shapes. Using this manipulation, the
researchers could examine whether or not people detect ostracism at a
level devoid of social information and in the absence of instructions to
create such information.
The effects of cyber-ostracism

The research on ostracism has examined two different types of reac-
tions: targets’ immediate reflexive reactions and their subsequent reflec-
tive reactions and behaviors (Williams, 2001, 2007). We will first discuss
the immediate effects of cyber-ostracism, and then discuss the reflective
effects of ostracism and how these may be affected by the situational
forces of electronic-based communication.
The reflexive reaction to ostracism is an experience of robust distress
and pain. These effects appear to be unaffected by individual differences,
attributions for the ostracism, or social/situational factors (Williams,
2009; cf. Wirth et al., 2010). Williams et al. (2000) found that the more
ostracism participants experienced in Cyberball, the worse they felt –
at no point did the participants become numb to the effects. Trait self-
esteem also did not seem to protect people from the immediate effects of
ostracism. The experience of ostracism is quite aversive; people seem to
prefer negative social attention (such as being scowled at) by comparison
(Williams, 2001; Williams et al., in preparation). The aversive power of
ostracism is not limited to those individuals who unexpectedly become

a target. Zadro et al. (2004) found that ostracized participants reported
strong decreases in need satisfaction and more negative mood, regardless
of whether or not they were led to believe the other players were instructed
to ostracize them. Even when told the other players were simply a com-
puter program designed to ostracize them, participants still reported the
same effects on need satisfaction and mood! Finally, other research sug-
gests that individuals do not have to be the targets of ostracism to feel its
reflexive effects; simply observing someone being ostracized can threaten
the observer’s need satisfaction and worsen mood (Wesselmann et al.,
2009).
Ostracism exerts its effect even under situations where people may
expect to find being ignored and excluded desirable. Even when par-
ticipants are ostracized by members of a despised out-group (i.e., the
Ku Klux Klan), they still report the same need satisfaction and mood
effects as participants ostracized by in-group members (Gonsalkorale
and Williams, 2007; see also Smith and Williams, 2004; Williams et al.,
2002). Other research demonstrated that participants still reported feel-
ing the negative effects of being ignored and excluded when their inclu-
sion in a Cyberball game could be detrimental, such as when receiving the
ball cost participants money or raised their chances of being eliminated
from the game (van Beest and Williams, 2006).
Ostracism maintains its sting under conditions of minimal social cues.
Simple non-verbal cues from a digital avatar (e.g., lack of eye contact)
are sufficient to induce ostracism’s negative effects (Wirth et al., 2010).
The effects are not limited to real-time interactions. Smith and Williams
(2004) found those participants who were ostracized during cell phone
text messaging reported the same effect on basic needs as the other
forms of cyber-ostracism that had real-time interactions. An even more
interesting finding was that ostracized participants did not attribute the
ostracism to technological problems; because they did not receive imme-
diate responses they assumed the lack of text was deliberate. Similar
results have been found examining people’s attributions towards lag time
in email exchange and other electronic-based communication (Bargh
and McKenna, 2004; Rintel and Pittam, 1997; Thompson and Nadler,
2002).
Kassner et al. (2010) hypothesized all that was necessary for ostracism
to wreak its proverbial havoc was the presence (or implication) of social
interaction. In Minimal World, participants were either instructed to
mentally visualize a “coherent story” about the movements of the shapes
they witnessed (similar to Heider and Simmel, 1944), or were simply
instructed to press one of two buttons when the sphere disappeared. In
the absence of instructions to infer social information, ostracized partici-

pants were no more likely to report the effects of ostracism than included
participants. To rule out the possibility that it was the instructions to
visualize something, and not the presence of real or inferred social cues,
the researchers replicated the study using the Cyberball paradigm (where
players are represented by human-looking avatars). In this paradigm the
avatars should imply social interaction, regardless of the instructions of
whether or not to construct a story. Results confirmed this hypothe-
sis; regardless of visualization instructions, ostracized participants always
reported the deleterious effects of ostracism in Cyberball. Kassner et al.
(2010) demonstrate ostracism requires merely the implication of social
interaction under minimal conditions; ostracism is likely to be detected
in any form of electronic-based communication because social interac-
tion is implied in myriad ways (e.g., email addresses, chat room screen
names, and online profiles).
Ostracism also has effects on individuals after they have had time to
reflect on their treatment. Zadro, Boland, and Richardson (2006) found
that, after forty-five minutes, the majority of ostracized participants had
recovered their normal levels of basic need satisfaction. An interesting
question arises: do need levels simply return to normal after a period of
time, or do participants behave a certain way to fortify their needs (or at
least speed up the recovery process)? Williams (2009) hypothesizes that
reflective reactions to ostracism (usually behaviors of some type) serve
to fortify the thwarted needs and recover from the deleterious effects of
ostracism. Are some types of behaviors more likely to occur than others in
reaction to ostracism? The research findings are conflicted on this issue –
people react to ostracism with both pro- and anti-social behavior, which
on the surface seems counterintuitive.
Research has demonstrated that ostracized people are more likely than
included people to act pro-socially, perhaps striving to regain social
acceptance. Socially excluded people, compared with those included,
are more likely to want to remain a member of a group (Snoek, 1962),
work harder on a collective group task (Williams and Sommer, 1997),
conform to group opinions (Williams et al., 2000), focus on strategies
for re-inclusion (Molden et al., 2009), and like or show interest in join-
ing new groups (Maner et al., 2007; Predmore and Williams, 1983;
Wheaton, 2001). Socially excluded persons, compared with included
persons, are also more likely to mimic pro-social behavior in organiza-
tions (Ouwerkerk et al., 2005), and engage in unconscious mimicry –
especially with in-group members (Lakin and Chartrand, 2005). Finally,
socially excluded people are more socially attentive (Gardner et al., 2000;
Pickett et al., 2004). Conversely, people are also more likely to behave
aggressively toward other people in response to exclusion and ostracism,

regardless of whether the person was involved or uninvolved in the tar-
gets’ rejection (Buckley et al., 2004; Chow et al., 2008; Twenge et al.,
2001; Twenge and Campbell, 2003; Warburton et al., 2006; Wesselmann
et al., 2010). Many of the students responsible for recent school shoot-
ings have written journals, citing perceived ostracism as the impetus for
their violent behavior (Leary et al., 2003; Williams, 2001).
How does one make sense of these disparate research findings?
Williams (2007) hypothesized the type of behavior utilized will depend
on which of the needs the person wishes to fortify: pro-social behaviors
are more likely to be linked to fortifying the needs for belonging and self-
esteem, and anti-social behaviors are more likely to be linked to the needs
for control and meaningful existence (see also Wesselmann and Williams,
in press). There are two studies that offer compelling support for the rela-
tion between control and anti-social responses to ostracism. Warburton
et al. (2006) found that ostracized participants who were given con-
trol over an unrelated task behaved no more aggressively than included
participants. Ostracized individuals who were not given this opportu-
nity to fortify their need for control responded aggressively. Wesselmann
et al. (2010) extended the argument that control needs influence the
ostracism→aggression relation by arguing that unpredictable ostracism
(the type usually experienced in experimental studies; see Twenge
et al., 2001) provides a double threat for targets – unpredictable ostracism
threatens need satisfaction (e.g., need for control; Williams, 2009), and
also threatens their predictive control over their inclusion in social inter-
actions. They found that individuals who were ostracized unexpectedly
responded with the most aggression, but individuals who still had pre-
dictive control and expected their ostracism were less inclined to respond
aggressively. These two studies suggest the basic ostracism→aggression
link may be more complex than previously thought, with a variety of
situational variables potentially important to this relation.
Anonymity, aggression, and cyber-ostracism

Williams (2009) argued situational cues may influence individuals’ attri-
butions for ostracism and ultimately which needs become more salient
for fortification. One situational aspect common to most studies that
examine the ostracism→aggression link is that participants are afforded
anonymity for their behavior. Participants are not being observed when
they choose to aggress, and they usually do not expect to interact
with their target. This degree of anonymity and lack of face-to-face
interaction (and perhaps accountability) is also found in many forms
of electronic communication. Whether individuals are simply identified

by a self-generated moniker or a known email address, electronic-based
communication lacks many of the standard social cues and norms that
govern face-to-face interactions (Bargh and McKenna, 2004; Christo-
pherson, 2007; McKenna and Bargh, 1999; Skitka and Sargis, 2006).
Regardless of the positives afforded by the anonymity/reduced social cues
of the internet, such as being able to express inner beliefs and emotions
with much less fear of disapproval and discrimination (McKenna and
Bargh, 1999, 2000), research suggests there are also several negative
effects that stem from this anonymity.
In general, research on deindividuation and subsequent uninhibited
behavior suggests anonymity is an important element in this link (Diener,
1979; Diener et al., 1976). The anonymous condition of electronic-based
communication often has deindividuating effects on individuals, which
can produce uninhibited behavior that tends to be more self-centered
and less socially regulated than in face-to-face types of communica-
tion (Kiesler et al., 1984; Sproull and Kiesler, 1986). Some examples
of behaviors that can be considered “uninhibited” in electronic com-
munication are increased “flaming” (i.e., insults or swearing) (Kiesler
et al., 1984; Sproull and Kiesler, 1986), risky or extreme group decision-
making (McGuire et al., 1987; Siegel et al., 1986), increased sexually
harassing comments and behaviors (Barak, 2005), various types of cyber
bullying (Kowalski et al., 2008), and even expressions of hatred, big-
otry, and calls to violence (Coffey and Woolworth, 2004; Lee and Leets,
2002). In one study, electronic-based interactions became so heated the
experimenters escorted participants separately from the building fol-
lowing the experiment for fear of potential conflicts; no such problems
arose in face-to-face communications (Dubrovsky et al., 1991). Based on
these research findings, it appears the lack of social cues and increased
anonymity of electronic-based communications present a double-edged
sword: it is a boon in decreasing social inhibitions and structural hierar-
chy detrimental to interpersonal communication, but can be a catalyst
for anti-social behavior if unchecked.
One phenomenon that emerged in several of the ostracism paradigms
lends support to the link between anonymity of electronic-based com-
munication and anti-normative behavior. In both the cell phone and
chat room paradigms, ostracized participants exhibited behavior that
we call virtual bravado. Many targets of cyber-ostracism responded with
an increase in bold and/or inflammatory language, instead of passively
accepting the ostracism and emotionally withdrawing from the situa-
tion as participants did in face-to-face ostracism (Williams and Sommer,
1997). Sometimes, ostracized participants would type nonsense
sentences simply to interject themselves into the confederates’ conver-

sation (Williams et al., 2002). This virtual bravado seems to be the
most significant difference between reactions to face-to-face and cyber-
ostracism, and fits closely with the research on uninhibited communica-
tion (Dubrovsky et al., 1991; Kiesler et al., 1984; Sproull and Kiesler,
1986). When examining bravado with the need threat model (Williams,
2009), it is possible this bravado was an attempt by participants to fortify
their threatened control and meaningful existence in the situation. It is
also possible that, under conditions of anonymity, needs such as control
or meaningful existence become easier to fortify, whereas under face-to-
face conditions where social norms and accountability are more salient,
needs such as belonging or self-esteem become easier (or more socially
acceptable) to fortify. Future research should examine these issues empir-
ically, and determine what types of situations and personality characteris-
tics lead ostracized people to choose aggressive behavior over other types
of behavior as a way of fortifying their threatened needs.
Future directions for research

There are many elements of ostracism and exclusion in electronic-based
communication that still need to be examined empirically. As previously
discussed, more research is needed on examining the contradictory find-
ings of reflective responses to ostracism: under which circumstances are
people more likely to respond in pro- vs. anti-social ways, and are certain
types of people more likely to choose one type of response than the other?
Williams (2007) proposes reflective reactions to ostracism (whether pro-
or anti-social) are ways in which ostracized individuals fortify their threat-
ened needs for belonging, self-esteem, control, and meaningful existence.
Future research should examine this hypothesis in more detail, as well as
the possibility that certain types of behaviors may fortify specific needs.
Researchers should also examine whether there are individual differences
in which need is more important to fortify. Because research suggests that
people who are socially anxious or lonely may be more likely than oth-
ers to seek acceptance and interactions on the internet (McKenna and
Bargh, 1999; McKenna et al., 2002; Peter and Valkenburg, 2006; Peter
et al., 2005; Reid and Reid, 2007), it is possible being ostracized in an
environment viewed as their last bastion of hope may lead these individ-
uals to redirect their focus from belonging and self-esteem to seek fortifi-
cation of control and meaningful existence instead. If Williams (2009) is
correct in hypothesizing these latter two needs are more likely to be for-
tified by anti-social and aggressive acts, then electronic-based ostracism
and rejection for socially anxious or lonely individuals may have serious
social implications for extreme anti-social behavior (see Williams and

Wesselmann, 2011).
Another interesting area to be researched is whether anonymity is
a factor that motivates people to engage in ostracism. Several studies
have demonstrated that the anonymity and deindividuation afforded
by electronic-based environments can lead to a variety of uninhibited
and often anti-social behaviors in social interactions – behaviors that
are significantly less likely to happen when these same interactions are
face-to-face. Kowalski et al. (2008) discuss several research findings that
indicated the anonymity of the internet gave people the flexibility and
confidence to engage in bullying. Several individuals in different stud-
ies claimed they engaged in cyber bullying because of the anonymity
it affords; these same people indicate they would not bully people in
face-to-face interactions where they could be easily identified and held
accountable. Because ostracism and rejection are such negative and
harmful experiences to people who are subjected to them, it is likely
there are many people who would be hesitant to ostracize others unless
they were confident there would be no negative penalties or backlash
for their behavior. Future research should examine ostracism and rejec-
tion under both face-to-face and electronic environments to see whether
characteristics such as anonymity and accountability affect tendencies to
ostracize others.
Conclusion
Ostracism – being ignored and excluded – is a powerful experience
that has many negative physical, emotional, and interpersonal conse-
quences (Williams, 1997, 2001, 2009). Immediate reactions to ostracism
seem to be robust, regardless of person or situation factors (cf. Wirth
et al., 2010), but reflective and behavioral responses seem to allow for
more variability. Whether people choose to behave pro- or anti-socially,
Williams (2007) suggests their responses serve to fortify their thwarted
needs (i.e., needs for belonging, control, self-esteem, and meaningful
existence). There is debate as to when and why some people respond
differently to ostracism, and considering anonymity (whether due to
experimental parameters or to the mode of communication, such as
electronic-based communication) as a determinant factor would make a
helpful contribution to examining this debate. Considerable research sug-
gests anonymity can lead to myriad anti-social and uninhibited behaviors
(see especially Dubrovsky et al., 1991; Kiesler et al., 1984; Sproull and
Kiesler, 1986); aggressive behavior could certainly be considered an anti-
social and uninhibited behavior. Experimentally, paradigms that study
cyber-ostracism have found evidence of such uninhibited and verbally

aggressive behavior that are absent in face-to-face ostracism paradigms.
We believe this bravado is an attempt to reestablish control over the inter-
action. Williams (2007) theorized anti-social reactions to ostracism are
more likely to fortify needs for control and meaningful existence, and
pro-social reactions are more likely to fortify needs for belonging and
self-esteem. Future research should examine these issues in more detail,
as well as the possibility of individual differences in which need is more
important for an individual to fortify consistently.
Because research suggests certain types of people (i.e., socially anx-
ious or lonely) may be more likely than others to seek acceptance and
interactions on the internet (McKenna and Bargh, 1999; McKenna
et al., 2002; Peter and Valkenburg, 2006; Peter et al., 2005; Reid and
Reid, 2007), it is possible ostracism and rejection in these environments
may affect these individuals in more extreme ways than other people,
having significantly different effects on their behavior. Another area yet
to be explored is under what conditions people are motivated to ostra-
cize or reject another person, and to what degree electronic-based media
facilitate or hinder these behaviors. By exploring these and other areas,
we can gain a richer understanding of the darker side of social relation-
ships and interactions in these forms of communication. The knowledge
gained from this research will in turn aid practitioners and policy-makers
in developing programs and interventions to safeguard individuals who
communicate with others over electronic-based channels.
References
ABC News (2007). Parents: Cyber bullying led to teen’s suicide. Retrieved
27 November 2007, from http://abcnews.go.com/print?id=3882520
ABC Primetime (2006). Primetime: Cyberbullying – cruel intentions [television
broadcast]. American Broadcasting Company, 14 September.
Barak, A. (2005). Sexual harassment on the internet. Social Science Computer
Review, 23, 77–92.
Bargh, J. A., and McKenna, K. Y. A. (2004). The internet and social life. Annual
Review of Psychology, 55, 573–590.
Baumeister, R. F., and Leary, M. R. (1995). The need to belong: Desire for
interpersonal attachments as a fundamental human motivation. Psychological
Bulletin, 117, 497–529.
Buckley, K. E., Winkel, R. E., and Leary, M. R. (2004). Reactions to acceptance
and rejection: Effects of level and sequence of relational evaluation. Journal
of Experimental Social Psychology, 40, 14–28.
Campbell, D. T., and Fiske, D. W. (1959). Convergent and discriminant valida-
tion by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–105.
Chow, R. M., Tiedens, L. Z., and Govan, C. L. (2008). Excluded emotions: The
role of anger in antisocial responses to ostracism. Journal of Experimental
Christopherson, K. M. (2007). The positive and negative implications of
anonymity in internet social interactions: “On the internet, nobody knows
you’re a dog.” Computers in Human Behavior, 23, 3038–3056.
Coffey, B., and Woolworth, S. (2004). “Destroy the scum, and then neuter their
families”: The web forum as a vehicle for community discourse? The Social
Science Journal, 41, 1–14.
Diener, E. (1979). Deindividuation, self-awareness, and disinhibition. Journal of
Diener, E., Fraser, S. C., Beaman, A. L., and Kelem, R. T. (1976). Effects
of deindividuation variables on stealing among Halloween trick-or-treaters.
making groups. Human–Computer Interaction, 6, 119–146.
Eisenberger, N. I., Lieberman, M. D., and Williams, K. D. (2003). Does rejection
hurt? An fMRI study of social exclusion. Science, 302, 290–292.
Friedland, N., Keinan, G., and Regev, Y. (1992). Controlling the uncontrol-
lable: Effects of stress on illusory perceptions of controllability. Journal of
Gardner, W., Pickett, C. L., and Brewer, M. B. (2000). Social exclusion and
selective memory: How the need to belong influences memory for social
events. Personality and Social Psychology Bulletin, 26, 486–496.
Gonsalkorale, K., and Williams, K. D. (2007). The KKK won’t let me play:
Ostracism even by a despised outgroup hurts. European Journal of Social
Psychology, 37, 1176–1186.
Greenberg, J., Pyszczynski, T., and Solomon, S. (1986). The causes and con-
sequences of the need for self-esteem: A terror management theory. In R.
F. Baumeister (Ed.), Public self and private self (pp. 189–212). New York:
Springer-Verlag.
Heider, F., and Simmel, M. (1944). An experimental study of apparent behavior.
American Journal of Psychology, 57, 243–259.
Kassner, M. P., Wirth, J. H., Law, A. T., and Williams, K. D. (2010). Effects of
mental visualization and social cues on detection and influence of ostracism.
Manuscript in preparation.
Kiesler, S., Siegel, J., and McGuire, T. W. (1984). Social psychological aspects of
computer-mediated communication. American Psychologist, 39, 1123–1134.
Kowalski, R. M., Limber, S. P., and Agatston, P. W. (2008). Cyber bullying:
Bullying in the digital age. Malden, MA: Blackwell Publishing.
Kraut, R., Patterson, M., Lundmark, V., Kiesler, S., Mukopadhyay, T., and
Scherlis, W. (1998). Internet paradox: A social technology that reduces social
involvement and psychological well-being? American Psychologist, 53, 1017–
1031.
Lakin, J. L., and Chartrand, T. L. (2005). Exclusion and nonconscious behavioral
mimicry. In K. D. Williams, J. P. Forgas, and W. von Hippel (Eds.), The social
outcast: Ostracism, social exclusion, rejection, and bullying (pp. 279–295). New
York: Psychology Press.
Leary, M. R., Kowalski, R. M., Smith, L., and Phillips, S. (2003). Teasing, rejec-
tion, and violence: Case studies of the school shootings. Aggressive Behavior,
29, 202–214.
Leary, M. R., Tambor, E. S., Terdal, S. K., and Downs, D. L. (1995). Self-
esteem as an interpersonal monitor: The sociometer hypothesis. Journal of
Leary, M. R., Twegne, J. M., and Quinlivan, E. (2006). Interpersonal rejection
as a determinant of anger and aggression. Personality and Social Psychology
Review, 10, 111–132.
Lee, E., and Leets, L. (2002). Persuasive storytelling by hate groups online:
Examining its effects on adolescents. American Behavioral Scientist, 45, 927–
957.
Maner, J. K., DeWall, C. N., Baumeister, R. F., and Schaller, M. (2007).
Does social exclusion motivate interpersonal reconnection? Resolving the
“porcupine problem.” Journal of Personality and Social Psychology, 92, 42–
55.
McGuire, T. W., Kiesler, S., and Siegel, J. (1987). Group and computer-mediated
discussion effects in risk decision-making. Journal of Personality and Social
McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the
internet: Identity “demarginalization” through virtual group participation.
(1999). Causes and consequences of social interaction on the internet: A con-
ceptual framework. Media Psychology, 1, 249–269.
(2000). Plan 9 from cyberspace: The implications of the internet for person-
ality and social psychology. Personality and Social Psychology Review, 4, 57–
75.
McKenna, K. Y. A., Green, A. S., and Gleason, M. E. J. (2002). Relationship
formation on the internet: What’s the big attraction? Journal of Social Issues,
58, 9–31.
Molden, D. C., Lucas, G. M., Gardner, W. L., Dean, K., and Knowles, M.
L. (2009). Motivations for prevention or promotion following social exclu-
sion: Being rejected versus being ignored. Journal of Personality and Social
Ouwerkerk, J. W., Kerr, N. L., Gallucci, M., and Van Lange, P. A. M. (2005).
Avoiding the social death penalty: Ostracism and cooperation in social dilem-
mas. In K. D. Williams, J. P. Forgas, and W. von Hippel (Eds.), The social
outcast: Ostracism, social exclusion, rejection, and bullying (pp. 321–332). New
York: Psychology Press.
Peter, J., and Valkenburg, P. M. (2006). Research note: Individual differences in
perceptions of internet communication. European Journal of Communication,
21, 213–226.
Peter, J., Valkenburg, P. M., and Schouten, A. P. (2005). Developing a model of
adolescent friendship formation on the internet. CyberPsychology & Behavior,
8, 423–430.
Pickett, C. L., Gardner, W. L., and Knowles, M. (2004). Getting a cue: The
need to belong and enhanced sensitivity to social cues. Personality and Social
Predmore, S. J., and Williams, K. D. (1983). The effects of social ostracism
on affiliation. Paper presented at the annual meeting of the Midwestern
Psychological Association, Chicago, May.
Reid, D. J., and Reid, F. J. M. (2007). Text or talk? Social anxiety, loneliness,
and divergent preferences for cell phone use. CyberPsychology & Behavior,
10, 424–435.
Rintel, E. S., and Pittam, J. (1997). Strangers in a strange land: Interaction
management on Internet Relay Chat. Human Communication Research, 23,
507–534.
Sheeks, M. S., and Birchmeier, Z. P. (2007). Shyness, sociability, and the use of
computer-mediated communication in relationship development. CyberPsy-
chology & Behavior, 10, 64–70.
Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. W. (1986). Group pro-
cesses in computer-mediated communication. Organizational Behavior and
Human Decision Processes, 37, 157–187.
Skitka, L. J., and Sargis, E. G. (2006). The internet as psychological laboratory.
Annual Review of Psychology, 57, 529–555.
Smith, A., and Williams, K. D. (2004). R U there? Effects of ostracism by cell
phone messages. Group Dynamics: Theory, Research, and Practice, 8, 291–301.
Snoek, J. D. (1962). Some effects of rejection upon attraction to a group. Journal
of Abnormal and Social Psychology, 64, 175–182.
Solomon, S., Greenberg, J., and Pyszczynski, T. (1991). A terror management
theory of self-esteem and its role in social behavior. In M. Zanna (Ed.),
Advances in Experimental Social Psychology (pp. 93–159). New York: Aca-
demic Press.
in organizational communication. Management Science, 32, 1492–1512.
Stritzke, W. G. K., Nguyen, A., and Durkin, K. (2004). Shyness and computer-
mediated communication: A self-presentational theory perspective. Media
Tesser, A. (1988). Toward a self-evaluation maintenance model of social behav-
ior. In L. Berkowitz (Ed.), Advances in Experimental Social Psychology,
Vol. 21 (pp. 181–227). San Diego, CA: Academic Press.
Thompson, L., and Nadler, J. (2002). Negotiating via information technology:
Theory and application. Journal of Social Issues, 58, 109–124.
Twenge, J. M., Baumeister, R. F., Tice, D. M., and Stucke, T. S. (2001). If you
can’t join them, beat them: Effects of social exclusion on aggressive behavior.
Twenge, J. M., and Campbell, W. K. (2003). “Isn’t it fun to get the respect
that we’re going to deserve?” Narcissism, social rejection, and aggression.
Personality and Social Psychology Bulletin, 29, 261–272.
van Beest, I., and Williams, K. D. (2006). When inclusion costs and ostracism
pays, ostracism still hurts. Journal of Personality and Social Psychology, 91,
918–928.
Warburton, W. A., Williams, K. D., and Cairns, D. R. (2006). When ostracism

leads to aggression: The moderating effects of control deprivation. Journal
of Experimental Social Psychology, 42, 213–220.
Wesselmann, E. D., Bagg, D., and Williams, K. D. (2009). “I feel your pain”: The
effects of observing ostracism on the ostracism detection system. Journal of
Wesselmann, E. D., Butler, F. A., Williams, K. D., and Pickett, C. L. (2010).
Adding injury to insult: Unexpected rejection leads to more aggressive
responses. Aggressive Behavior, 36, 232–237.
Wesselmann, E. D., and Williams, K. D. (in press). Ostracism and stages of
coping. In C. N. DeWall (Ed.), The Oxford handbook of social exclusion. New
York: Oxford University Press.
Wheaton, A. E. (2001). Ostracism and susceptibility to the overtures of socially
deviant groups and individuals. Unpublished Honors thesis, Macquarie Uni-
versity, Sydney.
Williams, K. D. (1997). Social ostracism. In R. M. Kowalski (Ed.), Aversive
interpersonal behaviors (pp. 133–170). New York: Plenum Press.
(2001). Ostracism: The power of silence. New York: The Guilford Press.
(2007). Ostracism. Annual Review of Psychology, 58, 425–452.
(2009). Ostracism: Effects of being excluded and ignored. In M. P. Zanna
(Ed.), Advances in experimental social psychology, Vol. 41 (pp. 275–314). New
York: Academic Press.
Williams, K. D., Cheung, C. K. T., and Choi, W. (2000). CyberOstracism:
Effects of being ignored over the internet. Journal of Personality and Social
Williams, K. D., Govan, C. L., Crocker, V., Tynan, D., Cruickshank, M., and
Lam, A. (2002). Investigations into differences between social- and cyber-
ostracism. Group Dynamics: Theory, Research, and Practice, 6, 65–77.
Williams, K. D., Selvarajah, K., and Richards, D. (in preparation). Ostracism in
a virtual environment: Cocktail Party World. Manuscript in progress.
Williams, K. D., and Sommer, K. L. (1997). Social ostracism by coworkers:
Does rejection lead to social loafing or compensation? Personality and Social
Williams, K. D., and Wesselmann, E. D. (2011). The link between ostracism and
aggression. In J. P. Forgas, A. W. Kruglanski, and K. D. Williams (Eds.),
The psychology of social conflict and aggression. New York: Psychology Press,
in press.
Williams, K. D., and Zadro, L. (2004). Ostracism: Empirical studies inspired
by real-world experiences of silence and exclusion. Paper presented at the
annual meeting of the Society for Personality and Social Psychology, Austin,
TX, January.
Wirth, J. H., Lynam, D. R., and Williams, K. D. (2010). When social pain is not
automatic: Personality disorder traits buffer ostracism’s immediate negative
impact. Journal of Research in Personality, 44, 397–401.
Wirth, J. H., Sacco, D. F., Hugenberg, K., and Williams, K. D. (2010). Eye gaze
as relational evaluation: Averted eye gaze leads to feelings of ostracism and
relational devaluation. Personality and Social Psychology Bulletin, 36, 869–882.
Zadro, L., Boland, C., and Richardson, R. (2006). How long does it last? The
persistence of the effects of ostracism in the socially anxious. Journal of
Zadro, L., Williams, K. D., and Richardson, R. (2004). How low can you go?
Ostracism by a computer is sufficient to lower self-reported levels of belong-
ing, control, self-esteem, and meaningful existence. Journal of Experimental
8 Opinion-based groups
(Racist) talk and (collective) action on the internet

Craig McGarty, Girish Lala, and Karen M. Douglas
In this chapter we take a broad view of the social psychology of intergroup

relations and apply it to the internet. We start by addressing two popular
propositions about the internet. The first of these is that the internet
is a safe haven for racists who, using the cover of anonymity, engage in
hostile flaming of members of other communities. We can call this first
proposition the anonymous cyberhate view of the internet. The second
popular proposition is the idea that the internet is a hotbed of social
activism where people develop plans to change the world using a device
that spans continental boundaries. We call this second proposition the
global activist view of the internet.
Let us be clear that there is some truth to both propositions and it is
easy to find evidence for both of them. Our contention, however, is that
both of the propositions are gross oversimplifications that tend to lead
commentators and observers to misunderstand the dynamics present in
this medium.
In this chapter we develop a response to both these propositions. We
address the first by considering the work of Karen Douglas and her
colleagues (Douglas and McGarty, 2001, 2002; Douglas et al., 2005).
Our argument here is that to understand the dynamics we need a
more sophisticated conceptualization of anonymity, and to achieve this
goal we build on the social identity model of deindividuation effects
(drawing on the work of Reicher et al., 1995). We then turn to the
global activist proposition using data recently collected by Girish Lala
(2008). In this work we explore the proposition that for social move-
ments the internet provides an obvious home for talk rather than
action. Before we turn to these propositions and our response it is
helpful for us to spell out some of the conceptual tools we will apply
to this domain. The first point we need to clarify is the ways in
which the internet provides a platform for intergroup and intragroup
relations.
145
146 Craig McGarty, Girish Lala, and Karen M. Douglas
Broadcast and gathering

Group life over the course of human civilization has been underpinned
by two broad modes of communication that we can term broadcast and
gathering. The earliest (pre-technological) form of broadcast is the speech
or public address. Such addresses (in particular to massed troops or
followers of religious leaders) are recorded in the oldest writings and pre-
literary traditions such as the Iliad, the Mahabharata, and the Bible. The
method is widely used for conveying information, for issuing instructions,
and for elaborating a shared group position (as a politician might do in
a campaign launch). With the rise of the printing press the range of
people able to use the broadcast mode increased rapidly, and the use of
political and religious pamphlets heralded an era of profound (sometimes
traumatic) change. The broadcast mode continued to develop through
radio and TV broadcasts to extend to static web page content (e.g., most
blogs).
The gathering or assembly mode has existed alongside the broadcast
mode for the same period (again there are detailed descriptions of assem-
blies of troops in the Iliad and other sources). Over time these gatherings
developed into parliaments, conventions, town hall meetings, and the
like. Such gatherings allow the possibility for a diversity of opinion and
disagreement, and for attempts to build consensus. Technology has con-
tributed to this form more slowly, so that in the twentieth century talk-
back radio, interactive TV, and teleconferencing started to emerge, along
with discussion boards and Usenet newsgroups, and in the twenty-first
century we have a suite of devices, sometimes called Internet 2.0, that
provides interactivity with website content (e.g., through the comments
page of blogs or through social networking software and the like).
Human social life rests upon both these broad modes and it is difficult
to imagine any complex human activity being successful without adopt-
ing both modes. They are particularly important for our chapter because
both modes are critically involved in group processes. This seems obvi-
ously true in relation to the discussion of the gathering mode, which is the
very stuff of group processes. We argue further, however, that the broad-
cast mode is central to the process of forming a specific type of group
that underpins our analysis of online (and other) actions. We consider
these opinion-based groups in the next section.
Social identity and opinion-based groups

The approach we take is firmly grounded in the social identity tra-
dition established by Tajfel and Turner (1979) and carried on in the
Opinion-based groups 147
self-categorization theory of Turner et al. (1987), and in the social iden-

tity model of deindividuation effects (Reicher et al., 1995; see also Spears
et al., this volume). The attraction in applying self-categorization theory
to online behavior including behavior on the internet is that self-
categorization theory seeks to account for both personal, private phe-
nomena and collective, public aspects of human life in terms of the same
principle. The internet would seem to provide an exquisite blend of
all these elements and nothing instantiates this better than the profu-
sion of web-logs or blogs. Everyday millions of people from around the
world publish what would otherwise be private thoughts and invite com-
mentary and exchange of opinion about their musings and about other
aspects of the world. Blogs are often intensely thematic; indeed, many
blogs can be seen as part of political projects or campaigns to promote
different ways for people to lead their lives and for society to organize its
affairs.
Self-categorization theory’s core principle is the idea of variable self-
perception. This surprisingly simple idea is that the way we think and act
in any setting varies powerfully with the current self-perception. Although
other variations are envisaged by the theory, the key variation involves
switching from seeing oneself as a unique individual who is different
from other individuals (and has a different personal identity) to seeing
oneself as a member of a group, and thus as relatively similar to other
members of the same group with whom we share a social identity (differ-
ent from other social identities). The phenomena of interpersonal liking,
perception, preference, attitudes, and so on are therefore believed to be
most applicable where self-perception is located at the personal identity
end of the spectrum, whereas the phenomena of collective action, inter-
group conflict and cooperation, stereotyping, and the like are associated
with the social identity end where perceivers come to see themselves as
interchangeable with other ingroup members.
According to this theory, context is all important. To the degree to
which context renders a particular social categorization to be salient
(switched on, activated, or psychologically prepotent), perceivers are
more likely to see themselves and others in these terms. Salience is
hypothesized to be a function of the fit between the stimuli and the cate-
gory specifications (including the social behavioral meaning of those cate-
gories) and the perceiver’s readiness to use that categorization (Turner
et al., 1994). For example, text advocating racial segregation fits the
specification of the category white supremacist and so the social cate-
gory “racist” may become salient, but we may be especially ready to use
such a social categorization when we are viewing the contents of a KKK
website.
We mentioned above that part of the attraction of the social identity

approach for us is its applicability to the domain of collective action. In
this regard the work of Klandermans et al. (2002) is instructive:
The basic hypothesis regarding collective identity and protest is fairly straightfor-
ward: A strong identification with a group makes participation in political protest
on behalf of that group more likely. The available empirical evidence overwhelmingly
supports this assumption (de Weerd and Klandermans, 1999; Kelly and Breinlinger,
1995; Simon et al., 1998; Stürmer, 2000). [emphasis added]
(Klandermans et al., 2002, p. 239)
Contrary to Klandermans et al., the available evidence for the proposi-

tion is actually pretty weak in the case of social categories. For example,
de Weerd and Klandermans (1999), Kelly and Breinlinger (1995), and
Simon et al. (1998) found in a range of contexts that identification with
social category memberships was weakly correlated with intentions to
take action.
These findings are important in relation to our argument as they sug-
gest that in order to better understand collective action it is necessary
to explore identification with more specific categories which are relevant
in the context of action rather than broader social categories. Activist
identities (as argued by Kelly and Breinlinger, 1995; Simon et al., 1998)
seem to get closer to the core of the issue of collective action by offering
a more direct route to predicting intentions to take or become involved
in collective action, but they contain other problems. Activists are rare;
mass movements (by definition) involve large numbers of people and
most of these people are not activists. That is, the participants in mass
action are not so much activists as supporters of that cause.
Given that mass action is not taken only by activists, do we need to
abandon the role of identification or commitment in predicting mass
action? Opinion-based groups are psychological groups in the sense used
by Turner (1982), but which have a social identity defined by a shared
opinion (see Bliuc et al., 2007). Merely holding the same opinion as
others is not sufficient for such a group to be said to exist; rather, the
shared opinion needs to become part of that social identity. In this way,
people can come to perceive and define themselves in terms of their
opinion group membership in the same way as with any other psycho-
logically meaningful social category or group. Where the opinion-based
group membership becomes switched on, group members should behave
in line with the norms of that group. For example, people who oppose
abortion or support a woman’s right to choose may come together to form
pro-life or pro-choice groups (which are not identical to their religious
or political affiliations).
A key interest for this chapter is the way in which internet technologies
have expanded the boundaries of interaction for many of these groups
that form around a shared opinion. Increasingly, many of these groups
are interacting online, and are heavily instantiated as online communities
(e.g., see McCaughey and Ayers, 2003; van de Donk et al., 2004). In
other words, it is apparent that, given the capacity for communication
and interaction that the internet presents, opinion-based groups and the
internet are very well suited to each other.
Opinion-based groups are particularly relevant for understanding col-
lective action for a number of reasons. First, they often form around
controversial issues, and oppositionally defined opinion-based groups
tend to champion opposing perspectives on key issues. In other words,
there are opinion-based groups that aim to change the social world in
some way and also groups that aim to preserve the status quo.
Opinion-based groups can be readily distinguished from social cate-
gories and action groups. Opinion-based groups can often be formed
within a broader social category (or can be broader than a social cate-
gory). One important idea in relation to opinion-based groups is that they
are often formed about the relations between social categories or groups
in general. This is the case with feminist and anti-feminist opinion-
based groups, which are groups formed around ideologies about relations
between social categories based on gender. In this way opinion-based
groups can also help to restructure problematic intercategory bound-
aries. It is easier for women and men to work together to promote gender
equality, or for people of different races to work to promote tolerance and
acceptance, if members of both categories share a relevant opinion-based
group membership.
Opinion-based groups can also be distinguished from action groups.
Most action groups are based on shared opinions, and opinion-based
groups are an incipient stage of action groups. Specifically, activist groups
can be considered as emerging from opinion-based groups. The “single-
issue pressure groups” studied by Kelly and Breinlinger (1995) and the
activist groups such as Gray Panthers, the gay movement, AIDS vol-
unteer service organization, and the fat acceptance movement, investi-
gated by Simon et al. (1998, 2000) and Stürmer et al. (2003), can be
regarded as directly emerging from opinion-based groups. All members
of an opinion-based group may have the potential to take spontaneous
collective action in specific circumstances, but only a small fraction of
the membership may be involved in organizing it.
Research that we have undertaken in Australia and Romania (e.g., by
Bliuc et al., 2007) shows that identification with a political opinion-based
group is an excellent predictor of intention to take politically relevant
behavior in research conducted in Australia and Romania. Our strongest

claim is that to talk of collective action is to talk of opinion-based groups
in action. In other words, collective action is the material crystallization or
expression of the existence of salient opinion-based group memberships.
Work by Musgrove and McGarty (2008) also suggests that opinion-based
group memberships about the War on Terror are excellent predictors
of contrasting group-based emotions and actions (cf. McGarty et al.,
2005, who found weak evidence for such relationships with national
identification; see also McGarty and Bliuc, 2004).
Collective action is strongly linked to opinion-based group identity for
a number of reasons. First, opinion-based groups are formed primarily
to convert broad ideologies or affinities into collective action. Unified
social action is aimed either at creating social change or at preserving
the status quo. Building on social identity (Tajfel and Turner, 1986) and
Klandermans’ (2000) ideas, such action is taken by people who share
an opinion about achieving or rejecting social change. Depending on the
opinion they hold, they can be considered to be members of one opinion-
based group or another (e.g., pro- or anti-change of the status quo). The
consequence of that is that members of such groups may be particularly
prone to take action.
As intimated earlier, activist groups can be seen as organized subcat-
egories of larger opinion-based groups or as factions organized to act in
terms of a certain subjective opinion-based group membership. Thus,
in terms of relevance to action, broad social categories should be the
least relevant, with opinion-based groups representing the missing link
between social categories and activist groups. In order to understand the
mechanisms and links between collective identities and action it is essen-
tial to look at the underlying nature of collective action and of the groups
participating in collective action. Industrial action provides a good exam-
ple of the levels of the process. The action tends to be organized by union
officials and organizers (the activist group) on behalf of the broad social
category (workers in the industry or workplace), but the action is taken by
that subset of workers (usually union members) who define themselves
as supporters of the action (see Taylor and McGarty, 2001).
Thus, activist identities can be considered to be opinion-based group
identities that are aligned with organizations constituted on the basis of
opinion-based group membership, with rules and a structure to promote
a certain stance or view of the world. In other words, activist identities
are opinion-based group identities which have achieved a more objective
consistency and which take responsibility for more organized forms of
collective action. In the case of spontaneous collective action, opinion-
based groups are identifiable in a more raw, less structured form. People
participating in spontaneous action need not be attached to specific

activist groups but share a common understanding and stance on a cer-
tain issue that ultimately defines an opinion-based group membership.
The strongest form of our claim is that collective action is essentially
the material expression of opinion-based group membership (McGarty
et al., 2009). We will return to this idea below, but before we do it is useful
to return to a focus on the contest of ideas that takes place on the internet.
The internet is the ideal home for opinion-based groups because the key
forms of communication lend themselves to the formation of argument
and counterargument. Moreover, many other possible social identities,
such as ethnicity, gender, and age, are easy to assert but are difficult to
detect on the internet. Although opinion-based group memberships can
be simulated almost as easily as any other group membership (viz. the rise
of concern trolls – http://en.wikipedia.org/wiki/internet troll#Concern
troll downloaded 27/11/07 – who are people who feign commitment to a
cause), the existence of shared opinions is nevertheless easy to detect (at
least in other active web users).
With this conceptual overview in place we will now address the first
proposition that we argue has widespread popular currency. We have
termed this the anonymous cyberhate view of the internet, or the idea
that the internet provides a haven for racists who, using the cover
of anonymity, engage in inflammatory and aggressive attacks on other
groups and communities.
The anonymous cyberhate proposition

There is a wealth of social psychological literature examining the conse-
quences of anonymity for talk and action. Indeed, many studies of online
communication have focused on the proposition that online behavior is
different than other modes of communication because people can con-
ceal their identity. This idea has been explored extensively in a variety of
settings since the very beginning of online communication, such as in task
performance at work (e.g., Siegel et al., 1986; Sproull and Kiesler, 1986),
the development of relationships (e.g., Lea and Spears, 1995; van Gelder,
1985), and the high levels of hostile, flaming communication that can be
exhibited in electronic communication (Douglas and McGarty, 2001,
2002; Kiesler et al., 1984; Lea et al., 1992), and it remains an important
issue for researchers today. Closely tied to the notion of deindividuation
(Diener, 1980; Zimbardo, 1969), it has been argued that anonymous
talk over the internet is often “disinhibited” because people are free from
constraints that would bring their behavior more in line with personal,
group, and societal norms. It is therefore unsurprising that the internet
is perceived to be the ideal medium for extreme opinion-based groups to

express their views of hate and hostility (e.g., Back, 2002) and for indi-
viduals to express themselves in ways that they would never consider in
face-to-face interactions. People can broadcast their opinions with very
little restriction and promote themselves and their actions without the
accountability that face-to-face talk or a public campaign would involve.
The social identity model of deindividuation effects (SIDE; Reicher
et al., 1995; see also Spears and Lea, 1994), however, takes a different
perspective on the effects of anonymity, especially when it comes to elec-
tronic communication involving groups. SIDE proposes that anonymity
will not always lead to negative talk and action. Instead, it is argued that
anonymity can facilitate the enactment of social identity, which will often
have positive consequences. For example, Spears and Lea (1994) pro-
posed that being personally anonymous to a powerful audience may be
liberating; anonymous communicators need not feel pressured to con-
form to the norms and expectations of powerful others (see also Reicher
and Levine, 1994a, 1994b), but may enact aspects of their own iden-
tity that the group would normally deem unacceptable, and potentially
punish. According to this perspective, there is therefore an important
purpose to anonymous behavior, especially for group members, than
other theories of deindividuation might suggest.
Following the SIDE model, it is therefore possible to make predictions
about the conditions under which talk on the internet may help opinion-
based groups achieve their objectives. Remaining anonymous from the
outgroup may be one option. Also, however, identifiability may facilitate
the enactment of identity when the intended audience is comprised of
members of the ingroup. In this case, the support provided by the pres-
ence of like-minded others can lead to heightened expression of ingroup
normative views (e.g., Reicher et al., 1998). Some of our own research
supports this idea.
To examine the impact of identifiability on hostile flaming behavior,
Douglas and McGarty (2001, 2002) examined the online comments that
people made about a white-power group member who had posted their
views on the internet. Research participants were asked to respond to the
comments of the extreme group member either anonymously (supplying
no personal details) or identifiably (by name and location – a condition
we called internet identifiability). Crucially, the audience to which
participants were making their comments was an ingroup audience of
like-minded (anti-racist) undergraduate students. Results revealed that
participants were more likely to stereotype the target by using abstract
adjectival descriptions (e.g., “racist,” “bigoted”) under conditions
of internet identifiability (not anonymity) to this ingroup audience.
Following the SIDE logic, the expression of stereotypes about the target
was facilitated by the supportive “presence,” however virtual, of their
opinion-based group. Therefore, it is also possible that extremists who
perceive their audience to be made up of like-minded individuals would
most effectively express their views having chosen to do so identifiably.
We investigated some other strategies that online extremist groups
use to promote their cause and recruit new members. Douglas et al.
(2005) examined the strategies that forty-three online white-supremacist
groups used on their websites. Drawing on social identity theory (SIT;
Tajfel and Turner, 1986), this research started from the position that
white-power groups feel superior to other groups but also feel in an
insecure position due to the threat posed by other groups (Green et al.,
1999). As the self-defined high-status group, white-supremacists have a
number of strategies available to them that enable them to effectively
sell their products, disseminate their opinions, and recruit new white-
supremacists to continue their “work.” Douglas et al. (2005) drew on
Haslam’s (2001) interpretation of social identity theory, arguing that
groups can adopt different self-enhancement strategies depending on
their status, the extent to which people are able to move freely between
groups, and the security of relations among groups. For white-power
groups, the threat from outgroups is the primary motivation for their
existence. In this context, Haslam’s reading of SIT would predict overt
racism, and the advocacy of hostility and conflict.
However, surprisingly, Douglas et al. (2005) did not find strong evi-
dence of advocated conflict, and especially little advocacy of violence,
on white-supremacist websites. Instead, these groups used more socially
creative strategies that redefine the elements of the intergroup context
without being openly hostile. These included arguments in favor of
white-superiority, and supremacist arguments that attempt to justify the
position of whites over non-whites. Of course, it is not surprising that
white-supremacist websites advocated less violence than other forms of
conflict and social creativity. After all, the sites can be shut down if they
violate the law. However, the fact that very little violence was advocated
overall, in what we call “hate” sites, was surprising. It seems that legal,
economic, and social grounds for conflict, and socially creative argu-
ments, are white-supremacists’ main weapons of persuasion (see also
McDonald, 1999).
It is important to consider how useful social creativity strategies can
be for online opinion-based groups such as these. First, social creativ-
ity may serve the function of helping to create conflict by making the
ingroup (usually whites) feel that they have a reason to feel disadvan-
taged. Second, without seeming overtly hostile or violent, socially creative
strategies may be a way to convert potentially sympathetic individuals to

the opinions of white-supremacist groups by overcoming their reluctance
to become involved. An excellent example of this strategy in action is in
the many Holocaust denial sites on the internet. These sites do not, in
general, advocate violence against Jews. Instead, they challenge the accu-
racy of historical claims about the Holocaust and suggest either that Jews
invented these claims or (even more ridiculously) that they committed
the atrocities themselves. The power of this tactic is that it serves to
redefine Jews from being worthy of sympathy to being liars or criminals.
The global activist proposition

We now turn our attention to investigating the second popular proposi-
tion about intergroup relations on the internet – the view that the medium
acts as a hotbed of social action and protest. As we have already noted,
we argue that, although there is some truth to this proposal, an uncritical
acceptance of the global activist proposition leads to an oversimplified
understanding of internet groups. We suggest that this crude view masks
the ways in which participants in online activist groups actually interact,
and does an injustice to group participants in that it underestimates the
level of insight participants themselves have for the role of their online
communities (here our work touches on ideas sympathetic to the strategic
and motivated user, expected and emergent effects framework – SMEE;
Joinson, 2003). Our analysis takes two parts. We will first review key the-
oretical ideas relevant to our general contentions about activist groups.
Next, we will discuss some of our empirical findings that shed light on
our more specific arguments about the nature and role of online activism.
Our interest is in understanding group-based internet activism, and to
do so we must first focus on the study of social movements. Although our
overall conceptualization of social movements draws on key ideas from
a range of disciplines, two assertions are particularly important for our
account of collective behavior on the internet. We contend that social
movements (groups of people which form to engage in collective action
around an issue or cause) should be regarded as legitimate psychologi-
cal groups, and that online communities (OLCs) associated with social
movement groups are genuine instantiations of social movements.
Consistent with the idea that social movements are real groups, Tajfel
(1981) argued that such movements were large numbers of people that
both defined themselves as members of a common group and were
defined by others as a group. Similarly, Snow et al. (2004) affirmed
the group-based nature of social movements, and Klandermans (1997)
noted the importance of the group for social movement research.
Lending weight to our second assertion, Foot and Schneider (2002)

observed that both individuals and groups used the internet to mobilize
action, and Jacobs (2005) has highlighted how online social movement
groups aimed to use web networks to interact with their participants, and
to influence their movements’ actions. Likewise, observations of OLCs
formed around social issues or causes demonstrate that these online
groups exhibit characteristics of face-to-face social movements. For
example, the Global Call to Action Against Poverty (GCAP) explicitly
identifies itself as a social movement on its website, and has a strong
focus on behavior and action, encouraging participants to “speak out”
and “get involved” (GCAP, 2005).
However, notwithstanding the centrality of the group for these col-
lectives, much of the effort in research on these social movements has
focused on understanding what drives individuals toward particular social
movements. A number of socio-economic analyses have noted that “ratio-
nal actor” models provide one popular explanation for individual partic-
ipation in protest groups or movements. For example, Cohen (1985)
details how utilitarian adherents of resource mobilization theory began
to stress the objective, strategic, and instrumental factors of social move-
ment groups (and see also Jenkins, 1983; McAdam et al., 1988).
Around the same time, others were considering similar themes from a
more explicitly social psychological perspective. For example, Klander-
mans (1984) argued that individuals weighed collective motives (over-
all collective benefits), selective motives (potential reactions of signifi-
cant others), and reward motives (individual benefits) when assessing
social movement participation. Klandermans also proposed two compo-
nents of movement mobilization – consensus mobilization (movements
attempting to gain support for their position) and action mobilization
(movements mobilizing people to act), suggesting that action mobiliza-
tion could not occur without consensus mobilization. His and others’
work foreshadowed a key contribution that social psychology added to
the theoretical discussion – recognizing the role of identity in social move-
ments. More specifically, this work began to explore how identity-related
motivations toward social action might interact or coexist with tradition-
ally recognized utilitarian motives.
More recently, researchers have begun to refine these ideas by propos-
ing models of collective action that posit multiple routes to engagement
and participation, or multiple pathways to collective action (Simon et al.,
1998; van Zomeren et al., 2004). These pathway explanations often pro-
pose two independent focuses – instrumental (where motives are primar-
ily related to cost/benefit analyses) and identity (where self-definition is
the central motive).
For example, Simon et al. (1998) proposed a dual-path model that

had identity as central to one pathway, but focused on more explicit
cost–benefit calculations as a second (instrumental) route to collective
action. Although acknowledging their instrumental path includes norma-
tive considerations (and thus a collective component), the key difference
between instrumental and identity pathways as conceptualized by Simon
et al. is that collective action via the instrumental path occurs because of
explicit judgments about costs or benefits, whereas participation through
the identity path results solely from a salient shared social identity that
speaks to self-definition.
van Zomeren et al. (2004) offered an alternative model when they
examined the role that emotion plays in promoting collective action.
Building on work by Tajfel and Turner (1979) and Lazarus (1991), van
Zomeren et al. argued that perceptions of unfair collective disadvantage
associated with a salient social identity can lead to group-based anger, and
that validation by collective opinion increases this anger, presenting an
identity-related, emotion-focused path to collective action. In contrast,
they argued that a problem-focused (instrumental) path resulted from
perceptions that other group members are willing to engage in collective
action, increasing a sense of group efficacy, and so leading to actual
collective action. In common with Simon et al. (1998), van Zomeren and
colleagues recognized the role of shared identity (i.e., that perceptions
of action support and efficacy are group-based), and acknowledged the
complementary and interacting nature of the instrumental and identity
pathways they described.
One issue that has sometimes arisen in models similar to those pro-
posed by Simon et al. (1998) and van Zomeren et al. (2004) is the place
of automatic versus controlled behavior. Although Simon and colleagues
specifically acknowledge this point, they draw no conclusions about auto-
maticity in their conceptualization of identity and instrumental paths.
While they do recognize the value of further research in this area, con-
trolled processes are employed in both instrumental and identity paths
in the specific research they describe – in other words, their participants
were attentive to their decisions when determining actions according to
cost/benefit and identity considerations.
Despite that research, at one level the connection between instrumen-
tal and identity paths and controlled and automatic behavior can appear
straightforward. For example, Simon et al. use the term “calculation” to
describe the processes that occur along the instrumental path, and note
that calculation “should be of less importance than identification or self-
definition” (p. 656) in the identity pathway. One interpretation, therefore,
is that action driven by identity considerations is always automatic, or
less thoughtful or attentive. In common with Simon and colleagues, our

research has not sought to investigate this issue and so we cannot draw
any specific conclusions on this point. Notwithstanding this, however, we
make the more general theoretical argument that there is no imperative
for behavior based on identity to be more or less “controlled” than that
based on instrumental motives. One motivation for the development of
dual-pathway models has been to redress the individual-oriented con-
ceptualization of self that has predominated in the literature on collective
action. Viewing cost/benefit calculations as necessarily more controlled
and explicit than (social) identity-based motivations misunderstands the
nature of identity in general, and reduces the relevance of social com-
pared with individual identity. Under the social identity approach, nei-
ther automaticity nor control should have primacy in social or personal
identities – in a similar way to the relationship between instrumental
and identity paths, so too is behavior the likely result of complementary
and interacting controlled and automatic processes (e.g., see Bargh and
Chartrand, 1999).
There are many potential instances of online social movements that
lend themselves to interpretation in rational actor or identity-based terms.
The Open Source (OSI, 1998) and Free Software (FSF, 1998) move-
ments provide two particularly good examples – one as an exemplar
of the rational actor approach, the other more in line with identity-
based principles. Broadly speaking, both movements advocate the use
of “open,” non-proprietary software; however, they approach their goal
from different philosophical positions. The Open Source movement
highlights the practical benefits of non-proprietary software and thus
falls squarely into the rational actor camp; in contrast, those who align
with the Free Software movement emphasize ethical and social rea-
sons for their position along the lines of an identity-based pathway
(Stallman, 2007).
Analyses drawn from other contexts seem applicable to internet move-
ments, but there are also complexities. In particular, as the role of iden-
tity comes more into focus, the complexity of the relationship between
identity, identification, and participation has also become increasingly
apparent.
Related to this complexity, we can identify at least two characteristics
that should be consistently important in determining the content of par-
ticipants’ social identity across a broad range of social movements. Social
movements exist to bring about change, and so action must be central
for these groups (Klandermans, 1997; Snow et al., 2004; Tajfel, 1981).
Consequently, then, themes of change and action should be integral to
the identity of a social movement, and one of the central features of any
social identity adopted by social movement participants should be an

action orientation.
Moreover, social movements are highly heterogeneous – social move-
ments are commonly “organizations of organizations” (Klandermans,
1997, p. 118; see also Gerlach and Hine, 1970; Oberschall, 1980; Simon
and Klandermans, 2001). But for social movements this prominent het-
erogeneity might lead to a high, and potentially problematic, level of
internal diversity. Where a collective is made up of distinct sub-groups
and organized around the pursuit of some action or goal, it follows that
dispute about how that goal should be attained will be more likely than
in groups where sub-group identity is less distinct. Tensions surrounding
group goals make it more difficult for group unity to develop and endure
because, by definition, debate and argument about what a group should
be and do implies the existence of conflicting group definitions (see, for
example, Sani and Reicher, 1998). In turn, volatile group definitions
make it more difficult for a clear social identity to become salient for
individual participants because of uncertainty about the nature and con-
tent of such an identity. Accordingly, the lack of a salient social identity
should add ambiguity for members of a social movement trying to engage
in ongoing group-based action-oriented behavior and, therefore, should
mean that the social movement might be less likely to achieve its aims,
or indeed to stay together at all.
Paradoxically, then, although social movements are clearly real psy-
chological groups, some aspects of their very nature would seem to mil-
itate against their ability to function as effective groups. These issues are
partly addressed in, for example, Haslam’s (2001) outline of organic
social identity (see also Durkheim, 1933) where he suggested that a
dual identity model as proposed by Gaertner et al. (2000) might resolve
issues related to disputed group definition and categorization. Thus,
Haslam argued that a superordinate group identity could accommodate
(as opposed to subsume) distinct other lower-level identities when differ-
entiation was perceived as an integral characteristic of the superordinate
identity. Social actions conducted under the ambit of anti-globalization
provide ideal contemporary illustrations of Haslam’s general point. For
example, large-scale protests against a World Economic Forum (WEF)
meeting in Melbourne, Australia, were driven by the broad “S11” anti-
globalization community, and involved protesters who were members of a
diverse range of pre-existing groups uniting for a specific common cause,
while at the same time retaining their distinctive sub-group identities. In
the words of the protesters themselves, the S11 community was “sym-
bolic of this difficult but ultimately productive diversity” (Price, 2000;
but see below for a more detailed analysis of the S11 community’s online
interaction).
On a related issue, Haslam et al. (1998) also highlighted the role
of consensus in groups. Elaborating the process of consensus develop-
ment in terms of the idea of consensualization (Haslam, 2001; Haslam
et al., 1999), they argued that a salient social identity in concert with
social influence through group interaction enhanced consensus about
the group. Put another way, consensualization works upon group mem-
bers to homogenize individual beliefs into group beliefs that come to
represent reality as defined by the shared social identity of the group.
Following a similar theme, Reicher (2000) contended that group-based
“collective assertions” are not simply the product of socially determined
perceptions of existing reality but may also represent statements about
the ideal reality that a group hopes to achieve. Moreover, and with par-
ticular relevance to social movements, Reicher made plain the reciprocal
relationship between identity and social action, arguing that not only
does identity lead to action, but it influences and is affected by the social
constraints that determine the limits of acceptable action (Reicher, 2000;
Reicher et al., 1995; Sani and Reicher, 1998; Stott and Reicher, 1998).
We will now turn to the analyses of two prominent social movements,
describing how closely our theoretical description of their expected online
behavior corresponded with our actual observations of those groups. We
will argue that our empirical evidence provides a more complete picture
of the role that online interaction plays in social movements, and leads us
towards a more subtle understanding of how OLCs enable and facilitate
social action.
The global activist – online social movements

It makes a lot more sense to meet people through people, through organizations,
and through activist work, than to meet them on the Net. As an organizer, you
want to know a lot more about a person than their opinions. Opinions are easy;
what an organizer looks for are dependable, steady, focused people who will do
political work. In fact, organizers spend a lot of time trying to avoid people whose
main desire is to express their opinions at length rather than work. These folks
make up one of the groups that is drawn to the internet.
(Yates, 1996)
We know that certain characteristics of organization and behavior should

be common across most forms of social movements – social move-
ments are typically superordinate organizations encompassing varied
sub-groups and individuals, and social movements have a strong action
orientation; in other words, they form with the intention to act, and
their members join and participate because they want to bring about
change. We have argued that social movements in general are psycho-
logical groups, and that online social movements are as real for their
participants as for their face-to-face counterparts; it follows, then, that
online and face-to-face social movements should share similar character-
istics and behave in the same ways.
As a consequence of the infancy of the medium itself, literature about
online interaction in general, and online protest and social movements in
particular, has a relatively short history. Academic interest in the medium,
meanwhile, has developed rapidly, mirroring the explosion of the internet
into public awareness. In this short but intensive tradition, researchers
have viewed online communities as including both face-to-face groups
that have adopted an online instantiation and groups that have formed
and exist exclusively online. Thus, for example, Olesen (2004) noted how
the pre-existing community of Zapatistas (a political movement com-
prised of indigenous peoples from the Chiapas region in Mexico seeking
control and autonomy over local resources and affairs) used the internet
to establish “solidarity networks,” whereas Earl and Schussman (2003)
looked at activist groups that formed online to facilitate strategic voting
during the US presidential elections. Though their origins were different,
both groups were legitimate examples of online social movements.
Regardless of where they originate, there would appear to be significant
qualitative differences in the nature of interaction in online groups com-
pared with groups that meet face-to-face. Thus, even now, communica-
tion on the internet is still predominantly asynchronous and text-based,
and, as we have already discussed, the consequences of online anonymity
have proved popular targets for comment and research. Furthermore,
activists themselves hold mixed views about the utility of the internet,
as we can see from Yates’s (1996) argument that movement organizers
actually try to avoid internet-based groups because of those groups’ focus
on argument and debate. In many ways, then, the internet would seem
to be precisely the sort of diffuse context where opinion-based groups
(OBGs) are extremely useful because online forums provide ideal envi-
ronments for the exchange of opinions, and thus for the formation of
OBGs. Along these lines, we can point to some famous examples of the
development of opinion-based activism online, for example within the
online anti-globalization movement (Clark and Themundo, 2006; Wall,
2007; see also Fisher et al., 2005; Postmes and Brunsting, 2002).
Notwithstanding the utility of online interaction for OBG formation,
popular analysis has often noted the way that protest groups have used the
internet as an organizational tool while ignoring the other roles that the
internet might play in these groups (e.g., Allard and O’Loughlin, 2000;
Lawrence, 2000; McDonald, 2000). At first blush, given the unmistak-
able action focus of social movements generally and the high online pro-
file of many anti-globalization movements in particular, this may seem a
reasonable focus. We should ask, though, how the formation of opinion-
based groups online reconciles with this action-centered model of online
social movements and, more broadly, with the disconnection between
commitment to groups and commitment to action that we discussed ear-
lier in this chapter. Put more bluntly, a nagging question remains: does
the internet help or hinder the active pursuit of goals in online social
movements? To investigate this question, we first looked at the actual traf-
fic within one online community pursuing an anti-globalization agenda,
namely the online movement that developed around protests against the
WEF meeting in Melbourne.
On 11 September 2000, members and associates of the WEF met in
the Crown Casino complex in Melbourne for a three-day Asia Pacific
Economic Summit. There to greet them were several thousand anti-
globalization protesters rallying under the S11 banner. One focus of the
lay analysis of the S11 protests, and, indeed, other protest actions taking
place around the world at similar times, was the role of the internet as
a medium for communication and organization. In fact, an important
mechanism through which interaction took place in the S11 movement
was an online community (for example, nearly 3,000 messages were
posted to S11 over the period leading up to the protest event). During
eleven months over which the S11 protest organized and unfolded, we
looked at the structure and content of the interaction in that online
community by analyzing a stratified random sample of 600 messages.
Following the pattern we expect from a social movement, S11’s stated
aim was to plan and conduct protest action to disrupt the Melbourne
meeting. We were especially interested in the degree to which the S11
online community (an open, unmoderated emailing list) grew into a
fully fledged group that developed principled positions and agreed on
norms of behavior. In particular, we reasoned that, as the WEF meeting
(the movement’s raison d’être) approached, there would be increasing
attempts to reach agreement on group behavior and action.
In fact, we found the opposite. As the date of the protest action loomed
closer there was proportionately less discussion that could be interpreted
as agreeing on action. Rather (adopting the parlance developed by Tuck-
man, 1965), as the online group approached the point where it should
be ready to perform, it continued to storm and storm (with such storm-
ing – and associated recriminations – continuing even after the actual
protest event had passed). Furthermore, contrary to our expectations
about consensualization, we found no evidence of consensus formation,

merely lengthy discussion, much of which was based around debates
between anarchism and socialism.1
Although contrary to our expectations, we realized that one important
factor could have tempered the willingness of movement participants to
focus their interaction as we had originally anticipated. The S11 OLC
was an open community – in other words, all community interaction was
public and available for view by anyone with internet access. As a result,
interacting parties were aware of the potential for their activities to be
monitored by outsiders, including police, intelligence organizations, and
political foes. We reasoned, therefore, that participants may have been
strategically avoiding action planning and that any actual organization of
action may have been taking place in private online forums hidden from
public scrutiny. We also observed that the S11 OLC used no formal
mechanism for polling members or taking votes about specific topics
or issues, even though this facility was readily available on the software
platform on which the community was based. In other words, the S11
OLC appeared to make no effort to formally gauge participants’ points
of view.
To assess the extent to which the interaction we observed on S11 may
have been influenced by participants’ strategic use of their community we
looked at a large number of other open online communities, less in the
public eye but still formed around social or political activism. We wanted
to see whether those communities engaged in decision making and action
planning through the use of polling or whether they, too, behaved as did
the S11 OLC. As with S11, we found almost no evidence for the use of
polls in these other communities, either to decide on courses of action
or to measure consensus about any issues relevant to community causes
(in fact, in the limited instances where we observed their use, polls were
really used only for procedural list management purposes).
Clearly, however, there are online social movement communities that
do attempt to measure consensus. One of the largest and, arguably, most
influential social movement OLCs is MoveOn.org (www.moveon.org).
A broad-based social movement, MoveOn.org has more than 3 million
members and encompasses campaigns and actions on a wide range of
social and political issues (most actions are related to US policy issues,
though some focus on international outcomes – e.g., environmental sus-
tainability, the genocide in Darfur). While the organization itself engages
1 Using Cohen’s Kappa, for categories relevant to discussion, average inter-rater reliability
between four coders ranged from κ=0.55 to κ=0.70 (scores falling between 0.40 and
0.75 show fair to good agreement; Armitage and Berry, 1994).
in a mix of face-to-face and online actions, the MoveOn.org website

and associated lists and forums are the main focal point for member-
ship recruitment, distribution of information, and announcements and
reviews of future and completed actions. Significantly, MoveOn.org oper-
ates an online “ActionForum” where posters are able to comment on,
suggest, and debate topics relevant to the wider MoveOn organization.
That forum is specifically structured to militate toward decision mak-
ing and action by arranging messages by importance ratings assigned to
them by other forum members. In other words, not only does the forum
encourage action content, it allows participants to explicitly evaluate the
worth of messages, and displays messages hierarchically according to
their merit as judged by other forum users. Thus, the structure of the
ActionForum itself is designed to facilitate consensual decision making
and action planning (see, for example, Lemus et al., 2004, about how
perceptions of agreement in computer-mediated groups influences intra-
group decision making).
We used stratified random sampling to select 400 messages from those
most highly rated in the ActionForum and assessed them for advocacy
of action compared with statements of general opinions or ideas. Similar
to the pattern of observations from our other studies of online commu-
nities, we found no clear evidence that MoveOn.org OLC participants
preferred action-related messages to general messages about opinions or
ideas.2 In other words, even when OLC participants were explicitly able
to express their preference for action, they did not clearly do so, measured
in either raw numbers of message postings or in the levels of importance
assigned to messages about action versus non-action. To clarify, we are
not suggesting that action-related messages were not present at all on
the forum, or that the broader MoveOn movement (i.e., its online and
face-to-face instantiations combined) did not engage in action – in fact,
the overall movement does conduct regular actions (see the movement
website for details).
Rather, we make the point that participants in the online forum did
not engage in significant levels of action-related interaction or decision
making, and so that forum cannot be seen to clearly drive the over-
all movement’s actions. In continuing work, we are conducting a more
fine-grained analysis of the MoveOn.org data that tracks more closely
the content of messages related to specific issues over time. This analy-
sis allows us to detect evidence of discussion and debate within the
2 Inter-rater reliability between four coders ranged from κ=0.65 to κ=0.90 (scores between
0.40 and 0.75 show fair to good agreement, and scores greater than 0.75 demonstrate
excellent agreement; Armitage and Berry, 1994).
forum, and to assess the development of intragroup consensus, another

key characteristic we would expect to find in this type of OLC. Again,
however, we have thus far found no evidence of this form of consensual-
ization. Rather than debate issues toward a conclusion, we have typically
observed participants present “stand-alone” statements or points of view
(usually, but not always, in sympathy with the broad aims of the com-
munity). So, for example, in our analysis of one message thread, we
detected only a single direct response to a previous comment. While
other messages in that thread did sometimes address common issues
(e.g., the war in Iraq), those messages did not reference previous posts,
nor directly solicit opinion or feedback from other participants in the
interaction. Moreover, although the chronology of messages displays
some characteristics of consensualization (e.g., early on, participants
posted messages both for and against the issue, whereas in the latter
part of the interaction only messages favoring one side were posted),
the content of messages does not refer to or rebut opposing arguments.
Significantly, in this follow-up analysis we are extracting and analyzing
the minority of messages particularly related to action about a specific
topic. In other words, even where we have identified interaction where
participants are focused on movement action, we have so far found
little evidence of engagement with others, or argument or consensus
development.
Overall, then, we found little evidence of debate, consensus, or action
in the online social movement groups we looked at. Nevertheless, the
OLCs we observed were active, functioning communities (e.g., 3,000
messages posted to the S11 forum and more than 40,000 individual
postings on MoveOn.org at the time of our study). So, if not planning or
conducting action or working toward intragroup consensus, what were
activists doing in their online groups?
Our findings suggested a paradox – even though the charters of social
movement OLCs usually involved a commitment to action, they seem
to settle for talk. At face value, this apparent contradiction could be
explained in the context of the online communities we looked at. Thus,
it is possible that in open communities participants are cognizant of
problems associated with internet identifiability. Notwithstanding recent
work contrasting “depersonalization” and “deindividuation” (Douglas
and McGarty, 2001; Spears et al., 2002; see also our discussion of cyber-
hate in this chapter), some have noted how openness and accessibility
combined with the relative anonymity of much online interaction can
be a double-edged sword for activist groups, providing both opportuni-
ties for participation and mobilization, while at the same time increasing
the potential threats of subversion and misrepresentation (e.g., Lebert,
2003). For example, it makes no sense to have an online vote in an open

forum on whether to conduct a street rally in Ohio because poll num-
bers can be manipulated by political opponents, or by people who would
never turn up to a rally in that geographical locality. Moreover, agent
provocateurs can sabotage progress toward consensualization in forums
where anyone is free to post (although the rating mechanism employed by
the MoveOn.org OLC effectively protected against that threat). Funda-
mentally, however, our observations showed that participants continued
to enthusiastically participate in action communities that did not, in fact,
engage in “action.”
How, then, should we account for this? One possible way is to broaden
our definition of what constitutes acceptable action, and thus also refine
our understanding of the role that some social movement OLCs play in
wider activist collectives. We have employed the metaphor of a newspa-
per opinion page to illuminate this idea further. The authors of letters
that appear on an opinion page write to present an opinion or point
of view. Opinion pages are neither mechanisms for ongoing debate nor
environments for consensualization to occur. Both writers and readers
understand and accept the role of the opinion page, and the boundaries
within which interaction on those pages occurs and is useful. From our
analysis of social movement OLCs, we argue that many of these types of
communities fulfill a similar role. In other words, we should expect that
many online groups will be no more likely to reach consensus over time
than will the content of the letters to the editor of a newspaper. Conse-
quently, the purpose of these online discussions is not to convince other
communicators to necessarily embrace a consensus, but rather to define,
capture, and make public the zeitgeist of the movement, and to convince
readers (or, in online terms, “lurkers”) of the merits of a position. In one
sense, then, these types of online communities are true instantiations of
opinion-based groups.
We make one final claim about our findings in relation to social move-
ment OLCs. Following our observations about the continued primacy of
action in the communities we have studied (where that primacy appears
to be in direct contrast to the reality of community interaction), we
are currently exploring the idea that new members of OLCs join with
the purpose of promoting action, but come to accept talk as a perfectly
acceptable substitute for action. One interpretation of our findings, sum-
marized as the community transformation hypothesis, is that the internet
acts as a tool for demobilization; it becomes, in effect, the new “opium
for the masses.” We would argue that this critique is too harsh. In our
view, the consciousness raising provided by communication itself can be
important and we contend that OLC participants recognize this role for
their communities. In this way we provide a hopeful counterpoint to the

community transformation hypothesis: namely, that talk is action, and
even collective action.
Conclusion
In this chapter we have considered and taken issue with two popular
propositions about the internet. We conclude that both propositions rep-
resent good places to start an analysis of the nature of online interac-
tion, but they represent poor places to finish. The true situation is both
more complex and more interesting than either the anonymous cyberhate
proposition or the global activist proposition would suggest.
We have presented an interpretation of online communities where,
in one sense, participants engage in a quite strategic use of interaction
and communication to shape and present their ideal conceptualization
of reality – both for their immediate community and for their wider
cause. Our work makes use of a number of recent advances in under-
standing of internet-based communication. For example, our ideas about
cyberhate and anonymity are based on the SIDE model (Reicher et al.,
1995; Spears and Lea, 1994), and our developing understanding of global
activism draws from ideas about the interaction between action and iden-
tity (Haslam, 2001; Reicher, 2000), and instrumental and identity-based
motivations (Simon et al., 1998; van Zomeren et al., 2004). Joinson’s
(2003) SMEE framework also emphasizes interactivity, describing the
“ongoing process whereby users’ on-line activities feed back to their on-
line and real-life characteristics” (p. 179), and some of our conclusions
clearly connect with his work.
A crucial theme that our work adds, however, is the concept of the
opinion-based group that we argue is useful for understanding action
and inaction on the internet. The internet is a supreme forum for the
exchange and formation of opinions. From that fertile ground on which
the seeds of so many opinions are sewn it seems inevitable that we will see
the explosive growth of opinion-based groups. The only historical prece-
dent we have for such an explosive growth of a tool for social organization
was the rise of the printing press and the use of political and religious
pamphlets in the Europe of the Reformation era. We sincerely hope that
twenty-first-century society is able to continue to apply these tools with-
out the chaos and hostility that accompanied the earlier developments
(e.g., the Thirty Years War). We believe that the continued development
of a social science of computer-supported interaction is one way that
such problems can be forestalled.
References
Allard, T., and O’Loughlin, T. (2000). The storming of Melbourne. The Sydney
Morning Herald, 31 August.
Armitage, P., and Berry, G. (1994). Statistical methods in medical research (3rd
edn.). Oxford: Blackwell Scientific Publications.
Back, L. (2002). Aryans reading Adorno: Cyber-culture and twenty-first century
racism. Ethnic and Racial Studies, 25, 628–651.
Bargh, J. A., and Chartrand, T. L. (1999). The unbearable automaticity of being.
American Psychologist, 54, 462–479.
Bliuc, A.-M., McGarty, C., Reynolds, K. J., and Muntele, D. (2007). Opinion-
based group membership as a predictor to commitment to political action.
European Journal of Social Psychology, 37, 19–32.
Clark, J. D., and Themundo, N. S. (2006). Linking the web and the street:
Internet-based “Dotcauses” and the “Anti-Globalization” movement. World
Development, 34, 50–74.
Cohen, J. L. (1985). Strategy or identity: New theoretical paradigms and con-
temporary social movements. Social Research, 52, 663–716.
de Weerd, M., and Klandermans, B. (1999). Group identification and politi-
cal protest: Farmers’ protest in the Netherlands. European Journal of Social
Psychology, 29, 1073–1095.
Diener, E. (1980). Deindividuation: The absence of self-awareness and self-
regulation in group members. In P. Paulus (Ed.), The psychology of group
influence (pp. 209–242). Hillsdale, NJ: Erlbaum.
Douglas, K. M., and McGarty, C. (2001). Identifiability and self-presentation:
Computer-mediated communication and intergroup interaction. British
(2002). Internet identifiability and beyond: A model of the effects of identifia-
bility on communicative behavior. Group Dynamics, 6, 17–26.
Douglas, K. M., McGarty, C., Bliuc, A. M., and Lala, G. (2005). Under-
standing cyberhate: Social competition and social creativity in online white
supremacist groups. Social Science Computer Review, 23, 68–76.
Durkheim, E. (1933). The division of labor in society. Glencoe, IL: Free Press.
Earl, J., and Schussman, A. (2003). The new site of activism: On-line organiza-
tions, movement entrepreneurs, and the changing location of social move-
ment decision making. In P. G. Coy (Ed.), Research in social movements,
conflict and change (Vol. 24) (pp. 155–187). Amsterdam: JAI.
Fisher, D. R., Stanley, K., Berman, D., and Neff, G. (2005). How do organiza-
tions matter? Mobilization and support for participants at five globalization
protests. Social Problems, 52, 102–121.
Foot, K. A., and Schneider, S. M. (2002). Online action in Campaign 2000: An
exploratory analysis of the U.S. political web sphere. Journal of Broadcasting
and Electronic Media, 46, 222–244.
FSF (1998). Free Software Foundation. Retrieved 4 December 2007, from www.
fsf.org
Gaertner, S. L., Dovidio, J. F., Banker, B. S., Houlette, M., Johnson, K. M., and
McGlynn, E. A. (2000). Reducing intergroup conflict: From superordinate
goals to decategorization, recategorization, and mutual differentiation.

Group Dynamics: Theory, Research, & Practice March, 4, 98–114.
GCAP (2005). Global Call to Action against Poverty. Retrieved 1 December
2007, from www.whiteband.org
Gerlach, L. P., and Hine, V. H. (1970). People, power, change: Movements of social
transformation. Indianapolis, IN: The Bobbs-Merrill Company Inc.
Green, D. P., Abelson, R. P., and Garnett, M. (1999). The distinctive political
views of hate-crime perpetrators and white supremacists. In D. A. Prentice
and D. T. Miller (Eds.), Cultural divides: Understanding and overcoming group
conflict (pp. 429–464). New York: Russell Sage Foundation.
Haslam, S. A. (2001). Psychology in organizations: The social identity approach.
London; Thousand Oaks, CA: Sage.
Haslam, S. A., Oakes, P. J., Reynolds, K. J., and Turner, J. C. (1999). Social
identity salience and the emergence of stereotype consensus. Personality and
Social Psychology Bulletin, 25, 809–818.
Haslam, S. A., Turner, J. C., Oakes, P. J., Reynolds, K. J., Eggins, R. A., Nolan,
M., et al. (1998). When do stereotypes become really consensual? Investi-
gating the group-based dynamics of the consensualisation process. European
Jacobs, D. (2005). Internet activism and the democratic emergency in the U.S.
ephemera, 5, 68–77.
Jenkins, C. J. (1983). Resource mobilization theory and the study of social move-
ments. Annual Review of Sociology, 9, 527–553.
Joinson, A. N. (2003). Understanding the psychology of internet behaviour: Virtual
worlds, real lives. Houndmills, Basingstoke, Hampshire; New York: Palgrave
Macmillan.
Kelly, C., and Breinlinger, S. (1995). Identity and injustice: Exploring women’s
participation in collective action. Journal of Community and Applied Psychol-
ogy, 5, 41–57.
Kiesler, S., Siegel, J., and McGuire, T. W. (1984). Social psychological aspects of
computer-mediated communication. American Psychologist, 39, 1123–1134.
Klandermans, B. (1984). Mobilization and participation: Social-psychological
expansions of resource mobilization theory. American Sociological Review,
49, 583–600.
(1997). The social psychology of protest. Oxford: Blackwell.
(2000). Identity and protest: How group identification helps to overcome col-
lective action dilemmas. In M. V. Vugt, M. Snyder, T. Tyler, and A. Biel
(Eds.), Cooperation in modern society: Promoting the welfare of communities,
states and organizations. London: Routledge.
Klandermans, B., Sabucedo, J. M., Rodriguez, M., and de Weerd, M. (2002).
Identity processes in collective action participation: Farmers’ identity and
farmers’ protest in the Netherlands and Spain. Political Psychology, 23, 235–
251.
Lala, G. (2008). Talking the talk: Do words speak louder than actions in socio-
political communities on the internet? Unpublished doctoral dissertation,
Australian National University, Canberra, Australia.
Lawrence, K. (2000). When a group becomes a mob – Who was behind S11?
The Daily Telegraph, 14 September.
Lazarus, R. S. (1991). Progress on a cognitive-motivational-relational theory of

emotion. American Psychologist, 46, 819–834.
Lea, M., O’Shea, T., Fung, P., and Spears, R. (1992). “Flaming” in computer-
mediated communication: Observations, explanations, implications. In M.
Lea (Ed.), Contexts of computer-mediated communication (pp. 89–112). New
York: Harvester Wheatsheaf.
Lea, M., and Spears, R. (1995). Love at first byte? Building personal relationships
over computer networks. In J. T. Wood and S. Duck (Eds.), Under-studied
relationships: Off the beaten track (pp. 197–233). Understanding Relationship
Processes: Vol. 6. Thousand Oaks, CA: Sage.
Lebert, J. (2003). Wiring human rights activism: Amnesty International and
the challenges of information and communication technologies. In M.
McCaughey and M. D. Ayers (Eds.), Cyberactivism: Online Activism in Theory
and Practice (pp. 209–232). New York: Routledge.
Lemus, D. R., Seibold, D. R., Flanagin, A. J., and Metzger, M. J. (2004). Argu-
ment and decision making in computer-mediated groups. Journal of Com-
munication, 54, 302–320.
McAdam, D., McCarthy, J. D., and Zald, M. N. (1988). Social movements. In
N. J. Smelser (Ed.), Handbook of sociology (pp. 695–737). London: Sage.
McCaughey, M., and Ayers, M. D. (Eds.) (2003). Cyberactivism: Online activism
in theory and practice. New York: Routledge.
McDonald, K. (2000). The new revolutionaries. The West Australian, 2 Septem-
ber.
McDonald, M. (1999). Cyberhate: Extending persuasive techniques of low cred-
ibility sources to the World Wide Web In E. Thorson and D. W. Schumann
(Eds.), Advertising and the World Wide Web (pp. 149–157). Mahwah, NJ:
Lawrence Erlbaum Associates.
McGarty, C., and Bliuc, A.-M. (2004). Collective guilt in Australia. In B. Doosje
and N. R. Branscombe (Eds.), Collective guilt: Antecedents, correlates and
consequences (pp. 112–129). Cambridge: Cambridge University Press.
McGarty, C., Bliuc, A.-M., Thomas, E. F., and Bongiorno, R. T. (2009). Collec-
tive action as the material expression of opinion-based group membership.
Journal of Social Issues, 65, 839–857.
McGarty, C., Pedersen, A., Leach, C. W., Mansell, T., Waller, J., and Bliuc,
A.-M. (2005). Group-based guilt as a predictor of commitment to apology.
British Journal of Social Psychology, 44, 659–680.
Musgrove, L., and McGarty, C. (2008). Opinion-based group membership as a
predictor of collective emotional responses and support for pro and anti war
action. Social Psychology, 39, 37–47.
Oberschall, A. (1980). Loosely structured collective conflict: A theory and an
application. Research in Social Movements, Conflict and Change, 3, 45–68.
Olesen, T. (2004). The transnational Zapatista solidarity network: An infrastruc-
ture analysis. Global Networks, 4, 89–107.
OSI (1998). Open Source Initiative. Retrieved 3 December 2007, from www.
opensource.org
Postmes, T., and Brunsting, S. (2002). Collective action in the age of the internet.
Social Science Computer Review, 20, 290–301.
Price, S. (2000). S11: “We made it work by all sticking together.” Green Left
Online, retrieved 15 May 2008, from www.greenleft.org.au/2000/421/22757
Reicher, S. D. (2000). Social identity definition and enactment: A broad SIDE
against irrationalism and relativism. In T. Postmes, R. Spears, M. Lea, and
S. D. Reicher (Eds.), SIDE issues centre stage: Recent developments in studies of
de-individuation in groups. Amsterdam: Royal Netherlands Academy of Arts
and Sciences.
Reicher, S. D., and Levine, M. (1994a). Deindividuation, power relations
between groups and the expression of social identity: The effects of visi-
bility to the out-group. British Journal of Social Psychology, 33, 145–163.
(1994b). On the consequences of deindividuation manipulations for the strate-
gic communication of self: Identifiability and the presentation of social iden-
tity. European Journal of Social Psychology, 24, 511–524.
Reicher, S. D., Levine, R. M., and Gordijn, E. (1998). More on deindividuation,
power relations between groups and the expression of social identity: Three
studies on the effects of visibility to the in-group. British Journal of Social
197.
Sani, F., and Reicher, S. D. (1998). When consensus fails: An analysis of the
schism within the Italian Communist Party (1991). European Journal of
Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. W. (1986). Group pro-
cesses in computer-mediated communication. Organizational Behaviour and
Human Decision Processes, 37, 157–187.
Simon, B., and Klandermans, B. (2001). Politicized collective identity. American
Psychologist, 56, 319–331.
Simon, B., Loewy, M., Stürmer, S., Weber, U., Freytag, P., Habig, C., et al.
(1998). Collective identification and social movement participation. Journal
of Personality & Social Psychology, 74, 646–658.
Simon, B., Stürmer, S., and Steffens, K. (2000). Helping individuals or group
members? The role of individual and collective identification in AIDS vol-
unteerism. Personality and Social Psychology Bulletin, 26, 497–506.
Snow, D. A., Soule, S. A., and Kriesi, H. (Eds.) (2004). The Blackwell companion
to social movements. Oxford: Blackwell Publishing.
Spears, R., and Lea, M. (1994). Panacea or panopticon? The hidden power
in computer-mediated communication. Communication Research, 21, 427–
459.
Spears, R., Postmes, T., Lea, M., and Wolbert, A. (2002). When are net effects
gross products? The power of influence and the influence of power in
computer-mediated communication. Journal of Social Issues, 58, 91–107.
in organizational communication. Communication Research, 32, 1492–1512.
Stallman, R. (2007). Why “Open Source” misses the point of Free Software.
Retrieved 4 December 2007, from www.gnu.org/philosophy/open-source-
misses-the-point.html
Stott, C., and Reicher, S. D. (1998). Crowd action as intergroup process: Intro-
ducing the police perspective. European Journal of Social Psychology, 28, 509–
529.
Stürmer, S., Simon, B., Loewy, M., and Jörger, H. (2003). The dual-path model
of social movement participation: The case of the fat acceptance movement.
Social Psychology Quarterly, 66, 71–82.
Tajfel, H. (1981). Human groups and social categories: Studies in social psychology.
Cambridge: Cambridge University Press.
Tajfel, H., and Turner, J. C. (1979). An integrative theory of group conflict. In W.
G. Austin and S. Worchel (Eds.), The social psychology of intergroup relations
(pp. 33–47). Monterey, CA: Brooks/Cole Publishing Company.
(1986). The social identity theory of intergroup behavior. In S. Worchel and
W. G. Austin (Eds.), Psychology of intergroup relations (pp. 7–24). Chicago,
IL: Nelson-Hall.
Taylor, N., and McGarty, C. (2001). The role of subjective group memberships
and perceptions of power in industrial conflict. Journal of Community &
Applied Social Psychology, 11, 389–393.
Tuckman, B. W. (1965). Developmental sequence in small groups. Psychological
Bulletin, 63, 384–399.
Turner, J. C. (1982). Towards a cognitive redefinition of the group. In H. Tajfel
(Ed.), Social identity and intergroup relations (pp. 102–140). Cambridge:
Cambridge University Press.
Basil Blackwell.
Turner, J. C., Oakes, P. J., Haslam, S. A., and McGarty, C. (1994). Self and
collective: Cognition and social context. Personality and Social Psychology
Bulletin, 20, 454–462.
van de Donk, W., Loader, B. D., Nixon, P. G., and Rucht, D. (Eds.) (2004).
Cyberprotest: New media, citizens and social movements. London: Routledge.
van Gelder, L. (1985). The strange case of the electronic lover. In C. Dunlop
and R. Kling (Eds.), Computerization and controversy: Value conflicts and social
choices (pp. 364–375). San Diego, CA: Academic Press.
van Zomeren, M., Spears, R., Fischer, A. H., and Leach, C. W. (2004). Put
your money where your mouth is! Explaining collective action tendencies
through group-based anger and group efficacy. Journal of Personality & Social
Wall, M. A. (2007). Social movements and email: Expressions of online identity
in the globalization protests. New Media & Society, 9, 258–277.
Yates, L. L. (1996). The internet: What it can and can’t do for activists. The
Internet Society Sixth Annual Conference. Retrieved 1 December 2007, from
www.isoc.org/inet/96/proceedings/e6/e6 1.htm
Zimbardo, P. G. (1969). The human choice: Individuation, reason, and order
versus deindividuation, impulse, and chaos. In W. J. Arnold and D. Levine
(Eds.), Nebraska symposium on motivation (Vol. 17) (pp. 237–307). Lincoln,
NE: University of Nebraska Press.
9 A juxtaposition of social influences
Web 2.0 and the interaction of mass, interpersonal,

and peer sources online
Joseph B. Walther, Stephanie Tom Tong, David C.
DeAndrea, Caleb T. Carr, and Brandon Van Der Heide
Contemporary web-based communication technologies often present

a juxtaposition of multiple sources. In January of 2009, for instance,
CNN.com and Facebook.com partnered so that internet users could see
and hear President Barack Obama’s inauguration in one window, and
read and write comments about it with myriad other viewers in another
window. YouTube.com presents not only videos but also the comments
that viewers have posted about the video. Facebook.com supports indi-
viduals’ self-descriptive information online, but also “wall postings” by
a profile owner’s friends, which may enhance, modify, or contradict the
self-presentations individuals attempt to make. Sellers on eBay advertise
goods online, and other buyers’ ratings of those sellers are a click away.
Travel sites offer hotel listings alongside ratings and reviews by previous
guests. And in addition to the overt comments and statements left online
by people, the computational systems supporting these interactions often
provide interesting statistical coefficients summarizing the evaluations or
even the electronic footprints users leave behind: How many people rated
a product evaluation as useful? How many times has a viral video been
viewed, or a journal article been downloaded? How many friends does an
individual have on Facebook? In all of these cases, there is a central mes-
sage, and there are socially generated responses. How do these mutual
sources of influence affect users?
New communication technologies are changing the manner of recep-
tion by which individuals acquire information from institutional, inter-
personal, and peer information sources. Technology changes the tem-
poral and contiguous presentations of these sources, and may in fact
change the information processing and social influence dynamics among
these sources; that is, the sequence with which sources are sampled
or the simultaneity with which they appear may have potent effects
on the information processing filters and biases. Such developments in
communication technologies are raising new questions and resurrecting
172
A juxtaposition of social influences 173
old questions about the interplay of interpersonal and mass commu-

nication. New technologies blur the boundaries between interpersonal
and mass communication events and/or the roles that communicators
take using new systems. Arguments have been made that the “conver-
gence” of old and new media demands new and unified perspectives on
traditionally segregated processes. “Media convergence” is a term that
has been used to connote several phenomena that are brought about by
advancements in telecommunication technology that may change some
aspect of the communication process. Sometimes the term refers to the
blending of previously individuated mass media: one can watch movies
on one’s computer, for example. We wish to discuss another kind of con-
vergence: the potential for simultaneous communication via computers
of both mass and interpersonal channels. Additionally, new and salient
sources of information are present, and potentially influential, via con-
temporary online information systems: peer communication and com-
putational summaries of social behavior. How these forms and sources of
information coalesce, confirm, or contradict one another asks for concep-
tual and analytic frameworks to address them. New technologies invite
research that will advance understanding of how individuals conceptual-
ize communication and interpret messages, and how their attitudes are
changed as a result of new mediated message forms and content.
The purposes of the present work are several. First, we review
approaches to the division and interaction of mass and interpersonal
communication processes, to see what questions and assertions have
been raised that may continue to guide understanding of these processes
as they unfold via new technologies. Second, we discuss a recent articu-
lation of an expanded perspective on the interplay of institutional, peer,
and interpersonal sources through contemporary communication tech-
nologies, originally introduced by Walther et al. (2010), whose work is
reflected in many of this chapter’s assertions. The present work expands
on theirs, however, in offering a typological view of competing influence
sources online, and the consideration of additional research results which
inform the frameworks they suggested. These expansions highlight how
the paradigm can foster understanding of the information processing
patterns that such convergent forms of new media sources make salient.
Perspectives on mass/interpersonal
divisions and mergers
Traditionally, mass communication has been conceptualized as one-way
message transmissions from an institutional source to a large, relatively
174 Walther, Tong, DeAndrea, Carr, and Van Der Heide
undifferentiated and anonymous audience. Interpersonal communica-

tion, by contrast, involves smaller numbers of participants who exchange
messages directed toward particular others. Interpersonal communica-
tion has been typified as bidirectional message exchange between two
or more individuals in which communication strategies are shaped by
knowledge about one another’s idiosyncratic preferences and the instru-
mental and relational goals of the individuals involved (see, for review,
Berger and Chaffee, 1989; Cappella, 1989). Some important scholarly
works have posited integrated, sequential processes involving both mass
and interpersonal communication. Mass media’s influence on individu-
als’ opinions about politics, style, and other cultural issues is well known
to involve a two-step flow of mass media messages and interpersonal dis-
cussions (e.g., Katz, 1957; Katz and Lazarsfeld, 1955; Lazarsfeld et al.,
1944). The integration of mass and interpersonal processes is explicit in
the theory of the diffusion of innovations, as well (Reardon and Rogers,
1988).
Despite these exceptions, the exploration of mass and interper-
sonal processes often takes place in isolation. Several researchers
have described the historical and institutional causes of disciplinary
divides between mass communication and interpersonal communication
research. Wiemann et al. (1988) attributed the division to historical and
academic/bureaucratic differences. Reardon and Rogers (1988) argued
that the division developed as a result of scholars’ efforts to define their
distinctive contributions to social science. Interpersonal communication
research followed the tradition of psychology and social psychology from
the 1920s–1930s. Sources such as Heider’s (1958) Psychology of Interper-
sonal Relations and the psychological, sociological, and anthropological
works of Argyle, Goffman, and Bateson, respectively, helped solidify the
relevance of social scientific research on face-to-face interaction and rela-
tionships (Reardon and Rogers, 1988), and helped to define the study
of interpersonal communication. Mass media research evolved primar-
ily from sociology and political science. Despite its historical utility, this
division has been lamented for a variety of reasons. The most prevalent
concern is that a lack of synthesis between mass and interpersonal com-
munication theories and research methods has developed so that schol-
ars with functionally similar interests may not be aware of the scientific
work being performed outside of their area of specialization (Berger and
Chaffee, 1988; Pingree et al., 1988; Reardon and Rogers, 1988).
For some time, however, advocates have argued that new communica-
tion technologies have the potential to merge the very processes conven-
tionally considered as pertaining to mass communication or interpersonal
communication, and that the merger of processes demands the merger of
approaches in order to understand such phenomena. For example, Rear-

don and Rogers (1988) suggested that new interactive media did not
neatly fit into pre-existing areas of study. They claimed that a new episte-
mological approach to communication research may be needed. Several
observers suggested that new technologies defy easy categorization as
either interpersonal or mass media channels because of their interac-
tive nature (Gumpert and Cathcart, 1986; Newhagen and Rafaeli, 1996;
O’Sullivan, 1999, 2005; Pingree et al., 1988; Reardon and Rogers, 1988).
Thus, commentators held out hope that “this technological change may
facilitate a long-needed paradigm shift in communication science” (Rear-
don and Rogers, 1988, p. 297) since analytic approaches from mass or
interpersonal communication traditions may be insufficient to grasp the
effects of new technologies on communication dynamics.
Likewise, O’Sullivan (1999, p. 580) argued that “the functional con-
vergence of mass and interpersonal channels, perhaps best represented by
the internet, is both a challenge and an opportunity for scholars to pursue
convergence of the two areas of study.” More recently, O’Sullivan (2005)
suggested that there are and have been unique blends of “masspersonal”
communication, not only in internet forms but through unconventional
appropriations of conventional media, when individuals use traditional
mass communication channels for interpersonal communication, tradi-
tional interpersonal communication channels for mass communication,
and new communication channels to generate mass communication and
interpersonal communication simultaneously. Newspaper classified ads
that send an interpersonal greeting from one specific party to another are
one such example; marriage proposals communicated by skywriting are
another.
Integrating mass and interpersonal dynamics may be easier said than
done. Adherents of each tradition who focus on new technology some-
times fail to realize their sub-disciplinary biases. For instance, interactiv-
ity, which is frequently mentioned in association with new technology,
may connote different things for different analysts: new media are rela-
tively more interactive than traditional sources, to mass communication
researchers; new media are less interactive than traditional sources, to
interpersonal communication researchers (Walther et al., 2005). Oth-
ers caution that analysis of emergent forms of internet communica-
tion defies a simplistic merger of traditional mass and interpersonal
perspectives altogether. Caplan (2001), for instance, argues that CMC
involves mixtures of traditional features of mass and interpersonal com-
munication in unique and recombinant ways: in CMC, senders can be
sources of mass communication (e.g., personal web pages, participating
in a large online forum) and an interpersonal communication partner
(e.g., instant messaging, online chatting) at the same time. Receivers in

CMC can be anonymous audience members (lurkers), and can also be a
target of an instant personalized message. Additionally, in CMC, message
processes are not constrained by time or physical space. Caplan argued
that these fundamental differences between CMC and traditional mass or
interpersonal communication systems cannot be understood by simply
“merging” or “bridging” mass and interpersonal perspectives; they are
fundamentally new processes that require a new paradigmatic approach.
Although most predate the study of contemporary electronic commu-
nication technologies, there have been some efforts to bring together spe-
cific mass and interpersonal processes over the years. These integrations
provide stimulating launching points for reconsideration of communica-
tion processes in light of recent changes in the media and interpersonal
landscapes. The following discussion reviews some exemplars, and sug-
gests extensions of their potential application with respect to new media.
Functional perspective on information seeking

Chaffee (1986) argued succinctly that information sources are less likely
to be selected based on whether they are mass or interpersonal channels,
but on other criteria. For instance, an interpersonal source may have
more or less credibility on a particular topic than a mass media source.
This criterion has taken on new significance given the high degree of
peer-generated information that is available online, an issue we will dis-
cuss at length momentarily. Additionally, mass media sources may not
provide the same degree of access to information on a particular topic as
might be available by asking an interpersonal acquaintance. Moreover, a
sequence of information seeking, interpretation, and discussion may be
repeated and refined any number of times: “The traditional concept of
a directional ‘two step’ or ‘multi step’ flow fails to capture the cyclical
and reciprocal nature of this process” (Chaffee, 1986, p. 76). Chaffee’s
views have important implications in the contemporary technological
landscape, which has fundamentally transformed the characteristics of
access and the potential for cyclical information use.
The access criterion that Chaffee (1986) identified has been trans-
formed radically, in several ways, with dramatic implications. In Chaf-
fee’s time, access considerations may have led an individual to choose an
interpersonal or media source depending on which source was more able
to deliver information on a specific topic most readily. If it was unlikely
that TV news or a newspaper would soon carry information on a topic
of interest, one might seek a knowledgeable friend. In the age of the
internet, however, a wide array of information is accessible on demand.
Information seeking is one of the primary motives driving internet use,

and users recognize that it is easier to do online than using other media
(Papacharissi and Rubin, 2000). Information that formerly appeared in
transitory form now persists in newspaper archives or broadcast repos-
itories such as Hulu.com and YouTube.com. Information is stored in
informal online discussions, Wikipedia entries, and numerous websites.
The internet’s databases, reference tools, and media archives, coupled
with sophisticated search tools, make information on most topics acces-
sible on demand, and Hargittai (2002) found that, with experience, inter-
net users employ increasingly sophisticated search strategies. Because of
the availability of the internet, traditional mass media or interpersonal
sources may be less likely to be easy-access starting points for informa-
tion seeking. The search engine puts a virtual encyclopedia on every
lap.
Furthermore, this radical degree of access seems to have obviated
traditional credibility concerns in terms of preferences and acceptability
of sources. Chaffee (1986) argued that credibility – the expertise and
trustworthiness of a source – rather than the channel plays the greatest
role in our acceptance of information. Yet the major components of
credibility traditionally have been the trustworthiness and expertise of
a source, and online peers excel in these characteristics. The potential
influence of a peer group has traditionally been considered a powerful
factor in individual opinion formation and expression, as decades of social
psychological research have shown (see Glynn and Park, 1997). Never
before have the actual comments and sentiments of specific individual
peers been aggregated and made visible in such a fashion as the internet
has rendered, as in, for example, recommender systems such as those on
product vendor sites, on which customers can post their evaluations of
products and vendors.
We may say that the peers on these sites are not simply peers, but peers
exhibiting “optimal heterophily” (Rogers and Shoemaker, 1971). They
are like us in terms of interests and in their shared perspective (e.g., also
customers rather than vendors, students rather than teachers) except for
one important difference: they have experience with the specific target
(vendor, professor, etc.) while we do not. Thus their trustworthiness and
relative expertise should be quite strong (see also Metzger et al., 2003).
Indeed, Sundar and Nass (2001) found that people more highly value
information presented on computers when they believe that the informa-
tion was selected by other (unidentified) computer users. In an experi-
ment that presented identical news stories on computers to subjects,
ostensibly peer-selected stories were preferred, as opposed to stories that
appeared to have been chosen by news editors, computer algorithms, or
even by the subjects themselves. When other users were perceived to be

the source of online news, the stories were liked more and perceived to
be higher in quality and more representative of news.
Optimal heterophily includes a strong degree of homophily, yet we tend
not to know much about online peers as individuals, suggesting a strong
influence of social identification processes in some online settings. Online
discussion spaces allow users to interact without having to disclose much
about one’s offline identity or qualifications. The sites are populated
by relatively anonymous peers. As such, they are prone to the kinds of
influence that social identification facilitates.
In sum, we argue that a dramatic shift brought on by electronic technol-
ogy’s changes in information access pertains not only to the convergence
of mass media (television, newspapers, and movies via the internet) but
also to the more fundamental convergence of mass, interpersonal, and
peer channels (mass media sources on the one hand, and synchronous or
asynchronous discussion with peers, family, and/or friends on the other).
As a result, questions arise about whether the simultaneous convergence
of information from multiple sources has the same degree of influence as
the traditional type of flow, in which information from one source pre-
cedes information from the other source in a distinct temporal order. The
internet and CMC subvert previous patterns with regard to the sequence
of communication flows among sources.
How may the confluence and juxtaposition of multiple sources of social
influence be understood? How may we conduct research on how social
influence transpires under various conditions where online peer discus-
sions co-appear with institutionally authored messages or other messages
that bear the conventional characteristics of mass media? We wish to sug-
gest that one of the most fruitful approaches to understanding new tech-
nology may be through consideration of the multiple and simultaneous
social influence agents embodied in the channels that these technologies
make salient. We suggest that any given source appearing in a contempo-
rary interface may be characterized using conventional concepts of source
classification: a product vendor may be considered to be a persuasive
speaker or advertiser. A YouTube viewer who posts a comment about
a video may be, to another viewer, an anonymous peer; a chat partner
may, depending on initially apparent characteristics such as pseudonym
or language style, be an attractive potential relational partner. By invok-
ing conventional labels for online sources, a wealth of traditional theories
and research is rendered potentially applicable with which to understand
the dynamics of that source’s possible influence mechanisms and its com-
parative persuasive potency with respect to co-present online sources of
other types.
• Self-presentation information
– interactional
– profile and/or web
• Third-party information
– your peers
– target’s peers
• System information
– computational
– sociometric
Figure 9.1 A typology of online influence sources
We will illustrate this approach by discussing some exemplar studies,

in which the characterization of sources and their dynamics may be said
to align with various classifications, as depicted in Figure 9.1.
Self-presentation: interactive or web-based

Earlier forms of internet communication provided a great degree of
control over others’ impressions to the individual making the self-
presentation. There is abundant research on how individuals form
impressions, develop acquaintanceships, and foster relationships via
interactive online communication, via email, discussion forums, chats,
and games (see, for review, Walther and Ramirez, 2009). When the basis
of impressions resides in the verbal communication that individuals gen-
erate, it is safe to say that those individuals more or less control the kinds
of impressions their voluntary verbal behavior arouses. Similarly, personal
and institutional web pages, whether they showed aspects of individuals
and groups (see Miller, 1995) or were designed to promote ecommerce
(see, e.g., Metzger, 2006; Nielsen, 1999), left control over content to the
designer. They were changeable but not dynamic (Papacharissi, 2002).
Feedback to a website’s content was made through other channels –

primarily email – if at all. The traditional web was a one-to-many
medium, and in that respect was similar to other mass communication
channels (Trenholm, 1999).
Your peers
In contrast to earlier internet systems, newer systems facilitate viewing
one source’s message while communicating with others about that mes-
sage simultaneously or at some later time. A most rudimentary form of
simultaneous sources is seen in a study in which users observed one of
several video-based public service announcements (PSAs) focusing on
the risks of marijuana use, and a real-time online chat among adoles-
cent peers. David et al. (2006) showed that online group interaction after
weak PSA exposure led to more pro-marijuana attitudes and beliefs than
those in the no-chat conditions. A sample of seventh and twelfth grade
students was assigned to four treatments crossing strong versus weak
PSAs with chat versus no-chat conditions, in groups of 10–20 at a time,
with participants using pseudonymous nicknames when they discussed
the PSAs. David et al. proposed that high sensation seekers were likely to
process the PSA messages in a biased manner. These individuals dom-
inated the online discussions, eclipsing others who might have favored
the PSA’s messages but who remained relatively silent. As a result, the
outspoken participants influenced others negatively with respect to the
PSA’s intended effect on marijuana attitudes. This study demonstrates
potent effects of online chat, but did not examine whether online discus-
sions offer dynamics which differ from those potentially garnered from
face-to-face discussions.
Other research on social discussion of PSAs has reached alternative
conclusions, but these studies employed face-to-face discussion rather
than online chat. Kelly and Edwards (1992) assigned female college stu-
dents to several groups, some who observed anti-drug PSAs without
discussion and others who observed the PSAs and engaged in discus-
sion afterwards. Results were mixed overall, but the discussion of PSAs
had a significant positive effect on some attitudinal outcomes. Warren
et al. (2006) also compared the utility of classroom videos on adoles-
cents’ substance use rates, alone versus with accompanying face-to-face
discussions. Only with discussion were videos effective in reducing drug
use in that sample. Comparing these results to those of David et al.
(2006), there appear to be differences in the effects of online versus
offline discussion of anti-drug PSAs.
Although David et al. (2006) did not consider online chats to provide
anything other than a methodological convenience for the capture of ado-
lescents’ discussions, there is reason to believe that CMC exerted some
effect. The research on social influence in online settings under the aegis
of the social identity and deindividuation effects (SIDE) model of CMC
(Reicher et al., 1995) sheds some light on the issue. Several studies offer
compelling evidence that short-term anonymous online chats bestow
extraordinary pressure on participants to conform to normative posi-
tions in group discussions (Sassenberg and Boos, 2003; see, for review,
Postmes et al., 1999), and that these dynamics are diluted in face-to-face
settings. Thus, effects of CMC in the discussion of PSAs or other media
messages should be expected to differ from offline discussions. David
et al. (2006) did note that the older and more influential teens were
generally considered to have higher social status than younger ones and
more likely to have had experience with marijuana. It is just such social
identification dynamics that should lead to more pronounced effects in
CMC than face-to-face interaction. Social identification and peer group
influence in CMC should be a useful element in explaining a variety of
influence effects in the new technological landscape, as we will illustrate
below.
A recent experiment examined the juxtaposition of anti-marijuana PSA
videos on a YouTube page and the comments ostensibly left by other
viewers appended to the page (Walther et al., 2010). Actual comments
that other viewers had posted to other YouTube PSA pages were tran-
scribed, and classified by researchers as being positive with respect to the
PSA (with regard to its features or in their anti-marijuana statements) or
negative with respect to the PSA (or pro-marijuana stance). A number of
stimulus YouTube pages were created as follows. Four existing PSAs were
uploaded to YouTube. A set of positive statements was appended under
each of the PSA videos, or a set of negative statements was appended
to each video as well. Subjects then viewed one of these eight stimulus
pages – the video and the comments – and completed assessments of the
PSA, measures of social identification with the anonymous authors of the
comments they had read, and scales assessing their perceptions of mar-
ijuana’s risk of harm. The results showed no differences on observers’
responses among the four videos themselves. The positive versus nega-
tive comments, however, significantly affected participants’ evaluations
of the PSAs: subjects who viewed the positive comments rated the PSA
more positively, whereas those who viewed the negative comments rated
the PSAs as poorer. The nature of the comments did not have any
direct effect on the subjects’ marijuana risk attitudes. However, the sub-
jects’ level of identification with the ostensible YouTube message posters
significantly mediated the effect of comments on the participants’ mari-

juana perceptions: The more that subjects identified with the anonymous
peers who appeared to have posted the comments, the more their per-
ceptions of marijuana’s adverse effects aligned with the nature of the
YouTube users’ comments.
Other Web 2.0 sites (O’Reilly, 2005) facilitate interactivity and the
co-creation of content by both web page authors and website visitors.
Casting Web 2.0 as an interface that presents multiple sources of influ-
ence demands that we explore whether and how peers’ (users’) additions
to web pages affect other users’ perceptions of the original author’s mass
media message. Several studies have begun to examine these kinds of
effects.
Such dynamics are clear in online recommender or reputation systems:
tools explicitly designed to display peers’ evaluations of various targets.
Their foci range from product reviews and vendor reviews to professor
reviews. In terms of vendor reviews, Resnick et al. (2000) established that
the quality of one’s ratings as a seller on eBay – ostensibly generated by a
prospective buyer’s peers – renders a demonstrable monetary influence
on the prices one is able to garner for the goods one sells.
Edwards et al. (2007) experimentally examined the impact of online
peer reviews of college faculty in RateMyProfessor.com on students’
perceptions of faculty. Edwards et al. proposed that online reviews are
believed to be authored by individuals similar to the receiver. After
reviewing contrived positive peer reviews for a professor on RateMyPro-
fessor.com, and watching a video showing a sample of the professor’s
lecture, students rated the instructor more attractive and credible. Yet
when students read experimental negative peer evaluations, they rated
the instructor as less attractive and less credible, despite watching the
identical lecture video. This research found similar results with respect to
attitudes toward course material and learning. Edwards et al. concluded
that the interactive web has the ability to manipulate offline beliefs and
actions by affecting students’ perceptions of credibility and attractiveness,
their affective learning, and state motivation in the educational process.
Reliance on online user-generated recommendation systems has
become a normal strategy by which prospective shoppers, healthcare
users, and hobbyists evaluate the credibility of online sellers or service
providers, according to research by Metzger et al. (2010). A series of focus
group discussions uniformly indicated that internet users frequently rely
on “tools such as feedback systems, testimonials, and reputation sys-
tems as ways to help them make credibility evaluations. . . . Many partici-
pants indicated that they looked at the number of testimonials or reviews
available online, paid attention to the proportion of negative to positive
reviews, or relied more heavily on negative versus positive reviews”

(p. 420).
The influence of web-based social comments on perceptions of individ-
uals extends beyond the inspection of recommender systems, and even
beyond the deliberate consideration of others’ comments. Peers’ online
comments can also influence readers’ attitudes and perceptions about
the news. In Lee et al.’s (2009) experiment, undergraduates viewed online
news stories addressing teacher compensation packages. Alongside the
stories, peers’ comments appeared either to agree or disagree with the
actions that the news story presented. Those who read comments oppos-
ing the issues rated the story more negatively. In addition to affecting their
own attitudes regarding the news, the online comments affected readers’
perceptions of public sentiment about the teacher compensation issue:
participants who read other ostensible readers’ comments perceived that
public sentiment about teacher compensation packages was more con-
gruent with the direction of attitudes appearing in the posted comments.
Taken together, these results indicate that online peers’ commentaries
influence not only individuals’ attitudes regarding the specific target of
others’ comments but also individuals’ perceptions of the attitudes of the
general online community.
A target’s peers
The effect of third-party comments also extends to perceptions of indi-
viduals who created online profiles in social networking systems. For
example, Facebook, a social networking site, allows users to place com-
ments on their friend’s “wall,” thereby co-creating the friend’s home page
(Levy, 2007). To be permitted to post such comments, a prospective
commenter must be a “friend” of the profile owner. While the meaning
of friendship is less clear in social network sites than offline, one thing
is certain: to be accepted as a friend grants permission to that friend
to write comments on an individual’s site, which others may see, who
recognize that such permission has been extended.
Walther et al. (2008) found that the content of friends’ postings on pro-
file owners’ “walls” on Facebook affects perceptions of profile owners’
credibility and attractiveness. The physical appearance of one’s friends,
as shown in those wall postings, affects the perceived physical appearance
of the profile owner, as well. Additional research shows that when there is
a discrepancy between a Facebook profile owner’s self-disclosed extraver-
sion and perceived attractiveness, and the imputation of those character-
istics implied by one’s peers via their wall postings, others’ comments
override the profile owner’s claims (Walther et al., 2009).
System information
Online systems aggregate information about users’ online behaviors, and
some systems provide aggregation coefficients to users. Since these aggre-
gations are the byproduct of users’ natural online behaviors, they may
have great credence when others see them. They appear unmanipulated,
spontaneous, and therefore very genuine. They can reflect choices or
they can reflect sociometric patterns.
Aggregated choice data reflect what other people are doing, and can
be interpreted as what users prefer. On CNET’s download.com site, for
example, coefficients display how frequently various software programs
that have the same general niche and function have been downloaded.
One infers that these download counts constitute tacit endorsements by
the community of users at large. Likewise, the websites of many profes-
sional academic journals offer information about the most downloaded
articles in the journal. Like citation counts, download frequencies of one’s
publications have recently appeared as an indicator of the impact of those
publications in tenure and promotion cases. Unlike citation counts, how-
ever, no affirmative effort beyond the click of a mouse is truly signaled
by these data. On the web-based DVD-by-mail system, Netflix, one can
see what the most popular movies are (in terms of the number of DVD
requests or streamed viewings), generally or by one’s geographical locale.
Netflix offers users the opportunity to gather information automatically
about what movies chosen friends have rented.
Other systems reflect who knows whom and how many connections
there are, i.e., the sociometric information about users. An aggrega-
tion of such a nature as this has been found to affect perceptions about
participants. Although popularity seems to be linearly related to social
attractiveness offline, Tong et al. (2008) found this not to be the case
when it comes to the number of friends one appears to have on Face-
book. The definition of “friend” is stretched rather thin on Facebook,
where the 250–275 average number of friends an individual specifies and
links with (Vanden Boogart, 2006; Walther et al., 2008) exceeds by far
the 10–20 close relationships people tend to sustain in traditional rela-
tionships (Parks, 2007). Nevertheless, there are apparently upper limits
to the credulity of extremely large numbers. Tong et al. (2008) experi-
mentally varied a male and a female Facebook profile only with respect to
the number of friends that the profile owner appeared to have: 102, 302,
502, 702, or 902. Each subject rated one of the ten profiles only, and
their ratings demonstrated a curvilinear relationship between number
of friends the profile owner had and the profile owner’s social attrac-
tiveness (with the drop-off occurring above 302). It appears that “too
much of a good thing” can exist, when profile owners seem to gather
friends indiscriminately or appear to spend too much time grooming
their online friend list, although the study found that subjects who rated
these profiles were generally unaware of the aspect of the profile that
led to their attractiveness assessments. Neither were social attractiveness
ratings a function of deviation from subjects’ own number of Facebook
friends.
To summarize, one important avenue of research for the convergence
of sources that new technology promotes will be to understand the various
avenues and interactions of social influence agents who co-appear (or are
closely within clicking reach) in Web 2.0 interfaces. Another potentially
important line of research goes beyond the impact of the overwhelming
presence of what friends and peers think and do on passive social influ-
ence on receivers. The dynamics we have considered so far have focused
on how individuals passively use the social information made manifest
by participative social technologies, in terms of how such information
shapes receivers’ perceptions and decisions.
If individuals come to guide their own media information seeking and
information processing in order to attempt to satisfy other social goals
through subsequent or simultaneous interactions with social partners,
convergent social technologies make possible a separate set of dynamics.
For example, do friends and family members watch broadcasted political
debates for the express purpose of gathering talking points with which
to deride certain parties’ candidates in interpersonal conversations with
relational partners? If so, do these motivations affect attention to and
processing of candidates’ messages? Other research on traditional com-
munication sets the stage for a contemporary re-examination of just such
possibilities.
“Communicatory utility” in media information seeking

The previous discussion has focused on the multiplicity of information
sources that becomes apparent to users of contemporary media, and how
the relationship of peers or interpersonal partners interplays with overt
advocacies in affecting viewers’ perceptions. The following discussion
focuses on a variation of the theme: What prompts individuals to go
seeking information, for relationally based purposes, and the effect that
a relationally motivated information search may ultimately have on the
seeker. In exploring this phenomenon, we again focus on dramatic shifts
in the way information is stored and searched online, compared with
traditional media, and how users’ enhanced capabilities are a part of a
long and complex causal loop. The loop starts with simple interpersonal
inclinations. It works through the manner in which people can seek infor-
mation online, and it ends up, hypothetically, altering the information
seeker’s perceptions and probably without the seeker’s awareness of what
happened. We begin this exploration by reviewing conventional views
of information seeking, and the alternative approach suggested by the
communicatory utility concept.
The predominant view of the use of mass media and interpersonal
encounters is the two-step flow: individuals garner information from the
media which they then elaborate in interpersonal encounters, to under-
stand the issues that the media discuss. The framework suggests that
people’s interest in issues drives their information seeking, and interper-
sonal discussions are subservient to the goal of information gathering
and decision making. In contrast, Atkin (1972) demonstrated how inter-
personal motivations drive mass media information seeking in order to
fulfill relationship goals. Atkin (1973) defined behavioral adaptation as
one of the primary motivations to seek information: because of an indi-
vidual’s “need [of] information that is useful for directing . . . anticipated
behavior” (p. 217), people garner information from mass media when
they anticipated future communication with others about some topic. As
such, while information garnered from mass media sources may provide
its consumers with matter related to the topic, it also provides communi-
catory utility – awareness about a topic about which the individual expects
to interact – with respect to further conversations.
In establishing these constructs, Atkin (1972) analyzed survey data
that revealed an association between the number of conversations people
had with others about the news and the number of news sources to which
one was exposed. Atkin also found a significant association between the
degree to which individuals discussed an ongoing presidential campaign
with their family and friends and the degree to which they sought infor-
mation about that campaign, even after controlling for individuals’ level
of interest in the campaign (as well as education level and socioeco-
nomic status of participants). In other words, even when people were not
interested in the presidential campaign, they sought information about
it because they knew they would be called upon to have interpersonal
discussions about it. To further establish the effect, Atkin conducted an
original experiment in which he led subjects to different levels of expected
future interaction on various news topics of a local or national relevance.
Expected future communication about a topic significantly predicted the
extent to which participants reported information seeking on that partic-
ular topic. Similar findings are reported by Wenner (1976), who found
that some people who watched television did so because it provided a
vehicle for conversation, and Lull (1980), who found that media were
often used relationally to facilitate interpersonal communication. Similar

effects have been found in more recent studies as well (e.g., Southwell
and Torres, 2006). In short, one drive to employ mass media informa-
tion is because of prospective discussion about it among interpersonal
acquaintances.
Atkin’s (1972) notion of communicatory utility is intriguing on several
counts. Clearly, it offers another insight into the merger of mass and
interpersonal events, but it connects the utilization of mass communi-
cation to a superordinate interpersonal functionality. It is intriguing in
terms of the questions it raises with respect to the availability of mass
and interpersonal sources in the current technological landscape: Do
individuals peruse electronic mass media, as well as websites or recom-
mendation systems online, in order to fuel discussions with friends? Do
these discussions precede or co-occur with the perusal of information
sources, rather than follow them the next day at lunch? That is, does a
question (or an anticipated question) in an online chat with a friend or
friends prompt an information search in situ? All of these variations are
germane to the notion of communicatory utility online, and they raise
information processing questions that pertain to the timing and specifi-
ability of information sought when interpersonal discussion and media
searching can take place contemporaneously.
Communicatory utility is a concept that helps explain an example
offered above: individuals might watch a political debate not in order to
gather information with which to make a voting decision but rather to
have ammunition with which to derogate some candidates. Yet Atkin’s
original formulation of the utility construct offered little in the way of
what kinds of interpersonal goals might be served by sampling media,
other than to be able to hold one’s own conversationally. By expand-
ing the range of interpersonal goals one may consider, the potential of
communicatory utility can go beyond helping us understand media con-
sumption, to help illuminate issues of media information processing.
We posit that the specific interpersonal goal(s) that prompt an individ-
ual’s media consumption shape attention to variations in the content and
features of the topical information one consumes, affecting its interpreta-
tion and recall. For instance, collectively derogating political candidates
or office-holders may be an activity that relational partners use to rein-
force the similarity of their attitudes. This, of course, is not restricted
to online news and online chats, but may be a general-purpose cross-
media communication function. As such, one may not watch a debate or
speech with an open mind in an effort to make political decisions; rather,
one may watch for the illogical assertions and dumb mistakes a disliked
speaker utters.
These notions raise the question: Is purposive sampling of mass media

information biased by specific interpersonal goals? If so, how? How does
biased sampling affect attention, repetition, inference, and retention?
Goals may vary in any number of dimensions with respect to instru-
mental, identity, or relational issues (Clark and Delia, 1979; Graham
et al., 1980) in the service of needs for inclusion, affection, and/or con-
trol (Schutz, 1966). The goals of an online chat may include the desire
to impress a conversational partner. This could take the form of a desire
to maintain status, as may have been the case in the adolescent chats
observed by David et al. (2006), consistent with Heider’s (1958) balance
theory.
Alternatively, interpersonal goals may reflect a desire to express attitu-
dinal agreement and convey interpersonal similarity in order to impress
a prospective relationship partner. If the expression of one’s attitude
becomes a strategy subordinated to a goal of expressing solidarity with
another person, one’s sampling of media messages is likely to be exercised
in a manner which allows one to express the socially utilitarian attitude.
Thus, when one pursues relational goals, they may focus the nature
of one’s media sampling and the potential counter-attitudinal advocacy
one generates. In this way relational goals affect the attention, selection,
interpretation, and retention of media information.
The currency of this proposition is that information seeking and pro-
cessing may be different in traditional environments, where media expo-
sure and interpersonal discussion are separated by some interval of time,
compared with the new media environment in which mass and interper-
sonal channels may be sampled (and re-sampled) simultaneously. Even in
offline group discussions, communicators share or withhold information
in a biased manner due to the social motives they bring to discussions,
such as maintaining good relations, obviating conflict, or gaining status;
validation from others further biases information sharing (see Witten-
baum et al., 2004). Computer-mediated communication may exacerbate
this tendency. CMC has particular dynamic properties that facilitate
selective self-presentation in the pursuit of relational goals, facilitated
by unique characteristics of the channel and the context in which it is
deployed (Walther, 1996). Studies show that CMC allows users fluidly
to adapt one’s self-presentation to one’s expectations or observations of
a conversational partner in order to facilitate impressions and positive
interactions, in both asynchronous statements (e.g., Thompson et al.,
2001; Walther, 2007) and adaptive synchronous interactions (e.g., Her-
ring and Martinson, 2004). Web users are well aware of the impressions
they construct in the pursuit of relationships, and consider carefully the
balance between honest disclosure versus socially desirable distortion in
selecting communication strategies to attract others online (Gibbs et al.,

2006). For these reasons it is important to improve understanding of how
these internet-magnified motivations affect message processing.
One recent study (Walther et al., 2010) provides a glimpse of these
potentials. Subjects came to a lab and were assigned to pairs, but with
each member in different rooms, each with its own computer. Subjects
were asked to write down their names and to rank and rate their favorite
hamburger chains in the area. Why hamburgers? It was a topic about
which everyone has some opinion, but in most cases not a strong enough
opinion that social forces could not change it. Researchers exchanged
the names and rankings between partners, and gave them ten minutes
before they started a CMC chat with one another. At that point one
partner in each pair was given an affinity or disaffinity request: to try to
get the other person to like him or her and desire further contact, or to
get the other person to dislike him or her and want to discontinue any
further interaction. Subjects were not told how to enact these behaviors,
only to try to do so. Consistent with hypotheses, during that ten-minute
waiting period, those subjects who had explicit relational goals looked
up information they could use in their conversations more often than
did the unprimed subjects. Those with the relational goals looked up
their partners’ favorite (and least favorite) hamburger chains more fre-
quently, and some of those with the disaffiliative goal even searched for
advice on how to act mean online. When the chats commenced, those
who had relational goals generated more arguments and expressed more
agreements or disagreements than did the naı̈ve subjects, in a manner
consistent with their goals. Finally, these arguments had an effect on
those who made them. All subjects rated the hamburger chains a second
time, and analyses of post-chat ratings revealed that those who argued
positions consistent with their naı̈ve partners’ burger rankings actually
came to like their partners’ favorite burger more than they had prior to
the chat; those who argued in such a way as to distance their partner –
by denigrating the partner’s favorite burgers – ended up liking those very
burgers less as a result. These results offer preliminary support for the
causal loop we suggested between relational goals, information seeking,
online discussion, and perceptions.
In conclusion, we reiterate a new perspective on the merger of vari-
ous communication processes in the common interface that some new
communication technologies provide. The first analytic keystone is to
recognize that new interfaces bring into proximity or simultaneity infor-
mation from several types of sources. Analysis can fruitfully proceed by
identifying and classifying sources such as institutional, interpersonal,
and/or peer, and to assess the sources’ basis of influence on observers
by bringing to bear extant theories of social influence. A second analytic

keystone is the recognition not only that interpersonal contacts moti-
vate media information seeking but that an expanded range of particular
interpersonal goals may be found to affect information processing in
potentially different ways; different relational motivations such as sta-
tus seeking, maintenance, or relationship initiation may bias informa-
tion sampling from various media and affect the ultimate interpretations
derived from them. These dynamics may be especially potent when con-
versations guide media consumption simultaneously, as the internet not
only allows but promotes.
References
Atkin, C. K. (1972). Anticipated communication and mass media information-
seeking. Public Opinion Quarterly, 36, 188–199.
(1973). Instrumental utilities and information-seeking. In P. Clarke (Ed.), New
models for communication research (pp. 205–242). Beverly Hills, CA: Sage.
Berger, C. R., and Chaffee, S. H. (1988). On bridging the communication gap.
Human Communication Research, 15, 311–318.
(1989). Levels of analysis: An introduction. In C. R. Berger and S. H. Chaffee
(Eds.), Handbook of communication science (pp. 143–145). Newbury Park,
CA: Sage.
Caplan, S. E. (2001). Challenging the mass-interpersonal communication
dichotomy: Are we witnessing the emergence of an entirely new com-
munication system? Electronic Journal of Communication, 11(1). Retrieved
26 October 2007 from www.cios.org/EJCPUBLIC/011/1/01114.HTML
Cappella, J. N. (1989). Interpersonal communication: Definitions and funda-
mental questions. In C. R. Berger and S. H. Chaffee (Eds.), Handbook of
communication science (pp. 184–239). Newbury Park, CA: Sage.
Chaffee, S. H. (1986). Mass media and interpersonal channels: Competitive,
convergent, or complementary? In G. Gumpert and R. Cathcart (Eds.),
Inter/media: Interpersonal communication in a media world (3rd edn., pp. 62–
80). New York: Oxford University Press.
Clark, R. A., and Delia, J. C. (1979). Topoi and rhetorical competence. Quarterly
Journal of Speech, 65, 187–206.
David, C., Cappella, J. N., and Fishbein, M. (2006). The social diffusion of
influence among adolescents: Group interaction in a chat room environment
about antidrug advertisements. Communication Theory, 16, 118–140.
Edwards, C., Edwards, A., Qing, Q., and Wahl, S. (2007). The influence of
computer-mediated word-of-mouth communication on student perceptions
of instructors and attitudes toward learning course content. Communication
Education, 53, 255–277.
Gibbs, J. L., Ellison, N. B., and Heino, R. D. (2006). Self-presentation in
online personals: The role of anticipated future interaction, self-disclosure,
and perceived success in internet dating. Communication Research, 33, 152–
177.
Glynn, C. J., and Park, E. (1997). Reference groups, opinion intensity, and
public opinion expression. International Journal of Public Opinion Research,
9, 213–232.
Graham, J., Argyle, M., and Furnham, A. (1980).The goal structure of situations.
European Journal of Social Psychology, 10, 345–366.
Gumpert, G., and Cathcart, R. (Eds.) (1986). Inter/media: Interpersonal commu-
nication in a media world (3rd edn.). New York: Oxford University Press.
Hargittai, E. (2002). Second-level digital divide: Differences in peo-
ple’s online skills. First Monday, 7(4). Retrieved 28 January 2010
from http://firstmonday.org/htbin/cgiwrap/bin/ojs/index.php/fm/article/
viewArticle/942/864#h4
Heider, F. (1958). The psychology of interpersonal relations. New York: Wiley.
Herring, S. C., and Martinson, A. (2004). Assessing gender authenticity in
computer-mediated language use: Evidence from an identity game. Jour-
nal of Language and Social Psychology, 23, 424–446.
Katz, E. (1957). The two-step flow of communication: An up-to-date report on
an hypothesis. Public Opinion Quarterly, 21, 61–78.
Katz, E., and Lazarsfeld, P. (1955). Personal influence. New York: The Free Press.
Kelly, K., and Edwards, R. (1992). Observations: Does discussion of advertising
transform its effects? Yes . . . sometimes – A case among college students and
their response to anti-drug advertising. Journal of Advertising Research, 32(4),
79.
Lazarsfeld, P., Berelson, B. R., and Gaudet, H. (1944). The people’s choice. New
York/London: Columbia University Press.
Lee, E.-J., Jang, J.-W., and Kim, M.-J. (2009). Interpersonal interactivity in online
journalism: What do readers’ comments on internet news sites tell us? Paper pre-
sented at the annual meeting of the National Communication Association,
Chicago, November.
Levy, S. (2007). Facebook grows up. Newsweek, 27 August, pp. 40–46.
Lull, J. (1980). The social uses of television. Human Communication Research, 6,
197–209.
Metzger, M. J. (2006). Effects of site, vendor, and consumer characteris-
tics on web site trust and disclosure. Communication Research, 33, 155–
179.
Metzger, M. J., Flanagin, A. J., and Medders, R. B. (2010). Social and heuristic
approaches to credibility evaluation online. Journal of Communication, 60,
413–439.
Metzger, M. J., Flanagin, A. J., and Zwarun, L. (2003). College student web use,
perceptions of information credibility, and verification behavior. Computers
in Education, 41, 271–290.
Miller, H. (1995). The presentation of self in electronic life: Goffman on the internet.
Paper presented at Embodied Knowledge and Virtual Space conference,
Goldsmiths’ College, University of London, London, June.
Newhagen, J. E., and Rafaeli, S. (1996). Why communication researchers
should study the internet: A dialogue. Journal of Communication, 46, 4–
13.
Nielsen, J. (1999). Designing web usability: The practice of simplicity. Thousand
Oaks, CA: New Riders Publishing.
O’Reilly, T. (2005). What is Web 2.0: Design patterns and business models for the next
generation of software. Retrieved 25 September 2007 from www.oreillynet.
com/pub/a/oreilly/tim/news/2005/09/30/what-is-web-20.html
O’Sullivan, P. B. (1999). Bridging mass and interpersonal communication: Syn-
thesis scholarship in HCR. Human Communication Research, 25, 569–588.
(2005). Masspersonal communication: Rethinking the mass interpersonal divide.
Paper presented at the annual meeting of the International Communication
Association, New York, May.
Papacharissi, Z. (2002). The self online: The utility of personal home pages.
Journal of Broadcasting & Electronic Media, 46, 346–368.
Papacharissi, Z., and Rubin, A. M. (2000). Predictors of internet use. Journal of
Broadcasting & Electronic Media, 44, 175–196.
Parks, M. R. (2007). Personal networks and personal relationships. Mahwah, NJ:
Lawrence Erlbaum Associates.
Pingree, S., Wiemann, J. M., and Hawkins, R. P. (1988). Editor’s introduction:
Toward conceptual synthesis. In R. P. Hawkins, J. M. Wiemann, and S. Pin-
gree (Eds.), Advancing communication science: Merging mass and interpersonal
processes (pp. 7–17). Newbury Park, CA: Sage.
Postmes, T., Spears, R., and Lea, M. (1999). Social identity, normative content,
and “deindividuation” in computer-mediated groups. In N. Ellemers, R.
Spears, and B. Doosje (Eds.), Social identity: Context, commitment, content
(pp. 164–183). Oxford: Blackwell.
Reardon, K. K., and Rogers, E. M. (1988). Interpersonal versus mass media
communication: A false dichotomy. Human Communication Research, 15,
284–303.
Reicher, S., Spears, R., and Postmes, T. (1995). A social identity model of
198.
Resnick, P., Zeckhauser, R., Friedman, E., and Kuwabara, K. (2000). Reputation
systems. Communications of the ACM, 43(12), 45–48.
Rogers, E. M., and Shoemaker, F. F. (1971). Communication of innovations. New
York: Free Press.
Sassenberg, K., and Boos, M. (2003). Attitude change in computer-mediated
communication: Effects of anonymity and category norms. Group Processes
& Intergroup Relations, 6, 405–422.
Schutz, W. C. (1966). The interpersonal underworld. Palo Alto, CA: Science and
Behavior Books.
Southwell, B. G., and Torres, A. (2006). Connecting interpersonal and mass
communication: Science news exposure, perceived ability to understand
science, and conversation. Communication Monographs, 73, 334–350.
Sundar, S. S., and Nass, C. (2001). Conceptualizing sources in online news.
Journal of Communication, 51, 52–72.
Thompson, R., Murachver, T., and Green, J. (2001). Where is the gender in
gendered language? Psychological Science, 12, 171–175.
Tong, S. T., Van Der Heide, B., Langwell, L., and Walther, J. B. (2008). Too
much of a good thing? The relationship between number of friends and
interpersonal impressions on Facebook. Journal of Computer-Mediated Com-

munication, 13, 531–549.
Trenholm, S. (1999). Thinking through communication: An introduction to the study
of human communication. Needham Heights, MA: Allyn and Bacon.
Vanden Boogart, M. R. (2006). Uncovering the social impact of Facebook on a college
campus. Unpublished master’s thesis, Kansas State University, Manhattan,
Kansas. Retrieved 5 July 2007 from http://krex.k-state.edu/dspace/bitstream/
2097/181/4/MatthewVandenBoogart2006.pdf
personal, and hyperpersonal interaction. Communication Research, 23, 3–43.
(2007). Selective self-presentation in computer-mediated communication:
Hyperpersonal dimensions of technology, language, and cognition. Com-
puters in Human Behavior, 23, 2538–2557.
Walther, J. B., Carr, C., Choi, S., DeAndrea, D., Kim, J., Tong, S. T., and
Van Der Heide, B. (2010). Interaction of interpersonal, peer, and media
influence sources online: A research agenda for technology convergence. In
Z. Papacharissi (Ed.), A networked self: Identity, community and culture on
social network sites (pp. 17–38). New York: Routledge.
Walther, J. B., DeAndrea, D., Kim, J., and Anthony, J. (2010). The influence of
online comments on perceptions of anti-marijuana public service announce-
ments on YouTube. Human Communication Research, 36, 469–492.
Walther, J. B., Gay, G., and Hancock, J. T. (2005). How do communication
and technology researchers study the internet? Journal of Communication,
55, 632–657.
Walther, J. B., and Ramirez Jr., A. (2009). New technologies and new directions
in online relating. In S. W. Smith and S. R. Wilson (Eds.), New directions
in interpersonal communication research (pp. 264–284). Newbury Park, CA:
Sage.
Walther, J. B., Van Der Heide, B., Hamel, L., and Shulman, H. (2009). Self-
generated versus other-generated statements and impressions in computer-
mediated communication: A test of warranting theory using Facebook. Com-
munication Research, 36, 229–253.
Walther, J. B., Van Der Heide, B., Kim, S., Westerman, D., and Tong, S. T.
(2008). The role of friends’ behavior on evaluations of individuals’ Facebook
profiles: Are we known by the company we keep? Human Communication
Research, 34, 28–49.
Walther, J. B., Van Der Heide, B., Tong, S. T., Carr, C. T., and Atkin, C. K.
(2010). The effects of interpersonal goals on inadvertent intrapersonal
influence in computer-mediated communication. Human Communication
Research, 36, 323–347.
Warren, J. R., Hecht, M. L., Wagstaff, D. A., Elek, E., Ndiaye, K., Dustman,
P., et al. (2006). Communicating prevention: The effects of the keepin’ it
REAL classroom videotapes and televised PSAs on middle-school students’
substance use. Journal of Applied Communication Research, 34, 209–227.
Wenner, L. A. (1976). Functional analysis of TV viewing for older adults. Journal
of Broadcasting, 20, 77–88.
Wiemann, J. M., Hawkins, R. P., and Pingree, S. (1988). Fragmentation in the

field – and the movement toward integration in communication science.
Human Communication Research, 15, 304–310.
Wittenbaum, G. M., Hollingshead, A. B., and Botero, I. C. (2004). From cooper-
ative to motivated information sharing in groups: Moving beyond the hidden
profile paradigm. Communication Monographs, 71, 286–310.
10 The virtual social world
The continually changing landscape

of social interaction
Garold Stasser, Beth Dietz-Uhler, and
Zachary Birchmeier
When the lights go out in a kid’s room, a familiar place can become many
places – a strange land inhabited by frightening creatures or a playground
of warm and fuzzy animals as playmates. The loss of visual cues permits
the experience of a place to hinge on subtle cues – unfamiliar noises
or the lingering fantasies of a bedtime story. The experience of social
interaction is often shaped by place. When the sense of place is not
well defined by a physical context, we, like the kid when the lights go
out, can experience any one of many places, and our behaviors and
the interpretations of others’ actions change depending on the place we
construct. One of the emerging themes in this volume is that place is
not well defined in computer-mediated interactions. In fact, the physical
place is often irrelevant. The meaning of an e-mail exchange does not
depend on whether the communicators are in their offices or at the
local coffee shop when they send and receive the messages. Whether the
exchange is viewed as a business negotiation or idle chit-chat depends
less on where the communicators are located than on other cues, often
subtle, associated with the interaction. Hence, physical place is often
irrelevant to the definition of situation in the person x situation frame.
Another theme that threads its way through the chapters in this book
is that the identity of the person is malleable. Actors’ identities are often
partly or completely masked by the medium. Moreover, what one reveals
about one’s self is controllable and the opportunity for portraying a
fraudulent self is great. I can become whomever my desires, imagina-
tions, and ambitions dictate. Thus, anonymity has been and continues
to be a variable of interest in the study of computer-mediated interac-
tion. The emphasis in much of the early work was on the dark side of
anonymous interactions – the bad things that happen when one escapes
the constraints of norms and accountability. However, a sampling of
the chapters in this volume reveals that the dark side is only part of
195
196 Garold Stasser, Beth Dietz-Uhler, and Zachary Birchmeier
the picture. For example, the social identity model of deindividuation

effects (SIDE model; Spears et al., this volume) provides a much more
nuanced and dynamic view of what can happen when actors’ identities
are hidden or filtered by electronic media. Spears and his colleagues show
how anonymity can lead to more or less normatively scripted behavior
depending on the other cues in the context.
The third theme is related to the second but the point of view shifts
from the self to the other(s). If the information that others know about self
is sparse and controllable, then the information that self knows about oth-
ers is also likely to be sparse and selective. What one knows about others
may be a mix of misinformation, distortion, and inference – often mostly
inference. Think of a time that you received a message from someone
you had never met. It is likely that, without much effort or thought, you
ascribed a gender and some sense of status, and formed an impression
of the trustworthiness of the message and by implication the messenger.
These inferences may have been based on very little information – the
“signature” at the bottom of the message, the e-mail address, and the
composition and vocabulary in the message. Nonetheless, you probably
did not question your impression of the sender despite knowing that the
information could be faked and that, even if the information were valid,
the inferences could easily be misguided. Moreover, it is likely that a
subsequent exchange of e-mails, if there were a subsequent exchange,
was experienced as meaningful interactions by both you and the sender.
What is remarkable is not the potential for fraud in electronic interactions
but the development of meaningful social exchanges, experienced as real
as face-to-face interactions, in spite of the potential for sending misin-
formation and forming ill-informed impressions (Green and Carpenter,
this volume; McKenna and Bargh, 2000).
Applications of the person by situation perspective typically emphasize
the person – how, and to what degree, personality is expressed across sit-
uations (Mishel, 2004). The situation is the backdrop for understanding
why people behave as they do (Kelley et al., 2003). In a social exchange,
the situation includes both the embedding context and the other actors
in the interaction. This emphasis on the person is natural in psychol-
ogy. However, when both the embedding context and others are largely
constructions of the person, the situation not only is a stage for action
but is also part of the action. Applying the person x situation frame to
computer-mediated interaction requires a fluid and dynamic representa-
tion of the situation.
We will elaborate on each of these points in turn, with frequent refer-
ences to chapters in this volume. First, we will discuss the two sides of
anonymity: when the self is anonymous, others are also likely anonymous.
The virtual social world 197
Second, we will briefly note that the emphasis on anonymity may obscure
another reality of the digital world – namely, electronic interactions often
leave traces of our identities and behavior that make anonymity more
illusory than real. Third, electronically mediated social interactions do
not occur in physical locations and the parties involved in an interaction
are often physically isolated. Fourth, for many types of electronic interac-
tions, time is elastic and social interactions can be more deliberative than
their face-to-face counterparts. Finally, we will explore two perspectives
on understanding computer-mediated social interaction. The first view
is that the virtual social world is an extension of the physical social world
and, as such, is a powerful and flexible tool for studying social behavior.
The second view is that the virtual social world is a new world presenting
unique varieties of social phenomena that are inadequately understood
if we do not carefully distinguish the difference between virtual and live
interactions.
Self anonymity
Anonymity has been prominent in the study of electronic interactions.
Anonymity denotes a state or feeling of being unidentifiable. However,
as several have noted (cf. Green and Carpenter, and Spears et al., this
volume), anonymity is not necessarily a feature of electronic communi-
cations, and personal information is often very accessible in computer-
mediated interactions (e.g., as on Facebook, personal websites, and the
like). Perhaps more important to understanding electronically mediated
social behavior is the degree of control people have over revealing per-
sonal information. One can send an e-mail from a professional or a Hot-
mail account. The professional account typically includes considerable
identifying information whereas the Hotmail account may carry none.
Moreover, one can compose very carefully the content of a personal
website to reveal little or much about oneself.
Heightened control over revealing personal information affords the
opportunity for deception but also permits active self-presentation. As
Green and Carpenter (this volume) note, the line between deception and
self-presentation is fuzzy. Depending on one’s goals, one can reveal some
information while strategically withholding other information without
being blatantly deceptive. For example, consider a person whose hob-
bies are collecting fine wine and hunting. Revealing one of these hobbies
but not the other evokes different impressions depending on which is
revealed. However, presenting only one would not be considered fraudu-
lent by most. Moreover, recent evidence suggests that people’s portrayals
of themselves on online networking sites are reasonably accurate. Back
et al. (2010) assessed the Big Five personality traits of users of online
networking sites in the US and Germany. They asked the respondents
to repeat the Big Five measures, responding the second time in terms
of their ideal self (“describe yourself as you ideally would like to be”).
Independent judges viewed the networking sites and, based on the con-
tent, rated the profile owners on each dimension of the Big Five. Judges’
ratings were more highly correlated with reported actual personality than
idealized personality for all dimensions except neuroticism. Hence, cre-
ators of online networking self profiles were apparently not slanting the
content to convey what they would like to be but to communicate who
they are – at least, in regards to global personality traits such as extraver-
sion, agreeableness, conscientiousness, and openness.
Anonymity of others
There are two sides to anonymity. When self is anonymous, others are
also often anonymous in online interactions (Sassenberg, this volume).
More to the point, electronic media that permit control over disclosing
personal information typically afford the same control to everyone. Put
differently, opportunities to withhold or distort personal information are
often symmetrical. Thus, if others know little about me, it is likely that I
know little about them. If the information that I reveal is highly selected,
then the information that I receive about others is probably also highly
selected.
However, knowing little about each other does not necessarily make the
interaction less social. A key idea embodied in the SIDE model (Spears
et al., this volume) is that we actively construct impressions of others
based on little information, and people can be strongly identified with
others even when they know little about them. If all one knows about
others is that they share one’s interest in wine, their religious and political
views matter little in the moment. In fact, if Bob knows that he shares
one interest with Ann and Charlie, he is inclined to see himself as similar
to them in many other respects.
Abele and Stasser (2008) demonstrated that matching simple choices
in a coordination task leads to a sense of having a lot in common with
others. In their study, two participants choose between simple bets with
equivalent expected values (e.g., 50 percent chance to win $1 or 25 per-
cent chance to win $2). In one version of the task, the payoffs were dou-
bled if the partners chose the same bet (matching), whereas in another
version the payoffs were doubled if they chose different bets (mismatch-
ing). Successful coordination led to increases in perceived similarity and
liking in the matching task but not in the mismatching task. Whereas
this coordination work involved matching in deliberative choices, even

unconscious matching in the form of behavioral mimicry increases rap-
port and liking (Chartrand and Bargh, 1999). If matching on choices of
simple bets or mimicking hand gestures result in feelings of being socially
connected, then it is likely that knowing that others are participating in
the same blog or social networking site sets the stage for having a strong
social identity with others.
In fact, as suggested in the SIDE model, knowing a little may often be
more potent than knowing a lot, particularly when a little information
underscores a common identity and knowing a lot reveals the many ways
in which others are different. Moreover, when we know little, we may be
prone to fill in the gaps by projecting our own characteristics on others
(Abele, this volume). If so, the fact that two people, by forethought
or accident, have “met” in cyberspace may be sufficient to trigger a
sense of shared identity. Hence, when little is known about others, the
default assumption may be that they are like us in many or most respects.
Whether or not one feels a strong affinity to others online may hinge on
subtle cues. Just as learning a shared characteristic may trigger a sense of
social inclusion, information that underscores one’s distinctiveness may
erode the sense of oneness (see, for example, Spears et al.’s work on
gender identity in this volume).
In summary, an important theoretical and research agenda is to under-
stand how we form impressions of others when there is little information
about others. Additionally, how do we factor in the knowledge that what
we know could be strategically selected or distorted? When do we focus
on what is known and fill in the gaps to promote our current objectives
and when do we focus on what is not known and discount what is known
because it could be highly selected? Another important agenda item is
to understand how individual differences affect one’s inclination to see
others as similar to or different from self. Work on social value orien-
tation suggests, for example, that people differ in their inclinations to
expect others to act as they do. People high in pro-self orientations are
motivated by self-interests and expect others to also act in self-interested
ways (Kelley and Stahelski, 1970). Pro-socially oriented individuals are
inclined to act in the interest of the collective and also expect that others
will act to benefit the collective. A difference, however, is that pro-socials
are less confident in their expectations and more likely to revise their
default expectations in the face of disconfirming evidence (van Lange,
1992). An intriguing possibility is that there are more general disposi-
tional differences among people regarding perceived self–other similarity.
We suggested earlier that, in general, people may be inclined to see oth-
ers as being similar to them unless the available information suggests
otherwise. Here we are raising the possibility that this tendency may not
be universal. Some people may start with an assumption that they are
distinctive and require more information to conclude that they are simi-
lar. These types of people would react quite differently to the scarcity of
social information that often characterizes online interactions.
The flip side of anonymity

Taking a different perspective, the flip side of anonymity is public visi-
bility. We may feel anonymous and in control of what we disclose about
ourselves in computer-mediated communications, but personal infor-
mation is more publically available than ever before. We can receive an
e-mail from a stranger and learn more about the sender in five minutes
as amateur internet sleuths than many face-to-face meetings would yield.
We leave hundreds of electronic traces every day of who we are, what we
are doing, and where we have been. Our awareness of the potential public
nature of our selves may constrain us to be honest and circumspect in
our computer-mediated presentations. If so, our virtual selves may be as
accurate as our face-to-face presentations. Whereas anonymity (illusory
or real) may affect how we behave in some circumstances, we suspect that
we are mostly “true to ourselves” when online, regardless of the degree of
felt anonymity. From a social psychological perspective, much variance in
behavior is accounted for by the situation, which includes others and the
embedding context. We have already noted that our impressions of oth-
ers are frequently based on little information. Likewise, the embedding
context may be ill-defined.
Place
Tangible place is typically an important element of situation in face-to-
face interactions. The meaning of communications often depends on
where they occur (Abele, this volume). Similarly, our impressions of
others are informed by place. We expect to encounter different people in
church than in a bar (Cantor et al., 1982). Or, perhaps more to the point,
we expect people to be different in church than in a bar. In electronic
communications, place is not objectively defined. Without the embed-
ding information of physical place, the meaning of an interaction can
depend on something as simple as a label. Abele draws on lessons from
experimental economics where the nature of interdependencies is defined
by the objective rules of a game. Whereas place is not specified, some-
thing as simple as the name of the game can change the meaning of the
interaction and the behavior of the players. Thus, without the contextual
information of a physical place, people readily derive context from other

information such as the name of a chat room site or the domain of an
e-mail address.
The absence of physical place changes the relevance of basic dimen-
sions of social interactions. Hollingshead (this volume) explores the emer-
gence of leadership in virtual groups. As she aptly notes, the traditional
model of group behavior is based on several people interacting in the
same place at the same time. Same place and same time limits the func-
tional group size and often requires coordination of sending and receiving
messages. Moreover, having everyone in the same physical space makes
salient (often visually salient) who is and who is not part of the group.
Nonetheless, she contends that many online groups fit traditional defi-
nitions of a group: interdependent and shared goals, interaction among
members, and a sense of belonging – in short, the members view them-
selves as a group. In face-to-face groups, participation rates are variable
across members, with one or a few members accounting for the lion’s
share of the participation, and the disparity of participation increases with
group size (Bonito and Hollingshead, 1997; Stasser and Vaughan, 1996).
These disparities in participation rates are often attributed to competi-
tion for limited access to airtime. Those with high levels of participation
in face-to-face groups emerge as leaders, at least in the judgment of their
peers. As groups move from “same place, same time” to “any place, any
time,” one might expect a reduction in competition for airtime, fewer sta-
tus markers, and more equality in participation and, thus, less clarity or
consensus regarding who the leaders are. Hollingshead reviewed evidence
from both small teams whose interactions were computer-mediated and
larger internet groups. Tentative conclusions are that those with high lev-
els of participation are likely to be seen as leaders in both face-to-face and
online groups but that online groups are likely to have several emergent
leaders.
The absence of physical location for an interaction also means that par-
ticipants are often physically isolated. Sassenberg (this volume) suggests
that physical isolation increases self awareness. He explores the implica-
tions of isolation and anonymity of self and others for social influence.
As he notes, anonymity of self and others often co-occurs with physical
isolation. Spears et al. (this volume) state that anonymity and physical
isolation are conditions that can result in strategic use of social identities.
In their studies, women were likely to choose masculine avatars when
discussing a masculine topic and feminine avatars when discussing a
feminine topic. Men favored masculine avatars regardless of topic. Thus,
Spears et al. contend that women used subtle cues to counter stereotypical
gender differences in status and influence. Interestingly, however, when
gender identity was made salient immediately prior to the interaction,

women changed their strategy and chose neutral or masculine avatars
when discussing feminine topics and feminine avatars when discussing
masculine topics. Spears et al. discuss possible reasons for this shift in
strategy. Whatever the reasons, the interesting point is that women were
clearly using the information conveyed in an avatar differently depend-
ing on whether their gender was salient in the context. The implication
is that the choice of avatar was deliberate and strategic. Face-to-face
interactions typically do not permit such easy manipulation of socially
meaningful markers, partly because there is not sufficient time to both
plan and implement chameleon-like appearances and behaviors.
Time and deliberation

McKenna and Bargh (2000) suggested that internet communications
differ from face-to-face exchanges in four ways: opportunity to maintain
anonymity, irrelevance of physical distance, absence of visual cues, and
stretching of time. The irrelevance of physical distance is a consequence
of physical place being irrelevant. Another consequence is that the timing
of communication is more under the control of the communicators. Face-
to-face communications occur in the “here and now” whereas electronic
communications occur wherever and whenever. As a result, there is ample
time for deliberation. Abele (this volume) notes the irony in having time
to tune our message but not knowing much about the audience or the
context. She concludes, however, “that controlling or constraining social
information does not prevent actors from engaging in this deliberative
process. Rather, it makes them more industrious in mining what they can
from the scant information available.”
Time to think before responding allows time to tune messages to fit the
(presumed) audience and the objectives in an interaction. The elasticity
of time in electronic communication makes pace a potentially informative
dimension of interpersonal behavior. In face-to-face communications,
there is relatively little variance in the pace of an interaction; one cannot
wait for an hour or a day to respond in a face-to-face discussion. Nonethe-
less, people who respond quickly and participate frequently appear con-
fident and are often more influential. There is some empirical justifi-
cation for these inferences. Confident and high-status people do claim
speaking turns more readily in face-to-face discussions. (For reviews, see
Bonito and Hollingshead, 1997; Stasser and Vaughan, 1996.) The links
among rate of responding, participation, status, and influence in face-to-
face discussion are due in part to the physical presence of participants.
Turn-taking is a necessary feature of effective face-to-face communica-

tion but results in a competition for airtime. Those who are quick to talk
get more airtime, and people often yield in an attempt to gain the floor
to someone of higher status.
Electronic media differ in their demands for quick responses. Phone
conversations, teleconferencing, instant messaging, and electronic meet-
ings often have pacing demands that are similar to face-to-face discus-
sions. In these media, not responding promptly is noticed. Even these
media differ in their pacing demands: people queue for speaking time in
teleconferencing much as they do in face-to-face meetings whereas elec-
tronic meeting software often permits simultaneous communication. In
contrast, one can wait for an hour or a day before responding to an e-mail
or posting to a thread on a newsgroup, and can update one’s home page
profile next month or next year. Thus, the felt need for time to deliberate
is an important consideration in choice of media. For those media that
permit flexibility in pacing, the latency of response can be a cue about the
characteristics of the other. However, the meaning of the cue is inherently
ambiguous. Both fast and slow responses can imply interest or a lack of
interest depending on the topic and the context.
Person in the situation or situation in the person?

Snyder and Ickes (1985) noted that individual differences are more influ-
ential when the situation does not provide clear guides for, or constraints
on, behavior (Mischel, 1977). When social information is sparse and pos-
sibility contrived, and place and time are relative, the situation is largely
a construction of the person. Like the kid in the room when the lights
go out, we make our own virtual world, which seems very real. On the
one hand, computer-mediated interaction is an extension of the actual
world and a laboratory for social psychologists to study how people inter-
act. On the other hand, it provides the opportunity to study how indi-
vidual differences shape the perception of the social environment. We
have been interacting electronically since the advent of the telephone,
but increasingly electronic interactions are dominating our social world.
Thus, understanding how people make sense out of computer-mediated
social interactions is a worthwhile goal for social psychologists. In this
pursuit, the person by situation framework is pushed to a new level.
It is no longer simply a question of how different personalities react
differently to situations; the question is how different personalities con-
struct different situations when there is little information to anchor their
perceptions.
The virtual world as an extension of the face-to-face

social world
In many ways, the online social world resembles the rest of the social
world. We bring the same desires and experiences to interactions whether
they are virtual or face-to-face. An insult stings and praise feels good
whether delivered by an e-mail or across the table during lunch. Our
response may differ depending on the medium and the time that we have
to consider the options, but the immediate emotional experience is much
the same. Wesselmann and Williams (this volume) nicely illustrate how
basic social needs affect our experience in much the same way in both
domains.
They review the effects of ostracism. They conclude that the immediate
effects differ little regardless of whether one is ostracized face-to-face or
electronically. In either case, the immediate reaction to being excluded is
distress and psychological pain. They contend that ostracism is so potent
because it thwarts the fulfillment of four fundamental human needs:
the desire for meaningful social relationships (the need to belong), the
need for positive self-regard (self-esteem), the need to feel in control of
the environment, and the need for a meaningful existence. Reactions
to threats to these basic needs can be pro-social in the form of striving
to gain or regain social acceptance, or anti-social in the form of ver-
bally aggressive and hostile behavior. Whereas the immediate emotional
reaction to ostracism is the same online as it is when others are phys-
ically present, the authors report that subsequent coping behaviors in
electronic interactions are more likely to be anti-social than in face-to-
face situations. For example, when people were ignored in chat rooms,
they reacted with what Wesselmann and Williams call virtual bravado –
inflammatory or disruptive communications. In part, they attribute the
tendency to respond in anti-social ways online to the anonymity of these
interactions. Moreover, they speculate that when others are anonymous,
people may address the need for control, realizing that attempts to bolster
feelings of belonging are not feasible.
It is striking how potent the reactions are to ostracism even when
the “others” in the interaction are represented as crude animated fig-
ures. Social psychology has been criticized for studying behavior in
contrived social environments where information is sparse and manipu-
lated. Whereas these criticisms have some merit, it is worth noting that
online social interaction often resembles the social contexts created in
the social psychology laboratory. For example, studies of social influence
often present statistical information about peers’ attitudes, not unlike
the information one can readily retrieve on the internet about products
and professors. Similarly, studies of impression formation often present

sketchy information about a target, manipulating information about a
critical behavior or trait. Such studies have been criticized because the
information is sparse and not representative of what we often know about
others. However, the similarity between these contrived experimental set-
tings and the virtual social world is striking. Additionally, as the work on
cyber ostracism demonstrates, immediate emotional reactions to virtual
social interactions are often as potent as their live counterparts. What
changes is the subsequent coping with the emotional event. With time
to deliberate, one can mentally simulate the consequences of several
possible responses. One might spontaneously respond to a face-to-face
insult with a self-deprecating joke in order to gain social favor and dispel
the awkwardness of the situation. But, given the luxury of time and the
buffer of physical distance, the same person might respond with a clever
and equally insulting retort in an e-mail exchange or blog posting. The
work on cyber ostracism is a clear demonstration of how studying social
behavior in the virtual world can inform and extend our understanding
of “live” interactions.
The virtual world as the new frontier of social behavior

Asserting that the virtual world is in many ways an extension of face-to-
face social interactions does not deny that many kinds of online behav-
ior are unique in their dynamics and scope. This new social world has
attracted the attention of many scholars, and initial attempts to charac-
terize and understand these new phenomena have often been shaped
by existing ideas about social interaction. As many have noted (cf.
Sassenberg, and Spears et al., this volume), early accounts of the effects
of anonymity in computer-mediated interactions focused on deindividu-
ation as an explanatory concept. These accounts emphasized what hap-
pens when people are not constrained by social norms and focused on
examples of anti-normative communication in anonymous interactions.
McGarty, Lala, and Douglas (this volume) examine two propositions
about internet groups that are good examples of how characterizations of
online behavior are often exaggerated and distorted. The first they call the
anonymous cyberhate view of the internet – the notion that the internet pro-
vides a protected platform for verbal attacks on racial and ethnic groups
and promoting extremely, and even violently, prejudiced behavior. To the
contrary, they found little evidence of promoting violence in the contents
of white-supremacist websites. Rather, the content focused more on jus-
tifying the white-supremacist views and less on attacking other racial and
ethnic groups. While not denying that cyberhate happens, “hate” groups
are seemingly less likely to promote hate on the internet than to use the
internet to solidify their identity, justify their existence, and win converts.
The second proposition that they examine is the global activist view of the
internet – the idea that the internet provides a means of promoting and
instigating social activism in the world community. They compare online
and face-to-face social action groups, noting the similarities in their rea-
sons for existence and their purposes. Nonetheless, the action that they
found on the websites of social action groups was primarily talk. Rather
than a forum for consolidating opinions and planning action, these web-
sites were places for expressing and exchanging ideas. McGarty et al.
used the metaphor of an editorial page in a newspaper to character-
ize these sites. They are places to exchange opinions and ideas but not
to assess and consolidate consensus or to mobilize the participants for
action.
The internet gives groups, whether their goals are evil or good, an
unprecedented ability to broadcast their messages and expand the rolls
of group membership. In short, the internet is a communication tool
that permits groups to access a global audience immediately and con-
tinuously. Walther et al. (this volume) contend that computer-mediated
communication blurs the distinction between mass and interpersonal
communication. Mass communication refers to a message from one to
many whereas interpersonal communication is an exchange of messages
among a few. In mass communication, the recipients typically comprise a
large, often diverse, group, and the recipients are not expected to respond.
In interpersonal communication, the target is typically identified and the
message is tailored to the recipient. Moreover, the recipient is usually
expected to respond. Walther et al. show how these two modes of com-
munication converge on the internet. A dialogue on a blog is available
for the masses to read. An ostensibly personal message posted to a Face-
book site is sent with both the sender and the recipient knowing fully that
it will be read by many. Friends exchange reactions to YouTube videos
and discuss news events as they unfold. Certainly, such melding of mass
and interpersonal communications has occurred before the spread of the
internet, but today they have become commonplace. The confluence of
mass and interpersonal communication is not only permitted but also
encouraged by current technology.
Communicatory utility refers to the use of mass media to serve inter-
personal communication needs. A person may watch a sports event not
because she is interested in the event itself but because she anticipates
discussing it with friends the next day. In the convergence of mass and
interpersonal communications, communicatory utility has much more
immediacy. The transition between mass and interpersonal messages is
immediate and seamless. One can surf the internet for information about
a current event while exchanging messages about it with friends. Or,
one can learn of a media event and immediately pass along the informa-
tion with commentary to a friend. Transitions between consuming mass
media and discussing with friends unfolded over days in the past but now
can occur in minutes – such is the flexing of time in computer-mediated
communication.
Walther et al. also discuss reputation systems in which people post
evaluations of products, events, and even other people (professors, ven-
dors on eBay) for mass consumption. They noted that we prefer optimal
similarity with a source in seeking others’ evaluations. We prefer that
they be like us in all respects but the critical one of experience with the
target of evaluation. When seeking information about products online,
people often seek ratings by other consumers over expert evaluations. If
Joe needs a new refrigerator, he wants to know about the experience of
others, ostensibly like him, with the array of brands and models under
consideration. Again, we see the melding of mass and interpersonal mes-
sages. The personal experience of others, which was often hard to gather
in the past, is now posted online for all to see.
Mapping a changing social landscape

We have identified several themes that emerge throughout the chapters
in this book: anonymity, irrelevance of physical place, physical isolation,
elasticity of time, and opportunity for deliberative and strategic interac-
tions. Underlying these themes is another message: the rapid evolution
of technology. As a result, it is becoming more difficult to describe the
nature of computer-mediated interactions and specify the differences
between virtual and face-to-face interactions. When most online activity
consisted of e-mail and posts to newsgroups or blogs, the emphasis on
anonymity was understandable because of the potential to communicate
without revealing much about one’s identity. But things are changing.
Since 2005, YouTube and Facebook have become popular means of
communication, and many of the social filters that were characteristic of
earlier online communications do not apply to these and other emerg-
ing communication interfaces. The challenge is to understand how these
new communication tools affect social interactions.
Articulation of a coherent and workable model of person by situation
interactionism depends partly on having a good typology of situations
that goes deeper than surface features and is based on psychologically
meaningful dimensions (Mischel, 2004). As new communication tech-
nologies emerge, it would be useful to classify them in a comprehensive
framework. The themes that we have identified suggest two dimensions

of such a framework. First, communication media differ in the degree of
control over the discloser of personal and identifying information (dis-
closer control). It is hard to hide or disguise one’s identity in face-to-face
communications (although not impossible with careful planning), but
easy in e-mail and other online media that permit the use of aliases.
As Walther et al. (this volume) noted, active use of the newer social
networking technologies makes it difficult to shield one’s identity and
protect one’s privacy. Second, some media afford little opportunity for
deliberation and strategic planning whereas others do (permissible pac-
ing). One can ponder a response more easily when e-mailing than when
instant messaging – the “instant” places an expectation of a quick reply.
Interestingly, video-conferencing software is becoming widely available
and falls close to face-to-face interactions on the dimensions of discloser
control and permissible pacing.
A comprehensive typology of the virtual social world would need to
include those characteristics of communication technology and software
that are important in shaping the social interactions. In part, decid-
ing what are socially relevant features of communication technologies
depends on the perceptions of the users. It is not clear, for example,
whether people react differently to text or voice media beyond the fact
that voice media tend to demand immediate responses whereas text
media do not. Also, we have stated that place is either not defined or
irrelevant in much computer-mediated interaction. However, even this
characteristic may be changing. Incorporating global positioning tech-
nology into communication devices may reintroduce a sense of physical
location into virtual communications. Indeed, video-messaging carries a
sense of location. It is an open question whether such place information
is meaningful in the same sense as being physically located in the same
place as in face-to-face communications.
Thus, an important research goal is to develop a more comprehen-
sive way of describing similarities and differences among communica-
tion technologies. Moreover, the similarities and differences should be
socially relevant and psychologically meaningful. Socially relevant and
psychologically meaningful imply that people perceive different types
of interactions as associated with different locations on a dimension.
A fundamental element in social relations is the nature of the interde-
pendencies among the participants, particularly whether the interests
and goals of the parties in an interaction are compatible or conflictual.
For example, do people prefer media with high personal visibility (little
discloser control) and fast turnaround (fast permissible pace) when per-
sonal and collective interests are compatible, but high discloser control
and lenient response deadlines (slow permissible pace) when conflict is

anticipated? Additionally, a useful dimension for distinguishing media
should be associated with perceptions of others. Do people expect to
encounter different types of people depending on the characteristics of
the media? For example, do low discloser control and fast permissible
pace promote impressions of trustworthiness? Finally, different types of
people should prefer different types of media. Do extraverted people pre-
fer high discloser and fast-paced exchanges and do introverts prefer low
discloser and slow-paced interactions?
The foregoing examples and questions suggest the main effects. Differ-
ent media are associated with cooperative and competitive interactions,
give rise to different interpersonal impressions, and are differentially pre-
ferred by different personalities. In the spirit of person by situation inter-
actionism, we should not be surprised to find that these associations are
moderated by each other. For example, a socially confident and verbally
facile person might prefer to resolve conflict in a medium that permits
little control over discloser and forces a fast pace. A shy and verbally slow
person would likely prefer the opposite. For an interaction with cooper-
ative overtones, these differences in media preference may disappear or
even reverse.
The advent of computer-mediated interaction has clearly changed two
things on the social landscape. First, we have many choices of groups
and people with whom to interact. Second, we have many choices
about how we interact with them. Understanding how people navi-
gate this new social landscape is the challenge for behavioral scientists –
even more challenging because the landscape will certainly continue to
change.
References
Abele, S., and Stasser, G. (2008). Coordination success and interpersonal per-
ceptions: Matching versus mismatching. Journal of Personality and Social
Back, M. D., Stopfer, J. M., Vazire, S., Gaddis, S., Schmukle, S. C., Egloff,
B., and Gosling, S. D. (2010). Facebook reveal actual personality, not self-
idealization. Psychological Science, 20, 1–3.
Bonito, J. A., and Hollingshead, A. B. (1997). Participation in small groups.
Communication Yearbook, 20, 226–261.
Cantor, N., Mischel, W., and Schwartz, J. C. (1982). A prototype analysis of
psychological situations. Cognitive Psychology, 14, 45–77.
Chartrand, T. L., and Bargh, J. L. (1999). The chameleon effect: The perception–
behavior link and social interaction. Journal of Personality and Social Psychol-
ogy, 76, 893–910.
Kelley, H. H., Holmes, J. G., Kerr, N. L., Reis, H. T., Rusbult, C. E., and
van Lange, P. A. M. (2003). An atlas of interpersonal situations. New York:
Cambridge University Press.
Kelley, H. H., and Stahelski, A. J. (1970). Social interaction basis of coopera-
tors’ and competitors’ beliefs about others. Journal of Personality and Social
McKenna, K. Y. A., and Bargh, J. A. (2000). Plan 9 from cyberspace: The
implications of the internet for personality and social psychology. Personality
and Social Psychology Review, 4, 57–75.
Mischel, W. (1977). The interaction of person and situation. In D. Magnus-
son and N. S. Endler (Eds.), Personality at the crossroads: Current issues in
interactional psychology (pp. 333–352). Hillsdale, NJ: Lawrence Erlbaum.
(2004). Toward an integrative science of the person. Annual Review of Psychol-
ogy, 55, 1–22.
Snyder, M., and Ickes, W. (1985). Personality and social behavior. In G. Lindzey
and E. Aronson (Eds.), The handbook of social psychology (pp. 883–947). New
York: Random House.
Stasser, G., and Vaughan, S. I. (1996). Models of participation during face-
to-face unstructured discussions. In E. H. Witte and J. H. Davis (Eds.),
Understanding group behavior: Consensual action by small groups (Vol. I, pp.
165–192). Hillsdale, NJ: Erlbaum.
van Lange, P. A. M. (1992). Confidence in expectations: A test of the triangle
hypothesis. European Journal of Personality, 6, 371–379.
Index
access criterion, 176 common bond groups, 68

accountability, 21 common identity groups, 68
action groups, 149 communication technologies, 108, 174
activist groups, 150 communicatory utility, 186, 187, 206
activist identities, 148, 150 compliance, 70
anonymity, 9, 18, 19, 21, 23, 24, 26, 28, computer-mediated communication, 2, 3,
29, 36, 44, 49, 53, 65, 84, 115, 135, 6, 16, 19, 44, 57, 63, 86, 103, 114,
136, 138, 151, 166, 195, 197, 198, 188, 206
201 computer-mediated interaction, 2, 84, 85,
anonymity of others, 66, 67, 72, 75 87, 100, 103, 105, 195, 197, 203
anonymity of the self, 66, 77 computer-supported interaction, i, 1, 11,
anonymous cyberhate view of the internet, 85
145, 151, 205 convergent social technologies, 185
anti-normative behavior, 136 coordination problem, 93
Atkin, C. K., 186, 187 coordination task, 198
attitude change, 73 cross-cultural, 33
avatar, 31, 32, 53, 201 cyber bullying, 138
cyber ostracism, 205
bargaining, 46 Cyberball, 130, 131, 132, 134
Battle of the Sexes, 91 cyberhate, 166
Baumeister, R., 128
behavioral adaptation, 186 David, C., 180
behavioral control, 92 deception, 46, 47, 197
behavioral game theory, 96 beliefs about internet deception, 49
Berger, J., 111 deception online, 47
Big Five personality traits, 198 deviancy, 3
blogs, 147 lie detection, 50
broadcast, 146 lies, 47
lying, 46
Cappella, J. N., 174 motivations for deception, 48
Chaffee, S. H., 176, 177 online deception, 52
chat rooms, 131 online dishonesty, 52
choice dilemmas, 67 reactions to internet deception, 48
chronic private self-awareness, 74 decision timing, 98
citation counts, 184 deindividuation, 18, 19, 21, 136, 138, 151
CNET, 184 deliberation, 104
Cocktail Party World, 131 deliberative action, 85
collaboration, 2 deliberative communication, 86
collective action, 24, 148, 149, 150, 151, deliberative process, 105
156 DePaulo, B., 47, 50
collective norms, 68 depersonalization, 19, 23
211
212 Index
Dictator Game, 99 informal roles, 111

discloser control, 208 leader, 111
disclosure, 43 maintenance roles, 111
disinhibition, 49 self-centered roles, 111
Douglas, K. M., 145 task-oriented roles, 111
dynamic interactionism, 9, 10 group-based “collective assertions,”
dynamic social context cues, 115 159
group-based internet activism, 154
Eagly, A., 111
Edwards, C. E., 182 Heider–Simmel paradigm, 132
Edwards, R., 182
electronic teamwork, 8, 9 Ickes, W., 7
emergence of leadership, 112, 116, 120, identity, 150
201 identifiability, 152
empowerment, 25 identity play, 48, 52
evolution of technology, 207 identity verification, 55, 56
expectation states theory, 111, 117 identity-based pathway, 157
experimental economics, 87, 88, impression formation, 53, 101, 103, 199,
200 205
experimental game theory, 88, 104 individual differences, 71
information processing
Facebook, 183, 184 information access, 178
fate control, 92 information exchange, 72
Fishbein, M., 190 information sources, 176
Free Software, 157 information-seeking, 177, 186, 188
fundamental attribution error, 100 interactionism, 63
interactivity, 175
Gackenbach, J., 2 interdependence, 89, 92
Galimberti, C., 5 internal characteristics, 1
game theory, 89 internet groups, 121
gathering, 146 interpersonal processes
gender, 16, 20, 26 interpersonal communication, 174, 206
gender concealment, 33 interpersonal goals, 187, 188, 190
gender differences, 30 interpersonal inclinations, 186
gender identity, 17, 27, 32 interpersonal influence, 70, 71
gender relations, 27 interpersonal relationships, 45
gender relevance, 27
gender salience, 27, 28, 30 Joinson, A., 3, 4, 154
gender-bending, 30
global activist proposition, 154 Kelly, K., 180
global activist view of the internet, 145, Kiesler, S., 115
206
group awareness, 72 labels, 96
group processes leader consensus, 117, 122
group cohesiveness, 29, 30 leader emergence, 109, 110, 111, 112,
group consensus, 115, 159 114, 116, 119
group identity, 23, 24 Leary, M. R., 128
group leadership, 122 Lee, E.-J., 191
group member participation, 113 Levine, J., 114
group norms, 18, 19, 23, 26 Lewin, K., 1, 7
group polarization, 22, 23 locus of control, 10
group size, 109, 113, 119
group stability, 119 Mantovani, 4
group roles, 110 mass communication, 173, 206
formal roles, 110 masspersonal communication, 175
Index 213
McGarty, C., 145 private self-awareness, 70, 71

McGrath, J., 113 pro-social behavior, 134, 135
McKenna, K., 2, 10, 44 prototypicality, 30
media convergence, 173 PSAs, 181
media usage, 72 pseudo-sequential order, 98
Minimal World, 132, 133 public goods game, 93, 97, 101
Moreland, R., 114 public self-awareness, 77
MoveOn.org, 162, 163, 165 public service announcements (PSAs),
180
Nash equilibrium, 90 public visibility, 200
needs
need for belonging, 128, 135, 204 Qing, Q., 190
need for control, 135
need for meaningful existence, 135 RateMyProfessor.com, 182
need for self-esteem, 135 rational actor approach, 155, 157
Netflix, 184 rationality, 89
newspaper opinion page, 165 Reardon, K. K., 175
norm-based influence, 70 reduced social cues, 136
reflexive control, 92
online relational goals, 189
online communication, 52 relationships on the internet, 128
online extremist groups, 153 reputation systems, 56, 182
online groups, 108, 109, 110 Ridgeway, C., 111
online relationship, 41 Riva, G., 2
online reputation systems, 55 Rogers, E. M., 175
online social interaction, 204
online social world, 204 sanctioning systems, 97, 104
online systems, 184 self, 17
Open Source, 157 self-awareness, 201
opinion-based groups, 148, 149, 150, self-categorization, 19, 20, 23, 29, 30,
160 147
optimal heterophily, 177, 178 self-efficacy, 45
optimal similarity, 207 self-esteem, 204
ostracism, 43, 127, 128, 129, 138, 204 self-focused individuals, 73
cyber-ostracism, 129 self-monitoring, 8, 9, 73
face-to-face ostracism, 129 self-perception, 147
immediate effects of cyber-ostracism, self-presentation, 21, 49, 52, 179, 197
132 signature files, 121
physical ostracism, 129 simultaneous game structures, 98
reflective reactions to ostracism, 134, Snyder, M., 7, 203
137 social attraction, 29
reflexive reaction to ostracism, 132 social capital, 45
social categorization, 29
Pareto efficient solution, 90, 97 social comparison orientation, 78
participation rates, 120, 201 social context, 87
payoff matrices, 94 social creativity strategies, 153
permissible pacing, 208 social cues, 84, 86, 87, 103, 104
person by situation framework, i, 1, 4, 6, 7, social exchange, 196
11, 20, 56, 65, 72, 74, 75, 95, 195, social identity, 5, 17, 18, 64, 65, 67, 68,
196, 203, 207, 209 146, 148, 157, 159, 201
personal websites, 43, 49 Social Identity model of Deindividuation
physical distance, 202 Effects (SIDE), 5, 16, 17, 18, 21, 65,
physical isolation, 66, 77, 201 66, 68, 84, 105, 145, 147, 152, 166,
priming, 23 196, 198, 199
Prisoner’s Dilemma Game, 90, 92, 97 social identity theory, 34
214 Index
social influence, 23, 63, 178 technological and social determinism

compliance, 67 social deterministic model, 3, 4, 8, 9
informational influence, 64 text messaging, 131
interpersonal influence, 65, 68 Thibaut, J. W., 92, 94
norm-based influence, 64, 67 trust, 40, 41, 51
types of social influence, 64 development of trust, 42
social interactions, 201 dispositional tendencies, 45
social interdependence, 4 dispositional trust, 45
social isolation, 19, 22 generalized trust of others, 46
social movements, 154, 157, 158, 159, global trust, 41
161, 162 progression of trust, 42
social networking, 122 specific trust, 41
social networking sites, 54, 108 training for trust, 56
social role theory, 111 trust scale, 42
social value orientation, 72, 199 Turner, J., 19, 147
socially creative strategies, 153
source credibility, 177 Ultimatum Bargaining Game, 99
Sproull, L., 115 uninhibited behavior, 136
status differences, 69
stereotypes, 112 virtual bravado, 136, 204
stigmatized identities, 128 virtual reality, 2, 131
strategic and motivated user, 4 virtual selves, 200
strategic and motivated user, expected and virtual worlds, 53
emergent effects framework (SMEE),
154 Wagner, D. G., 123
strategic decision-making situations, Wahl, S., 190
89 Web 2.0, 182, 185
strategic uses of social technologies, Webcams, 54
4 word-of-mouth marketing, 57
superordinate group identity, 158 World Economic Forum (WEF), 161
Tajfel, H., 5, 146 YouTube, 181

tangible place, 200
task complexity, 113 Zaccaro, S. J., 109
team performance, 2 Zelditch, M., 123

Birchmeier - 2011 - Strategic Uses of Social Technology An Interactive Perspective of Social Psychology

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Birchmeier - 2011 - Strategic Uses of Social Technology An Interactive Perspective of Social Psychology

Uploaded by

Copyright:

Available Formats

Strategic Uses of Social Technology

An Interactive Perspective of Social Psychology

On an everyday basis, we communicate with one another using var-

Z a c h a r y B i r c h m e i e r has a Ph.D. in Psychology from Miami Uni-

B e t h D i e t z - U h l e r is Professor of Psychology at Miami Univer-

G a r o l d S t a s s e r is Professor of Psychology at Miami University. His

This publication is in copyright. Subject to statutory exception

First published 2011

Printed in the United Kingdom at the University Press, Cambridge

Library of Congress Cataloging in Publication data

ISBN 978-0-521-89926-0 Hardback

Cambridge University Press has no responsibility for the persistence or

List of figures page vii

1 Introduction: A social psychological analysis of

8 Opinion-based groups: (Racist) talk and (collective)

5.1 The Prisoner’s Dilemma Game page 90

4.1 Examples for moderators of the impact of media

This book grew out of the Conference on Group Processes in Computer-

A social psychological analysis of

The use of computer-supported social interaction (CSI) has become

psychology can now be applied to the communication forums that have

Contemporary literature reviews

Context may be co-constructed by social actors, but they use communication

Joinson’s (2003) text cast a wide net in documenting the pervasive

Technological and social determinism

to ask deeper questions and adhere more strongly to norms of disclosure

information, to develop a shared knowledge base, and to connect like

Person-by-situation interaction defined

In a set of papers, social psychologists Mark Snyder and William Ickes

impacted greatest by which situations? Thus, it can be observed that

manifestations of their characteristic dispositions, these individuals will come to

Thus, they acknowledged how a person’s choice of situations to experi-

of groups to join are limited by what is available and permeable. These

(2004b). Personality and the internet. In Y. Amichai-Hamburger (Ed.), The

Power and the gender divide

when and why the technology helps disempowered groups to transcend

The SIDE model: theoretical foundations

perception of others and the self. Depersonalization (Turner, 1982), a

members to differentiate themselves from others in the group and any

geographically isolated from each other. Although anonymity can

Empirical evidence for the SIDE model

differences in email use (individual identity salient). Our prediction, in

participants primed with an efficiency norm transferred this norm to two

aspect of SIDE by separating group members by means of screens or

The gender divide: SIDE effects affect power relations

potential of these communications technologies and see cyberspace as

Cognitive effects: gender salience

Based on the cognitive dimension of SIDE we predicted that the men

salience of these social categories, as well as providing the opportunity

Gender and the strategic SIDE

other words, do women use the gender-bending potential of cyberspace

In a second study (De Wolf, 2002; Spears et al., in prep), we were

differences in this context. Italian women also reported being reliably

Final comments, conclusions, and some

extends the reach of identities, and anonymity can sometimes intensify

in terms of the distinctive properties of technology, such as anonymity,

Reicher, S. D. (1987). Crowd behaviour as social action. In J. C. Turner, M. A.

Melanie C. Green and Jordan M. Carpenter

The internet is a defining feature of modern life. Individuals increas-

Our focus here is on social interactions: for example, conversing with

Global versus specific trust

home. Responses to these items were highly intercorrelated and formed

communicator motivations in evoking trust from others is a potentially