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 communica...

Author: Zachary Birchmeier | Beth Dietz-Uhler | Garold Stasser

19 downloads 452 Views 879KB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form

DOWNLOAD PDF

Strategic Uses of Social Technology An Interactive Perspective of Social Psychology

On an everyday basis, we communicate with one another using various 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 interactions. In this volume, the editors explore how aspects of a situation interact with characteristics of a person to help explain our technologically supported social interactions. The person-by-situation interaction 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, including 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 communicate 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 University. He has published scientific studies about online social relationships, 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 University. 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 Personality and Social Psychology, Psychological Inquiry, Journal of Personality and Social Psychology, Journal of Experimental Social Psychology, and Organizational 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 List of tables Notes on the contributors Acknowledgments

page vii viii ix xii

1 Introduction: A social psychological analysis of computer-supported social interaction zachar y birchmeier, beth dietz-uhler, and garold stasser 2 A SIDE look at computer-mediated interaction: Power and the gender divide 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 melanie c. g reen and jordan m. car penter

1

16

40

4 An interactional approach to social influence in computer-mediated communication kai sassenberg

63

5 Social interaction in cyberspace: Social construction with few constraints susanne abele

84

6 Dynamics of leader emergence in online groups 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 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

108

127

v

vi

Contents

8 Opinion-based groups: (Racist) talk and (collective) action on the internet 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 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 garold stasser, beth dietz-uhler, and zachar y birchmeier Index

145

172

195

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 Department of Psychology 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 Department of Psychology Miami University karen m. douglas School of Psychology University of Kent melanie c. g reen Department of Psychology 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


Department of Psychology University of Southern California girish lala Department of Psychology The Australian National University mar tin lea The School of Psychological Sciences The University of Manchester craig mcgar ty School of Psychology Murdoch University tom postmes Department of Social and Organizational Psychology University of Groningen kai sassenberg Knowledge Media Research Center ¨ Tubingen, Germany russell spears School of Psychology Cardiff University/University of Groningen garold stasser Department of Psychology 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


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

xi

Acknowledgments

This book grew out of the Conference on Group Processes in ComputerSupported 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 constraints) and psychological features (e.g., anonymity). As a consequence, many social scientists have investigated the social processes in computer-supported interactions, including online impression formation, 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 identified 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 recognizes 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 internal 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 resistant 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


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 communication, 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), Gibson and Cohen (2003), Hertel et al. (2005), Hinds and Kiesler (2002), Kreijns et al. (2003), and other treatises. These works present theories from industrial and organizational psychology, as well as in human resource management. Broader psychological issues of online impression 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 (AmichaiHamburger, 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 Galimberti’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 psycholinguistic 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 thereafter, Thurlow et al. (2004) published a textbook for undergraduate audiences 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 interdependence. 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 conclusions 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 presence, 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, technology 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

4


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 construct 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 others 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 commonly enhanced by specific communication modes. Personal influences on online cognitions and behaviors are noted where the authors discuss individuals’ strategic uses of social technologies (i.e., connecting to anonymous, online forums populated with like-minded others) as having 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 selfawareness and social identity salience; behavioral changes in language use and self-disclosure, etc.). These changes in users’ cognitive and behavioral states are then construed to influence future choices of communication media. Elements of these ideas were also addressed in Mantovani’s (2001) chapter in discussing the uses of and reactions to the computermediated 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 communication 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) chapter 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. AmichaiHamburger’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

6


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 interactionist framework in analyzing computer-mediated communication. These inferences have been drawn from topic-specific research findings 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 systematically 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, domainlevel 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 speculates as to how other sources of influence that have not yet been tested may affect individuals who communicate online. By identifying the influences 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 theoretical 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 outlined 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 explanations 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 combine 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 contemporary 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 influences 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

8


other factor (as when the extreme factor is near zero) or increase the weight of the other factor geometrically. Snyder and Ickes (1985) differentiated cases when personal variables moderate the impact of the situation, as well as the converse. They illustrated their points with observations 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 behavior (e.g., a Yahoo! American Idol chat room) can allow for individuals to express aspects of their self-concepts that are central to their identities 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 people’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 constraints 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 characterized 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 judgments 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 likelihood of participation from minority factions (McLeod et al., 1997). Dennis et al. (1999) also found that, in group tasks requiring greater interdependence (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 participation 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 selfawareness. Research on electronic teamwork has also found support for the technology x social x personal interaction perspective. For example, Hollingshead (this volume) shows how the amount of information shared during computer-mediated group discussions depends on the idiosyncratic goals that each group member may enter with. Competitive member 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 information that might disconfirm their belief. Moreover, Abele (this volume) discusses how the nature of dependencies affects interpersonal impressions. 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 individuals 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 technology 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

10


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 experience, 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 apprehensive 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 general, 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 opinionated group discussion to concerted effort is often peppered with discontent, 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 personal and social influences has been shown to account for significant variability 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 congruent 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 interests, prevent ostracism, or facilitate healthy, productive interdependence among people using networked computers. The potential for social technologies to connect remote minds, to communicate emergent ideas, and to empower individuals should establish the relevancy of the CSI perspective for interested readers in all communities and cultures. Theoretical 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.

12


(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 information 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: Findings, 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 differences in the effects of media richness. Small Group Research, 30(4), 405– 437. Gackenbach, J. (Ed.) (1998). Psychology and the internet: Intrapersonal, interpersonal, 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 (pp. 247–274). New York: Oxford University Press. 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: Oxford University Press. 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 environments: 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. Galimberti (Eds.), Towards cyberpsychology: Mind, cognition, and society in the internet age (pp. 39–51). Washington, DC: IOS Press.

14


McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the internet: 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 communication: The present and future of technology-based interactions. CyberPsychology & Behavior, 5, 581–598. Riva, G., and Galimberti, C. (Eds.) (2001). Towards cyberpsychology: Mind, cognition, 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 personality. 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 telecommunications. 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 (pp. 139–162). New York: Oxford University Press. 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 communication. 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, interpersonal, 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 exploring 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 anticipate 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 settings: the SIDE model. We have found this model useful in helping to correct a tendency, in the literature on CMC in particular, to underestimate 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 findings 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 submissive 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 strategies and effects similar for other groups and categories? After outlining our model and providing some evidence of empirical support, we concentrate 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) developed 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 selfconcept 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 management 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 theoretical 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 relevant 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 “deindividuated” 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. However, we have questioned the applicability of deindividuation theory to the context of CMC and developed a social identity critique of deindividuation 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 deindividuation 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 theoretical 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 associated with a deindividuated state, it proposes that anonymity can enhance the salience of group identities and further “depersonalize” social


19

perception of others and the self. Depersonalization (Turner, 1982), a concept central to self-categorization theory, is not the same as deindividuation; 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 individual 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 psychological 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 sometimes, 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 characterized 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

20


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 selfcategorization principles of salience (Oakes, 1987) and demonstrates the flexibility of the SIDE model in being able to account for a range of different 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 contextual conditions as comparable to the person x situation (PxS) interaction. 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, selfcategorization theorists have used this metaphor to apply to the sensitivity and power of the self-categorization theory (SCT) more generally in taking 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) identity; 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 identity differently to an identity that is not visibly cued (e.g., nationality).


21

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 otherwise be punished by powerful others or out-groups, and therefore constrained. 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 strategic behavior here may take into account the potential disapproval of the audience, it does not have to be negative or antisocial from the perspective 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 accountability 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)

22


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 sensitive 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 participating, 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 psychology students, not as individuals (group identity salient), whereas for the other half we stated that our focus was on individual personality


23

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 individual identity, where participants may be expected to differentiate their opinions from the group, asserting the uniqueness of the personal identity. 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 explanation of group polarization and social influence (Turner, 1991). However, 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 collaborative group task participants then engaged in group discussions of a scenario. 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 dimension 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

24


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 attraction 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 grouprelated 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 communication and cyberspace. Indeed, from early on many influential theorists 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 theoretical 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


25

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 tolerated 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 disempowered 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 anonymous 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

26


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 endorsement of behavior that was normative for students and also acceptable to the staff – further evidence for the cognitive dimension of SIDE increasing 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 unpopular 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 producing open resistance to the proposal. In sum, although CMC can isolate individuals and potentially disempower 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 gender 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


27

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 discussed four topics via email. Results revealed evidence of clear genderbased power differences emerging in this context: men initiated more autonomous comments than did women, and women asked more questions than men, indicating that men were more dominant in the discussions. 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 predict. 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 participants 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 profiles 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 participants 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).

28


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 visually 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 withhold 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 gender 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 interrupt 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


29

salience of these social categories, as well as providing the opportunity for individuation. This requires us to refine the predictions for the cognitive dimensions 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 “comparative fit” or “meta-contrast”; see Oakes, 1987). So where there are clear visual cues to category differences in appearance, as there are with gender, 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 depersonalization 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 cohesiveness 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

30


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 prototypicality 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 making 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 gender 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 differences 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 identities (“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 neutral 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


31

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 gender 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 masculine (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, depending 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.

32


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 conditions 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 expectations 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


33

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 organizations (COMMORG). As well as providing insights into the role of how men and women manage their gender identity using email in the organization, 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 following 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 reasons 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 communication 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 European 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 important in Italy than our northern European comparison countries. As well as this cross-cultural moderator variable we reasoned that gender concealment 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

34


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, recreational, and organizational contexts, to strategically manage their gender identity in ways that protect them from the status and power differentials 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 identity 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 strategies 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 influence 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 framework (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) technology, 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!)


35

extends the reach of identities, and anonymity can sometimes intensify their effects. The same seems to be true for the reach of power differentials associated with these identities – these do not necessarily disappear or get diluted in cyberspace, as the example of gender shows. Moreover, 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 property. 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

36


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, selfselection, and sex). Unpublished Masters thesis, University of Amsterdam. Diener, E. (1980). Deindividuation: The absence of self-awareness and selfregulation 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 phenomenon: Status effects in computer-mediated and face-to-face decisionmaking 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 information. Oxford: Westview. Haraway, D. (1990). A manifesto for cyborgs: Science technology, and socialist 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 online compliance. Paper presented at Conference on Group Processes in Computer Supported Interaction: Technological and Social Determinism. Miami University, Oxford, OH.


37

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, deindividuation, 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.), Rediscovering 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 internet: Mass communication and online mobilization. Social Science Computer Review, 20, 290–301. Postmes, T., and Spears, R. (1998). Deindividuation and anti-normative behavior: 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 boundaries? 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 selfregulation of behavior. In P. B. Paulus (Ed.), The Psychology of Group Influence (2nd ed., pp. 86–109). Hillsdale, NJ: Lawrence Erlbaum.

38


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.), Rediscovering 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 visibility to the out-group. British Journal of Social Psychology, 33, 145–163. (1994b). On the consequences of deindividuation manipulations for the strategic considerations of self: Identifiability and the presentation of social identity. 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 interaction. 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.), Contexts of computer-mediated communication (pp. 30–65). Hemel Hempstead: Harvester-Wheatsheaf. (1994). Panacea or panopticon? The hidden power in computer-mediated communication. 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 communication: 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. Joinson, 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.


39

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 Dynamics: 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 increasingly conduct social interactions online or use other forms of mediated communication, such as text messaging. Understanding how this communication medium affects interpersonal interactions and relationships is a key task for psychologists. Specifically, researchers have been interested in the extent to which individuals are willing to trust others 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 example, 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 behaviors; 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 characteristics 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 Messenger (IM) or social networking sites such as Facebook. These interactions are in the service of relationship building, entertainment, or sharing information 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 philosophy, economics, marketing, and psychology), and researchers have not always shared a common understanding of the term. From a psychological perspective, trust can be broadly defined as “confident expectations 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 situation 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 domainlimited 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 orientation toward another person (the judgment that the person will act with one’s best interests at heart), trust is not an all-or-nothing decision. 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 communication change some aspects of this process. For example, revealing personal 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 accelerate the progression of trust. For example, one study found that inperson friendships are more trusting and generally higher in quality than


43

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” perspective, 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 individual’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 discloses 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 embarrassment, 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 themselves self-disclosing (e.g., Joinson, 2001), may also be those that evoke greater trust. However, this previously identified progression from relatively impoverished forms of information (pseudonyms, text) to richer and more personal forms of contact is likely to change with new technological developments 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, without 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 individuals may hope to create trust through their disclosures as a means of building relationships with others, whereas others may be seeking validation, attention, or other goals that do not require trust. The role of

44


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 computermediated communication: that its relative anonymity allows easy deception 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 internet 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 members 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 themselves 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 somewhat complicated, however. For one thing, not only is the potential discloser 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 members is becoming increasingly common, especially among young people.


45

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; Valkenburg 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 capitalrelevant 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 behaviors on the website Facebook (i.e., messaging a friend to schedule dinner) emerged out of a curiosity about people’s perspectives, while more isolating 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.

46


Generalized trust of others “on the internet” The context of the internet itself is a powerful situational factor influencing individuals’ generalized trust in others. Perhaps because of the difficulty 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 bargaining 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 contexts 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 outlets 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 sophisticated view of lying. Fabrications can emerge from a variety of motivations, and the perception of these motivations can often determine the acceptability 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


47

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 exaggerate their positive emotions and avoid expressing negative ones. Lies require little cognitive preparation, and liars generally feel minimal anxiety about their lies being exposed, despite the fact that most lies are blatant 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 featuring 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 deception. One important situational factor is that cyberspace communication allows for deception that would be easily caught in face-to-face interactions, 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 deception, 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 different importance in predicting online versus face-to-face deception. For instance, Lu (2008) found that individuals high in attention-seeking and

48


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 generally 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


49

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 disinhibition 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 generally 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 dramatically 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 individuals feel suspicion very easily toward information that cannot easily be

50


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 exacerbates people’s already poor ability to spot lies. Obviously, successful deception depends on the listener having a limited amount of information 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’ unconscious 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 patterns 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 influence the behavior of others through communication, in online deception.


51

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 recognizing 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 hospital), 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 retiring). 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 Messenger or the telephone. Individuals appeared to be trading off the convenience 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. Individuals are more likely to choose email over the telephone or face-to-face

52


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 opportunities in computer-mediated communication and specific personality traits that may lead to deception. Among the situational factors, the invisibility of the internet gives people the chance to lie about a wider variety of things than would be possible otherwise, while the distance of textbased 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 selfprotective 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


53

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 pursue those goals is influenced by the norms and capabilities of a given online environment. Thus, the person factors (goals to create a particular 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 attention 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

54


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) interacting 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 unfortunate 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 themselves 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 information about its owner than an everyday face-to-face interaction could,


55

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 veracity 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 attractive 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 reputation 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 profile on an employer’s website, or other independent verification (Green,

56


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 provide 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 interaction. 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 business, 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).


57

Word-of-mouth marketing Although this chapter has focused primarily on interpersonal relationships, one area for future attention is the rise of viral or word-of-mouth marketing online. These practices may blur the lines between interpersonal sharing and a business pitch. That is, when a fellow website user recommends a product or service, that person may be doing so simply 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. Computermediated communication often comes with distinct levels of information, 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 freedom also inspires honesty and self-disclosure, and despite highly publicized 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 individuals’ 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 communication offered by the internet. The situational forces that are present online – the extent of information available (e.g., text versus video), the

58


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 increasingly 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 openness 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 interpersonal 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 forming a normative response to a deception revealed online. Social Science Computer 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: Individual 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: Preferences 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.


59

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 computermediated 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.

60


McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the internet: Identity “de-marginalization” through virtual group participation. Journal of Personality and Social Psychology, 75, 681–694. 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 perceptions 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: Denigration 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 “disconnecting” 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: Questioning 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. CyberPsychology & Behavior, 10(1), 64–70. Shklovski, I., Kraut, R., and Rainie, L. (2004). The internet and social participation: 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 telecommunications. 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.


61

Valkenburg, P. M., and Peter, J. (2007). Online communication and adolescent 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. Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 23(1), 3– 43. Walther, J. B. (2007). Selective self-representation in computer-mediated communication: Hyperpersonal dimensions of technology, language, and cognition. 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). Commitment, pro-relationship behavior, and trust in close relationships. Journal of Personality and Social Psychology, 77, 942–966. 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 cyberostracism. 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. Organizational 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

62


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 communication. Computers and Human Behavior, 20, 381–402. Zhou, L., and Zhang, D. (2007). Typing or messaging? Modality effect on deception detection in computer-mediated communication. Decision Support Systems, 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. Internet users send out information, summaries of subjective experiences, and opinions using communication tools. Recipients acquire unconsciously 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 summary 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 interindividual differences between users on the effects of situational characteristics (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 distinguished by Deutsch and Gerard (1955) and have been discussed subsequently by many others (for a summary, see Wood, 1999) are normative influence (the conformity to others’ expectations in order to reap social rewards and avoid social punishments) versus informational influence (conformity to others that results from the assumed validity of provided 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 Sassenberg 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 pressure 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 fostered 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 interpersonal influence. The following introduces the features of CMC relevant 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, paralinguistic 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 volume) suggests that the scarce accessibility of information under conditions 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 communication 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 resistant to influence in CMC than in FTF communication. Taken together, SIDE assumes that when social identity is salient, anonymity of others increases social influence (i.e., norm-based influence), and that when

66

Kai Sassenberg

personal identity is salient, anonymity reduces social influence (i.e., interpersonal 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). Therefore, 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


67

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 compliance to others (e.g., Smilowitz et al., 1988). Recently, research found evidence for the impact of anonymity of self to different targets. Participants 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 outgroup 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 demonstrated 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., abbreviations, 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, several 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 conflict 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 partners 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


69

categorization. Under these conditions, the communicators tend to follow 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 identity 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 process assumptions put forward by SIDE – namely, more striving to fulfill personal needs. Anonymous interaction partners (compared with nonanonymous 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 conforming 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 interpersonal influence that is moderated by CMC. Generally in communication, 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 interactions; 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 status 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 expertise 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 communication media on private self-awareness will most likely result from physical isolation. At the same time, private self-awareness hinders social influence (Froming and Carver, 1981; Scheier, 1980). Therefore, the heightened private self-awareness in CMC mediates the media differences in interpersonal influence (Sassenberg et al., 2005). To conclude, physical 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 characteristics and mechanisms that show how CMC can lead to increased or decreased social influence compared with FTF communication. Compliance, 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 especially salient when the communication partners are anonymous. Preconditions 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.


71

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 selfawareness (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 current 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 differences 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 moderating 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 normbased 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 isolated 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 AmichaiHamburger, 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 concerning media choice: introverted as well as emotionally unstable individuals 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 information 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 personal 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 selfmonitoring 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


73

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 cooperate less when getting this information (see also Kimmerle and Cress, 2008). In sum, individual differences moderate the impact of characteristics 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 specific 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 differences 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 attitude 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 foster social influence but also lead to particularly strong resistance against

74

Kai Sassenberg

social influence. My own work on the effect of chronic private selfawareness 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 private 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 demonstrates that individual differences matter in the context of media effects on social influence. There is evidence concerning a wide range of phenomena (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 mediabased social influence. The person x situation interaction approach For a long time, it has been widely accepted in psychology that characteristics of individuals, characteristics of the environment, and the interaction between the two determine human cognition and behavior (e.g., Cronbach, 1957; Lewin, 1936). Nonetheless, theoretical as well as empirical 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 situation interaction; that means person and situation characteristics which are functionally equivalent do not just add up in their impact on mental 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 differences which are activated by situational characteristics, similar to a


75

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 resolution 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 undergoing this treatment (compared with usual self-judgment). The authors apply two different metaphors to explain the divergent responses: narcissists 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 different processes (search for truth and search for self-enhancement) are necessary to understand disordinal interactions. Taken together, person x situation interactions can result from cognitive 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 interpersonal 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 collection of individuals whereas prosocials perceived them as their ingroup. Therefore, proselfs were more likely to experience interpersonal influence, and consistent with this assumption they showed less social influence (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 interactions underlying the disordinal interaction found by Cress (2005) are based on cognitive processes, because the impact of social value orientation 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 literature 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 overrules 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 trustworthiness 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 partners and therefore evaluate the incoming information more. Hence, the expected trust x anonymity of others interaction would be driven by a motivational process.


77

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 person 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 orientation 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 difference 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 mechanism 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 influence: interpersonal influence. The interaction between physical isolation and private self-awareness reported above (Sassenberg et al., 2005) suggests 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 isolated 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. Interactions 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 accessibility 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 influence (i.e., compliance, interpersonal influence, and norm-based influence) are moderated by individual difference variables that affect either message processing (i.e., cognitive process) or the expectancies and values in the situation that are relevant for action (i.e., motivational process). Anonymity of the self and physical isolation are more likely to show ordinal 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 individual difference variables, because it is relevant for two social influence processes. This type of interaction depends on the relevance of the individual difference variable for one or both social influence processes. The list of potential moderators in Table 4.1 is certainly not exhaustive. All person characteristics related to the chronic accessibility of the


79

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 characteristic 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 summarized 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 integrative 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. AmichaiHamburger (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 crossculturally? 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 identifiability on communicative behavior. Group Dynamics, 6, 17–26.

80

Kai Sassenberg

Dubrovsky, V. J., Kiesler, S., and Sethna, B. N. (1991). The equalization phenomenon: Status effects in computer-mediated and face-to-face decisionmaking 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 selfconsciousness: Assessment and theory. Journal of Consulting and Clinical Psychology, 43, 522–527. ¨ ¨ ¨ Franke, G. H. (1997). Uber die Moglichkeiten der computerunterstutzten ¨ Darbietung beim revidierten Freiburger Personlichkeitsinventar: Zwei experimentelle Studien [Computer-based administration of the Freiburg Personality Inventory: Two experimental studies]. Zeitschrift für Experimentelle 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 comparison: 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 communication. 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 preference in threatening and non-threatening situations. Social Psychology, 39, 231–243. Hiltz, S. R., Johnson, K., and Turoff, M. (1986). Experiments in group decision 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 Psychology, 31, 177–192. (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 Computer Supported Collaborative Learning, 3, 85–97.


81

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, deindividuation 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 language 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 communication. 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 personality: 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 boundaries? 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 attachments to groups and to their members: Distinguishing between commonidentity 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. Communication Research, 32, 193–234. Reicher, S. D., Spears, R., and Postmes, T. (1995). A social identity model of deindividuation phenomena. European Review of Social Psychology, 6, 161– 198. ¨ Renner, K.-H., Schutz, A., and Machilek, F. (Eds.) (2005). Internet und ¨ Persönlichkeit [Internet and personality]. Gottingen: 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¨ lemlosekompetenz: Der Einfluss von Expertise auf den Informationsaustausch in computervermittelter Kommunikation [Knowledge and problemsolving 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: Oxford University Press. 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 Psychology, 39, 514–521. Schmitt, M., Eid, M., and Maes, J. (2003). Synergistic person x situation interaction 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.),


83

Contexts of computer-mediated communication (pp. 30–65). Hemel Hempstead: Harvester-Wheatsheaf. (1994). Panacea or panopticon? The hidden power in computer-mediated communication. Communication Research, 21, 427–459. Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., and ten 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 communications. British Journal of Social Psychology, 29, 121–134. Spears, R., Postmes, T., Lea, M., and Watt, S. E. (2001). A SIDE view of social 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. 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: Blackwell. Van Lange, P. A. M. (1999). The pursuit of joint outcomes and equality in outcomes: 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 anonymous 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. Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal 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 of Personality and Social Psychology, 56, 70–80. 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 interested in studying interactions that are mediated by modern technologies such as the internet. Research topics have focused on how computermediated 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 Deindividuation 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 technologies, 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 studied 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 different in nature than the ones that are salient in face-to-face interactions (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 opportunity for deliberative action. Second, once we identify important distinctions 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 possibilities to be deliberative. Whereas in a lot of face-to-face interactions our behavior and responses to others need to be immediate, and consequently 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 communicating 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 partner 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 differently, 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 faceto-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 looking 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 bigger 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, people 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 deliberative when it is computer-mediated. When communication is deliberative, 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-toface 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, intonation of voice, and emphasis of words all refer to characteristics of the


87

person, or the way the communicators package their message. In contrast to social cues that are derived from characteristics of the person, features of the situation are equally present in face-to-face and computersupported 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 interaction. 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 consideration of how others will react to our behavior. On the other hand, there are fewer social and situational cues on which to base our deliberations. 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 others 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 economics 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 interdependency, 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 somebody 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 studying situations where the independent variables clearly and unambiguously define interdependencies. As a result, carefully constrained interactions 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 technologysupported 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 theory invokes a similar projection assumption in the form of common knowledge rationality. That is, game theory assumes that players are perfectly 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


89

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 situation, 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 perceptions 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 subjective. What one person might perceive and hence categorize as hostile might not be perceived that way by another. More technical classifications, which do not refer to affective experiences and can be objectively applied to interactive situations, are provided by game theory (von Neumann 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. Therefore, the experiments that have been conducted in these areas consist of clearly defined situations and have high anonymity between the interaction 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, situations where decision makers are interdependent, by describing the solutions 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 B C

D

20

10 C 0

10 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 situations. 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 precisely, the chosen alternatives and their corresponding pay-offs constitute a Nash equilibrium. In contrast, a Pareto efficient solution to an interaction 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. According 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 efficient. Consider the classic Prisoner’s Dilemma Game (PDG) illustrated in Figure 5.1. The Nash equilibrium is reached if both players defect. It is


91

H 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 behavioral 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 players 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 behavioral 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 component 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 defecting 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 person’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 predominantly 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.


93

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 conflict 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 provide opportunities for integrative solutions (Thompson, 2005). Nobody would make the inference that findings from one type of negotiation situation 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 partners are provided with an endowment. They can either keep this endowment 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 substitution, 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 narrow 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 representative in the graduate council. But if either side could convince a certain proportion of the other side to vote for their favorite, your department 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 equilibria 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, however. Kelley and Thibaut (1978; Thibaut and Kelley, 1959) argued that people can reinterpret pay-off matrices in social relations. Repeated interactions 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 conflict 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.


95

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 cooperativeness 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 individual 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 differences in the perception of cooperative and competitive interpersonal situations (King and Sorrentino, 1983). The authors showed that different 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 orientations 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 competitive 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 thoughtful 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 characteristics 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 na¨ıve theory 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 computermediated 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 interactions. 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 situation in the absence of strong cues from the physical environment. 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 information 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 personalities 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


97

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 nothing 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 noneconomic decision frame, in which participants were in the role of a representative of a company and had to decide how much of a contribution 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 conceptualize 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 significantly more cautious decisions under simultaneous decision timing compared with the pseudo-sequential timing structure. Abele et al. hypothesized 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 interactions, 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 division of an endowment between yourself and another party, whereby that other party remains anonymous to you, and their role consists of merely


99

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 fairness 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 participants 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 varied 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. Furthermore, 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 interaction 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 perception of whether the allocator seemed to have deemed distributive justice as important. The responder’s behavior again will affect subsequent 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.


101

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 impression 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: pleasantness (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 experimenter, 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 meanness 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.


103

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 computermediated 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 systems). 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 making 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 technologies 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 BlackBerries 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 improving 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 technologysupported. However, as others have noted, computer-mediated interactions often contain fewer and different social cues than do face-to-face interactions. Both of these characteristics of computer-mediated interactions 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 classify 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 interaction 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 examine these limited situations, game theory and interdependence theory can provide a useful classification of situations. Behavior is a function of personalities and situations. Moreover, different 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 sanctioning 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


105

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 understanding how it differs from face-to-face interactions and how those differences 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 computermediated interaction affords, even promotes, a deliberative process. In this deliberation, actors try to anticipate the reaction of others to alternative 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 information 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? Genderpairing, 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 perceived 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 goaloriented 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 Psychology, 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.


107

Spears, R., Lea, M., Corneliussen, R. A., Postmes, T., and ten 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 communications. British Journal of Social Psychology, 29, 121–134. Spears, R., Postmes, T., Lea, M., and Watt, S. E. (2001). A SIDE view of social 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 Psychology, 36, 1–25. 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. Computers 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 faceto-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 functionality (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 (Hollingshead 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, similar interests, or complementary goals from across the country or around the world to connect and, in some cases, form discussion groups, friendship 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 interdependent goals, interacting with one another, and having a sense of belonging associated with their membership (Hollingshead and Contractor, 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 description 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-toface 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-toface 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 psychologist, my work focuses more (but not exclusively) on situation-based rather than on person-based explanations and determinants of behavior, as will be evident in this chapter. However, there is an extensive literature 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 empirical 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 improvements 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 capacities, personality orientation, motives and values, social appraisal skills, problem-solving competencies and general and domain-specific expertise” p. 120. However, they also acknowledged that situational and contextual 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 chapter. I begin with an overview of social psychological research on leader emergence in face-to-face groups, then I present data from two unpublished studies about leaders in online groups. The first study, on which I was a collaborator, compares the process of leader consensus in faceto-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 Language 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 perform (Biddle and Thomas, 1969; Katz and Kahn, 1966).


111

In contrast, informal roles are negotiated by group members and emerge through interaction. Benne and Sheats (1948) provided a classification scheme that is still widely used today to describe three types of informal roles: task-oriented, maintenance, and self-centered. Members 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, feelings, 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 members (e.g., high participation, dominance, assertiveness); or through the behavior of group members toward them (the person whom group members address most). Leaders can also be identified by asking group members, 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 expectation 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 creating performance expectations that affect the degree to which group members participate, have influence, and are evaluated by others in their group.

112


The behavior of group members often conforms to performance expectations based on these shared cultural beliefs. Thus, shared cultural beliefs can have self-fulfilling effects and can influence leader emergence in groups: people from socially advantaged groups are more likely to emerge initially as leaders than people from socially disadvantaged 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 hierarchy, 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 example, 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 leadership traits was .93, suggesting a strong and positive association between participation and leader emergence. The researchers varied group composition across the sessions and did not measure whether the same people emerged as leaders across sessions.


113

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 partialled 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 contributions 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’ participation 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 relation 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 participating group members were chosen to be leaders. However, participation 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 predictor 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 participation 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 participation 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

114


members contribute proportionately less to the group discussion. For example, in groups with three members, the most active member participates 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 participates 3 percent. A similar pattern was discovered in much larger groups. Mitchell (2006) investigated the relation between meeting size and member 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 attendees 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 particular 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 opportunities 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 communication 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).


115

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 depersonalize 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 contact. 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 interactions 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 information cues that provide feedback regarding the impact of their behavior on other group members (e.g., nonverbal sanctioning acts: glares, interruptions; nonverbal signs of approval: nods). Consequently, group members engaged in computer-mediated computer interaction tend to feel more anonymous and detect less individuality 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 individual 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

116


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 decisions. It is important to note, however, that several studies observed no differences in participation patterns between face-to-face and computersupported groups (e.g., Poole et al., 1991; Walther and Burgoon, 1992). More recently, research has examined characteristics of members identified by the majority of their peers as informal leaders of computermediated groups. Wickram and Walther (2007) found that emergent leadership in computer-mediated groups was associated with high levels of participation, perceived intelligence, encouragement, and authoritarianism 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 characteristics 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” (Contractor 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 emergent 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 distributed work teams, committees, and the like collaborate across many points in time. Does leader emergence just take longer in computermediated groups or is it unlikely to happen at all? Study 1 addresses this issue.


117

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 organization and each week they would carry out assignments for different “client” organizations. Group size ranged between three and five members. The computer-mediated teams interacted through a computer network, which enabled only text-based communication. After each week’s assignments, group members completed a questionnaire individually that examined their reactions to the task, their group, and one another. Leader consensus was measured in ten face-to-face and ten computermediated 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 participants 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 faceto-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 consensus 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 significant. Neither were the differences in leader consensus between faceto-face and computer-mediated groups in weeks 2, 3, 4, or 5. Overall, the mean leader consensus across the five sessions did not differ significantly 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 characteristics (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

118


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 interactions. Consequently, face-to-face groups showed decreasing levels of leader consensus over time. The attenuated pattern of role consensus over time for computermediated groups in this study may be due to the reduced salience of members’ manifest characteristics (Kiesler and Sproull, 1992). In addition 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 nonsystematic differences in the relative magnitude of leader role consensus 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 development process. Since the majority of studies were one-shot experimental 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 computermediated work groups would endure over time. However, the results of the present study indicate that the difference in leader role consensus


119

90 80

Group size

70 60 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 computermediated interactions where members can participate asynchronously or simultaneously. In addition to few geographical and temporal constraints, 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 support groups. The support group topics included anxiety, attention deficit disorder, breast cancer, depression, diabetes, eating disorders, fibromyalgia, 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

120


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): computermediated 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 periods. 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 positively associated with emergent leaders in online groups as it is in faceto-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.


121

Number of groups with same leader(s)

10

8

6

4

2

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 people ask and answer technical questions and seek information, the contributors’ 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 selfpresentation and messages can be anonymous to some extent, so members may have somewhat more freedom to decide on the roles they will take within the group than their face-to-face counterparts. They can identify their messages with their given names, a pseudonym, or no name to maintain a sense of anonymity. Members also express and develop identities through signature files, which appear at the end of messages and may consist of quotes, pictures, professional titles, or other defining information. 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.

122


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, gender, 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 communication 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 participation, 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 differences 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 candidates, CEOs, and other leaders have Facebook and MySpace pages. How do leaders represent themselves on those sites, and in turn, how do those


123

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 longitudinal 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 communication in context, Vol. 2. Hillsdale, NJ: Lawrence Erlbaum Associates. The American Heritage Dictionary of the English Language, Fourth Edition C 2000 by Houghton Mifflin Company. Boston, MA: (2009). Copyright Houghton Mifflin Company. Bass, B. H. (1949). An analysis of leaderless group discussion. Journal of Applied Psychology, 33, 527–533. 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. R

124


Easton, A., Vogel, D. R., and Nunamaker, J. F. (1992). Interactive versus standalone 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 communication 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 consensus 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 applications. 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 England 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 management. American Psychologist, 45(2), 200–208.


125

Poole, M. S., and DeSanctis, G. (1990). Understanding the use of group decision support systems. In C. Steinfield and J. Fulk (Eds.), Organizations 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 decision 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). Picturing 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 decision support systems: Determinants of success. Proceedings of the 7th International 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 facilitate 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.

126


Yoo, Y., and Alavi, M. (2002). “Electronic mail usage pattern of emergent leaders in distributed teams.” Sprouts: Working Papers on Information Environments. 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 reasons 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) interpersonal 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 studies involving electronic-based interactions. Both immediate and longterm effects of ostracism on targets’ behaviors are discussed, focusing specifically on how people can be negatively impacted by being ostracized 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 similarities and differences between face-to-face and electronic-based interactions. 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 selfesteem (Leary et al., 1995; Tesser, 1988). Humans also have more actionfocused 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 important of these needs is belonging. They review considerable literature from various academic fields, demonstrating how important social relationships 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). Belonging 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 communication 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 possible 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 characteristics 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., nonmainstream 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 negative 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 psychological 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 studies 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 experiencing physical pain is also activated while experiencing ostracism (Eisenberger 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

130


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 cyberostracism on individuals and their subsequent behavior. The majority of these programs have focused on experimental methods, using a variety of electronic media, constructs, and manipulations of ostracism to gain a more thorough understanding of the intricacies of these phenomena (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 manipulate 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 negative 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 conditions). 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, participants 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


131

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/outgroup 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 interacting 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 (confederates) 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 confederates 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

132


compare two different degrees of ostracism (being unacknowledged vs. walked through as if one does not exist). Minimal World, conversely, examined the minimal conditions necessary 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 participants’ 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 reactions: targets’ immediate reflexive reactions and their subsequent reflective 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 selfesteem 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


133

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 computer program designed to ostracize them, participants still reported the same effects on need satisfaction and mood! Finally, other research suggests 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 participants 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 feeling the negative effects of being ignored and excluded when their inclusion 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 immediate 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

134


the absence of instructions to infer social information, ostracized participants 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 hypothesis; 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 interaction 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 joining 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 organizations (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


135

aggressively toward other people in response to exclusion and ostracism, regardless of whether the person was involved or uninvolved in the targets’ 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 shootings 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 selfesteem, 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 relation between control and anti-social responses to ostracism. Warburton et al. (2006) found that ostracized participants who were given control over an unrelated task behaved no more aggressively than included participants. Ostracized individuals who were not given this opportunity 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 interactions. They found that individuals who were ostracized unexpectedly responded with the most aggression, but individuals who still had predictive 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’ attributions 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

136


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; Christopherson, 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 communication (Kiesler et al., 1984; Sproull and Kiesler, 1986). Some examples of behaviors that can be considered “uninhibited” in electronic communication are increased “flaming” (i.e., insults or swearing) (Kiesler et al., 1984; Sproull and Kiesler, 1986), risky or extreme group decisionmaking (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, bigotry, 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 following 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 hierarchy 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 communication 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 situation as participants did in face-to-face ostracism (Williams and Sommer, 1997). Sometimes, ostracized participants would type nonsense


137

sentences simply to interject themselves into the confederates’ conversation (Williams et al., 2002). This virtual bravado seems to be the most significant difference between reactions to face-to-face and cyberostracism, and fits closely with the research on uninhibited communication (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-toface 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 empirically, and determine what types of situations and personality characteristics 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 findings 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 proor anti-social) are ways in which ostracized individuals fortify their threatened 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 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 being ostracized in an environment viewed as their last bastion of hope may lead these individuals to redirect their focus from belonging and self-esteem to seek fortification of control and meaningful existence instead. If Williams (2009) is correct in hypothesizing these latter two needs are more likely to be fortified by anti-social and aggressive acts, then electronic-based ostracism and rejection for socially anxious or lonely individuals may have serious

138


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 studies 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 rejection 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 consequences (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 suggests 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 antisocial and uninhibited behavior. Experimentally, paradigms that study


139

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 interaction. 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 anxious 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 ostracize 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 relationships 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 validation by the multitrait–multimethod matrix. Psychological Bulletin, 56, 81–105.

140


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 Social Psychology, 44, 896–903. 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 Personality and Social Psychology, 37, 1160–1171. Diener, E., Fraser, S. C., Beaman, A. L., and Kelem, R. T. (1976). Effects of deindividuation variables on stealing among Halloween trick-or-treaters. Journal of Personality and Social Psychology, 33, 178–183. Dubrovsky, V. J., Kiesler, S., and Sethna, B. N. (1991). The equalization phenomenon: Status effects in computer-mediated and face-to-face decisionmaking 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 uncontrollable: Effects of stress on illusory perceptions of controllability. Journal of Personality and Social Psychology, 63, 923–931. 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 consequences 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


141

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, rejection, 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). Selfesteem as an interpersonal monitor: The sociometer hypothesis. Journal of Personality and Social Psychology, 68, 518–530. 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 Psychology, 52, 917–930. McKenna, K. Y. A., and Bargh, J. A. (1998). Coming out in the age of the internet: Identity “demarginalization” through virtual group participation. Journal of Personality and Social Psychology, 75, 681–694. (1999). Causes and consequences of social interaction on the internet: A conceptual framework. Media Psychology, 1, 249–269. (2000). Plan 9 from cyberspace: The implications of the internet for personality 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 exclusion: Being rejected versus being ignored. Journal of Personality and Social Psychology, 96, 415–431. 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 dilemmas. 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.

142


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 Psychology Bulletin, 30, 1095–1107. 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. CyberPsychology & Behavior, 10, 64–70. Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. W. (1986). Group processes 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: Academic Press. Sproull, L., and Kiesler, S. (1986). Reducing social context cues: Electronic mail in organizational communication. Management Science, 32, 1492–1512. Stritzke, W. G. K., Nguyen, A., and Durkin, K. (2004). Shyness and computermediated communication: A self-presentational theory perspective. Media Psychology, 6, 1–22. Tesser, A. (1988). Toward a self-evaluation maintenance model of social behavior. 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. Journal of Personality and Social Psychology, 81, 1058–1069. 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.


143

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 Experimental Social Psychology, 45, 1308–1311. 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 University, 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 Psychology, 79, 748–762. Williams, K. D., Govan, C. L., Crocker, V., Tynan, D., Cruickshank, M., and Lam, A. (2002). Investigations into differences between social- and cyberostracism. 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 Psychology Bulletin, 23, 693–706. 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.

144


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 Experimental Social Psychology, 42, 692–697. 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 belonging, control, self-esteem, and meaningful existence. Journal of Experimental Social Psychology, 40, 560–567.

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 movements 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 preliterary 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 assemblies 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 contributed to this form more slowly, so that in the twentieth century talkback 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 adopting both modes. They are particularly important for our chapter because both modes are critically involved in group processes. This seems obviously true in relation to the discussion of the gathering mode, which is the very stuff of group processes. We argue further, however, that the broadcast 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 tradition 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 identity 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 selfcategorization theory seeks to account for both personal, private phenomena 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 profusion of web-logs or blogs. Everyday millions of people from around the world publish what would otherwise be private thoughts and invite commentary 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 selfperception. 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 (different 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, intergroup 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 category specifications (including the social behavioral meaning of those categories) 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 category “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.

148


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 straightforward: 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; Stu¨ rmer, 2000). [emphasis added] (Klandermans et al., 2002, p. 239)

Contrary to Klandermans et al., the available evidence for the proposition 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 suggest 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 psychologically 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).


149

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 collective 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 categories and action groups. Opinion-based groups can often be formed within a broader social category (or can be broader than a social category). 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 opinionbased 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 boundaries. 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 “singleissue pressure groups” studied by Kelly and Breinlinger (1995) and the activist groups such as Gray Panthers, the gay movement, AIDS volunteer service organization, and the fat acceptance movement, investi¨ gated by Simon et al. (1998, 2000) and Sturmer 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

150


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 opinionbased 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 subcategories 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 essential to look at the underlying nature of collective action and of the groups participating in collective action. Industrial action provides a good example 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, opinionbased groups are identifiable in a more raw, less structured form. People


151

participating in spontaneous action need not be attached to specific activist groups but share a common understanding and stance on a certain 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 consequences 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 conceal 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

152


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 individuals 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 electronic 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) proposed that being personally anonymous to a powerful audience may be liberating; anonymous communicators need not feel pressured to conform to the norms and expectations of powerful others (see also Reicher and Levine, 1994a, 1994b), but may enact aspects of their own identity 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 opinionbased 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 presence 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.


153

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 whitesupremacists 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 evidence 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 arguments, 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 creativity may serve the function of helping to create conflict by making the ingroup (usually whites) feel that they have a reason to feel disadvantaged. Second, without seeming overtly hostile or violent, socially creative

154


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 accuracy 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 proposition 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 theoretical 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 psychological 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.


155

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 collectives, 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 “rational actor” models provide one popular explanation for individual participation 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 movement 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, Klandermans (1984) argued that individuals weighed collective motives (overall collective benefits), selective motives (potential reactions of significant others), and reward motives (individual benefits) when assessing social movement participation. Klandermans also proposed two components of movement mobilization – consensus mobilization (movements attempting to gain support for their position) and action mobilization (movements mobilizing people to act), suggesting that action mobilization 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 movements. More specifically, this work began to explore how identity-related motivations toward social action might interact or coexist with traditionally recognized utilitarian motives. More recently, researchers have begun to refine these ideas by proposing 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 propose two independent focuses – instrumental (where motives are primarily related to cost/benefit analyses) and identity (where self-definition is the central motive).

156


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 normative 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 proposed 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 automaticity in their conceptualization of identity and instrumental paths. While they do recognize the value of further research in this area, controlled 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 instrumental 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 selfdefinition” (p. 656) in the identity pathway. One interpretation, therefore, is that action driven by identity considerations is always automatic, or


157

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 conceptualization 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 compared with individual identity. Under the social identity approach, neither 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) movements provide two particularly good examples – one as an exemplar of the rational actor approach, the other more in line with identitybased 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 reasons for their position along the lines of an identity-based pathway (Stallman, 2007). Analyses drawn from other contexts seem applicable to internet movements, but there are also complexities. In particular, as the role of identity 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 participants’ 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

158


social identity adopted by social movement participants should be an action orientation. Moreover, social movements are highly heterogeneous – social movements 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 heterogeneity 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 content 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 psychological groups, some aspects of their very nature would seem to militate 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 differentiation 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” antiglobalization 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 “symbolic of this difficult but ultimately productive diversity” (Price, 2000;


159

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 development 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 members 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 particular 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 movements are typically superordinate organizations encompassing varied sub-groups and individuals, and social movements have a strong action

160


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 psychological 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 characteristics 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 comprised 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 compared with groups that meet face-to-face. Thus, even now, communication 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 environments 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


161

internet might play in these groups (e.g., Allard and O’Loughlin, 2000; Lawrence, 2000; McDonald, 2000). At first blush, given the unmistakable action focus of social movements generally and the high online profile of many anti-globalization movements in particular, this may seem a reasonable focus. We should ask, though, how the formation of opinionbased 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 earlier 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 traffic 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 antiglobalization 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 Tuckman, 1965), as the online group approached the point where it should be ready to perform, it continued to storm and storm (with such storming – and associated recriminations – continuing even after the actual protest event had passed). Furthermore, contrary to our expectations

162


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 sustainability, 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).


163

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 membership recruitment, distribution of information, and announcements and reviews of future and completed actions. Significantly, MoveOn.org operates 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 making 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 intragroup 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 communities, 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 overall 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 analysis 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).

164


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 consensualization. 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 community). 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 cyberhate 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 opportunities for participation and mobilization, while at the same time increasing the potential threats of subversion and misrepresentation (e.g., Lebert,


165

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 numbers 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). Fundamentally, 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 newspaper 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. Consequently, 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 movement 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, summarized 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

166


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 represent good places to start an analysis of the nature of online interaction, 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 understanding 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 identity (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 online 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 precedent 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 without 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.


167

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). Opinionbased 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 contemporary social movements. Social Research, 52, 663–716. de Weerd, M., and Klandermans, B. (1999). Group identification and political protest: Farmers’ protest in the Netherlands. European Journal of Social Psychology, 29, 1073–1095. Diener, E. (1980). Deindividuation: The absence of self-awareness and selfregulation 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 Journal of Social Psychology, 40, 399–416. (2002). Internet identifiability and beyond: A model of the effects of identifiability on communicative behavior. Group Dynamics, 6, 17–26. Douglas, K. M., McGarty, C., Bliuc, A. M., and Lala, G. (2005). Understanding 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 organizations, movement entrepreneurs, and the changing location of social movement 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 organizations 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

168


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? Investigating the group-based dynamics of the consensualisation process. European Journal of Social Psychology, 28, 755–776. 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 movements. 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 Psychology, 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 collective 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 sociopolitical 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.


169

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 computermediated 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). Argument and decision making in computer-mediated groups. Journal of Communication, 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 September. McDonald, M. (1999). Cyberhate: Extending persuasive techniques of low credibility 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). Collective 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 infrastructure 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.

170


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 visibility to the out-group. British Journal of Social Psychology, 33, 145–163. (1994b). On the consequences of deindividuation manipulations for the strategic communication of self: Identifiability and the presentation of social identity. 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 Psychology, 37, 15–40. Reicher, S. D., Spears, R., and Postmes, T. (1995). A social identity model of deindividuation phenomena. European Review of Social Psychology, 6, 161– 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 Social Psychology, 28, 623–645. Siegel, J., Dubrovsky, V., Kiesler, S., and McGuire, T. W. (1986). Group processes 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., Sturmer, 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., Sturmer, S., and Steffens, K. (2000). Helping individuals or group members? The role of individual and collective identification in AIDS volunteerism. 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. Sproull, L., and Kiesler, S. (1986). Reducing social context cues: Electronic mail 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-sourcemisses-the-point.html


171

Stott, C., and Reicher, S. D. (1998). Crowd action as intergroup process: Introducing the police perspective. European Journal of Social Psychology, 28, 509– 529. ¨ ¨ Sturmer, S., Simon, B., Loewy, M., and Jorger, 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. 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., 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 Psychology, 87, 649–664. 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 individuals’ 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 message, and there are socially generated responses. How do these mutual sources of influence affect users? New communication technologies are changing the manner of reception by which individuals acquire information from institutional, interpersonal, and peer information sources. Technology changes the temporal 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 communication. 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 “convergence” 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 convergence: 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 contemporary online information systems: peer communication and computational summaries of social behavior. How these forms and sources of information coalesce, confirm, or contradict one another asks for conceptual and analytic frameworks to address them. New technologies invite research that will advance understanding of how individuals conceptualize 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 articulation of an expanded perspective on the interplay of institutional, peer, and interpersonal sources through contemporary communication technologies, 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 communication, by contrast, involves smaller numbers of participants who exchange messages directed toward particular others. Interpersonal communication 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 instrumental 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 individuals’ opinions about politics, style, and other cultural issues is well known to involve a two-step flow of mass media messages and interpersonal discussions (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 interpersonal 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 Interpersonal 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 relationships (Reardon and Rogers, 1988), and helped to define the study of interpersonal communication. Mass media research evolved primarily 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 communication theories and research methods has developed so that scholars 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 communication technologies have the potential to merge the very processes conventionally considered as pertaining to mass communication or interpersonal communication, and that the merger of processes demands the merger of


175

approaches in order to understand such phenomena. For example, Reardon and Rogers (1988) suggested that new interactive media did not neatly fit into pre-existing areas of study. They claimed that a new epistemological 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 interactive 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” (Reardon 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 convergence 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, traditional 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 sometimes fail to realize their sub-disciplinary biases. For instance, interactivity, which is frequently mentioned in association with new technology, may connote different things for different analysts: new media are relatively 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). Others caution that analysis of emergent forms of internet communication 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 communication 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

176


(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 communication technologies, there have been some efforts to bring together specific mass and interpersonal processes over the years. These integrations provide stimulating launching points for reconsideration of communication processes in light of recent changes in the media and interpersonal landscapes. The following discussion reviews some exemplars, and suggests 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 discuss 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 transformed radically, in several ways, with dramatic implications. In Chaffee’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.


177

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 repositories 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 accessible on demand, and Hargittai (2002) found that, with experience, internet 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 information 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 information was selected by other (unidentified) computer users. In an experiment 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

178


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 technology’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 precedes 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 discussions co-appear with institutionally authored messages or other messages that bear the conventional characteristics of mass media? We wish to suggest that one of the most fruitful approaches to understanding new technology 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 contemporary 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 invoking 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 comparative persuasive potency with respect to co-present online sources of other types.


•

•

•

179

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 selfpresentation. 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 generate, 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).

180


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 message 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 adolescent 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 dominated 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 discussions 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 students to several groups, some who observed anti-drug PSAs without discussion and others who observed the PSAs and engaged in discussion 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 adolescents’ 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.


181

Although David et al. (2006) did not consider online chats to provide anything other than a methodological convenience for the capture of adolescents’ 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 positions 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 transcribed, 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 marijuana’s risk of harm. The results showed no differences on observers’ responses among the four videos themselves. The positive versus negative 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 subjects’ level of identification with the ostensible YouTube message posters

182


significantly mediated the effect of comments on the participants’ marijuana perceptions: The more that subjects identified with the anonymous peers who appeared to have posted the comments, the more their perceptions 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 influence 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 RateMyProfessor.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 systems as ways to help them make credibility evaluations. . . . Many participants indicated that they looked at the number of testimonials or reviews available online, paid attention to the proportion of negative to positive


183

reviews, or relied more heavily on negative versus positive reviews” (p. 420). The influence of web-based social comments on perceptions of individuals 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 opposing 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 congruent 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 individuals who created online profiles in social networking systems. For example, Facebook, a social networking site, allows users to place comments 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 profile 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 extraversion and perceived attractiveness, and the imputation of those characteristics implied by one’s peers via their wall postings, others’ comments override the profile owner’s claims (Walther et al., 2009).

184


System information Online systems aggregate information about users’ online behaviors, and some systems provide aggregation coefficients to users. Since these aggregations 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 professional 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, however, 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 aggregation 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 Facebook. 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 relationships (Parks, 2007). Nevertheless, there are apparently upper limits to the credulity of extremely large numbers. Tong et al. (2008) experimentally 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 attractiveness (with the drop-off occurring above 302). It appears that “too


185

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 influence 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 communication 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

186


inclinations. It works through the manner in which people can seek information 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 understand the issues that the media discuss. The framework suggests that people’s interest in issues drives their information seeking, and interpersonal discussions are subservient to the goal of information gathering and decision making. In contrast, Atkin (1972) demonstrated how interpersonal 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 individual’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 communicatory 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 information about that campaign, even after controlling for individuals’ level of interest in the campaign (as well as education level and socioeconomic 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 particular 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


187

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 information 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 communication 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 recommendation 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 specifiability 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 expanding the range of interpersonal goals one may consider, the potential of communicatory utility can go beyond helping us understand media consumption, to help illuminate issues of media information processing. We posit that the specific interpersonal goal(s) that prompt an individual’s media consumption shape attention to variations in the content and features of the topical information one consumes, affecting its interpretation and recall. For instance, collectively derogating political candidates or office-holders may be an activity that relational partners use to reinforce the similarity of their attitudes. This, of course, is not restricted to online news and online chats, but may be a general-purpose crossmedia 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.

188


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 instrumental, identity, or relational issues (Clark and Delia, 1979; Graham et al., 1980) in the service of needs for inclusion, affection, and/or control (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 attitudinal 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 processing may be different in traditional environments, where media exposure and interpersonal discussion are separated by some interval of time, compared with the new media environment in which mass and interpersonal 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 Wittenbaum 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., Herring 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


189

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 frequently, 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 na¨ıve 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 na¨ıve 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 various 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 information 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

190


by bringing to bear extant theories of social influence. A second analytic keystone is the recognition not only that interpersonal contacts motivate 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 status seeking, maintenance, or relationship initiation may bias information sampling from various media and affect the ultimate interpretations derived from them. These dynamics may be especially potent when conversations guide media consumption simultaneously, as the internet not only allows but promotes. References Atkin, C. K. (1972). Anticipated communication and mass media informationseeking. 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 communication 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 fundamental 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.


191

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 communication in a media world (3rd edn.). New York: Oxford University Press. Hargittai, E. (2002). Second-level digital divide: Differences in people’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. Journal 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 presented 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 characteristics 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.

192


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: Synthesis 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. Pingree (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 deindividuation phenomena. European Review of Social Psychology, 6, 161– 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


193

interpersonal impressions on Facebook. Journal of Computer-Mediated Communication, 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 Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 23, 3–43. (2007). Selective self-presentation in computer-mediated communication: Hyperpersonal dimensions of technology, language, and cognition. Computers 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 announcements 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). Selfgenerated versus other-generated statements and impressions in computermediated communication: A test of warranting theory using Facebook. Communication 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.

194


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 cooperative 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, imaginations, and ambitions dictate. Thus, anonymity has been and continues to be a variable of interest in the study of computer-mediated interaction. 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 others 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 misinformation 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 situations (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 psychology. 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 representation of the situation. We will elaborate on each of these points in turn, with frequent references 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 interactions, 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 communications, and personal information is often very accessible in computermediated 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 personal information. One can send an e-mail from a professional or a Hotmail 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 hobbies 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 fraudulent by most. Moreover, recent evidence suggests that people’s portrayals of themselves on online networking sites are reasonably accurate. Back

198


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 content, 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, creators 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 extraversion, 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 percent chance to win $2). In one version of the task, the payoffs were doubled if the partners chose the same bet (matching), whereas in another version the payoffs were doubled if they chose different bets (mismatching). Successful coordination led to increases in perceived similarity and liking in the matching task but not in the mismatching task. Whereas


199

this coordination work involved matching in deliberative choices, even unconscious matching in the form of behavioral mimicry increases rapport 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 understand 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 orientation 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 dispositional differences among people regarding perceived self–other similarity. We suggested earlier that, in general, people may be inclined to see others as being similar to them unless the available information suggests

200


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 similar. 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 visibility. We may feel anonymous and in control of what we disclose about ourselves in computer-mediated communications, but personal information 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 others 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-toface 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 embedding 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, something 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


201

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 dimensions of social interactions. Hollingshead (this volume) explores the emergence 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 functional 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 definitions of a group: interdependent and shared goals, interaction among members, and a sense of belonging – in short, the members view themselves 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 competition 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 status 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 levels 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 participants are often physically isolated. Sassenberg (this volume) suggests that physical isolation increases self awareness. He explores the implications 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

202


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 depending 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. Faceto-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. Nonetheless, people who respond quickly and participate frequently appear confident and are often more influential. There is some empirical justification 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-toface discussion are due in part to the physical presence of participants.


203

Turn-taking is a necessary feature of effective face-to-face communication 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 meetings often have pacing demands that are similar to face-to-face discussions. 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 electronic 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 influential when the situation does not provide clear guides for, or constraints on, behavior (Mischel, 1977). When social information is sparse and possibility 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 interact. On the other hand, it provides the opportunity to study how individual 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 construct different situations when there is little information to anchor their perceptions.

204


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 verbally aggressive and hostile behavior. Whereas the immediate emotional reaction to ostracism is the same online as it is when others are physically present, the authors report that subsequent coping behaviors in electronic interactions are more likely to be anti-social than in face-toface 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 figures. Social psychology has been criticized for studying behavior in contrived social environments where information is sparse and manipulated. 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


205

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 settings 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-toface social interactions does not deny that many kinds of online behavior are unique in their dynamics and scope. This new social world has attracted the attention of many scholars, and initial attempts to characterize 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 deindividuation as an explanatory concept. These accounts emphasized what happens 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 provides 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 justifying the white-supremacist views and less on attacking other racial and ethnic groups. While not denying that cyberhate happens, “hate” groups

206


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 reasons 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 websites were places for expressing and exchanging ideas. McGarty et al. used the metaphor of an editorial page in a newspaper to characterize 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 continuously. 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 communication converge on the internet. A dialogue on a blog is available for the masses to read. An ostensibly personal message posted to a Facebook 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 interpersonal 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


207

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 information 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, vendors 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 messages. 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 interactions. 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 emerging 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 technologies emerge, it would be useful to classify them in a comprehensive

208


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 (discloser 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 pacing). 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, deciding 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 technology 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 comprehensive way of describing similarities and differences among communication 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 interdependencies 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 personal and collective interests are compatible, but high discloser control


209

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 prefer 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. Different media are associated with cooperative and competitive interactions, give rise to different interpersonal impressions, and are differentially preferred by different personalities. In the spirit of person by situation interactionism, 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 cooperative 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 navigate 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 perceptions: Matching versus mismatching. Journal of Personality and Social Psychology, 95, 576–592. 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 selfidealization. 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 Psychology, 76, 893–910.

210


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 cooperators’ and competitors’ beliefs about others. Journal of Personality and Social Psychology, 16, 66–91. 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. Magnusson 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 Psychology, 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 faceto-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 accountability, 21 action groups, 149 activist groups, 150 activist identities, 148, 150 anonymity, 9, 18, 19, 21, 23, 24, 26, 28, 29, 36, 44, 49, 53, 65, 84, 115, 135, 136, 138, 151, 166, 195, 197, 198, 201 anonymity of others, 66, 67, 72, 75 anonymity of the self, 66, 77 anonymous cyberhate view of the internet, 145, 151, 205 anti-normative behavior, 136 Atkin, C. K., 186, 187 attitude change, 73 avatar, 31, 32, 53, 201 bargaining, 46 Battle of the Sexes, 91 Baumeister, R., 128 behavioral adaptation, 186 behavioral control, 92 behavioral game theory, 96 Berger, J., 111 Big Five personality traits, 198 blogs, 147 broadcast, 146 Cappella, J. N., 174 Chaffee, S. H., 176, 177 chat rooms, 131 choice dilemmas, 67 chronic private self-awareness, 74 citation counts, 184 CNET, 184 Cocktail Party World, 131 collaboration, 2 collective action, 24, 148, 149, 150, 151, 156 collective norms, 68

common bond groups, 68 common identity groups, 68 communication technologies, 108, 174 communicatory utility, 186, 187, 206 compliance, 70 computer-mediated communication, 2, 3, 6, 16, 19, 44, 57, 63, 86, 103, 114, 188, 206 computer-mediated interaction, 2, 84, 85, 87, 100, 103, 105, 195, 197, 203 computer-supported interaction, i, 1, 11, 85 convergent social technologies, 185 coordination problem, 93 coordination task, 198 cross-cultural, 33 cyber bullying, 138 cyber ostracism, 205 Cyberball, 130, 131, 132, 134 cyberhate, 166 David, C., 180 deception, 46, 47, 197 beliefs about internet deception, 49 deception online, 47 deviancy, 3 lie detection, 50 lies, 47 lying, 46 motivations for deception, 48 online deception, 52 online dishonesty, 52 reactions to internet deception, 48 decision timing, 98 deindividuation, 18, 19, 21, 136, 138, 151 deliberation, 104 deliberative action, 85 deliberative communication, 86 deliberative process, 105 DePaulo, B., 47, 50 depersonalization, 19, 23

211

212

Index

Dictator Game, 99 discloser control, 208 disclosure, 43 disinhibition, 49 Douglas, K. M., 145 dynamic interactionism, 9, 10 dynamic social context cues, 115 Eagly, A., 111 Edwards, C. E., 182 Edwards, R., 182 electronic teamwork, 8, 9 emergence of leadership, 112, 116, 120, 201 empowerment, 25 evolution of technology, 207 expectation states theory, 111, 117 experimental economics, 87, 88, 200 experimental game theory, 88, 104 Facebook, 183, 184 fate control, 92 Fishbein, M., 190 Free Software, 157 fundamental attribution error, 100 Gackenbach, J., 2 Galimberti, C., 5 game theory, 89 gathering, 146 gender, 16, 20, 26 gender concealment, 33 gender differences, 30 gender identity, 17, 27, 32 gender relations, 27 gender relevance, 27 gender salience, 27, 28, 30 gender-bending, 30 global activist proposition, 154 global activist view of the internet, 145, 206 group awareness, 72 group processes group cohesiveness, 29, 30 group consensus, 115, 159 group identity, 23, 24 group leadership, 122 group member participation, 113 group norms, 18, 19, 23, 26 group polarization, 22, 23 group size, 109, 113, 119 group stability, 119 group roles, 110 formal roles, 110

informal roles, 111 leader, 111 maintenance roles, 111 self-centered roles, 111 task-oriented roles, 111 group-based “collective assertions,” 159 group-based internet activism, 154 Heider–Simmel paradigm, 132 Ickes, W., 7 identity, 150 identifiability, 152 identity play, 48, 52 identity verification, 55, 56 identity-based pathway, 157 impression formation, 53, 101, 103, 199, 205 individual differences, 71 information processing information access, 178 information exchange, 72 information sources, 176 information-seeking, 177, 186, 188 interactionism, 63 interactivity, 175 interdependence, 89, 92 internal characteristics, 1 internet groups, 121 interpersonal processes interpersonal communication, 174, 206 interpersonal goals, 187, 188, 190 interpersonal inclinations, 186 interpersonal influence, 70, 71 interpersonal relationships, 45 Joinson, A., 3, 4, 154 Kelly, K., 180 Kiesler, S., 115 labels, 96 leader consensus, 117, 122 leader emergence, 109, 110, 111, 112, 114, 116, 119 Leary, M. R., 128 Lee, E.-J., 191 Levine, J., 114 Lewin, K., 1, 7 locus of control, 10 Mantovani, 4 mass communication, 173, 206 masspersonal communication, 175

Index McGarty, C., 145 McGrath, J., 113 McKenna, K., 2, 10, 44 media convergence, 173 media usage, 72 Minimal World, 132, 133 Moreland, R., 114 MoveOn.org, 162, 163, 165 Nash equilibrium, 90 needs need for belonging, 128, 135, 204 need for control, 135 need for meaningful existence, 135 need for self-esteem, 135 Netflix, 184 newspaper opinion page, 165 norm-based influence, 70 online online communication, 52 online extremist groups, 153 online groups, 108, 109, 110 online relationship, 41 online reputation systems, 55 online social interaction, 204 online social world, 204 online systems, 184 Open Source, 157 opinion-based groups, 148, 149, 150, 160 optimal heterophily, 177, 178 optimal similarity, 207 ostracism, 43, 127, 128, 129, 138, 204 cyber-ostracism, 129 face-to-face ostracism, 129 immediate effects of cyber-ostracism, 132 physical ostracism, 129 reflective reactions to ostracism, 134, 137 reflexive reaction to ostracism, 132 Pareto efficient solution, 90, 97 participation rates, 120, 201 payoff matrices, 94 permissible pacing, 208 person by situation framework, i, 1, 4, 6, 7, 11, 20, 56, 65, 72, 74, 75, 95, 195, 196, 203, 207, 209 personal websites, 43, 49 physical distance, 202 physical isolation, 66, 77, 201 priming, 23 Prisoner’s Dilemma Game, 90, 92, 97

213 private self-awareness, 70, 71 pro-social behavior, 134, 135 prototypicality, 30 PSAs, 181 pseudo-sequential order, 98 public goods game, 93, 97, 101 public self-awareness, 77 public service announcements (PSAs), 180 public visibility, 200 Qing, Q., 190 RateMyProfessor.com, 182 rational actor approach, 155, 157 rationality, 89 Reardon, K. K., 175 reduced social cues, 136 reflexive control, 92 relational goals, 189 relationships on the internet, 128 reputation systems, 56, 182 Ridgeway, C., 111 Riva, G., 2 Rogers, E. M., 175 sanctioning systems, 97, 104 self, 17 self-awareness, 201 self-categorization, 19, 20, 23, 29, 30, 147 self-efficacy, 45 self-esteem, 204 self-focused individuals, 73 self-monitoring, 8, 9, 73 self-perception, 147 self-presentation, 21, 49, 52, 179, 197 signature files, 121 simultaneous game structures, 98 Snyder, M., 7, 203 social attraction, 29 social capital, 45 social categorization, 29 social comparison orientation, 78 social context, 87 social creativity strategies, 153 social cues, 84, 86, 87, 103, 104 social exchange, 196 social identity, 5, 17, 18, 64, 65, 67, 68, 146, 148, 157, 159, 201 Social Identity model of Deindividuation Effects (SIDE), 5, 16, 17, 18, 21, 65, 66, 68, 84, 105, 145, 147, 152, 166, 196, 198, 199 social identity theory, 34

214

Index

social influence, 23, 63, 178 compliance, 67 informational influence, 64 interpersonal influence, 65, 68 norm-based influence, 64, 67 types of social influence, 64 social interactions, 201 social interdependence, 4 social isolation, 19, 22 social movements, 154, 157, 158, 159, 161, 162 social networking, 122 social networking sites, 54, 108 social role theory, 111 social value orientation, 72, 199 socially creative strategies, 153 source credibility, 177 Sproull, L., 115 status differences, 69 stereotypes, 112 stigmatized identities, 128 strategic and motivated user, 4 strategic and motivated user, expected and emergent effects framework (SMEE), 154 strategic decision-making situations, 89 strategic uses of social technologies, 4 superordinate group identity, 158

technological and social determinism social deterministic model, 3, 4, 8, 9 text messaging, 131 Thibaut, J. W., 92, 94 trust, 40, 41, 51 development of trust, 42 dispositional tendencies, 45 dispositional trust, 45 generalized trust of others, 46 global trust, 41 progression of trust, 42 specific trust, 41 training for trust, 56 trust scale, 42 Turner, J., 19, 147

Tajfel, H., 5, 146 tangible place, 200 task complexity, 113 team performance, 2

YouTube, 181

Ultimatum Bargaining Game, 99 uninhibited behavior, 136 virtual bravado, 136, 204 virtual reality, 2, 131 virtual selves, 200 virtual worlds, 53 Wagner, D. G., 123 Wahl, S., 190 Web 2.0, 182, 185 Webcams, 54 word-of-mouth marketing, 57 World Economic Forum (WEF), 161

Zaccaro, S. J., 109 Zelditch, M., 123

Strategic Uses of Social Technology: An Interactive Perspective of Social Psychology

Essentials of Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology of Visual Perception

Catholic Social Teaching: An Historical Perspective

Disappearance social in american social psychology

Psychology and social care

Doing Social Psychology Research

Categorization in Social Psychology

Social Psychology and Discourse

Social psychology and health

Gender and Social Psychology

Social Psychology 12th Edition

Social Psychology, 7th Edition

Social Psychology and Health

Psychology of Grandparenthood: An International Perspective

Social Psychology, 8th Edition

Small Groups [Social Psychology

Social Psychology and Discourse

Critical Social Psychology

Understanding Critical Social Psychology

Representations of Death A Social Psychological Perspective

Representations of Death: A Social Psychological Perspective

Social Facilitation (European Monographs in Social Psychology)

Exploring Social Psychology

Social Psychology, 10th Edition

Strategic Uses of Social Technology: An Interactive Perspective of Social Psychology

Essentials of Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology

Social Psychology of Visual Perception

Catholic Social Teaching: An Historical Perspective

Disappearance social in american social psychology

Psychology and social care

Doing Social Psychology Research

Categorization in Social Psychology

Social Psychology and Discourse

Social psychology and health

Gender and Social Psychology

Social Psychology 12th Edition

Social Psychology, 7th Edition

Social Psychology and Health

Psychology of Grandparenthood: An International Perspective

Social Psychology, 8th Edition

Small Groups [Social Psychology

Social Psychology and Discourse

Critical Social Psychology

Understanding Critical Social Psychology

Representations of Death A Social Psychological Perspective

Representations of Death: A Social Psychological Perspective

Social Facilitation (European Monographs in Social Psychology)

Exploring Social Psychology

Social Psychology, 10th Edition

Recommend Documents