_ mess

Q00?

This is to certify that the
dissertation entitled

CONFORMITY AND DISSENT IN COMPUTER-
MEDIATED GROUP DECISION-MAKING:
INTEGRATING INDIVIDUAL DIFFERENCES IN
SOCIAL IDENTITY RESEARCH

presented by

JUNGHYUN KIM

has been accepted towards fulfillment
of the requirements for the

DOCTOR OF degree in Telecommunications,
PHILOSOPHY Information Studies and Media

 

 

 

V Major P’rofess’or’s Signature
3 / 3 /0[
I /

Date

 

MSU is an Alfinnative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

_.-.-._._.

- '4 fl. ..

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/C|RC/Date0ue.indd-p.1

CONFORMITY AND DISSENT IN COMPUTER-MEDIATED GROUP
DECISION-MAKING:
INTEGRATING INDIVIDUAL DIFFERENCES IN SOCIAL IDENTITY
RESEARCH

BY

JUNGHYUN KIM

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Telecommunications, Information Studies and Media

2006

ABSTRACT
CONFORMITY AND DISSENT IN COMPUTER-MEDIATED GROUP DECISION-
MAKING:
INTEGRATING INDIVIDUAL DIFFERENCES IN SOCIAL IDENTITY RESEARCH
BY
JUNGHYUN KIM

The present study addressed critiques on manipulations and conceptual confusion
surrounding anonymity and group identity in most of social identity research. It also
relaxed the tacit assumption of the undifferentiated individuals and broke new ground
regarding the role of individual differences in social identity processes. Moreover, this
study compared Social Identity model of Deindividuation Effects (SIDE) and the
compromise view (i.e. uniqueness theory and optimal distinctiveness theory) to see if the
latter provided a better explain for some of the social identity research findings that were
inconsistent with SIDE.

Findings suggested that the similarity in self-representation with others, one of the
confounding concepts of anonymity, contributed to amplified group identity even without
a salient social identity shared by group members. This study also found that perceived
deindividuation and conformity intention showed a negative curvilinear relationship,
which supported the compromise view (i.e. uniqueness theory and optimal distinctiveness
theory). Finally, this study indicated that personality traits had a stronger impact on
conformity intention when there was an intense situational pressure. This research takes
the interactionist framework, which looks at the interplay between situational factors —

avatar conditions - and personality traits - Need for Uniqueness (N FU) and

Interdependent Self-Construa1(ISC) - in explaining human behavior. Findings suggested

that the relationships between these two personality traits and conformity intention were
moderated by the degree of similarity in the way people were represented in computer-

mediated groups.

Copyright by
Junghyun Kim
2006

ACKNOWLEDGMENTS

Looking back when I first came to Michigan State University to join the doctoral
program, I was a naive Korean girl who did not have any clue what she was getting into
and was at a loss in interacting with Americans. After five years, the same girl achieved
a doctoral degree with a little bit of knowledge in performing independent research,
teaching students, academic culture of American schools, and most of all, with an
acknowledgement that getting a doctoral degree is only the beginning of scholarly life.
This achievement would not have been possible without help and support from
wonderful people that I have met during the last five years.

First of all, I want to thank God for helping me finish up this dissertation through
all the difficulties and challenges I faced. He has been faithful all through the five years
in my Ph.D. program and always has been there for me even when I could not feel His
presence. I am really gratefiil to my family for their support and trust in me. I thank my
parents, Mr. Kim and Ms. Lee, for letting me make my own decisions in all the
important moments of my life, even though they have been concerned about me being
away from them. Thanks to my younger brother, Sungyeop, for his gentle and kind heart.
I am away from my parents and studying here in the US with peace, only because I
know that he is taking good care of my parents in Korea.

It has been such a blessing to have wonderful and supportive faculty members. I
want to thank Dr. Charles Steinfield, my advisor, for his kindness and support. During
my Ph.D. program, he always has been patient with my research questions and
administrative requests. Even with his busy schedule, Chip has allowed me to share my

academic concerns and gave me valuable insights on academic life. Especially, I want to

thank him for encouraging me to explore my own research areas and helping me grow
as an independent researcher.

I want to express my warm and sincere gratitude to Dr. Steve Wildman, a
wonderful mentor and fi-iend. I have not only learned how to do thorough research from
Steve by working with him, but also learned how to be an approachable and kind
teacher to students. Furthermore, I am grateful for sparing his precious time to listen to
me and answer my trivial questions with patience, encouraging comments, and humor.

I also thank Dr. Robert LaRose, who has been challenging me a with sharp mind
and sincere concern for my growth as a solid social science researcher. It has been a
great joy to work with him and observe his inquisitive and innovative mind in studying
social phenomena.

I have to thank Dr. Johannes Bauer. I was able to learn how to be a good scholar
and manager at the same time from his amazing capability to multitask. Even though he
was not on my dissertation committee, he was willing to make time for listening to my
difficulties and agony in my academic life.

I also want to thank Dr. Joseph Walther for his valuable advice and critique on
my experimental design. His work was one of the reasons that I decided to write my
dissertation on computer-mediated communication. Furthermore, I was so happy to have
chances to talk with Joe face-to-face and find out how kind and sincere a person he was.
I must thank Dr. Hee Sun Park for sharing her expertise on experimental design and data
analysis methods. She has been so generous and kind to make time for answering my
research related questions and giving me practical advice on surviving as a young

minority scholar. Thanks to Dr. Nojin Kwak for his help with data analysis and sincere

vi

advice on finding a job as a minority student. Thanks to Dr. Hairong Li who provided a
valuable critique to my dissertation.

Finally, I cannot leave out my dear friends. Thanks to Su who designed and
built up a wonderful website for my dissertation experiment. Thanks to Anand, Finny,
Grace, Maureen, Rachel, and Rex & Vangie for their beautiful friendship and
motivating me to become a better person. Thanks to Dr. Bell, Chica, David & Dillon,
Dozie, John & Inge, Jonathan & Kimmie, Raymond & Shella, and Songyi for their
limitless encouragement and prayers. Without their emotional and spiritual support, I

would not be able to overcome my discouragernents and difficulties in life.

vii

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... x
LIST OF FIGURES ................................................................................. xi
INTRODUCTION .................................................................................... 1
LITERATURE REVIEW ........................................................................... 5
Concepts ....................................................................................... 5
Computer-Mediated Communication (CMC) ................................... 5
Conformity as a group norm in computer-mediated discussions ............. 5
Deindividuation theory and Social Identity model of Deindividuation Effects
(SIDE) ........................................................................................ 7
Challenges to SIDE and social identity research ........................................ 9
Anonymity manipulations ......................................................... 9
The ambiguity of group identity ................................................ 10
Deindividuation and perceived deindividuation .............................. 12
The compromise view ..................................................................... 13
Uniqueness theory ................................................................. 14
Optimal distinctiveness theory (ODT) .......................................... 14
The inverted U-shape curve ...................................................... 17
Interactions between perceived deindividuation and avatar conditions ............ 18
Personality variables ....................................................................... 19
Need for Uniqueness (N FU) ..................................................... 20
Interdependent Self-Construal (ISC) ........................................... 21
Interactions between personality traits and avatar conditions ............... 22
METHODS .......................................................................................... 26
Procedure .................................................................................... 26
Measures .................................................................................... 33
FINDINGS ........................................................................................... 35
Main effect of avatar conditions on group identity .................................... 35
Moderated Multiple Regression (MMR) ................................................ 36
Conformity intention as a function of group identity and perceived
deindividuation ................................................................................ 3 8
Perceived deindividuation and avatar conditions ...................................... 39
Personality traits and avatar conditions ................................................. 44
DISCUSSION ....................................................................................... 51
CONCLUSIONS .................................................................................... 55
Summary and implications ............................................................... 55

viii

Limitations and future studies ............................................................ 58

BIBLIOGRAPHY ................................................................................... 62

ENDNOTE ........................................................................................... 69

ix

LIST OF TABLES

Table 1: Inter-scale correlations and reliability of five scales ................................. 34
Table 2: Mean values of group identity of three different avatar conditions ................ 36

Table 3: The significance test of group mean differences in group identity across three
different avatar conditions .......................................................................... 36

Table 4: Conformity intention as a function of group identity and perceived

deindividuation ....................................................................................... 39
Table 5: Interaction between perceived deindividuation and avatar conditions ............ 41
Table 6: Interaction between NFU and avatar conditions ..................................... 45
Table 7: Interaction between ISC and avatar conditions ....................................... 49

LIST OF FIGURES

Figure 1. The inverted U-shaped curve between perceived deindividuation and conformity

intention ............................................................................................... 18
Figure 2. A screenshot of the different-avatar condition ....................................... 28
Figure 3. A screenshot of the same-avatar condition ........................................... 29
Figure 4. A screenshot of the no-avatar condition .............................................. 29
Figure 5. A Screenshot of the chat room window of the different-avatar condition ....... 32
Figure 6. A Screenshot of the chat room window of the same-avatar condition. . . . . . . . ....33

Figure 7. Conformity intention by perceived deindividuation and avatar conditions. . ....44
Figure 8. Conformity intention by NFU and avatar conditions ............................... 47

Figure 9. Conformity intention by ISC and avatar conditions ................................. 50

xi

INTRODUCTION

Developments in communication technology have extended human
communication in many ways. Communication mediated by electronic computing
technology not only extends human communication by helping overcome limitations of
time and space, but also alters the ways in which people interact with each other. Indeed,
the rapid dissemination of computer-mediated communication (CMC) technologies has
generated many research questions, especially, related to how CMC might produce
different interaction patterns in group processes (Walther & Burgoon, 1992).

One of the most important factors that affect human interactions in CMC is the
lack of social cues (Sproull & Kiesler, 1991). Text-based CMC lacks physical cues, such
as gesture, voice tone and facial expressions, which can supplements interactions
between people in face-to-face (FtF) communication. In the past, this lack of social cues
in CMC was expected to reduce social influences coming from social norms and
standards, and in turn, encourage uninhibited behavior (Sproull & Kiesler, 1991).
Additionally, the lack of such cues was associated with a democratization of hierarchical
relationships, enabling people to express themselves more openly due to their anonymity
and depersonalization (Kim, 2000). However, a prominent theory, the social identity
model of deindividuation effects (SIDE) argues that the reduced social cues afforded by
CMC can actually make people be more responsive to a computer-mediated group’s
norm rather than behaving in anti-nonnative ways. A precondition of this conformity to
group norms is that people should identify with other members of the computer-mediated

group. According to SIDE, reduced individual differences caused by the limited social

cues in CMC heighten the feeling of group membership (i.e. group identity) and lead to
increased conformity to group norms.

Much CMC research on group identity and conformity focuses on effects linked
to the lack of social cues and anonymity, but has not paid attention to the influence of
individual differences on group identity and conformity (Postrnes, & J etten, 2006).
However, several social identity researchers has argued for relaxing the assumption that
computer-mediated features have affect all people in the same way in computer-mediated
groups, without regard for individual differences (Brewer, 1993; Lee, 2004; Postmes &
Jetten, 2006). They suggested that peoples' perceptions of the situation and their inherent
personality traits would affect group identity and conformity in computer-mediated
groups.

Therefore, the purpose of this dissertation is to investigate how computer media
features, people’s perception of the situations in which they are located and their inherent
personality traits might affect group identity and conformity in group discussions. This
study expands previous social identity research in CMC contexts, which mainly focused
on the impact of computer-mediated features and environments on human behaviors, into
the realm of looking at the interaction between human factors and situational factors. It
also points out challenges to the existing social identity research - to clarify confusions in
operationalizing core concepts, anonymity and group identity, and to integrate individual
differences in social identity research — and suggests solutions.

In attempts to resolve these issues, first of all, this study distinguishes between
confounding constructs of anonymity, unidentifiability (i.e. no release of personal

information) and similarity (i.e. uniform self-representation among group members)

(Baron, 1971; Lee, 2004; Snyder & Fromkin, 1980), and focuses on the latter rather than
the former. “Avatars,” icons or representations of a person in a shared virtual space
(Wikipedia.com, 2006), are used as a way to vary the level of similarity in virtual self-
representation while holding absolute unidentifiability across different similarity
conditions. This study also distinguishes a stable group identity from a transient group
identity (Lea, Spears, Watt, & Rogers, 2000) in order to focus on the impact of similarity
in self-representation on the transient group identity.

In order to find a better theoretical framework integrating individual differences
in computer-mediated group discussions, this study compares two different theoretical
fiameworks, SIDE and the “compromise view” offered by Blanton and Christie (2003).
As with other social identity theories, like self-categorization theory (Turner, Hogg,
Oakes, Reicher, & Wetherell, 1987), SIDE exaggerates inter-group contrasts while
minimizing intra-group differences (Brewer, 1993; Postmes & J etten, 2006). However,
assimilation within an in-group has received less empirical support (Brewer, 1993). On
the contrary, some research has indicated that people who were made highly similar to
others within the in-group pursued higher differentiation from others (Codol, 1984;
Cooper & Jones, 1969; Kim, 2006; Lee, 2004; Maslach, 1974). These research results
highlight the existence of different individual needs that could exhibit diverse reactions to
a highly deindividuated condition (Lee, 2004). They diverge from SIDE, but conform
more to the compromise view (Blanton & Christie, 2003), which consists of uniqueness
theory (Snyder & Fromkin, 1980) and Optimal Distinctiveness Theory (ODT) (Brewer,
1991). The compromise view promotes individual uniqueness within intensely

standardized conditions. Thus, the present study seeks to compare SIDE with this

compromise view in order to see if the latter is able to provide a better argument for such
inconsistent findings and integrating individual differences into social identity research.

Based on these theoretical frameworks, hypotheses are introduced and tested to
assess the effect of situational factors, people’s perception of the situation in which they
are located, and individual personality traits (i.e. need for uniqueness and need for
similarity) on conformity intention in a group discussion. Moreover, this study examines
interactions between situational factors and human factors in affecting conformity
intention.

Research methods and results are discussed in detail, which is followed by the
theoretical implications and contributions of this study. This dissertation found that not
only computer media features, reduced individual cues and uniform representation of
people, but also people’s perception of how indistinguishable they were from others and
their desire to be different from or similar to others could affect conformity intention in
computer-mediated group discussions. In addition, it was found that people could dissent
more rather than conform when their uniqueness was oppressed within standardized self-
representation formats in computer-mediated environments. Such findings suggest the
existence of the effect of individual differences in social identity process and contradict
to SIDE, which puts limits on individual differences allowed in social identity research
and focuses more on environmental factors’ effects on human interaction in group
settings. This study compared the dominant theoretical framework in social identity
research, SIDE, with competing theoretical frameworks, uniqueness theory and optimal
distinctiveness theory, hoping to improve our understanding in conformity or dissent

behavior in computer-mediated group processes.

LITERATURE REVIEW

Concepts

Computer-Mediated Communication (CMC)

“Computer-mediated communication”, CMC means communication mediated by
machine, computer, which is relatively indirect means of interaction. Usually the medium
in existing CMC studies is text based rather than visual or verbal communication
channels. According to Rafaeli, CMC could be either interpersonal or group, but not mass

(Rafaeli & Sudweeks, 1997).

Conformity as a group norm in computer-mediated discussions

Among many human interactions, conformity has been one of the most studied
behaviors in computer-mediated group contexts. Conformity or consensus reaching could
be a type of group norm in the context of group decision-making. A group norm is
defined as “a shared expectation about how the members of a group ought to behave”
(Levine & Moreland, 1990, p.600). Members’ behavior in accordance with a group norm
is interpreted as conformity to the norm, while irregular behavior is considered as
deviance from the norm. Group norms are situationally and locally defined, so they tend
to be independent from existing social norms (Postrnes & Spears, 1998). Computer-
mediated groups tend to build up new rules and norms, since the formation of those
groups are not always based on the predefined social norms. The most common way to
compensate for the absence of preexisting norms is to infer latent group norms from the

common or predominant behavior of other members (Postrnes & Spears, 2000; Turner,

1982). Therefore, computer mediated group norms are formed through the interaction
among group members and the observation of other members’ behavior.

The influence of norms on group decision-making has not been investigated in
many studies (Postmes, Spears, & Cihangir, 2001). However, it is possible to infer the
functions of norms on a group decision-making based on some of the previous small
group decision-making literature in both social psychology and some of group decision-
making support system (GDSS) studies. Most of the GDSS literature is based on the
assumption that group decision-making performance would be improved when any kind
of social influences is diminished through anonymity. All the factors that are thought to
prevent individual members from freely expressing their opinions are named as
“dysfimctions” in the group decision-making process. For this reason, pressure to achieve
consensus or conformity in the group decision-making process has been considered as
one of the negative dysfunctions (Steiner, 1972).

On the other hand, a group of social psychology research (Baron, Kerr & Miller,
1992; Levine & Moreland, 1990; Postmes, Spears, Sakhel, & de Groot, 2001) perceives
the regulatory fimction of group norms as a positive element. For example, some groups
promote consensus and emphasize members’ conformity to achieve members’ agreement
on issues central to the group (Turner, 1991). In addition, norms promoting consensus
and conformity can facilitate group performance, such as willingness to invest effort on
achieving group’s goal and augmented group identity (Postrnes & Lea, 2000). The
present research used consensus as a group norm in order to create an environment that
can highlight individuals’ conformity and dissent decision. Even though these two groups

of research focus on contradictory effects of CMC, they both focus on the “reduced social

cues”, which has been known as the most important factor in accentuating or decreasing

individual voices in CMC research.

Deindividuation Theory and Social Identity model of Deindividuation Effects

(SIDE)

Deindividuation is defined as a psychological state of decreased self-awareness
that reduces the influence of social norms and induces an antinorrnative behavior (Diener,
1977; 1980; Lee, 2004; Zimbardo, 1969). Following such definition, deindividuation
theory (Zimbardo, 1969) argues that individuals behave in anti-normative ways when
they are located in an anonymous environment. Based on this theoretical framework,
some of early studies on computer-mediated groups showed that a lack of social context
cues in CMC led to an anonymous environment in which people’s apprehension for
evaluation decreased and finally led to uninhibited behaviors (Kiesler, Siegel, & McGuire,
1984; Sproull & Kiesler, 1986).

However, Reicher et a1. (1995) proposed that the lack of individuating cues or
anonymity should reduce the focus on individual differences while increasing
identification with a situational group, which became a foundational idea of the social
identity model of deindividuation effects (SIDE). Postrnes and Spears’s meta-analysis
(1998) supported Reicher et al.’s research by showing that people who were deprived of
their individuating cues conformed to situational norms (i.e. norms of the group to which
people were assigned), which did not necessarily conform to general social norms.

SIDE is based on social identity theory (Tajfel, 1978; Tajfel & Turner, 1986) and

self-categorization theory (Turner, 1981). According to the social identity theory, a

person’s self-conception consists of two parts: social and personal identities (Taj fel,
1978). The social identity refers to self-descriptions related to group memberships, while
the personal identity refers to personality traits and attributes of individuals. Personal
identity and social identity are different levels of self-categorization and these two levels
relate to each other in an antagonistic way. If people focus more on one level of the two,
the less they focus on the other (Turner et al., 1987). This is why the social identity
theory defines the group identification as a “shift toward the perception of self as an
exemplar of some social category and away from the perception of self as a unique
person” (Turner et al., 1987, p.50). Once people categorize themselves as group members,
similarities among ingroup members are accentuated and they seek to favor their ingroup
identities (Terry & Hogg, 2001).

SIDE explains social identity and social influences in groups. It has been applied
to computer-mediated communication (CMC) in group settings (Lea & Spears, 1991;
Postrnes & Lea, 2000; Postrnes & Spears, 1998; Postmes, Spears, & Cihangir, 2001;
Postmes, Spears, & Lea, 1998; Postmes, Spears, & Lea, 1999; Postmes, Spears, Sakhel,
& de Groot, 2001) where reduced contextual cues accentuate conformity to group norms
when a social identity is salient (Postrnes et al., 1999; Reicher, 1987; Reicher, Spears, &
Postmes, 1995; Spears, Lea, & Postmes, 2000). However, SIDE has been criticized for
confusion in operationalizing core variables, anonymity and group identity (Lea, Spears,
Watt, & Rogers, 2000; Lee, 2004; Snyder & Fromkin, 1980), and for its failure to
integrate individual differences in in-group settings (Brewer, 1993; Postrnes & Jetten,

2006)

Challenges to SIDE and Social Identity Research

Anonymity Manipulations

Visual anonymity, which is caused by a lack of social context cues (Kiesler et al.,
1984), has been the most important predictor of social influence in SIDE research. There
are two different constructs in anonymity: unidentifiability and similarity (Baron, 1971;
Snyder & Fromkin, 1980; Lee, 2004). Most social identity studies have focused on the
unidentifiability aspect of anonymity, but their manipulations of anonymity tended to
confound unidentifiability and similarity aspects. For example, one of the most popular
anonymity manipulations of social identity research involved dressing participants in the
same mask or overall while leaving participants in their normal clothing for an
individuated condition (Zimbardo, 1969). In CMC studies, anonymity was manipulated
by allowing participants to communicate with each other only through standardized text-
only CMC environments. However, these anonymity operationalizations not only
deprived people of their individual information (i.e. unidentifiability) but also presented
them in visually similar forms (i.e. similarity).

Therefore, in order to distinguish the effects of unidentifiability from those of
similarity, the present study used avatars, as a way to vary the level of similarity in virtual
self-representation, while holding absolute unidentifiability across different similarity
conditions. Given that avatars can play the role of clothing in cyberspace, being
represented by the same avatar as other members should decrease group members’
individualities while increasing the level of deindividuation. However, if people are
distinguished by their unique virtual self-representations, they can preserve more

individuality in CMC. Participants in most of the previous CMC experiments were

destined to lose their individualities, since they were placed in text-only CMC conditions
and did not use any unique virtual symbols differentiating them from others. Studies by
Lee (2004) and Lee & Nass (2002) found that the similarity in visual representation
contributed to accentuated deindividuation in computer-mediated groups. In keeping with
this distinction, this study manipulated three different avatar conditions in order to vary
the degree of situational deindividuation: 1) the same-avatar condition: a group in which
all members share the same avatar, 2) the no-avatar condition: a group in which members
do not use any avatar, and 3) the different-avatar condition: a group in which each
member has unique avatar.

Certain visual cues that reflect social categories, such as gender or ethnicity, are
known to facilitate social identification processes (Postrnes et al., 1998). However, the
present study excluded such socio-categorical information from avatars, as was done in
Lee’s study (2004). Uniform self-representation in cyberspace might not be as explicit or
as powerfiil as labeling people according to their common social category (i.e. a salient
social identity), but the similarity in the way people are represented can be a basis for
psychological group formation and might still trigger identification with ad-hoc groups
(Lee, 2004; Turner, 1984).

H1: People in the same-avatar condition will show the highest group

identify, followed by those in the no-avatar condition and then
by those in the different-avatar condition.

The Ambiguity of Group Identity

SIDE predicts that a lack of individual cues will heighten conformity to group

norms only when there is a common social identity shared among group members

10

(Postrnes et al., 1999). However, the mediation effect of the salient social identity
connecting anonymity and conformity needs further clarification and investigation (Lea
et al., 2000; Postmes, 1997; Postrnes et al., 1999). Group identity refers to cognitive
identification with a social category without being dependent upon interactions between
individual group members (Reicher et al., 1995; Turner, 1982). This definition implies
that the formation of group identity does not necessarily require actual interaction of
group members. For example, one of the manipulation methods was to tell one group of
participants that researchers were interested in them as members of certain groups (i.e.
departments or schools) vs. to tell another group of participants that researchers were
interested in them as individuals (Spears, Postmes, Lea, & Watt, 2001).

Social identity studies in CMC have included a wider level of social
categorization as a factor mediating the impact of anonymity on conformity to group
norms. In such cases, whether anonymity in CMC had a primary effect on people’s
identification with their common social category (i.e. a stable group identity such as
nationality, gender, political association, etc.) or with an immediate interacting
experimental group (i.e. a transient group identity) was not clear].

In attempts to resolve these issues, Lea et a1. (2000) discerned the effect of a
stable group identity from that of a transient group identity. They had one group in which
the stable group identity (i.e. German nationality) was aligned with the experimental
group identity (i.e. all Germans in the same group) and the other group in which the
stable group identity cut across the experimental group identity (i.e. participants of

diverse nationalities in the same group). The stable group identity did not have a

 

I While a stable group identity, or a pre-existing longstanding category, refers to an identity that is likely to
endure and less likely to be affected by contextual conditions, a transient group identity can be easily
formed and changed depending on changes in communication environments (Lea et al., 2000).

11

significant effect on participants’ identification with the experimental group (Lea et al.,
2000). Meanwhile, Lee and Nass (2002) also found that people could associate with
others who shared the same visual self-representation (i.e. avatar) as theirs without a
common prior group identity. That is, a group composed of participants from different
schools nevertheless showed a greater group identity and conformity to group norms than
another group composed of participants from the same school (Lee, 2004; Lee & Nass,
2002). According to these studies, the transient group identity could even overcome the
social categorical differences (Lee, 2004), which challenges the notion that a salient
common group identity shared by participants is a necessary condition for increased

conformity.

H2: Group identity that is formed by visual self-representations without
a salient social identity will promote conformity intention.

Deindividuation and Perceived Deindividuation

Deindividuation has been treated as a categorical variable in most social identity
research (Kim, 2006; Lee, 2004). This is because participants in an identifiable group are
assumed to feel the same level of lower deindividuation and those in an unidentifiable
group are assumed to feel the same level of higher deindividuation. However, such an
indiscriminate effect of deindividuation on individuals, one of the tacit assumptions made
by SIDE, was challenged by a group of studies indicating that even identical
deindividuation manipulations could induce various psychological states for different
individuals (Cooper & Jones, 1969; Kim, 2006; Lee, 2004; Maslach, 1974). In attempts

to explain these findings, which were inconsistent with SIDE, this study used a

12

“perceived deindividuation.” It was derived from Maslach’s (1974) definition of
deindividuation as “a state in which people feel indistinguishable from other people.”
Perceived deindividuation is clearly distinct from avatar conditions, since the former
reflects individual differences in perception while the latter refers to the level of objective

similarity in the way participants are represented.

The Compromise View

SIDE focuses on communication contexts and their effects on a social
identification process but limits individual differences that might affect the process
(Postrnes & Jetten, 2006). On the other hand, uniqueness theory (Snyder & Fromkin,
1980) and Optimal Distinctiveness Theory (ODT) (Brewer, 1991) accept the existence of
individual differences and take them into consideration in explaining the social identity
process. These two theoretical frameworks are called the “compromise view” (Blanton &
Christie, 2003) and suggest that a person feels uncomfortable if the situation in which
he/she is located either boosts or oppresses his/her individuality too much. According to
this perspective, a person would resist against the situation in which he/ she loses too
much of his/her individualities and tries to be distinguishable from others. As Maslach
(1974) said, “being similar or different is all right up to a point, but beyond that it is
considered deviant and bad” (Maslach, 1974, p.413). In line with Maslach’s argument,
Snyder and F romkin (1980) proposed that there is an appropriate level of uniqueness,
which differentiates people not too much or not too little from others. Based on such

principle, Snyder and Fromkin came up with uniqueness theory that defined “unique” as

13

a moderate degree of similarity instead of extreme differentiation from others (Blanton &

Christie, 2003).

Uniqueness Theory

While SIDE suggests that increased deindividuation, through unidentifiablity or
similarity, promotes identification with a group and eventually conformity to the group
norms, uniqueness theory argues that people show the most positive emotional reactions
or conformity to others when they feel a moderate amount of similarity relative to others
(Snyder & Fromkin, 1980). This is because uniqueness theory considers deindividuation,
or undistinctiveness of self, as an attack on ego identity (Duval & Wicklund, 1972;
Snyder & Fromkin, 1980; Zimbardo, 1969). Loss of identity can occur by being in a
crowd, disguised, masked, or dressed in a uniform (Zimbardo, 1969). In those conditions,
people try to manifest their uniqueness by disagreeing with others if they perceive a large
amount of similarity relative to others. This results in an inverted U—shaped curve
between their similarity to others and their positive reactions (i.e. agreement) to others

(See Figure 2 in Snyder & Fromkin, 1980, p. 35).

Optimal Distinctiveness Theory (ODT)

Optimal Distinctiveness Theory (ODT) is an “amalgamation of self-categorization
theory and uniqueness theory” (Brewer, 1991 , p. 3). Brewer (1991) developed the
Optimal distinctiveness theory based on uniqueness theory and tried to find out why

people looked for equilibrium between extreme deindividuation and extreme

l4

personalization. According to ODT, there are two conflicting human motivations that
drive people’s conformity behavior: the need to belong (Baumeister & Leary, 1995;
Maslow, 1968) and the need to be different (Snyder & Fromkin, 1980). The former refers
to “a need for inclusion of the self into larger social collectives”, while the latter refers to
“a need for differentiation of the self from others” (Brewer, 1993, p.3). The need for
intimacy or belonging is universal and primitive, so to be rejected or ostracized by other
people is one of the highly negative experiences for human beings (Williams, 2001). On
the other hand, there is another polar of the need to see oneself as a distinctive being.
Upholding the sanctity of the self and being true to the self have been cultural norms in
Western society (Baumeister, 1991). So, if people submerge their individualities in order
to firlfill group responsibilities or maintain the peace within the group, such behaviors
have been considered as negative elements to personal growth (Baumeister, 1991;
Wallach & Wallach, 1983). When people are located in a group, the existence of such
opposing forces creates tension in constructing social identity in the group, so they try to
find the optimal point where they can satisfy both needs (Brewer, 1991). The person is
not comfortable if he/she is too distinctive from others or too assimilated to others. Thus,
he/she tries to find the optimal point where he/ she can satisfy both needs for similarity
and uniqueness at an appropriate level (Brewer, 1991).

Then how can a person find a balance between the two conflicting needs?
Homsey and Jetten (2004) suggest several ways to balance the needs to belong and the
needs to be different. One way is to identify with a unique group that is different from
other mainstream groups. By joining the group that is clearly distinctive from other

groups and has a strong sense of cohesiveness, people can experience their sense of

15

(ingroup) belonging and promote (intergroup) distinctiveness at the same time. Another
way is to use subgroups within a subordinate group. For example, a large group can be
divided into smaller subgroups based on profession, socioeconomic status, gender,
religion, or ethnicity, etc., which can promote individual members’ sense of
distinctiveness while maintaining subordinate group identification. These two strategies
of balancing the two conflicting needs are based on ODT (Brewer, 1991) arguing that
people can achieve the need for inclusiveness by being a member of a certain group while
maintaining the need for distinctiveness by selecting a group that is distinctive from other
groups. Thus, ODT’s solution for balancing the need to belong and the need to be
different is to achieve distinctiveness in the group level rather than in the individual level
within a group. However, in reality, people would not have many choices if balancing
need for distinctiveness and inclusiveness can be resolved only through finding and
identifying with a unique group (Homsey & Jetten, 2004). This group level solution in
balancing the two conflicting needs misses an important research area: How do
individuals satisfy their need to be different from others within their own groups?

Even though ODT is rooted in self-categorization theory and uniqueness theory,
ODT is different from these two theories. Both ODT and uniqueness theory predict that
people pursue moderate levels of differentiation from and similarity to others. However,
ODT indicates that the need for differentiation is met through increasing inter- group
distinctiveness (Brewer, 1991), while uniqueness theory applies the same logic to
individual levels of self-representation (V ignoles, Chryssochoou, & Breakwell, 2000).
Meanwhile, ODT is also different from self-categorization theory in the sense that the

former concentrates on motives or drives (i.e. the need for differentiation and the need for

16

inclusion), rather than personal identity and social identity. According to SIDE, personal
identity refers to the perception of one’s self as a unique individual, while social identity
corresponds to the perception of one’s self as an example of some social category (Turner
et al., 1987, p.50; Reicher et al., 1995). According to ODT, personal identity represents
maximal satisfaction of the need for differentiation, while social identity indicates

satisfaction of the need for inclusion (Brewer, 1993, p. 3).

The Inverted U-shaped curve

Uniqueness theory and ODT are termed as the “compromise view” (Blanton &
Christie, 2003) because they suggest that people feel most comfortable when they
achieve equilibrium between their conflicting needs for inclusion (similarity) and for
uniqueness (difference). Applying the compromise view to a group decision-making,
people experience two competing influences: conformity and dissent (Baumeister, 1982;
Guerin, 1986; Snyder & Fromkin, 1977). Conformity is a type of behavior reflecting a
person’s desire to be similar to others, while dissent is a type of behavior reflecting a
person’s desire to be different. Therefore, according to the compromise view, if people
perceive extremely high or low levels of deindividuation in a situation, they should not
show as much conformity intention as they would if they are to perceive a moderate level
of deindividuation (V ignoles et al., 2000). This results in an inverted U-shaped
curvilinear relationship between people’s perceptions of how similar they are to others
and their positive reactions to others (Figure 1). In this study, conformity intention was

used to measure peoples’ positive reactions to others.

17

H3: The relationship between perceived deindividuation and conformity
intention will be an inverted U-shaped curve.

Figure 1. The inverted U-shaped curve between perceived deindividuation and
conformity intention

Confon'nity Intention

 

 

Perceiwd Deindividuation

Interactions between Perceived Deindividuation and Avatar Conditions

Similarity in self-representation and perceived deindividuation are treated as two
different variables, provided that the former is an objective situational condition given to
participants, while the latter is participants’ perceptions on how similarly they are
represented to others in the condition. Therefore, it is possible to look at the moderating
effect of avatar conditions on the relationship between perceived deindividuation and

conformity intention. That is, the inverted U-shaped curve between perceived

l8

deindividuation and conformity intention can be altered in its shape and direction as
situational conditions change.

Because the intensity of objective similarity is highest in the same-avatar
condition, participants should be more sensitive toward the highly deindividuated
situation and try to differentiate themselves from others more than those in conditions of
less similarity (i.e. the different-avatar condition or the no-avatar condition). Therefore,
while people should show an increase in their conformity intention as their perceived
deindividuation increases up to a moderate level, as shown in Figure 1, those who are in a
condition of intense objective similarity among participants will rather show a decrease in
their conformity intention, even with the slight increase of their perceived

deindividuation.

H4: The relationship between perceived deindividuation and conformity
intention will be modified by avatar conditions such that the inverted
U-shaped curve will be clearest in the least deindividuated condition
(the different-avatar condition) and become less manifest as the level
of objective similarity increases in the no-avatar condition and in the
same-avatar condition.

Personality Variables

In Lee’s (2004) study, both being represented by the same virtual uniform and
being members of the same social group led participants to differentiate themselves from
the rest of the group instead of increasing conformity among in-group members.
Participants' conformity decreased because of the heightened objective similarity, but

such situational pressure could have also intensified participants’ needs for uniqueness

l9

(Lee, 2004). This logic is in accordance with uniqueness theory suggesting that striving
for uniqueness is influenced by both perceptions of similarity between self and others and
dispositional individual differences in uniqueness motivation (Breakwell, 1987; From,
1941; Maslow, 1968; Snyder & Fromkin, 1980, p. 77). Therefore, this study took into
consideration inherent differences in motivation, an aspect which has been restricted in

most social identity research (Brewer, 1993; Lee, 2004; Postrnes & J etten, 2006).

Need for Uniqueness (NF U)

Need for Uniqueness (N FU) is also called a ‘distinctiveness principle,’ which is
defined as “a motive within identity pushing toward establishing and maintaining a sense
of differentiation from others” (Vignoles et al., 2000, p.337). Researchers such as
Breakwell (1987), From (1941), Maslow (1968), and Snyder and Fromkin (1980) argue
that a person has a need to realize and express his/her unique and individual self. In this
argument, distinctiveness or uniqueness is considered as a fundamental human need as it
is. On the other hand, the distinctiveness principle has been also considered as one of the
antecedents of self-enhancement, in which positive self-esteem is achieved by
establishing individual distinctiveness (Breakwell, 1987; Wills, 1991). Either as a mean
of self-enhancement or as a fundamental human need, distinctiveness is an important
property of self-definition (Vignoles, et al., 2000). In order to fulfill this distinctiveness
need, people try to differentiate their appearance and behavior from those of others,
which is believed to attract attention and enable them to stand out from the crowd. People
who have higher need for uniqueness (N FU) (Fromkin & Snyder, 1980; Snyder, 1992)

get satisfaction from the perception that they are unique, special, and separable from the

20

masses (Fromkin & Snyder, 1980). Therefore, individuals with high NFU tend to be more
sensitive to the degree to which they are seen similar to others and want to exhibit

behaviors that could establish their distinctiveness from others (Snyder, 1992).

H5: Need for uniqueness will have a negative relationship with conformity
intention.

Interdependent Self-Construal (ISC)

As much as people of high NFU pursue a dissonance with others by resisting
highly standardized situations, there must be those who want to be assimilated.
Interdependent Self-Construal (ISC) measures people’s need to belong. Self-construal
refers to “the extent to which people view themselves either as individuated entities or in
relation to others” (Markus & Kitayarna, 1991, p. 226). There are two types of self-
construals: independent and interdependent self-construals. Independent self-construal is
a trait that makes people view themselves as autonomous, unique and distinctive beings,
which encourages unique attributes of individuals and the pursuit of one’s own goals
(Markus & Kitaymama, 1991; Singelis, 1994). On the other hand, interdependent self-
construal emphasizes one’s connectedness with others and sensitivity to situation or
social contexts (Singelis, 1994). Therefore, while NFU is a trait urging uniqueness and
difference from others, ISC emphasizes people's connectedness with others and their
acceptance by others (Singelis, 1994). People with strong ISC define themselves by their

reference groups and try to behave in accordance with the thoughts and behaviors of

21

others. Therefore, people high in ISC should show a higher conformity intention than

those low in ISC.

H6: Interdependent self-construal will have a positive relationship with

conformity intention.

ISC was chosen as a personal trait opposing to NFU. While NFU emphasizes the
individuality and uniqueness of individuals that rather promotes dissonance with others,
the core characteristic of ISC emphasizes the relatedness of individuals to others and
attachment to others. As a proof of the contradictory relationship between the two scales,
the Pearson correlation between NFU scale and ISC scale turned out to be significantly
negative (r = -.45, p < .01). Then why was not the independent self-construal scale used
as the opposing personality trait to ISC? According to Markus and Kitayama (1991),
independent self-construal and ISC are viewed as orthogonal scales rather than opposing
ones. In the present study’s data analysis, the Pearson correlation between independent
self-construal and ISC scales turned out to have no significant relationship in either
positive or negative direction. Since the present research had to find opposing personality
traits to each other, independent self-construal was not a proper choice considering its

orthogonal relationship with ISC.

Interactions between Personality Traits and Avatar Conditions

The notion that people’s reactions are not uniform, even under the same highly
deindividuated situation, and that such discrepancy could be caused by both situational

influences and differences in individual dispositions leads to the possibility of interaction

22

between situational factors and personality traits. Both situational factors (i.e. avatar
conditions) and personality traits (i.e. NFU and ISC) are crucial predictors of a person’s
conformity intention in a computer-mediated group decision-making. Naturally, there has
been an ongoing debate between the situationalist model and the trait model (Monson,
Hesley, & Chemick, 1982, p. 385) concerning which factor is more influential in
predicting a person’s behavior. The Situationist model advocates that human behavior can
be best predicted by situations in which a person is momentarily located, while the trait
model argues that human behavior is best predicted by a person’s innate personalities or
characteristics that are not changed by external factors (Endler & Magnusson, 1976;
Monson et al., 1982). However, recognizing that both elements are important in
predicting human behavior, recent social psychology studies have been focusing on the
interaction between situations and personality traits in predicting human behavior (Endler
& Magnusson, 1976; Magnusson & Endler, 1977; Monson et al., 1982).

In response to the promotion of the interactionist perspective, a considerable
amount of research evidence showed the existence of the interaction between situations
and personality traits. However, most of them investigated the interaction in a post-hoe
manner instead of specifying the pattern of the interaction before analyzing data (Mischel,
1973; 1977; Monson, et al., 1982: p.386). Among many efforts to find principles of the
interaction between situations and personality traits, Mischel (1973) focused on the
intensity of situations and the predicting the influence of personality traits under various
situational pressures. According to Mischel, people should show little variance in their
behavior, which is affected by their personality traits, in a situation of high-level

constraints (e. g. church or job interview), since there is much pressure to behave in a

23

certain way. In addition, Price and Bouffard (1974) argued that when there was a little
situational constraint (e.g. in a person’s own room), people would exhibit more wide
range of behavior according to their inherent personality traits. Monson et al.’s study
(1982) also supported these two arguments, showing that the variance of a person’s
behavior explained by his/her personal characteristics decreased when situational
pressure was strong.

On the contrary, Santee and Maslach (1982) found that personality traits
explained more variance of people’s conforming or dissenting behavior in the situation
with a strong group pressure rather than in the situation with a weak group pressure. In
their study, Santee and Maslach (1982) predicted that people’s decision to dissent in a
group discussion might be explained not only by individual differences in willingness to
call attention to themselves, but also by situations in which people were motivated to
stand out from the crowd. The research results were consistent with their prediction
proving that personal characteristics were more influential predictors of dissent and
conformity behavior when there was a strong situational pressure of reaching consensus
in the group (Santee & Maslach, 1982).

Even though the interaction between situations and personality traits has been
studied in the traditional social psychology research, most of the CMC studies did not
pay attention to the interaction between situations and personality traits. Rather, those
CMC studies had the assumption that people in the same computer-mediated
environment would react to the condition in the same way. This perspective focuses
mainly on the effects of situations created by computer media, which is based on

technological determinism. SIDE is an example of the framework assuming that

24

individuals who are stripped of their personal attributes in computer-mediated
environments automatically show high degree of group identity and conformity to group
norms.

The direction of this interaction could be inferred from the predictions of the
interactionist model of social psychology (Monson et al., 1982; Santee & Maslach, 1982)
and uniqueness theory. In predicting the interaction between personality traits and avatar
conditions in the computer—mediated group, this study borrows the interactionist
framework of the traditional social psychology research. If applying Mischel (1973;
1977)’s argument, the impact of personality traits might be more explicit when the level
of situational deindividuation is low (i.e. the different-avatar condition). However, if
applying Santee and Maslach’s argument (1982), the impact of personality traits might be
stronger when the level of situational deindividuation is high (i.e. the same-avatar
condition). Between the two contradictory perspectives, the present study adopts Santee
and Maslach’s perspective, which suggests that a strong situational pressure makes
people reveal their personality traits more clearly than a weak situational pressure does.

Uniqueness theory predicts that people who are high in NFU will experience
greater negative affects and exhibit greater changes in their direction of dissimilarity
under high situational similarity than those who are low in NF U (Snyder & F romkin,
1980). This prediction is in accordance with Santee and Maslach’s finding (1982) that
personality traits explain more variance of a person’s conforming or dissenting behavior
in a situation with strong situational pressure than in a situation with weak situational
pressure. This finding was based on the interactionist model, which was a dialectic

combination of the Situationist model and the trait model (Monson et al., 1982, p. 385).

25

The Situationist model advocates that human behavior can be best predicted by situations
in which a person is momentarily located, while the trait model argues that human
behavior is best predicted by a person’s innate personality or characteristics that are not
changed by external factors (Endler & Magnusson, 1976; Monson et al., 1982).
According to the interactionist model and uniqueness theory, people with high
levels of NFU are expected to exhibit more dissent when they are located in conditions of
high objective similarity (i.e. the same-avatar condition). In the same way, if people high
in ISC are placed in the same-avatar condition, they are expected to agree with other
members more than if they are placed in the conditions of lower similarity. To agree with
others sharing the same avatar as theirs would be a more conspicuous expression of
people's desires to belong to a group than to agree with others represented by different

avatars.

H7a: Need for uniqueness will show a stronger negative relationship with
conformity intention in the same-avatar condition than in the
different-avatar condition or in the no-avatar condition.

H7b: Interdependent self-construal will show a stronger positive

relationship with conformity intention in the same-avatar condition
than in the different-avatar condition or in the no-avatar condition.

METHODS

Procedure

A total of 557 students at a large Midwestern university voluntarily participated in

an online experiment. All participants were directed to visit the online experiment

26

website and participate in a discussion on resolving a controversial issue. The discussion
took place in an online chat room that was specifically set up for the experiment. All
participants participated in the discussion with the same topic and with the same format.
The only difference was in the way they were represented in the discussion chat room; 1)
the same-avatar condition: a condition in which participants were represented by the
same avatar as other discussants, 2) the different-avatar condition: a condition in which
participants were represented by different avatars from other discussants, and 3) the no-
avatar condition: a condition in which all participants were represented by text identities
without any avatar image. In order to control the effects of unidentifiability, no personal
or autobiographic information was released within experimental groups (Lee, 2004).
With the help of a random assignment program placed in the log-in page, participants
were randomly assigned to one of the three conditions.

Those who were assigned to the same-avatar condition and the different-avatar
condition were asked to choose their own avatars from four possible choices. A focus
group and several computer graphic designers were asked to choose virtual images that
would not contain negative cultural or social connotations. This process resulted in eight
different avatar images — animals that did not have any negative cultural or social
stereotypes, cartoon and movie characters, and other images that were perceived as
“cute” or “neutral” without causing any negative reaction.

In the different-avatar condition, each of the five avatars representing five
discussants was different fi'om each other, no matter what avatar was selected by each
participant. This was possible because four of the discussants were not real people but

programmed ones. Whatever avatar was chosen by a real participant, the other avatars

27

chosen by the other four virtual discussants were programmed to be different from each
other as well as from that of the one real participant. Once participants had chosen their
avatars, they were able to see what avatars had been chosen by the four other discussants
(See Figure 2). In order to reduce any confounding effects from virtual nicknames, all
discussants were assigned with uniform textual identities: Member 1, Member 2, Member

3, Member 4, and Member 5 (a real participant).

Figure 2. A screenshot of the different-avatar condition

mace

Your team members have CD056" the IOIIOWng avatars::

$ Member 1

' “"'" Member 2

 

Member 3
Member 4

a)
Member 5 (You)

 

 

 

A participant who was assigned to the same-avatar condition was also allowed to
choose one avatar. Again, whatever avatar was chosen by the participant, the four avatars
chosen by the other discussants in this condition were programmed to be the same (See

Figure 3).

28

Figure 3. A screenshot of the same-avatar condition

AlfiAQCD

Your team members have chosen the following avatars:

 

Member 1

Member 2

Member 3

Member 4

®®®@@

Member 5 (You)

 

 

Finally, participants who were assigned to the no-avatar condition were assigned
to uniform textual identities (i.e. Member 1, Member 2, Member 3, Member 4, and
Member 5) without any avatar image. In this no-avatar condition, as well as the other two
avatar conditions, all five discussants were to maintain absolute anonymity (i.e.

unidentifiability) and no personal or autobiographic information was released.

Figure 4. A screenshot of the no-avatar condition

 

Individuals participating in the discussion will be represented like this:

Member 1
Member 2
Member 3
Member 4
Member 5 (You)

 

 

 

29

 

Before they moved to the discussion page, participants were asked to answer how
much they felt similar to other discussants in the way that they were represented (i.e.
perceived deindividuation) and how much they could identify with other discussants as
members of the same group (i.e. group identity).

On the next page, participants were asked to read a hypothetical scenario in which
a person was faced with a dilermna and had to choose one option out of two. The scenario

was borrowed fi'om Lee’s (2004) experiment.

Ms. E, a college senior, has studied the piano since childhood. She has won
amateur prizes and given small recitals, suggesting that she has considerable
musical talent. As graduation approaches, she has the choice of taking a
medical school scholarship to become a physician, a profession which would
bring certain financial rewards, or entering a conservatory of music for
advanced training with a well-known pianist. She realizes that even upon
completion of her piano studies, success as a concert pianist would not be
assured (p.240).

After reading, participants were asked to join a discussion with four other
discussants in a chat room. The conversation window was designed like a typical online
chat room that showed participants in the discussion at the right side of the chat room
with the only real participant’s avatar (Member 5) at the top of the chat room box (Figure
4).

Once participants had finished reading the scenario, they were asked to read the
four other discussants’ opinions. At the time estimated for participants to be finished
reading the scenario, one of the other four virtual discussants presented its opinion on
what the person in the scenario should do, followed by the other discussants. Four virtual
discussants were programmed to take turns in order to give the impression that each

discussant was expressing its own opinion after reading the prior discussant’s opinion

30

(Lee, 2004). All the opinions presented by the programmed discussants were taken from
the previous focus group.

The four programmed discussants shared the same opinion, which implied the
existence of a conformity norm within the group. Once participants were finished reading
the other discussants’ opinions, a pop up questionnaire appeared asking them to choose
what they would like to say to the other four discussants out of seven possible answers: I
strongly agree with you, I agree with you, I slightly agree with you, not sure, I slightly
disagree with you, I disagree with you, and I strongly disagree with you. The participants
were also asked to provide reasons for their decisions in a text box. Next, both the
participants' selected answers and typed explanations were shown up in the chat room

window following the other four programmed discussants’ answers.

31

Figure 5. A Screenshot of the chat room window of the different-avatar
condition

Ms. E, a college senior, has studied the piano since childhood. She has won amateur prizes and given small recitals.
suggesting that she has considerable musical talent. As graduation approaches, she has the choice oftaking a medical
school scholarship to become a physician. a profession which would bring certain financial rewards, or entering a
conservatory of music for advanced training with a well-known pianist. She realizes that even upon completion of her piano
studies. success as a concert pianist would not be assured.

 

 

a Member5

*Memberlil think she should go to medical school. mm mm
Member 4:1 agree with you. It is better for her to go to med school. * Mamba”
Member3:WelI,.. maybe medical school. You can make more money. %ugmbe,2

fiiMember 21Yeah. definitely medical school! More money! Member 3

  

Member4
$5 Members
4.4
Q] What is your decision?
I strongly agree I agree I slightly agree I slightly disagree I disagree lstronglv disagree
With you With you wr h you with you with you with you
r r" c __—r‘ ————r —r —-————r~

 

Not Sure

 

 

 

32

Figure 6. A Screenshot of the chat room window of the same-avatar condition

 

Ms. E. a college senior. has studied the piano since childhood. She has won amateur prizes and given small recitals.
suggesting that she has considerable musical talent. As graduation approaches, she has the choice of taking a medical
school scholarship to become a physician. a profession which would bring certain financial rewards. or entering a
conservatory at music for advanced training with a well—known pianist She realizes that even upon completion of her piano

studies. success as a concert pianist would not be assured.

 

ig Member 5

@Member 1| think she should go to medical school

IQMember 4.I agree with you. It is better for her to go to med school.
‘gMember 3tWerl maybe medical school You can make more money

,gMember 21 Yeah. definitely medical school! More money!

In this room

.9 Member 1
9 Member 2
.9 Member 3

.9 Member 4
.@ Member 5

Q] What IS your decrsron?
I strongiy agree I agree I slightlv agree 1 slightly disagree I disagree I strongly disagree
With you With you Wllh you with you With you with you

r .~

 

(
Not Sure

F————-r’-—-————-—————-(

 

 

 

Finally, under the assumption that the first discussion would be followed by other
discussion sessions, participants were asked whether they would be willing to conform to
other discussants in coming discussions (i.e. conformity intention), under the assumption

that all the other four discussants were to share the same opinion as in the first discussion.

Measures

Six newly invented items were used to measure the perceived deindividuation
construct.l Need for Uniqueness (N FU) was measured by eight items2 selected from the

original NFU scale developed by Snyder and Fromkin (1980). Interdependent Self-

33

Construal (ISC) construct was measured by six items3 from the original ISC scales
(Markus & Kitayama, 1991; Singelis, 1994). Group identity was measured by three
modified items4 from Cheney’s organizational identification scale (1983). All of the
items were rated on seven-point Likert-type scales (scored 7 for strongly agree and 1 for
strongly disagree). Finally, conformity intention5 consisted of four items measuring
participants’ intentions to agree to a dominant opinion promoted by the other discussants.
The inter-scale correlations and reliability of all five scales are presented in Table 1.

All the constructed scales were tested for measurement model using confirmatory
factor analysis (CFA) of the structural equation modeling (SEM). The data analysis using
EQS program ver. 6.1 (Bentler, 2004) showed a good fit of the measurement model (Chi-
square 8.54, df = 4, p = .07). Other goodness-of-fit indices also showed a strong fit of the

measurement model: NFI = .98, NNFI = .94, CFI = .99, RMSEA = .05.

Table 1:1nter-scale correlations and reliability of five scales

 

 

1 2 3 4 5
1. Perceived Deindividuation .90
2. Group identification .01 .88
3. Conformity Intention -. l 7" .24" .75
4. Need for Uniqueness .01 -.08 -.28** .77
5. Interdependent self-construal -.04 .17" .30" -.45** .75
Mean 3.78 4.58 4.67 4.24 4.40
Standard Deviation 1.52 1.12 1.19 .94 .90

 

Note: Diagonal numbers show reliability (Cronbach alpha), N = 516
*p < .05, ”p < .01

34

FINDINGS

Out of 516 valid data collections, 173 participants were assigned to the same-
avatar condition, 172 participants were assigned to the different-avatar condition, and 171

participants were assigned to the no-avatar condition.

Main Effect of Avatar Conditions on Group Identity

Hypothesis 1 predicted that participants in the same-avatar condition would show
a higher level of group identity than those in the different-avatar condition as well as
those in the no-avatar condition. A one-way ANOVA showed a significant effect of the
three different avatar conditions on participants’ levels of group identity, F (2, 513) =
32.67, p < .001. The level of group identity was highest in the same-avatar condition (M
= 4.96, SD = 1.13), which was significantly higher than that of the different-avatar
condition (M = 4.07, SD = 1.08, p < .001) but was not significantly different from that of
the no-avatar condition (M = 4.72, SD = .95, p = n.s.). The reason why there was no
significant difference between the same-avatar condition and the no-avatar condition was
that the standardized text-only self-representation could induce as high level of similarity
as the same-avatar condition. In short, increased similarity in self-representation reduced
the focus on individuality and led to an elevated group identity, which follows the logic

of SIDE.

35

Table 2: Mean values of group identity of three different avatar conditions

 

 

Condition Mean Std. Deviation N

1.00 4.96 1.13 173
2.00 4.07 1.08 172
3.00 4.72 .95 171
Total 4.58 1.12 516

 

Table 3: The significance test of group mean differences in group identity across
three different avatar conditions

 

Experimental Method

 

Sum of Squares df Mean Square F Sig.

 

Avatar Between groups 72.75 2 36.38 32.67 .00
conditions Within groups 571.27 513 1.11
Total 644.02 515

 

 

Moderated Multiple Regression (MMR)

While avatar condition is a categorical variable, perceived deindividuation is a
continuous variable. One way to analyze the interaction between categorical variables
and continuous variables is using ANOVA, in which continuous variables are divided
into two or more groups at certain score points that were randomly assigned by
researchers (e. g. median as a dividing point for high and low group). However,
artificially divided continuous variables are known to reduce the power of statistical tests
and affect the tests of main effects and interaction effects. In case there are two or more

continuous variables in a model, dividing continuous variables into categories makes it

36

difficult to detect any curvilinear effects of the continuous variables on the dependent
variable (West, Aiken & Krull, 1996; Aiken & West, 1991).

Because of such problems, moderated multiple regression (MMR) method has
increasingly replaced the AN OVA in analyzing the interaction between continuous
variables and categorical variables. In MMR, categorical variables are represented by one
or more code variables that assign a unique value to each group. This study used a
dummy coding for categories. Three experimental conditions — the same-avatar condition,
the different-avatar condition, and the no-avatar condition - were dummy- coded in order
to see if there was any effect of different avatar conditions on people’s conformity
intention. Dummy coding is useful when a researcher is interested in comparing one
reference group with the other groups. The present study was interested in two
comparisons: 1) comparison between the same-avatar condition and the different-avatar
condition, and 2) comparison between the same-avatar condition and the no-avatar
condition. Therefore, the same-avatar condition was set as the reference group and
received a value 0, whereas the other two groups were assigned a value 1 on the code
variable that contrasted each of the groups with the comparison group and a value 0
otherwise (Aguinis, 2004; p. 119). That is, when the different-avatar condition was
compared with the same-avatar condition, the different-avatar condition was assigned 1
while the no-avatar condition was assigned 0. When the no-avatar condition was
compared with the same-avatar condition, the assigned number was converted. The first
comparison was represented in Z1 code variable, while the second comparison was

represented in Z2 code variable.

37

Interactions terms were represented as the products of independent variables
involved in the interactions (e. g. the interaction between X and W is X*W) and the
curvilinear relationship with a dependent variable were expressed through higher order
functions of the independent variables (e. g. X2 for the quadratic term of X) in MR
(West et al., 1996). The present study used a hierarchical MMR method that started from
main effect terms of each predictor in the first block, followed by the second block with
two-way interaction terms, and the third block with three-way interaction terms (Stone &

Hollenbeck, 1984).

Conformity Intention as a Function of Group Identity and Perceived Deindividuation

Hypothesis 2 predicted that the transient group identity without any salient
common social identity shared by participants would have a positive effect on conformity
intention. On the other hand, Hypothesis 3 predicted that perceived deindividuation and
conformity intention would show an inverted U-shaped curve. A hierarchical regression
analysis was done for conformity intention with group identity, the linear term of
perceived deindividuation and the quadratic term of perceived deindividuation. As shown
in Table 2, the transient group identity contributed to an increase in conformity intention
(,B = 2.61, p < .01), which supports Hypothesis 2. The linear term of perceived
deindividuation was a significant predictor of conformity intention (,8 = -.10, p < .01) and
accounted for 2.8% of the variance in conformity intention. The quadratic term of
perceived deindividuation was also significant in predicting conformity intention (,6 = -

.05, p < .05) and accounted for an additional 0.8% of the variance in conformity

38

intention. In other words, perceived deindividuation and conformity intention showed a

negative curvilinear relationship.

Table 4: Conformity intention as a function of group identity and perceived

 

 

 

deindividuation
Variables Conformity Intention
B t value
Transient Group Identity 2.61“ 5.82
Perceived Deindividuation -.10** -2.70
Perceived Deindividuationz -.05* -2.16

 

Note. Total R2 = 3.4% (F: 9.02"); AR’ ofthe transient group identity = 5.7% (AF = 31.12"); ARI of
perceived deindividuation = 2.8% (AF = 15.98”); AR2 of the squared perceived deindividuation = .8% (AF
= 4.68“)

*p < .05, "p < .01

Perceived Deindividuation and Avatar Conditions

Hypothesis 4 tested the moderating impact of avatar conditions on the curvilinear
relationship between perceived deindividuation and conformity intention. For the avatar
conditions, dummy coding was used and the same-avatar condition was set up as a
reference group with Z; for the comparison between the same-avatar condition and the
different-avatar condition and Z; for the comparison between the same-avatar condition
and the no-avatar condition. The interaction terms were represented by the products of
independent variables involved in the interactions (e. g. the interaction between X and W
is X*W) (West et al., 1996). A squared term of perceived deindividuation was also added

to reflect the curvilinear relationship with conformity intention. The present study used a

39

hierarchical regression analysis that started fi'om main effect terms of avatar conditions in
the first block, followed by the second block with the linear and squared terms of
perceived deindividuation, and the third block with interaction terms between avatar

conditions and perceived deindividuation;

Y = a + b1Z1+ bzzz + b3PD + b41002 + b5Z1*PD + b6Z2*PD + b721*P02 bgzz*PD2
(1)

(Y: Conformity intention, PD: perceived deindividuation, PDZ: squared perceived
deindividuation)

As shown in Table 3, the perceived deindividuation turned out to be the only
significant predictor ([3 = -.28, p < .01) in main effect blocks. As predicted in Hypothesis
4, avatar conditions moderated the curvilinear relationship between perceived
deindividuation and conformity intention. Interaction terms with avatar conditions and
squared perceived deindividuation were significant predictors of conformity intention.
The incremental R2 by the entire interaction term block was marginally significant (AR2 =

l.6%,p<.l).

40

Table 5: Interaction between perceived deindividuation and avatar conditions

 

 

 

Variables Conformity Intention
B t value
Step 1
Z, -.23 -1.29
22 .17 .94
Step 2
Perceived Deindividuation -.28** -3. 16
Perceived Deindividuation 2 .05 1.00
Step 3
le Perceived Deindividuation -.04 -.23
sz Perceived Deindividuation .19 1.66
Z. X Perceived Deindividuation 2 -. l 7'" -2. 17
sz Perceived Deindividuation 2 -.16** -2.58

 

Note. Total R2 = 6.6% (F= 4.48"); R2 ofthe avatar condition block (Step 1) = 0.3% (AF = .83); R2 of the
perceived deindividuation block (Step 2) = 4.7% (AF = 12.59"); R2 of the interaction terms block (Step 3)
= 1.6% (AF= 2.171)

'ip<.l,*p<.05,**p<.01

In order to make it easier to interpret this interaction, this study used graphical
presentations. For the graphical display of multiple regression analysis results, this study
followed the procedures suggested by West, Aiken, and Krull (1996). Inserting all the
unstandardized coefficients of the <Table 5> in to the equation (1), the MMR model with

conformity intention as a dependent variable turned out to be:

41

Y = 4. 777 -. 2332, +.168Z2 -.283X +.045X2 -.0352,X +. 1912211 -. 1 722,17 —. 11502.1(? (1.1)

This equation should be differed according to the three different avatar conditions;

I) The same avatar condition (21, Z2 = 0): Y = 4.777 -.283X + .045 X2 (1.1.1)
2) Diflerent avatar condition (21:1, 22 = 0): Y = 4.544 -.3I8X-.127X2 (1.1.2)
3) No avatar condition (21:0, 22 = 1): Y = 4.945 - .092X-.115X2 (1.1.3)

Figure 5 is a graphical representation of the three equations (1.1.1), (1.1.2.), and
(1.1.3). It shows that the multiple regression analysis supported Hypothesis 4, since the
shape of the inverted U-shaped curve between perceived deindividuation and conformity
intention was sharpest in the different-avatar condition and became less manifest as the
level of objective similarity in self-representation increases. As predicted in Hypothesis
4, people in the same-avatar condition showed a decrease in their conformity intention
even with the slight increase of their perceived deindividuation. Both in the no-avatar
condition and the different-avatar condition, people showed a mild increase in their
conformity intention when they perceived low or moderate level of deindividuation from
their situations. However, when they perceived more than a moderate level of
deindividuation, people showed a drastic decrease in their conformity intention.

In the different-avatar condition, when people’s perceived deindividuation is at its
lowest (score 1), the score of the conformity intention was predicted to be 4.45. As the
perceived deindividuation increased, participants’ conformity intention increased to 4.74
at the point of 2.48 in the perceived deindividuation. After this point, people’s conformity

intention started to decrease as their perceived deindividuation increased. When the

42

perceived deindividuation was at its highest (score 7), the score of participants’

conformity intention was predicted to decrease to 2.20.

In the no-avatar condition, people started from the score of 4.31 on conformity
intention at the lowest point of the perceived deindividuation (score 1). As the perceived
deindividuation increased, participants’ group identification with other members
increased to 4.96 at the point of 3.38 in the perceived deindividuation. After this point,
people’s conformity intention started to decrease with increasing perceived
deindividuation. When the perceived deindividuation is at its highest (score 7), the score

on people’s conformity intention was predicted to increase to 3.45.

On the other hand, people in the same-avatar condition showed an almost
negative relationship between perceived deindividuation and conformity intention.
People in the same-avatar condition showed a decrease in their conformity intention even
with the slight increase in their perceived deindividuation. However, people’s conformity
intention continued to decrease even when they perceived an intense deindividuation in

their situation.

43

Figure 7. Conformity intention by perceived deindividuation and avatar conditions

 

 

7
5.
r: 5‘
g
44
E
2?
'g 3.
§ + Some avamr condition 'x.\
g 2. +Noavatarcondition 3

-*- Different avatar condition

 

 

0 . - .
0 2 4 6 8

Perceived Deind'widuation

 

Personality Traits and Avatar Conditions

Hypothesis 7a predicted that avatar conditions would moderate the
relationship between NFU and conformity intention. The MMR equation with NFU as

the independent variable is expressed as:

Y = a + b121+ bZZZ + b3NFU+ b4ZI*NFU + b522*NFU (2)

However, the multiple regression analysis failed to detect a significant interaction
between avatar conditions and NFU. While the main effect of NF U turned out to be

significant (,8 = -.48, p < .01), supporting Hypothesis 5, the increased R2 by interaction

44

terms NFU X avatar conditions (0.7 %) was not significant (F (2, 510) = 1.90, p = .15).
More specifically, the relationship between NFU and conformity intention in the same-
avatar condition was marginally different from that of the different-avatar condition (6 =

.23, p < .1) but was not significantly different from that of the no-avatar condition (,6 =

.20, p = .12).

Table 6: Interaction between NFU and avatar conditions

 

 

 

Variables Conformity Intention
B I value
Constant 4.64" 4.07
Step 1
Z. -.O4 —.31
Z; .11 .91
Step 2
NFU -.48*"‘ -5.46
Step 3
21 x NFU .23‘1r 1.76
ZZX NFU .20 1.56

 

Note. Total R2 = 8.5% (F = 9.47"); R2 of the avatar condition block (Step 1) = 0.3% (AF = .83); 1c2 of the

perceived deindividuation block (Step 2) = 7.5% (AF = 41.6"); R2 of the interaction terms block (Step 3) =
0.7% (AF = 1.90)
1'p<.l,"‘p<.05, "p<.01

In order to present the MMR results into a graphical representation,

unstandardized coefficients of all the elements in <T able 8> were inserted into the

equation (4):

45

Y = 4.64-.04Z, +.11Z2 -.48X +.23Z,X +.20Z2X (2.1)

(X = NFU, Y = Conformity Intention)

Then, the equation (2.1) was divided into three equations according to the three

different avatar conditions:

1) The same-avatar condition (21, Z2 = 0): Y = 4.641 -.482X (2.1.1)
2) The diflerent-avatar condition (21:1, Z2 = 0): Y = 4.641 -.038 —.482X + .23X (2.1.2)

3) The no-avatar condition (21:0, Z2 = 1): Y = 4.641 + .113 -.482X+ .2X (2.1.3)

Figure 6 is a graphical representation of the relationship between NFU and
conformity intention across the three different avatar conditions. Even though the R2
increased by the interaction term block was not statistically significant, Figure 6 supports
the pattern of the interaction predicted in Hypothesis 7a. Participants in the same-avatar
condition showed a stronger negative relationship between NFU and conformity intention
than those in the different avatar-condition or in the no-avatar condition. At the lowest
score of NFU (score 1), people in the different-avatar condition and people in the no-
avatar condition were predicted to have 5.42 and 5.67 respectively in their conformity
intention. At the same score of NFU, people in the same-avatar condition showed 6.2
point. Meanwhile, at the highest score of the NFU (score 7), people in the different—avatar

condition and people in the no-avatar condition were predicted to have 3.91 and 3.98

46

respectively in their conformity intention. At the same score of NFU, people in the same-
avatar condition showed 3.31 point.

In short, Even though the interaction term block was not statistically significant,
participants in the same-avatar condition showed a stronger negative relationship between
NFU and conformity intention than those in the different-avatar condition and in the no-
avatar condition (Figure 6). That is, as the situational deindividuation intensified, a
person’s NF U showed a clearer negative linear relationship with his/her conformity
intention. Even though the interaction between NFU and avatar conditions did not turn
out to be significant, if there were more samples, the non-parallel pattern caused by the

interaction might have been much clear and statistically significant.

Figure 8. Conformity intention by NFU and avatar conditions

 

 

 

 

 

7

6.
c 5‘
é
w d
g 4
2::
E 3.
Q + Some avatar condition
§ 2‘ + No avatar condition

*ir- Different avatar condition
1.
0 .
0 2 4 6 8

47

Hypothesis 7b predicted that avatar conditions would moderate the
relationship between ISC and conformity intention in such a way that the high level of
objective similarity caused by sharing the same avatar with others increased the effect of
ISC on conformity intention. The MMR equation with ISC as the independent variable is

expressed as:

Y = a + blZ, + 19222 + b 315C + 1242, *ISC + b5Z2*ISC (3)

(Y = Conformity Intention)

As shown in Table 5, the main effect of ISC was significant (F (1, 512) = 50.55, p
< .01), as was predicted by Hypothesis 6. The interaction terms were also significant
predictors of conformity intention and increased 1.9% of the variance in conformity
intention (F (2, 510) = 5.55, p < .01). The relationship between ISC and conformity
intention in the same-avatar condition was significantly different from that of the
different-avatar condition ()3 = -.44, p < .01) and from that of the no-avatar condition (,6 =

-.33,p < .01).

48

Table 7: Interaction between ISC and avatar conditions

 

 

 

Variables Conformity Intention
B t value
Constant 4.65” 4.07
Step 1
Z, -.04 -.31
Z2 .08 .64
Step 2
ISC .66" 6.76
Step 3
le ISC -.44** -3.20
Z2>< ISC -.33* -2.43

 

Note. Total R2 = 11% (F = 12.88"); R2 of the avatar condition block (Step 1) = 0.3% (AF = .83); R2 of the

perceived deindividuation block (Step 2) = 9.3% (AF = 50.55"); R2 of the interaction terms block (Step 3)
= 1.9% (AF: 5.55")
Tp<.1,*p<.05,*"‘p<.01

Based on the unstandardized coefficients of all the elements in Table 7, the MMR

model turned out to be:

Y = 4.65 -.04Z, + .08Z2 +.66X—.44Z1X-.3322X (3.1)

(X = ISC, Y = Conformity Intention)

49

This equation should be differed according to avatar conditions;

1) The same avatar condition (21, Z2 = 0): Y = 4.65 +.66X (3.1.1)
2) Diflerent avatar condition (2] =1, ZZ = 0): Y = 4. 65 -.04 +.66X -. 44X (3.1.2)
3) No avatar condition (ZI=0, 22 = 1): Y = 4.65 + .08 +.66X -.33X (3.1.3)

The positive relationship between ISC and conformity intention was strongest
when people were represented by the same avatar (Figure 7). Participants were more
conspicuous about their desire to be accepted in a highly deindividuated situation than in

a less deindividuated situation (Santee & Maslach, 1982).

Figure 9. Conformity intention by ISC and avatar conditions

 

 

 

 

 

7 .
54
.5 5‘
E
:5 4.
E
E 3.
.9
C
8 2. -+- Same avatar condtion
«4:— No avatar condition
1- + Different avatar condition
0 T . r
o 2 4 6 8

ISC

50

DISCUSSION

The present study tried to address critiques on manipulations and conceptual
confusion surrounding important concepts of social identity research. It also relaxed the
tacit assumption of the undifferentiated individuals seen in most social identity research
and broke new ground regarding the role of individual differences in perception and
personality traits. Moreover, using the compromise view, this study tried to explain some
of the social identity research findings that have been inconsistent with SIDE. In short,
the present paper made contributions towards expanding the theoretical perspectives on
social identity research in CMC in several respects.

First of all, this study focused on similarity, one of the confounding constructs of
anonymity, rather than unidentifiablity, which has been the main independent variable in
most social identity studies. It also distinguished between a salient social identity and a
transient group identity that coexisted in the “group identity” concept of social identity
research. Avatars were used in this study for two purposes: in order to focus on the effect
of similarity in self-representation while holding the same level of unidentifiability across
different experimental conditions and in order to induce the transient group identity (i.e.
experimental group identity) without introducing any salient common social category.
This study found that the similarity in self-representation with others contributed to
amplified group identity even without a salient social identity shared by group members.
This suggests that identification with pre-existing social groups in the off-line world can
be overcome by transient online group memberships, instead of being fortified as SIDE

indicates.

51

This study also acknowledged the existence of “a consensus that limits the
different selves” (Postrnes & J etten, 2006) in social identity research. However, the
assumption of uniform individuals could not explain the existence of different identity
needs and defiant reactions against group norms in a group of social identity studies (i.e.
Codol, 1984; Cooper & Jones, 1969; Maslach, 1974; Lee, 2004). This could mean that
prior social identity studies employing the assumption of standardized individuals within
the same group had incomplete and erroneous findings. Thus, in order to supplement the
denial of individual differences and explore the roles of differentiated perceptions and
needs at an intra—group level, uniqueness theory and ODT were used.

In regard to deindividuation, this study distinguished between objective similarity
conditions and individual perceptions of the situation. While the former was
operationalized by the three different avatar conditions (a categorical variable), the latter
was measured by a continuous variable that appraised people’s perceptions of how
deindividuated or similar they were to others. These two variables have been considered
equal in most social identity research and have been treated as a dichotomous variable,
since SIDE assumes a uniform level of deindividuation to be perceived by people when
they are under the same anonymity manipulation. On the other hand, the compromise
view assumes that people use their perceptions of the conditions in which they are located
in order to judge whether they are too distinctive from others or too similar to others. In
short, SIDE predicts that intensified situational deindividuation caused by reduced social
cues can encourage a person’s positive reaction to group norms, while the compromise
view suggests that heightened perceived deindividuation can discourage a person’s

positive reactions toward a group. Therefore, this study’s finding that perceived

52

deindividuation and conformity intention showed a negative curvilinear relationship
supports the compromise view and tends to undermine the assertions of SIDE.

In addition, the present study suggests that the curvilinear relationship between
perceived deindividuation and conformity intention should differ depending on the level
of similarity with others in self-representation. This is based on the assertion that
objective situational similarity was a different variable from participants’ perception of
the given similarity condition. The curve became less manifest and even changed from an
inverted U-shape to a slight U-shape as the objective similarity in self-representation
increased. In general, heightened perceived deindividuation seemed to depress people’s
conformity intention across the all three avatar conditions. When the level of objective
similarity was low, conformity intention increased as people felt low or moderate level of
similarity between them and other group members. On the other hand, conformity
intention decreased even with the slight increase in perceived deindividuation when the
objective similarity was high. This is because expressing dissent or giving different
opinions can be used as a way to promote people’s individualities when they feel that
their unique self-concepts are threatened by being made highly similar to others (Cooper
& Jones, 1969; Fromkin, 1970).

Finally, this study integrated individual needs and motivations into the social
identity research domain and investigated how these individual differences would interact
with situational factors. As Santee and Maslach (1982) predicted, personality traits had a
stronger effect on human behavior when there was a high level of situational pressure -
similarity in the way people were represented. Even though the interaction was not

statistically significant, NFU showed a stronger negative relationship with conformity

53

intention in the same-avatar condition than in the different-avatar condition or in the no-
avatar condition. As indicated by uniqueness theory and ODT, intensely deindividuated
conditions made people more concerned about their unique self-presentations (Cheek &
Hogan, 1983) and led them to behave in ways in which they could establish and define
their unique selves (Alexander & Lauderdale, 1977; Maslach, Santee, & Wade, 1987).
Such concerns about their unique selves matter more to people inherently high in NF U
than to those low in NFU. If people high in NFU perceive an intensified similarity from
the same visual representation shared with other discussants and the existence of a
dominant opinion, the only way for them to prove their unique selves is to express dissent.
Meanwhile, in a situation where people high in NFU can be distinguished from others by
representing themselves with distinctive self-representations, they will not feel such a
strong urge to express dissent and prove their uniqueness as they would in the same-
avatar condition. In the same way, the influence of ISC on conformity intention was more
explicit when people were represented by the same avatar than when they were
represented by different avatars or no avatar. When the pressure to conform was already
high both from sharing the same avatar with other members and from the existence of a
dominant opinion promoted by the entire group, the only option left for participants who
were high in ISC and wanted to express their desire for acceptance was to conform to the

dominant opinion.

54

CONCLUSIONS

Summary and Implications

The present study started with describing the increasing use of computers as a
communication channel and how computer media features could change interpersonal
and group communication patterns. This study focused on group identification and
conformity in computer-mediated group discussion settings, addressing critiques on
manipulations and conceptual confusion surrounding the important concepts of
anonymity and group identity. In addition, this study acknowledged that social identity
research should take individual differences into consideration.

Using three different avatar conditions as a way to manipulate the level of
similarity in self-representation, this study found that the similarity in self-representation
with others contributed to amplified group identity even without a salient social identity
shared by group members. This suggests that identification with pro-existing social
groups in the off-line world can be overcome by transient online group memberships,
instead of being fortified as SIDE indicates. When an online community is built based on
a salient group identity, the existing social identities are replaced by the situational group
membership. For example, many online communities composed by people who are
suffering from the same disease focus on sharing information and comforting each other
while minimizing their social or cultural backgrounds.

This study further investigated how individual differences, which were not taken
into consideration in most social identity research, could affect people’s intention to
conform. The two individual differences taken into account were perceptions of similarity

to others in the group, and individuals' needs to be either similar or different from others.

55

Perceptions regarding the level of similarity, termed perceived deindividuation, was
invented and distinguished from objective similarity in the way people were represented.
The significance of perceived deindividuation in this study relies on the fact that it
measures what degree of similarity with others in a CMC group participants perceive,
enabling its use as continuous variable that could be regressed with a continuous
dependent variable, intention to conform. The regression curve turned out to be negative
curvilinear (i.e. inverted U-shaped curve), which implies that people agreed with others
most when they perceived a moderate similarity between themselves and others. On the
other hand, it also suggests that people can disagree more if they perceive high level of
similarity with others, which is contradictory to SIDE but in accordance with uniqueness
theory and optimal distinctiveness theory. In practice, this suggests that anti-normative
behavior or “flaming” in cyberspace might be explained by people’s pursuit of their
individualities as well as resistance against highly standardized self-representation online,
not necessarily due to freedom from the norms and regulations of the off-line world.
Therefore, one way to reduce this flaming in online communities might be to allow a
certain amount of individual choice in peoples’ virtual self-representations so that they
are less likely to choose a strategy of disagreeing with others simply to demonstrate their
uniqueness.

This study also integrated individual needs and motivations into the social identity
research domain and investigated how these individual differences interacted with
situational factors. The results showed that intensified similarity among group members
magnified the influence of personal traits — people with a high need for uniqueness in a

high similarity condition disagreed more while people with a high need for acceptance in

56

such conditions agreed more. This could be interpreted as an indication that contributors
who have an inherent need to be different on online discussion boards (e. g. political
blogs), where there is a high degree of situational pressure both from standardized self-
representation and consensus norms, might disagree with a dominant opinion in order to
preserve their unique self-concepts and not necessarily because they disagree with the
dominant opinion itself. On the other hand, those who have an inherent need to be similar
to others could agree with the dominant opinion in order to assimilate and be accepted,
but not necessarily because they agree with the opinion. Therefore, increasing
conformity among online communities could be hindered by standardizing the way
people are represented in cyberspace, a implication that directly conflicts with the SIDE
model. Instead, allowing a moderate degree of individually customized self-
representation can reduce the possibility of disagreement due to the need to be different
from others as well as agreement due to the need to be similar to others in cyberspace. In
cyberspace where people’s identities can be detached from their existing social status and
easily replaced by situational group membership, virtual self-representation plays an
important role in affecting people’s perception and attitude toward others in group
discussions.

Collectively, the findings suggest revisions to the SIDE model by suggesting not
only that situational factors in CMC environments (i.e. the lack of social cues and
anonymity) should be taken into consideration in explaining group identity and
conformity, but also human factors as well (i.e. people’s perception of the situation and

their need to be different or similar) .

57

Limitations & Future Studies

This study includes some aspects that require further investigation. First of all,
using a programmed discussion could have reduced the actuality of real-time
conversations. In future studies, five real participants could be asked to discuss and reach
a decision within a limited time. This might allow more time and interactions for
participants to naturally affiliate with their own avatars and to recognize their similarity
to (the same-avatar condition) or difference from (the different-avatar condition) the
avatars of other discussants.

Secondly, allowing participants to choose avatars might have confounded
similarity in virtual self-representation and similarity in their tastes for virtual self-
representation. However, randomly imposed avatars could have created repulsion in
participants, even though avatars were designed so as to reduce any negative reaction as
much as possible. Therefore, the present study gave participants a limited number of
choices, so that it would not allow too much variation in participants’ personal
preferences for avatar images. In addition, these images were created to be the same size
and were similar in the impressions they gave, so participants’ reactions toward avatars
did not vary much due to the nature of the avatars. Still, it would be interesting to
compare a condition where participants are randomly assigned to avatars and a condition
where participants have choices to pick fi'om.

Thirdly, the present study looked only at the situation in which there was one
dominant opinion shared by other discussants. This might have brought a confounding
effect to the same-avatar condition, since participants could have conformed to the other

discussants’ opinion either because they identified with those who shared the same avatar

58

or because they were complying with the consensus pressure coming from others. In
future studies, these two potentially confounding influences could be disentangled by the
dividing of experimental groups into three new conditions: a group with two equally
numbered subgroups sharing different avatars (Figure 10.1), a group with a larger
subgroup of one opinion sharing the same avatar as that of the participant and a smaller
subgroup of the other opinion with a different avatar (Figure 10.2), and a group with a
larger subgroup of one opinion with a different avatar from that of the participant and a
smaller subgroup of the other opinion with the same avatar as that of the participant
(Figure 10.3). These groups can distinguish the influence of the same avatar fi'om that of

the majority-opinion subgroup pressure on participants’ conformity decision.

Figure 10.1. A screenshot of a group with two equally numbered subgroups sharing
different avatars

 

 

Subgroup l Subgroup 2

Q A Participant

 

 

 

59

Figure 10.2. A screenshot of a group with a larger subgroup of one opinion sharing
the same avatar as that of the participant and a smaller subgroup of the other
opinion with a different avatar

 

 

A larger subgroup A smaller subgroup

Q A Participant

 

 

 

Figure 10.3. A screenshot of a group with a larger subgroup of one opinion with a
different avatar from that of the participant and a smaller subgroup of the other
opinion with the same avatar as that of the participant

 

 

A larger subgroup A smaller subgroup

9 A Participant

 

 

 

60

Finally, this study used conformity intention instead of actual conformity decision
as a dependent variable. Since there was only one discussion, it was difficult to use this
decision as a dependent variable and generalize its finding to other decision-making cases.
Thus, this study chose conformity intention as a dependent variable based on the
assumption that behaviors are strongly influenced by motivations and intentions (Aj zen
& F ishbein, 1980). As a solution, conformity decision could have been measured by
several discussion sessions, but the reliability among different conformity decisions
tended to be very low (Lee, 2004) because participants’ personal involvements for each
discussion topic might vary across different discussions. Other types of tasks, such as
games or arriving at a group consensus on ranking a list of items (Hiltz, Johnson, &
Turoff, 1986; Schroeder & Benbassat, 1975) could be used as group assignments that are

relatively neutral and unrelated to personal interests.

61

BIBLIOGRAPHY

Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting
interactions. Newsbury Park, CA: Sage.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social Behavior.
Englewood Cliffs, NJ: Prentice-Hall, Inc.

Alexander, C. N. Jr., & Lauderdale, P. (1977). Situated identities and social influence.
Sociometry, 40, 225-233.

Baron, R. S. (1971). Anonymity, deindividuation and aggression. Paper presented at the
Western Psychological Association, San Francisco, CA.

Baron, R. S., Kerr, N. L., & Miller, N. (1992). Group Process, Group Decision, Group
Action. Buckingham: Open University Press.

Baumeister, R. F. (1982). A self-presentational view of social phenomena. Psychological
Bulletin, 91 (January), 3-26.

Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal
attachments as a fundamental human motivation. Psychological Bulletin, 11 7,
497 -529.

Bentler, P. M. (2004). EQS structural equations modeling software (Version 6.1)
[Computer software]. Encino, CA: Multivariate Software.

Blanton, H., & Christie, C. (2003). Deviance regulation: A theory of identity and action.
Review of General Psychology, 7(2), 115 - 149.

Breakwell, G. M. (1987). Identity. In H. Beloff & A. Coleman (Eds.), Psychology Survey
6 (pp.94-114). Leicester, England: British Psychological Society.

Brewer, M. (1991). The social self: On being the same and different at the same time.
Personality and Social Psychology Bulletin, 1 7(5), 475-482.

Brewer, M. (1993). The role of distinctiveness in social identity and group behavior. In M.
Hogg & D. Abrams (Eds.), Group motivation: Social psychological perspectives
(pp. 1 -16). New York: Harvester-Wheatsheaf.

Cheek, J. M., & Hogan, R. (1983). Self-concepts, self-presentations, and moral
judgments. In J. Suls & A. G. Greenwald (Eds.), Psychological perspectives on
the self: Vol. 2. (pp. 249-273). Hillsdale, MJ: Erlbaum.

Cheney, G. (1983). On the various and changing meanings of organizational
membership: Field study of organizational identification. Communication

62

Monographs, 50, 342-362.

Codol, J. P. (1984). Social differentiation and non-differentiation. In H. Tajfel (Ed.), The
social dimension: European developments in social psychology: Vol. 1. (pp.314-
337). Cambridge, UK: Cambridge University Press.

Cooper, J ., & Jones, E. E. (1969). Opinion divergence as a strategy to avoid being
miscast. Journal of Personality and Social Psychology, 13, 23-30.

Diener, E. (1977). Deindividuation: Causes and consequences. Social Behavior and
Personality, 5, 143-155.

Diener, E. (1980). Deindividuation: The absence of self-awareness and self-regulation in
group members. In P. B. Paulus (Ed.), The psychology of group influence (pp.
209-242). Hillsdale, NJ: Lawrence Erlbaum.

Duval, T. S., & Wicklund, R. A. (1972). A theory of objective self-awareness. New York:
Academic Press.

Endler, N. S., & Magnusson, D. (l 976). Interactional psychology and personality.
Washington, DC: Hemisphere.

Fromkin, H. L. (1970). Effects of experimentally aroused feelings of undistinctiveness
upon valuation of scarce and novel experiences. Journal of Personality and Social
Psychology, 16, 521-529.

Fromkin, H. L., & Snyder, C. R. (1980). The search for uniqueness and valuation of
scarcity. In K. J. Gergen, M.S. Greenberg, & R. H. Willis (Eds.), Social
exchange: Advances in theory and research (pp. 57-75). New York: Plenum.

Fromm, E. (1941). Escape from freedom. New York: Avon.

Guerin, B. (1986). Mere presence effects in humans: A review. Journal of Experimental
Social Psychology, 22(January), 38-77.

Hiltz, R. S., Johnson, K., & Turoff, M. (1986). Experiments in group decision making:
Communication process and outcome in face-to-face versus computerized
conferences. Human Communications Research, 13, 225-252.

Homsey, M. J ., & J etten, J. (2004). The individual within the group: Balancing the need
to belong with the need to be different. Personality and Social Psychology Review,
8(3), 248-264.

Kiesler, S., Siegel, J ., & McGuire, T. (1984). Social psychological Aspects of computer-
mediated communication. American Psychologist, 39, 1123-1134.

63

Kim, J. -H. (2006). Conformity in the computer-mediated group decision-making: An
online experiment investigating the interaction between virtual self-representation,
perceived deindividuation, and personality traits. Unpublished doctoral
dissertation, Michigan State University, Michigan.

Kim, J. (2000). Social interaction in computer-mediated communication: America society
for information science. Bulletin of the American society for information science.

Lea, M., & Spears, R. (1991). Computer-mediated communication: De-individuation and
group decision-making. International Journal of Man-Machine Studies, 34, 283-
301.

Lea, M., Spears, R., Watt, S. E., & Rogers, P. (2000). The INSIDE story: Social
psychological processes affecting on-line groups. In T. Postmes, R. Spears, M.
Lea, & S. Reicher (Eds.), SIDE issues centre stage: Recent developments in
studies of deindividuation in groups (pp.47-62). Royal Netherlands Academy of
Arts and Sciences. Amsterdam: The Netherlands.

Lee, E. (2004). Effects of visual representation on social influence in computer-mediated
communication: Experimental tests of the Social Identity Model of
Deindividuation Effects. Human Communication Research, 30(2), 234-259.

Lee, E., & Nass, C. (2002). An experimental test of normative group influence and
representation effects in computer-mediated communication: When interacting
via computers differs from interaction with computers. Human Communication
Research, 28, 349-381.

Levine, J. M., & Moreland, R. L. (1990). Progress in small group research. Annual
Review of Psychology, 41, 585-634.

Magnusson, D., & Endler, N. S. (1977). Personality at the crossroads: Current issues in
interaction psychology. Hillsdale, N.J.:Erlbaum.

Markus, H. R., & Kitayama, S. (1991). Culture and the self: Implications for cognition,
emotion, and motivation. Psychological Review, 98, 224-253.

Maslach, C. (1974). Social and personal bases of individuation. Journal of Personality
and Social Psychology, 29(3), 411-425.

Maslach, C., Santee, R. T., & Wade, C. (1987). Individuation, gender role and dissent:
Personality mediators of situational forces. Journal of Personality and Social
Psychology, 53(6), 1088-1093.

Maslow, A. H. (1968). Toward a psychology of being. New York: Van Nostrand.

Mischel, W. (1973). Toward a cognitive social learning reconceptualization of

64

personality. Psychological Review, 80, 252-283.

Mischel, W. (1977). The interaction of person and situation. In D. Magnusson and NS.
Endler (Eds.), Personality at the crossroads: Current issues in Interactional
psychology. Hillsdale, N.J.: Erlbaum, 1977.

Monson, T. C., Hesley, J. W., & Chemick, L. (1982). Specifying when personality traits
can and cannot predict behavior: An alternative to abandoning the attempt to

predict single-act criteria. Journal of Personality and Social Psychology, 43(2),
385-339.

Postrnes, T. (1997). Social influence in computer-mediated groups. Javnost - the Public,
4, 98-99.

Postmes, T., & J etten, J. (2006). Reconciling individuality and the Group. In T. Postrnes
& J. J etten (Eds.), Individuality and the group: Advances in social identity (pp.
258- 269). London: Sage.

Postrnes, T., & Lea, M. (2000). Social processes and group decision-making: anonymity
in group decision support systems. Ergonomics, 43(8), 1252-1274.

Postmes, T., & Spears, R. (1998). Deindividuation and antinorrnative behavior: A meta-
analysis. Psychological Bulletin, 123, 238-259.

Postmes, T. & Spears, R. (1999, July). Refining the cognitive definition: Deindividuation
eflects in common bond vs. common identity groups. Paper presented at the
meeting of the SIDE Issues Centre Stage: Recent Development in Studies of
Deindividuation in Groups, Amsterdam, the Netherlands.

Postmes, T., Spears, R., & Cihangir, S. (2001). Quality of decision making and group
norms. Journal of Personality and Social Psychology, 80, 918-930.

Postmes, T., Spears, R., & Lea, M. (1998). Breaching or building social boundaries?:
SIDE-effects of computer-mediated communication. Communication Research,
25(6). 689-715.

Postmes, T., Spears, R., & Lea, M. (1999). Social identity, normative content and
“deindividuation” in computer-mediated groups. In N. Ellemers, R. Spears, & B.
Doosje (Eds.), Social identity: Context, commitment, content (pp. 174-183).
Oxford, UK: Blackwell.

Postmes, T., Spears, R., Sakhel, K., & 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.

Price, R. H., & Bouffard, D. L. (1974). Behavioral appropriateness and situational

65

constraint as dimensions of social behavior. Journal of Personality and Social
Psychology, 30, 579-586.

Rafaeli, S., & Sudweeks, F. (1997). Networked interactivity. Journal of Computer-
Mediated Communication, 1(4).
http://www.ascusc.org/jcmc/volZ/issue4/rafaeli.sudweekshtml

Reicher, S. D. (1987). Crowd behaviour as social action. In J. C. Turner, M. A.
Hogg, P. J. Oakes, S. D. Reicher, & M. S. Wetherell (Eds.), Rediscovering the

social group: A self-categorization theory (pp. 171-202). Oxford, UK: Basil
Blackwell.

Reicher, S., Spears, R., & Postmes, T. (1995). A social identity model of deindividuation
phenomena. In W. Stroebe & M. Hewstone (Eds.), European Review of Social
Psychology: Vol. 6. (pp. 161-198). Chichester: Wiley.

Santee, R. T., & Maslach, C. (1982). To agree or not to agree: Personal dissent amid
social pressure to conform. Journal of Personality and Social Psychology, 42(4),
690-700.

Schroeder, R. G., & Benbassat, I. (1975). An experimental evaluation of the relationship
of uncertainty in the environment to information used by decision makers.
Decision Sciences, 6, 556-567.

Singelis, T. M. (1994). The measurement of independent and interdependent self-
construals. Personality & Social Psychology Bulletin, 20(5), 580-591.

Snyder, C. R. (1992). Product scarcity by need for uniqueness interaction: A consumer
catch-22 carousel? Basic and Applied Social Psychology, 13 (March), 9-24.

Snyder, C. R., & Fromkin, H. L. (1977). Abnormality as a positive characteristic:
The development and validation of a scale measuring need for uniqueness.
Journal of Abnormal Psychology, 86, 51 8-527.

Snyder, C. R., & Fromkin, H. L. (1980). Uniqueness: The human pursuit of difierence.
New York: Plenum.

Spears, R., Lea, M. & Postmes, T. (2000). One side: Purview, problems and prospects. In
T. Postmes, R. Spears, M. Lea, & S. Reicher (Eds.), SIDE issues centre stage:
Recent developments in studies of deindividuation in groups (pp. 1-16).
Amsterdam: Royal Netherlands Academy of Arts and Sciences.

Spears, R., Postrnes, T., Lea, M., & Watt, S. E. (2001). A SIDE view of social influence.

In J. Forgas & K. Willams (Eds.), Social Influence: Direct and indirect processes
(pp. 331-350). Philadelphia: Psychology.

66

Sproull, L., & Kiesler, S. (1986). Reducing Social Context Cues: Electronic Mail in
Organizational Communication. Management Science, 32(11), pp. 1492-1512.

Sproull, L., & Kiesler, S. (1991). Connections : new ways of working in the networked
organization. Cambridge, Mass: MIT Press.

Steiner, I. D. (1972). Group process and productivity. New York: Academic Press.

Stone, E. F., & Hollenbeck, J. R. (1984). Some issues associated with the use of
moderated regression. Organizational Behavior and Human Performance, 34,
195-213.

Tajfel, H. (1978). Differentiation between Social Groups. London: Academic Press, 1978.

Tajfel, H. (1981). Human Groups and Social Categories. Cambridge: Cambridge
University Press.

Tajfel, H. & Turner, J. C. (1986). The social identity theory of inter-group behavior. In S.
Worchel and L. W. Austin (eds.), Psychology of Intergroup Relations. Chigago:
Nelson-Hall.

Terry, D. & Hogg, M. (2001) ‘Attitudes, Behavior and Social Context: The Role of
Norms and Group Membership in Social Influence Processes’, in J. Forgas and K.
Williams (eds.) Social Influence: Direct & Indirect Processes (pp.253-70).
Psychology Press.

Turner, J. C. (1982). Towards a cognitive redefinition of the group. In H. Tajfel (Ed),
Social identity and intergroup relations (pp. 15-40). Cambridge, UK: Cambridge
University Press.

Turner, J. C. (1984). Social identification and psychological group formation. In H. Taj fel
(Ed.), The social dimension: European developments in social psychology (pp.
518-538). Cambridge, UK: Cambridge University Press.

Turner, J. C., Hogg, M. A., Cakes, P. J ., Reicher, S. D., & Wetherell, M. S. (1987).
Rediscovering the social group: A self-categorization theory. Oxford, England:
Blackwell.

Vignoles, V. L., Chryssochoou, X., & Breakwell, G. M. (2000). The distinctiveness
principle: Identity, meaning and the bounds of cultural relativity. Personality and
Social Psychology Review, 4(4), 337-354.

Wallach, M. A., & Wallach, L. (1983). Psychology’s sanction for selfishness: The error
of egoism in theory and therapy. San Francisco: Freeman.

67

Walther, J. B., & Burgoon, J. K. (1992). Relational communication in computer-mediated
interaction. Human Communication Research, 19(1), 50-88.

West, S. G., Aiken, L. S., & Krull, J. L. (1996). Experimental personality designs:
Analyzing categorical by continuous variable interactions. Journal of Personality,
64(1), 1-48.

Wikipedia. (2006). Retrieved March 15, 2006, from
http://en.wikipedia.org/wiki/Avatar_(virtual_reality)

Zimbardo, P. (1969). The human choice: Individuation, reason, and order versus

deindividuation, impulse and chaos. In W. Arnold & D. Levine (Eds.), Nebraska
symposium on motivation (pp. 237-307). Lincoln: University of Nebraska Press.

68

ENDNOTE

 

l Perceived deindividuation items

1) I think I could be who I am in this group

2) I think I saw myself predominantly as an individual in this group

3) I think I considered myself as an individual different from other members in this group
4) I think my distinctive individuality was oppressed in this group

5) I think I was not considered as a distinctive individual in this group

6) I think members in this group were not represented as unique individuals in this group

2 Need for uniqueness items

1) I find that criticism affects my self-esteem.

2) I sometimes hesitate to use my own ideas for fear they might be impractical.

3) Others’ disagreements make me uncomfortable.

4) Feeling “different” in a crowd of people makes me feel uncomfortable.

5) I would rather be just like everyone else than be called a “freak.”

6) It is better always to agree with the opinions of others than to be considered a disagreeable
person.

7) I tend to keep quiet in the presence of persons of higher rank, experience, etc.

8) In most things in life, I believe in playing it safe rather than taking a gamble.

3 Interdependent self-construal items

1) I feel uncomfortable disagreeing with my group.

2) My happiness depends on the happiness of those in my group.

3) I am careful to maintain harmony in my group.

4) When with my group, I watch my words so that I will not offend anyone.

5) It is important to consult close friends and get their ideas before making decisions.
6) The security of being an accepted member of a group is very important to me.

4 Group identity items

1) I have a lot in common with other group members

2) I find it easy to identify with this group

3) I find that my values and the values of other group members are very similar

5 Conformity intention items

1) I am willing to agree with them

2) I am willing to follow their opinion

3) I am willing to join them as the same group member

4) I am willing to share the same opinion as the same group member

69