Review of Virtual Environment Studies on Social and Group Phenomena

Published in Small Group Research (2011), 42(4): 405-457

2

Review of Virtual Environment Studies on Social and Group Phenomena

Anu Sivunen

BIT Research Centre Aalto University School of Science and Technology Finland

[email protected]

Marko Hakonen

BIT Research Centre Aalto University School of Science and Technology Finland

Abstract

This article provides a review of previously published studies on virtual environments (VEs),

focusing especially on empirical articles on social and group phenomena in VEs and their

methodological and theoretical trends. Virtual environments can be defined as

communication systems in which interactants share the same three-dimensional digital space

and can navigate, manipulate objects, and interact with one another via avatars. When

examining the methodological and theoretical choices of these studies, four trends could be

identified that characterize group studies on virtual environments: 1) testing the applicability

of real-life, social behavior norms in VEs, 2) a lack of work group studies using VEs, 3) the

micro-level treatment of social and groups and 4) a lack of covering theory. Finally,

propositions for future research are presented.

Keywords: computer-mediated communication, social psychology, virtual environment,

virtual group, virtual world


2

Review of Virtual Environment Studies on Social and Group Phenomena

Virtual environments (VEs), such as Second Life and Entropia Universe, have

gained massive popularity during the past decade (e.g. Gillath, McCall, Shaver, Blascovich,

2008). Millions of users interact daily in 3D VEs, which have been developed to resemble the

real world more and more and simultaneously provide possibilities to surpass real-life

obstacles (e.g., flying, teleporting). Virtual environments can be defined as communication

systems in which multiple interactants share the same three-dimensional digital space despite

occupying remote physical locations and can navigate, manipulate objects, and interact with

one another via avatars (Sallnäs, 2005; Yee & Bailenson, 2007). Avatars are flexible and

easily transformed digital self-representations in a graphic 3D form (Yee & Bailenson, 2007).

In addition to 3D worlds developed primarily for massive multi-user online game-

playing, there are also environments in the market that have no specific focus on particular

activities but that provide their users with a modifiable environment for collaboration. This

relatively new forum for human interaction has inspired scholars to study the social aspects of

VEs. In their seminal article, Blascovich et al. (2002) suggest that VEs provide an intriguing

new platform for empirical research in social psychology. They argue that the limitations of

real life studies such as the trade-off between realism and control, the impossibility of perfect

replications and the use of non-representative samples, especially in experimental studies, can

be avoided in VEs. Current VE technology enables realistic set-ups which are nevertheless

controllable, in which the conditions can be simply copied over time, and the population

interacting in VEs is highly representative. These points underline that the research potential

of VEs is enormous.

VE research has taken many directions, from technology-driven approaches to

individual psychology to studies of educational possibilities in VEs. Studies have been

https://www.researchgate.net/publication/227516065_The_Proteus_Effect_The_Effect_of_Transformed_Self-Representation_on_Behavior?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/220089751_Effects_of_Communication_Mode_on_Social_Presence_Virtual_Presence_and_Performance_in_Collaborative_Virtual_Environments?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


3

conducted on themes such as presence and co-presence felt in VEs (Axelsson, Abelin, Heldal,

Schroeder & Wideström, 2001; Hoyt, Blascowich & Swinth, 2003; Slater, Sadagic, Usoh &

Schroeder, 2000), the verbal and nonverbal communication of the avatars operating in them

(Bailenson & Yee, 2006), the effects of avatar appearance on individual perceptions (Yee &

Bailenson, 2007), and the effects of different communication media (voice, text) on virtual

environments (Williams, Caplan & Xion, 2007). Despite the growing interest in the area,

little is known on the costs and benefits of using these environments for group work,

especially for organizational groups and teams.

The goal of this article is to integrate the findings of studies of virtual environments

and to identify directions for future research. We do this by examining peer-reviewed articles

published in academic journals, concentrating especially on studies that cover social

(psychological) and group phenomena in VEs. However, our review is not a conventional one

in that, rather than categorizing the articles studied according to their contents, we focus more

on methodological and theoretical trends to be found in them. The reason for this is that

research on virtual environments is not yet very robust, and conclusive classifications based

on study results are still difficult to make. Instead, our aim is to review articles in a

constructive manner and find the opportunities and challenges for further research within this

rather young field of study.

One of the main currents in VE studies deals with massively-multiplayer online role-

playing games (see e.g., Yee, 2006). We do not fully exclude game-playing studies, but due

to our interest in broadening the scope of research towards the potential of VEs for work

groups, the fully anonymous, fun-motivated game-playing VEs are not in the center of our

scrutiny.

Despite the burgeoning literature on virtual environments, there are no

comprehensive reviews on the use of VEs from a social scientific point of view. Our article


4

tries to fill this gap by providing a thorough exploration of empirical articles published in the

area. The main contribution of our review arises from the scrutiny of VE studies on social

and group phenomena within the context of four trends we observed in the literature

reviewed. Here we must highlight two points: we do not wish by any means to downplay

current VE research, and we recognize that steps have already been taken in the directions we

propose later. However, in the light of our review, these four trends prove to be rather

dominant and raise many questions and propositions for future research. First, many studies

assume or try to demonstrate that findings from real-life studies and “laws of social behavior”

apply also in VEs. Notwithstanding the importance of this endeavor, this trend raises a need

to study VE-specific dynamics in order to understand the possibly unique nature of these new

platforms of social action. Second, the lack of attention to VEs as potential platforms for

work groups, especially for virtual teams, prompted us to discuss the importance of this

notion and to give examples of how to broaden VE research to this realm. Third, it seems that

the dominant approach to social phenomena in current research is that researchers focus on

micro-level social phenomena (e.g., proximity, gaze, etc.), sometimes at the expense of

broader group-level phenomena like social identification with, or leadership in, groups.

Fourth, following from the last point, we argue that this concentration on interesting

phenomena has often overshadowed the efforts to bridge the research and findings with

covering and well-established (even dominant) theoretical approaches in social sciences.

The structure of this review is as follows. In section two we present the criteria and

methods of our literature review. Section three is devoted to supporting the argument on the

four trends observed and illustrating them with examples from the studies reviewed. In

section four, we suggest how to go further from each dominant trend and build propositions

for further VE research into social and group phenomena. In the last section, we summarize

our findings and discuss the conclusions of our review.


5

Methodology

The studies presented in this review were identified by various means. First, we

searched the main electronic databases (e.g. Ebsco, PsycINFO, Sage, JSTOR, Emerald) for

studies in communication, social psychology, business, and information systems using an

extensive list of relevant terms (e.g. “virtual environment and collaboration”, “virtual world

and communication”). From the articles found, we selected only the peer-reviewed empirical

studies. Next, we scanned through several leading journals in psychology, communication,

information systems, and management using similar key words. Third, we manually scanned

the reference lists of the articles identified by the first two methods. This procedure provided

us with 67 articles, of which 47 were selected for the final review (see Appendix). Articles

that were left out from the final review were exclusively game-, education- or clinical

psychology-related or focused only on individual, not social, phenomena (see a review on

e.g. clinical psychology and training-related VE studies on Fox, Arena & Bailenson, 2009).

The results of the review are presented in the following section.

Trends Identified in the Literature

Appendix presents the 47 articles studied in this review, including a short summary

of their methods and results. From these 47 articles, we identified four different trends which

characterize studies on virtual environments: 1) testing the applicability of real-life, social

behavior norms in VEs, 2) a lack of work group studies using VEs, 3) an emphasis on micro-

level treatment of social and groups and 4) a lack of covering theory. Table 1 shows how we

categorized each of the 47 articles in terms of these four trends.


6

Table 1 Classification of the Reviewed Articles According to the Four Trends Outlined Above

Authors Year Trend 1 Trend 2 Trend 3 Trend 4 Antonijevic 2008 X Axelsson, Abelin, Heldal, Schroeder & Wideström 2001 X X Axelsson, Abelin & Schroeder 2003 X Bailenson, Beall & Blascovich 2002 X X X X Bailenson, Beall, Loomis, Blascovich & Turk 2005 X X X Bailenson, Blascovich, Beall & Loomis 2003 X X X X Bailenson, Blascovich & Guadagno 2008 X X Bailenson & Yee 2005 X X X X Bailenson & Yee 2006 X X X Bailenson, Yee, Merget & Schroeder 2006 X X Bente, Rüggenberg, Krämer & Eschenburg 2008 X X Diehl & Prins 2008 X Dotsch & Wigboldus 2008 X X X Feldon & Kafai 2008 X X Fiedler 2009 X X Fiedler & Haruvy 2009 X X X Fox & Bailenson 2009 X X Friedman, Garmiel & Dinur 2009 X X X Gillath, McCall, Shaver & Blascovich 2008 X X X Gong & Nass 2007 X X Guadagno, Blascovich, Bailenson & McCall 2007 X X X Hindmarsh, Heath & Fraser 2006 X X Hoyt, Aguiular, Kaiser, Blascovich & Lee 2007 X X Hoyt & Blascovich 2003 X X Hoyt & Blascovich 2007 X X Hoyt, Blascovich & Swinth 2003 X X X Jackson, Zhao, Qui, Kolenic, Fitzgerald, Harold & Von Eye 2008 X X Jung & Lee 2009 X X X Lantz 2001 X Linebarger, Janneck & Kessler 2005 X X X Nakanishi 2004 X X X Nowak 2004 X X Nowak & Biocca 2003 X X Nowak & Rauh 2005 X X Nowak & Rauh 2008 X X Peña, Hancock & Merola 2009 X X X Sallnäs 2005 X X Scheck, Allmendinger & Hamann 2008 X X Slater, Sadagic, Usoh & Schroeder 2000 X X Talamo & Ligorio 2001 X Vasalou & Joinson 2009 X X X Vasalou, Joinson, Bänzinger, Goldie & Pitt 2008 X Vinayagamoorthy, Steed & Slater 2008 X X X X Yee & Bailenson 2007 X X X Yee & Bailenson 2009 X X Yee, Bailenson & Duchenaut 2009 X X X Yee, Bailenson, Urbanek, Chang & Merget 2007 X X X X

We estimated the accuracy of our categorization of the articles by conducting an

inter-rater agreement test. After our own coding we recruited two researchers who had not


7

been involved in the writing of this article. We gave them both all 47 articles, explained our

four trends and their inclusion/exclusion criteria and asked them to read and code the articles

independently. We used our own ratings as a third “coder” in the test. We had to count the

inter-rater agreement for each trend separately because the trends are not mutually exclusive.

However, by coding each article dichotomously as either belonging to the trend in question

(value = 1) or not belonging to it (value = 0) we were able to create a table for the calculation

of three-rater, two-category Fleiss’s Kappa for each trend (Fleiss, 1971). The Kappas were

.60 (SE = .08) for trend one, 1.0 (SE = .08) for trend two, .71 (SE = .08) for trend three .75

(SE = .08) for trend four. This indicated moderate agreement for trend one, perfect one for

trend two and strong agreement for trends three and four. The strongest disagreements

pertained to the trend one which our external raters noted to be sometimes slightly

ambiguous. Trend two was, as expected, totally unambiguous due to its rather factual nature.

Finally, the rather small disagreements in coding were negotiated between us and the two

external raters and they resulted in no changes. Overall, the inter-rater agreement was very

satisfactory. We trust that these agreement rates are sufficient for the purposes of categorizing

articles into trends which is our way of pointing out interesting lines in current VE research –

not finding clear-cut facts as often is the case when many coders are used.

Trend One: Testing the Applicability of Real-Life, Social Behavior Norms in VEs

As virtual environments make it possible to study social phenomena that would be

difficult or even impossible to study in real life, several scholars have made an effort to

construct complex set-ups and conduct laboratory studies to find out whether social norms

and behavior transfer from real life to VEs. Sixteen studies out of 47 explored the transfer of

known real-life social norms to a virtual environment. These norms include, for example,

interpersonal distance and eye contact (Yee, Bailenson, Urbanek, Chang & Merget, 2007),


8

interpersonal distance and discrimination (Dotsch & Wigboldus 2008), the presence of others

and its impact on task performance (Hoyt et al., 2003), and reactions to persons in need

(Gillath et al., 2008). This trend does not seem to be very strong based on the number of

articles categorized to it. Had we included all the articles that referred to real life-based

theories, studies or phenomena, almost all reviewed articles would have been included in this

category. Instead, we decided to include only those articles which clearly tested whether

some phenomenon found in real life (e.g., interpersonal distance) applied also to VE. Hence,

for example, most articles on avatar characteristics were not included even though their

theory sections included references to real-life research.

In the study of Yee and his colleagues (2007), the authors found that in a virtual

environment, a decrease in interpersonal distance was compensated for by avoiding eye-

contact with the interacting partner. In Dotsch and Wiboldus’s study (2008), subjects

maintained a larger personal distance in a VE with avatars that belonged to a different ethnic

background and looked different from their own avatars (and themselves in real life). A

questionnaire also revealed the subjects’ implicit prejudice against this same ethnic minority.

In a classic replication of the impact of the presence of others on the performance of

individuals, Hoyt and co-authors (2003) showed that social inhibition effects also occur in

VEs. Participants performed a novel task significantly worse when other avatars were

observing their performance than when performing the same task alone. Furthermore, basic

emotions, such as empathy towards persons in need, can be triggered in the same vein in

experimental set-ups in VE’s as in real life (Gillath et al., 2008).

Despite the eagerness to test whether social laws persist in VEs, there are also

several studies that have explored phenomena that would be very difficult or even impossible

to test in real life, such as avatar androgyny (Nowak & Rauh, 2008) or interaction with one’s

photographic self-representation (Bailenson et al., 2008). A good example is the seminal

https://www.researchgate.net/publication/240276771_Transformational_and_Transactional_Leadership_in_Virtual_and_Physical_Environments?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/220495628_Choose_your_buddy_icon_carefully_The_influence_of_avatar_androgyny_anthropomorphism_and_credibility_in_online_interactions?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


9

study of Yee and Bailenson (2007) on the appearances of avatars and their effects on social

behavior. The influence of facial attractiveness and height showed that looks matter in social

relationships. Participants interacting in a VE with attractive-looking avatars walked closer to

the confederate and revealed more information than unattractively represented avatars.

Moreover, participants with tall avatars more often split money in their own favor than

participants with short avatars, who were more likely to accept an unfair offer than

participants in normal and tall condition. These phenomena occurred independently of the

real height and appearance of the participants.

In conclusion, scholars have put much effort into proving that several social

phenomena familiar from real environments also exist in virtual environments. Our need for

personal space, gaze transfer and perceptions of persons of specific appearance in real life

seem to persist when we interact as avatars in a virtual world. However, on the basis of the

studies reviewed, we do not know much about new types of behavior or about the

characteristics that would be specific to virtual world interaction, as the studies have focused

less on the new phenomena occurring in these environments.

Trend Two: The Lack of Work Group Studies Using VEs

A massive trend in the articles reviewed was the use of student subjects and the lack

of groups that had a common history or background, such as organizational teams. Thirty-six

out of 47 studies used student or pupil participants, and five studies reported that their

subjects had been recruited from a university campus, making the total number of studies

using university or school-related subjects as high as 41.

Only one of the 47 articles reviewed studied organizational work groups. This study,

conducted by Lantz (2001), observed the implementation of a virtual environment for a

workgroup that had four geographically distributed members and had earlier been using only

https://www.researchgate.net/publication/240680546_Meetings_in_a_distributed_groups_of_experts_Comparing_face-to-face_chat_and_collaborative_virtual_environmentsto_be_published_in_BIT?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


10

face-to-face meetings. The study found that meetings held in a VE were considered to be

more task-oriented than face-to-face meetings. It was also discovered that graphic avatars

played an important role in turn-taking and making the meetings more enjoyable in the virtual

environment.

As virtual environments can be seen as new collaboration tools or workspaces for

geographically dispersed work teams and organizations, the absence of empirical, peer-

reviewed studies on VE use in such contexts is rather surprising. The development of VEs,

such as Second Life, provides many opportunities, especially for geographically distributed

work groups. Such groups are also known as virtual teams, that is, groups of people striving

toward a common goal, dispersed in many locations, and communicating with each other

predominantly via information and communication technology (e.g., Axtell, Fleck, & Turner,

2004). VEs can be seen as a new media for virtual teams that struggle with communication

problems over distance (see Hertel, Geister, & Konradt, 2005; Martins, Gibson, & Maynard,

2004 for reviews of virtual team literature). The tools available even in many desktop-based

VEs, such as chat, voice and whiteboards, not to mention the avatars that can point, express

emotions and use tools, are powerful competitors to more traditional means of remote

interaction through expensive and static videoconferencing systems or simple but

asynchronous e-mails.

As working life is becoming more and more global and as companies are

simultaneously striving to minimize their travel costs and to find new ways of collaboration,

we attribute great importance to the research on VEs as potential media for distributed

organizational teams. Virtual environments can also provide a more inspiring experience for

collaborators than standard practices by enabling the team members to do several

unconventional activities together in a 3D environment, such as flying, diving or sitting

around a camp fire. Such activities might serve as team building functions and further


11

strengthen the participants’ identification with their virtual team (Sivunen, 2006).

Furthermore, meetings in VEs can be more enjoyable as the team members can flexibly

change and modify the meeting environment as well as their avatars’ graphic appearance.

Trend Three: The Micro-Level Treatment of Social and Group Phenomena

We classified 33 out of 47 articles as representing a trend in which the scope of

social phenomena could be characterized to be at micro- or individual level rather than at

macro- or group level. Our argument is based on two notions. First and most importantly, all

of these 33 articles studied some social tendency (e.g., interpersonal distance) of individuals

or some more VE-specific social construct (e.g., avatar appearance) of individual behavior or

attitudes in the presence of one or two others, thus representing a somewhat limited view of

groups. Second, the fact that 31 out of these 33 studies were conducted in laboratory settings

gave indirect support to our claim as to the micro-social trend because these set-ups are rarely

used in the study of larger groups and macro-level social or group phenomena. Nevertheless,

our identification of a trend does not imply that these studies may not be relevant or of high

quality. For example, many studies (e.g., on communication modalities) ingeniously utilized

the specific features and conditions of VEs to broaden the scope and findings of traditional

computer-mediated communication (CMC) literature (e.g. Daft & Lengel 1984; Rice &

Gattiker, 2001). Nevertheless, the trend reveals the need for more studies of broader-scale

social phenomena (e.g. leadership or social identity) occurring in real and larger groups than

dyads within VEs.

The two examples of this mainstream trend and one exception described below

illustrate the difference between micro and macro levels or individual- and group-centered

approaches. An example of the micro-level scrutiny of social behavior is the excellent study

on interpersonal distance and eye gaze transfer in VEs (Bailenson, Blasovich, Beall &

https://www.researchgate.net/publication/226342450_Strengthening_Identification_with_the_Team_in_Virtual_Teams_The_Leaders'_Perspective?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


12

Loomis, 2003). In this laboratory experiment, a total of 160 undergraduate students

participated under two experimental conditions. Using fine-grained and impressive data

collection techniques, including head-mounted displays (HMDs) and tracking devices, the

researchers found that participants, through their avatars, were inclined to keep interpersonal

distance in VE (as in real life) and that they compensated for shorter interpersonal distance by

avoiding looking at the other avatar or agent.

Another example of this trend is a study that examined the reliance on visual

characteristics in online avatar perception and especially the relationships between avatar

androgyny, anthropomorphism and credibility (Nowak & Rauh, 2008). In a two-step research

process, the first sample of university student participants selected eight maximally differing

avatars in terms of the above-mentioned dimensions, which were used for the step 2

experiment. In the second step, a different set of student participants rated their text chat

partners, represented visually by the previously-chosen face images, based on the three

dimensions. Path analysis revealed that perceived avatar androgyny negatively predicted

perceived avatar anthropomorphism, which in turn positively predicted perceptions of partner

credibility.

A counter-trend example is the study by Talamo and Ligorio (2001) on the

construction of identities in a VE using ethnographic and conversational methodology. The

task of the multicultural participant group (N = 48) was to build a community in a VE called

Active Worlds in eight months. Two types of data were collected: 1) open responses to a

questionnaire, 2) recorded chat. It was found that identity construction was heavily context-

dependent and identities were not static but negotiated in interactions. Many strategies for

playing with different identities and making them distinct were observed – most commonly

that of trying different avatars. Here the scope was behavior in a rather large group. The set-


13

up did not limit the observations to any particular level (individual or group) and the

longitudinal approach ensured rich data and variations in behavior over time.

All three examples are well-designed and rigorously reported; in other words, they

represent sophisticated academic research. However, the given examples of the micro-social

trend clearly focus on individual reactions to rather limited social stimuli. For example, in

Nowak and Rauh’s (2008) study, the stimuli were the visual cues provided by static pictures

of individual faces. The micro-level, individual-centered treatment of social phenomena in

these and the other 31 studies showing a similar trend simply raises the question of how we

could be better informed about larger-scale, group-level dynamics in VEs. The counter-

example presented above gives some clues as to how to enrich research on social and group

phenomena in VEs. However, we would like to go further and encourage VE scholars to

study, for instance, whether group-level social identification predicts pro-social behavior in

VEs just as it has been shown to do in the real world (e.g., Tyler & Blader, 2001) and in

distributed teams (Hakonen & Lipponen, 2007).

Trend Four: The Lack of Covering Theory

The question of the lack of larger-scale theoretical approaches among the articles

reviewed (18 out of 47) is related to, but not the same as, the previous trend, namely the

concentration on micro-social phenomena. It seems to us that often the concentration of

interest upon an in-depth scrutiny of fairly similar, interesting but rather narrow research

fields (e.g., proximity) tends to lead to theoretical discussion that is more concerned with

previous findings and small-scale theories than with bridging the findings to more covering

and even dominant larger-scale theoretical approaches in social sciences. Again, this is not a

value statement. Deepening research certainly produces valuable accumulating knowledge on

the constructs studied. Furthermore, it must also be noted that different journals have

https://www.researchgate.net/publication/220495628_Choose_your_buddy_icon_carefully_The_influence_of_avatar_androgyny_anthropomorphism_and_credibility_in_online_interactions?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


14

differing demands and space available for theoretical discussion. Hence, the lack of covering

theory in an article should not be attributed to lack of theoretical understanding on the part of

the authors but rather to the style of particular journals publishing VE research. Still,

interestingly, in their seminal conceptual article, Blascovich and his co-authors (2002) do

consider VEs to be a “methodological tool for social psychology.” The authors consider how

VE might help the study of different social and group phenomena but pay fairly little

attention to how to link this methodological promise to theoretical discourses. Our point here

is that research on social and group phenomena in VEs might benefit and be enriched by

linking future studies to well-established theoretical discussions and paradigms within social

sciences.

The study of cultural differences in morality in real and virtual worlds (Jackson et

al., 2008) is an example of this lack-of-theory trend. The authors cite Kohlberg’s theory of

moral development (1984) and briefly refer to cross-cultural research (e.g., Bond, 1986) in

the introduction. They found that Chinese youth emphasized conventional moral off-line

behaviors but accepted morally questionable on-line behavior more than their US

counterparts. However, while discussing this counterintuitive result showing inconsistent

moral norms on-line and off-line, the authors do not refer to any of the theories or sources

they presented in their introduction. The interesting result is left open to speculation and not

rooted in, for example, discussion on dissonance of social cognitions. In fact, one rather

commonly used theoretical framework was the social cognition research and its different

varieties (e.g., Gong & Nass, 2007; Hoyt et al., 2007). These tend to fall in the mid-range of

our trend: the hypotheses and results are grounded in a well-specified theory, but their

anchoring to broader theoretical approaches is not always very clear. However, the potential

of VEs for social cognition research was noted as early as 2002 by Groom, Sherman and

Conrey. We coded social cognition-based articles as having a covering theory, because social

https://www.researchgate.net/publication/222954273_The_self-protective_and_undermining_effects_of_attributional_ambiguity?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


15

cognition is a major theoretical field in communication and social psychology. Consequently,

this trend does not seem to be very strong. Had we decided otherwise and judged whether an

article had rooted its social cognition theorizing "well enough" to be merited as having a

covering theoretical basis, the coding criteria would have become too ambiguous.

A good but unfortunately rare example of the type of theory-use we recommend VE

scholars to adopt is exploring of the construction of cultural identity and intercultural literacy

in Second Life (Diehl & Prins, 2008). The authors based their study firmly on Cultural

Historical Activity Theory (e.g., Engestrom, 2001) and Heyward’s model of cultural literacy

(2002). It was found that Second Life interaction increased intercultural literacy and that

avatar appearance was used in cultural identity construction. Throughout the article, the

reasoning and findings are rooted to these two theoretical frameworks. Cultural Historical

Activity Theory may not be the best known tradition in social sciences, but it certainly fulfills

the criteria of a covering (meta)theory that has been used to explain various social

phenomena. Another example comes from the experimental research tradition and goes a

long way in the direction we would recommend: the study by Yee and Bailenson (2007) on

the so called Proteus Effect – that is, the effects of different self-representations (e.g., avatar

height) on behavior. The introduction and theoretical discussion goes deeply into CMC

literature (e.g., Postmes, Spears, & Lea, 1998; Walther, 1996), the hypotheses are firmly built

on well-developed theory, and the results are discussed in terms of the theoretical framework

presented.

Notwithstanding the promising counter-examples to this trend presented above, we

want to go further and suggest a particular, concrete theoretical approach for the study of

social and group phenomena in VEs. The Social Identity Approach (SIA; Tajfel & Turner,

1979) provides a theoretically coherent and covering view of many social psychologically

relevant phenomena, such as conformity, cohesion and leadership. Thus, we see no reason

https://www.researchgate.net/publication/227516065_The_Proteus_Effect_The_Effect_of_Transformed_Self-Representation_on_Behavior?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/242531588_Computer-Mediated_Communication?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


16

why the SIA should not be utilized in the study of VEs. More specifically, the SIA has been

found to be useful in understanding social phenomena in work contexts (Haslam, 2004;

Riketta 2005), and as such it provides a theoretical approach to our interests in broadening

VE research to work settings (trend two) and to group-level studies. We elaborate this

argument and present the SIA in detail in the next section.

Agenda for Further Research

In the space of a few years, there has been an increase in research interest in virtual

environments and their effects on our collaboration and communication. As virtual worlds

develop and gain wider popularity among groups and organizations, studies in the area will

probably also increase in the future. Researchers are only now beginning to understand the

effects of virtual environments on groups’ collaboration, and our review suggests that more

work remains to be done to understand the full potential of virtual environments for social

science as well as for distributed groups and organizations. We will now elaborate on the four

trends found in our review and propose directions for further research.

Exploring Virtual Environment-Specific Phenomena

The review shows that there is an accumulating body of knowledge in the area of

studies related to the transfer of real-life social norms to virtual environments. People act

similarly in many ways whether they collaborate with one another in real life or in virtual

worlds through avatars. However, there is a gap in research focusing on virtual environment-

specific features and their effects on group behavior. Therefore, virtual environments still

present great opportunities for empirical social science research.

One virtual environment-specific phenomenon that has not been studied

comprehensively is the development of groups in virtual environments. Studying groups that


17

collaborate solely in a virtual world for a longer period of time and do not see one another

face-to-face might reveal something new related to the development of roles, leadership and

power relations in such groups.

Another aspect that has been studied in more traditional virtual contexts, such as in

computer-mediated groups and communities, but not in virtual environments, is the effect of

anonymity on group members’ identities, trust and perception of others (see e.g. Jarvenpaa &

Leidner, 1998; Lea & Spears, 1991). It would be interesting to explore what kind of impact

the presentation of others as anonymous, 3D graphical avatars has on the development of

trust and identities between the members of a group working together in a virtual

environment. In addition to experimental studies, research is needed on longitudinal and

naturally occurring group behavior where members build their trust and perceptions of others

on the interactions taking place in a virtual environment.

Furthermore, there are several social occurrences that could be studied as virtual

environment-specific phenomena. The existing research has focused more on replicating the

“laws of real-life social behavior” in VEs, such as whether the presence of avatars can trigger

social inhibition effects (Hoyt et al., 2003) or certain emotions (Gillath et al., 2008).

Fascinating as the results of such studies are, it would be valuable to give attention to virtual

environments as their own type of communication environment and not only as a substitute or

comparison for real-life settings. Instead of comparing VEs to real-life contexts, future

research should focus on the specific social affordances these environments provide for

groups.

One example of a virtual environment-specific phenomenon that can be very

different from real-life experience is self-disclosure and impression management via 3D

avatars. Research on text-based computer-mediated communication shows that mediated self-

presentation differs from unmediated self-disclosure (e.g. Tidwell & Walther, 2002).

https://www.researchgate.net/publication/240276771_Transformational_and_Transactional_Leadership_in_Virtual_and_Physical_Environments?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/220106860_Computer-Mediated_Communication_De-Individuation_and_Group_Decision-Making?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


18

However, the way we present ourselves when collaborating via graphic 3D avatars and the

mechanisms by which first impressions are formed from others’ graphic representations may

uncover some unique features that could not be found by studying mediated settings with text

or video-based communication technology – nor real-life face-to-face groups. Studying

anonymous groups collaborating for the first time in a virtual world might illustrate some

social mechanisms that do not exist outside graphic 3D space. The circumstances of such a

collaboration environment may allow or promote certain social, cognitive and communicative

processes that could not occur in other environments (see also Walther, 2007).

Studying virtual environments as a collaboration space for distributed teams

One of the findings of our review that has direct practical implications is the

surprising lack of study of work groups using VEs as a collaboration space. Only one of the

articles reviewed studied a geographically distributed team that used virtual world for

common meetings (Lantz, 2001). It is understandable that obtaining data from field settings

can be very difficult, especially as many organizations do not use virtual environments yet in

their daily activities. However, it is clear that distributed teams and organizations would

benefit greatly from using virtual environments for performing their real-life, organizational

tasks and that studying such naturally-occurring teams would advance the literature through

asking and answering questions that cannot be sufficiently tested in laboratory settings. Some

interesting questions for future research include: What are the implications of collaborating in

a virtual environment for the productivity of distributed teams? Can virtual environments be

efficiently used for organizational training or team-building activities? How do virtual

environments affect the distributed team members’ identification with their team?

Studying the use of virtual environments in organizational settings would also

provide new insights for inter-group and inter-organizational collaboration research. Thus far,

https://www.researchgate.net/publication/222652602_Selective_Self-presentation_in_Computer-mediated_Communication_Hyperpersonal_Dimensions_of_Technology_Language_and_Cognition?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/240680546_Meetings_in_a_distributed_groups_of_experts_Comparing_face-to-face_chat_and_collaborative_virtual_environmentsto_be_published_in_BIT?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


19

micro-social behavior occurring in isolated laboratory groups has received a fair amount of

attention in VE studies (see also trend three). If VE research were more oriented toward

conducting field studies on distributed teams and organizations, the relations and networks of

the groups and organizations under study could better be taken into consideration. As virtual

worlds, such as Second Life, consist of public and private collaboration spaces, these

environments can provide better possibilities for inter-organizational collaboration than the

closed private collaboration tools existing today. For social science research, this would give

a tremendous opportunity to study inter-group as well as inter-organizational collaboration in

naturally occurring contexts.

Finally, virtual environments provide new methods for studying distributed teams.

Thus far, studies of distributed teams have tended to use either survey or interview methods

to collect data for team members’ perceptions and attitudes (see e.g. Chidambaram, 1996;

Mortensen & Hinds, 2001; Walther & Bunz, 2005). Other methods, such as observation, have

been used much more rarely due to the geographical distance between the team members.

When studying distributed teams that use VEs for collaboration, observing the members’

nonverbal and verbal behavior through avatars is easier, as the members and the researcher

can share the same virtual collaboration space, which can be accessed through the internet all

over the world. Moreover, team members’ textual communication in VE is saved

automatically by the system regardless of where they are or at what time they work. This

gives richer opportunities for data collection than traditional survey and interview methods.

Toward a Broader Social Scope

The articles reviewed do include studies of groups in VEs that broadened the usual

micro-social trend. Many of these used qualitative methodologies, like ethnography. In

addition to the above-mentioned study by Talamo and Ligorio (2001), another encouraging


20

example is Antonijevic’s (2008) study of naturally-occurring nonverbal communication,

which covers larger groups than dyads. Such approaches allow researchers to observe larger

groups than is usually possible in laboratory settings. Moreover, qualitative methods lend

themselves well to the study of naturally occurring groups in VEs, and since VEs are a

relatively new area of scientific research, it would be advisable (see e.g. Denzin & Lincoln,

1994 on the benefits of qualitative research) to use qualitative and exploratory set-ups to find

out whether there is anything unique or novel in these new contexts of human interaction (see

also trend one). This call for broadening the methodological spectrum by no means

downplays the importance of laboratory studies. An intriguing example of socially broader

scale laboratory studies is that of Hoyt and Blascovich (2003) comparing groups under

transactional and transformational leaders, measured in different communication modalities

(face-to-face, audio and IVE). This study also demonstrates that the methodology is not

necessarily an obstacle if the aim is to extend the framing of the phenomena studied to group

level. This study could have been conducted, for example, using individual subjects instead

of triads, but this would have excluded the interaction aspect of “subordinates”. This, in turn,

would have made the measurement of, for example, cohesion rather awkward if not

impossible, since cohesion is by nature a group phenomenon (e.g., Hogg, 1993).

The above-mentioned examples also underline the importance of researchers’ own

interests in shaping the nature of studies. If the community of VE scholars becomes more

interested in broader-level phenomena, such as inter-group relations in VEs, the social scope

will broaden in a natural way. The environment or methods are not the obstacles. For

instance, the study of prejudice by Dotsch and Wigboldus (2008) could quite easily have

been reframed from an individual level of scrutiny and explanation to a more social level.

The study found that white student participants kept a greater distance from Moroccan-

https://www.researchgate.net/publication/247505889_Group_Cohesiveness_A_Critical_Review_and_Some_New_Directions?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


21

looking avatars than white-looking avatars. This avoidance was explained by participants’

implicit prejudices and skin conductance level. An alternative way of studying prejudice

could have included, for example, a longitudinal observation of interaction of avatar groups

representing two different ethnic groups. Repeated measurements of in-group bias and the

scope of group identification could have informed the researchers about the development and

dynamics of prejudice rather differently from the actual study. Again, the point of the

example is by no means to downplay the high quality of the study in question but to show

that changing the scope and framing might have yielded different interpretations of the same

very interesting phenomenon. Given the dominant trend of micro-level studies of social and

group phenomena, we suggest that broader conceptualizations might provide more variety to

the domain of VE studies of this type. This might also attract a larger community of social

scientists to engage in discussion and research on the subject of groups in this new

environment.

Use of Larger Scale Theoretical Approaches – The Social Identity Approach as an

Example

One of the dominant theoretical “umbrellas” in social psychology that has developed

in recent decades is the Social Identity Approach (SIA). We argue that it could be a useful

(meta)theoretical lens for the study of social and group phenomena in the context of VEs.

The SIA provides a theoretical framework for the relationship between individual

and group. Specifically, it consists of two distinct theories: the original social identity theory

(e.g., Tajfel & Turner, 1979), and the more recent self-categorization theory (Turner, Hogg,

Oakes, Reicher, & Wetherell, 1987). Despite certain differences, both theories share the same

fundamental assumption that individuals define themselves in terms of their social group

memberships and that group-defined self-perception produces distinctive effects on social


22

behavior and inter-group relations (Hogg & Abrams, 1988; Turner, 1999). This means that

the more an individual conceives of him- or herself in terms of membership in a group (i.e.,

identifies with the group), the more his or her attitudes and behavior are governed by this

group membership (Hogg & Abrams, 1988; Van Knippenberg & van Schie, 2000).

During the past ten years, social identity principles have been increasingly applied to

the study of organizational psychological processes (e.g., Haslam, 2004; Hogg & Terry,

2001). In this context, organizational or team membership is understood to reflect on the self-

conception in the same way as other social memberships do (Ashforth & Mael, 1989; Hogg

& Terry, 2001). Thus, organizational identification is defined as “the perception of oneness

with or belonging to a group” (Ashforth & Mael, 1989, p. 34). Moreover, it is proposed that

this group-based self-conception leads to activities that are congruent with this identity.

Accordingly, empirical studies have shown that group identification is linked to various

important outcomes, such as high levels of extra-role behaviors (e.g., Tyler & Blader, 2001;

see Riketta, 2005, for a review).

The work context orientation of the SIA also means that it could be usefully applied

to remedying the lack of VE studies of collaboration in work settings as underlined above

(trend two). Recent studies of geographically distributed teams have shown the applicability

of SIA to studies of this organizational form. For example, it has been found that fairness

perceptions are important antecedents of virtual team-level identification and that this social

identification, in turn, leads to advancement of group goals beyond work-role descriptions

(Hakonen & Koivisto, 2008; Hakonen & Lipponen, 2007), as predicted by the SIA. Scrutiny

of antecedents and consequences of group-level identification in general would certainly be

interesting in VE settings, too.

Insofar as the applicability of these and related findings in different settings has not

been thoroughly examined in the VE context, we would urge VE scholars to use the SIA or

https://www.researchgate.net/publication/233729899_Social_Identity_Theory_and_Organization?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/226768500_Social_Identifications_A_Social_Psychology_of_Group_Relations_and_Group_Processes?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==

https://www.researchgate.net/publication/226768500_Social_Identifications_A_Social_Psychology_of_Group_Relations_and_Group_Processes?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


23

other broad theoretical frameworks when considering further studies of social and group

phenomena in VEs. Covering theoretical framings of interesting group dynamics in VEs

might result in innovative cross-fertilization of ideas and publishing forums and hence

broaden and enrich the discussion within social psychology and across other disciplines.

Conclusions

Our goal in this article has been to present and evaluate the existing research on VEs

related to social and group phenomena. Based on our review, we have identified four trends

that prevail in social science research on these environments. It seems that scholars have

focused more on proving how social norms transfer from real life to virtual worlds than on

trying to identify phenomena specific to these environments. There has been a paucity of

field research using, for example, distributed teams and of studies conducted in

organizational settings. Instead, the study of micro-social behavior and experimental research

using student samples have gained more attention in VE studies. Furthermore, the use of

larger theoretical frameworks has not been as dominant in VE studies as in traditional social

science research.

Our review shows several openings for further research. The study of virtual

environment-specific phenomena, such as certain group behaviors that can be triggered only

in virtual environments, would be a fascinating possibility for social scientists. Field research

conducted in organizational settings with qualitative, longitudinal set-ups would also raise

new possibilities for virtual environment scholars. The framing of research questions

focusing on macro-level phenomena such as leadership and intergroup relations could

broaden the field even further. Finally, utilization of covering (meta)theories of social

sciences, such as the Social Identity Approach, would contribute to the development of

virtual environment research in the future.


24

An interesting question that this review raises is whether there are any social

phenomena about which VE research could not inform us or that would be extremely difficult

to study in this new environment. Even though this question cannot, naturally, be extensively

answered by speculation, we found a couple of illustrative examples. First, and perhaps most

clearly, it might be very difficult to study the natural development of social interaction

between a parent and his/her (genetic) offspring (e.g., attachment; see Bowlby, 1980) when

the child is very young. One could test parents’ reactions by programming babys’ movements

and voices to a VE, but that hardly informs us about the natural social process, which is

highly emotional and based on physical touch. If this example seems to be slightly off the

scope of our review, our second example of a hard-to-study-in-VE phenomenon is the long-

term development of social coping mechanisms for work-related stress or work-life balance

(see e.g., Theorell & Karasek, 1996). Work, as we have shown, is a much understudied

context in VE research, but that could change over time. What is problematic here is the long-

term process, which involves many actors and complicated processes at workplace and at

home. However, this speculation was a hard exercise, which suggests that VEs can indeed

inform us of a very broad spectrum of social phenomena relevant to human behavior.

Considering the fact that research on social and group phenomena in VEs is a

relatively new endeavor, it is natural that the scope of studies in the field is not as extensive

as in more mature research areas. Consequently, our review has presented four trends in

current VE research. More importantly, we have tried to open new perspectives and research

agendas that go beyond these trends. It is our hope that these openings encourage and inspire

VE researchers in a constructive manner. In the future, as the technology develops further,

VE platforms can be expected to provide even more opportunities for social scientists to

study communication and collaboration dynamics in several different contexts.


25

References

Antonijevic, S. (2008). From text to gesture online: A microethnographic analysis of

nonverbal communication in the Second Life virtual environment. Information,

Communication & Society, 11(2), 221–238.

Ashforth, B. E., & Mael, F. A. (1989). Social identity theory and the organization. Academy

of Management Review, 14, 20-39.

Axelsson, A.-S., Abelin, A. & Scroeder, R. (2003). Anyone speak Spanish? New Media &

Society, 5(4), 475–498.

Axelsson, A.-S., Abelin, Å, Heldal, I., Schroeder, R. & Wideström, J. (2001). Cubes in the

cube: A comparison of a puzzle-solving task in a virtual and a real environment.

CyberPsychology & Behavior, 4(2), 279–286.

Axtell, C. M., Fleck, S. J. & Turner, N. (2004). Virtual teams: Collaborating across distance.

In C. L. Cooper & I. T. Robertson (Eds.) International Review of Industrial and

Organizational Psychology, vol. 19. Chichester: John Wiley, 205-48.

Bente, G., Rüggenberg, S., Krämer, N. C. & Eschenburg, F. (2008). Avatar-mediated

networking: Increasing social presence and interpersonal trust in net-based

collaborations. Human Communication Research, 34(2), 287-318.

Bailenson, J. N., Beall, A. C. & Blascovich, J. (2002). Gaze and task performance in shared

virtual environments. The Journal of Visualization and Computer Animation, 13(5),

313-320.


26

Bailenson, J. N., Beall, A. C., Loomis, J., Blascovich, J. & Turk, M. (2005). Transformed

social interaction, augmented gaze, and social influence in immersive virtual

environments. Human Communication Research, 31(4), 511–537.

Bailenson, J. N., Blascovich, J., Beall, A. C. & Loomis, J. M. (2003). Interpersonal distance

in immersive virtual environments. Personality and Social Psychology Bulletin,

29(7), 819–833.

Bailenson, J. N., Blascovich, J., Guadagno, R. E. (2008). Self-representations in immersive

virtual environments. Journal of Applied Social Psychology, 38(11), 2673–2690.

Bailenson, J. N. & Yee, N. (2005). Digital chameleons: Automatic assimilation of nonverbal

gestures in immersive virtual environments. Psychological Science, 16(10), 814-819.

Bailenson, J. N. & Yee, N. (2006). A longitudinal study of task performance, head

movements, subjective report, simulator sickness, and transformed social interaction

in collaborative virtual environments. Presence, 15(6), 699–716.

Bailenson, J. N., Yee, N., Merget, D., Schroeder, R. (2006). The Effect of Behavioral

Realism and Form Realism of Real-Time Avatar Faces on Verbal Disclosure,

Nonverbal Disclosure, Emotion Recognition, and Copresence in Dyadic Interaction.

Presence, 15 (4), 359–372.

Blasovich, J., Loomis, J. M., Beall, A. C., Swinth, K. R., Hoyt, C. L. & Bailenson, J. N.

(2002). Immersive virtual environment technology as a methodological tool for social

psychology. Psychological Inquiry, 13(2), 103-124.

Bond, M. H. (1986). Lifting one of the last bamboo curtains: review of the psychology of

Chinese people. Hong Kong: Oxford University Press.


27

Bowlby, J. (1980). Attachment and loss: Loss, sadness, and depression. Basic Books: New

York.

Chidambaram, L. (1996). Relational development in computer supported groups [1]. MIS

Quarterly, 20(2), 143–165.

Daft, R. L. & Lengel, R. H. (1984). Information richness: A new approach to managerial

information processing and organization design. In L. L. Cummings & B. M. Staw

(Eds.) Research in organizational behaviour, Vol. 6. Greenwich: JAI Press, 191–234.

Denzin, N. K. & Lincoln, Y. S. (1994). Introduction: Entering the field of qualitative

research. In N. K. Denzin & Y. S. Lincoln (Eds.) Handbook of Qualitative Research.

London: Sage, 1–17.

Diehl, W. C. & Prins, E. (2008). Unintended outcomes in Second Life: Intercultural literacy

and cultural identity in a virtual world. Language & Intercultural Communication,

8(2), 101–118.

Dotsch, R. & Wigboldus, D. H. J. (2008). Virtual prejudice. Journal of Experimental Social

Psychology, 44(4), 1194–1198.

Engestrom, Y. (2001). Expansive learning at work: Toward an activity theory

reconceptualization. Journal of Education at Work, 14(1), 133-156.

Feldon, D. F. & Kafai, Y. B. (2008). Mixed methods for mixed reality: Understanding users’

avatar activities in virtual worlds. Educational Technology Research and

Development, 56, 575–593.

Fiedler, M. (2009). Cooperation in Virtual Worlds. Schmalenbach Business Review, April

2009, 173-194.


28

Fiedler, M. & Haruvy, E. (2009). The lab versus the virtual lab and virtual field—An

experimental investigation of trust games with communication. Journal of Economic

Behavior & Organization, 72, 716–724.

Fleiss, J. L. (1971). Measuring nominal scale agreement among many raters. Psychological

Bulletin, 76, 378–382

Fox, J., Arena, D. & Bailenson, J. N. (2009). Virtual reality: A survival guide for the social

scientist. Journal of Media Psychology, 21(3), 95–113.

Fox, J. & Bailenson, J. N. (2009). Virtual virgins and vamps: The effects of exposure to

female characters’ sexualized appearance and gaze in an immersive virtual

environment. Sex Roles, 61(3–4), 147–157.

Friedman, D., Karniel, Y. & Dinur, A. L. (2009). Comparing group discussion in virtual and

physical environments. Presence, 18(4), 286–293.

Gillath, O., McCall, C., Shaver, P. & Blascovich, J. (2008). What can virtual reality teach us

about prosocial tendencies in real and virtual environments? Media Psychology,

11(2), 259–282.

Gong , L. & Nass, C. (2007). When a talking-face computer agent is half-human and half-

humanoid: Human identity and consistency preference. Human Communication

Research, 33(2), 163-193.

Groom, C. J., Sherman, J. W. & Conrey, F. R. (2002). What Immersive Virtual Environment

Technology Can Offer to Social Cognition. Psychological Inquiry, 13, 125–145.

Guadagno, R. E., Blascovich, J., Bailenson, J. N. & McCall, C. (2007). Virtual humans and

persuasion: The effects of agency and behavioral realism. Media Psychology, 10(1),

1–22.


29

Hakonen, M. & Lipponen, J. (2007). Antecedents and consequences of identification with

virtual teams: Structural characteristics and justice concerns. The Journal of E-

working, 1, 137-153.

Hakonen, M. & Koivisto, S. (2008). Antecedents of identity in virtual teams. Nordic

Organization Studies, 2/08, 54-73.

Haslam, S. A. (2004). Psychology in organizations: The social identity approach. (2nd

edition) London: Sage.

Heyward, M. (2002). From international to intercultural: Redefining the international school

for a globalized world. Journal of Research in International Education, 1(1), 9-32.

Hertel, G., Geister, S. & Konradt, U. (2005). Managing virtual teams: A review of current

empirical research. Human Resource Management Review, 15(1), 69–95.

Hogg, M. A. (1993). Group cohesiveness: A critical review and some new directions.

European Review of Social Psychology, 4, 85-111.

Hogg, M. A., & Abrams, D. (1988). Social identifications: A social psychology of intergroup

relations and group processes. London: Routledge.

Hogg, M. A., & Terry, D. J. (2001). Social identity theory and organizational processes. In

M. A. Hogg & D. J. Terry (Eds.), Social identity processes in organizational contexts

(pp. 1-12). Philadelphia, PA: Psychology press.

Hoyt, C. A., Aguiular, L., Kaiser, C. R., Blascovich, J., & Lee, K. (2007). The self-protective

and undermining effects of attributional ambiguity. Journal of Experimental Social

Psychology, 43, 884–893.

Hoyt, C. L. & Blascovich, J. (2003). Transformational and Transactional Leadership in Virtul

and Physical Environments. Small Group Research, 34(6), 678–715.


30

Hoyt, C. L. & Blascovich, J. (2007). Leadership efficacy and women leaders' responses to

Stereotype Activation. Group Processes & Intergroup Relations, 10(4), 595–616.

Hoyt, C. L., Blascovich, J. & Swinth, K. R. (2003). Social inhibition in immersive virtual

environments. Presence, 12(2), 183–195.

Jackson, L. A., Zhao, Y., Kiu, W., Kolenic, A., Fitzgerald, H. E., Harold, R. & Von Eye, A.

(2008). CyberPsychology & Behavior, 11(3), 279–286.

Jung, K. H. & Lee, J.-H. (2009). Implicit and explicit attitude dissociation in spontaneous

deceptive behavior. Acta Psychologica, 132, 62–67.

Jarvenpaa, S. L. & Leidner, D. E. (1998). Communication and trust in global virtual teams.

Journal of Computer-Mediated Communication, 3(4). doi: 10.1111/j.1083-

6101.1998.tb00080.x

Kohlberg, L. (1984). The psychology of moral development. San Francisco: Harper & Row.

Lantz, A. (2001). Meetings in a distributed group of experts: comparing face-to-face, chat

and collaborative virtual environments. Behaviour & Information Technology, 20(2),

111–117.

Lea, M. & Spears, L. (1991). Computer-mediated communication, de-individuation and

group decision making. International Journal of Man-Machine Studies, 34(2), 283–

301.

Linebarger, J. M., Janneck, C. D. & Kessler, G. D. (2005). Leaving the world behind:

Supporting group collaboration patterns in a shared virtual environment for product

design. Presence, 14(6), 697–719.

Martins, L. L., Gilson, L. L. & Maynard, M. T. (2004). Virtual Teams: What do we know and

where do we go from here? Journal of Management, 30(6), 805–835.


31

Mortensen, M. & Hinds, P. (2001). Conflict and shared identity in geographically distributed

teams. The International Journal of Conflict Management, 12(3), 212–238.

Nakanishi, H. (2004). Free Walk: A social interaction platform for group behaviour in a

virtual space. International Journal of Human-Computer Studies, 60(4), 421–454.

Nowak, K. L. (2004). The Influence of Anthropomorphism and Agency on Social Judgment

in Virtual Environments. Journal of Computer-Mediated Communication, 9. doi:

10.1111/j.1083-6101.2004.tb00284.

Nowak, K. L. & Biocca, F. (2003). The effect of the agency and anthropomorphism on users’

sense of telepresence, copresence, and social presence in virtual environments.

Presence, 12(5), 481–494.

Nowak, K. L., and Rauh, C. (2005). The influence of the avatar on online perceptions

of anthropomorphism, androgyny, credibility, homophily, and attraction. Journal of

Computer-Mediated Communication, 11(1). doi: 10.1111/j.1083-

6101.2006.tb00308.x

Nowak, K. L & Rauh, C. (2008). Choose your 'buddy icon' carefully: The influence of avatar

androgyny, anthropomorphism and credibility in online interactions. Computers in

Human Behavior, 24(4), 1473-1493.

Peña, J., Hancock, J. T., & Merola, N. A. (2009). The priming effects of avatars in virtual

settings. Communication Research, 36, 838-856.

Postmes, T., Spears, R., & Lea, M. (1998). Breaching or building social boundaries?

Communication Research, 25, 689–699.


32

Rice, R. E. & Gattiker, U. E. (2001). New media and organizational structuring. In F. M.

Jablin & L. L. Putnam (Eds.) The new handbook of organizational communication.

London: Sage, 544–581.

Riketta, M. (2004). Organizational identification: A meta-analysis. Journal of Vocational

Behavior, 66(2), 358–384.

Sallnäs, E.-L. (2005). Effects of Communication Mode on Social Presence, Virtual Presence,

and Performance in Collaborative Virtual Environments. Presence, 14(4), 434–449.

Scheck, S., Allmendinger, K. & Hamann, K. (2008). The effects of media richness on

multilateral negotiations in a collaborative virtual environment. Journal of Media

Psychology, 20(2), 57–66.

Sivunen, A. (2006). Strengthening identification with the team in virtual teams: The leaders’

perspective. Group Decision and Negotiation, 15, 345–366.

Slater, M., Sadagic, A., Usoh, M. & Schroeder, R. (2000). Small-Group Behavior in a Virtual

and Real Environment: A Comparative Study. Presence, 9(1), 37–51.

Tajfel, H. & Turner, J. (1979). An integrative theory of intergroup conflict. In W. G. Austin

& S. Worchel (Eds.) The social psychology of intergroup relations. Monterey:

Brooks/Cole, 33–47.

Talamo, A. & Ligorio, B. (2001). Strategic identities in cyberspace. CyberPsychology &

Behavior, 4(1), 109–122.

Theorell, T., & Karasek, R.A.. (1996). Current issues relating to psychosocial job strain and

cardiovascular disease. Journal of Occupational Health Psychology, 1, 9-26.


33

Tidwell, L. C., & Walther, J. B. (2002). Computer-mediated communication effects on

disclosure, impressions, and interpersonal evaluations: Getting to know one another a

bit at a time. Human Communication Research, 28(3), 317-348.

Turner, J. C. (1999). Some current issues in research on social identity and self-categorization

theories. In N. Ellemers, R. Spears, & B. Doosje (Eds.), Social identity: Context,

commitment, content (pp. 6-34). Oxford, UK: Blackwell.

Turner, J. C., Hogg, M. A., Oakes, P. J., Reicher, S. D., & Wetherell, M. S. (1987).

Rediscovering the social group: A self-categorization theory. Oxford, UK: Blackwell.

Tyler, T. R., & Blader, S. L. (2001). Identity and cooperative behavior in groups. Group

Processes and Intergroup Relations, 4, 207-226.

Van Knippenberg, D., & van Schie, E. C. M. (2000). Foci and correlates of organizational

identification. Journal of Occupational and Organizational Psychology, 73,137-147.

Vasalou, A., Joinson, A. N. (2009). Me, myself and I: The role of interactional context on

self-presentation through avatars. Computers in Human Behavior, 25, 510–520.

Vasalou, A., Joinson, A., Bänziger, T., Goldie, P. & Pitt, J. (2008). Avatars in social media:

Balancing accuracy, playfulness and embodied messages. International Journal of

Human-Computer Studies, 66 (11), 801–811.

Vinayagamoorthy, V., Steed, A. & Slater, M. (2008). The impact of a character posture

model on the communication of affect in an immersive virtual environment. IEEE

Transactions on Visualization and Computer Graphics, 14(5), 965–981.

Walther, J. B. (1996). Computer-mediated communication: Impersonal, interpersonal, and

hyperpersonal interaction. Communication Research, 23(1), 3–43.


34

Walther, J. B. (2007). Selective self-presentation in computer-mediated communication:

Hyperpersonal dimensions of technology, language, and cognition. Computers in

Human Behvior, 23(5), 2538–2557.

Walther, J. B. & Bunz, U. (2005). The rules of virtual groups: Trust, liking, and performance

in computer-mediated communication. Journal of Communication, 55(4), 828–846.

Yee, N. (2006) The demographics, motivations, and derived experiences of users of

massively multi-user online graphical environments. Presence, 15(3), 309–329.

Yee, N. & Bailenson, J. (2007). The Proteus Effect: The effect of transformed self-

representation on behavior. Human Communication Research, 33(3), 271-290.

Yee, N. & Bailenson, J. N. (2009). The difference between being and seeing: The relative

contribution of self-perception and priming to behavioral changes via digital self-

representation. Media Psychology, 12, 195–209.

Yee, N., Bailenson, J. N. & Ducheneaut, N. (2009). The proteus effect: Implications of

transformed digital self-representation on online and offline behavior.

Communication Research, 36, 285–312.

Yee, N., Bailenson, J., Urbanek, M., Chang, F. & Merget, D. (2007). The unbearable likeness

of being digital: The persistence of nonverbal social norms in online virtual

environments. CyberPsychology & Behavior, 10(1) 115-121.

Published in Small Group Research (2011), 42(4): 405-457 Appendix

Studies included in the review Authors Year Aim / Research

question Methods Subjects / Sample Main findings Journal

Antonijevic 2008 Nonverbal

communication in

a virtual

environment.

Six month-long, ethnographic field

study in the virtual world Second

Life. Data were gathered from 108

publicly available areas by

observing and recording the

naturally-occurring interaction in

both dyadic and group settings.

Approximately 1000

users of Second Life.

Nonverbal communication in

Second Life can be divided into

four categories of nonverbal cues:

1) user-defined, 2) predefined, 3)

blended, and 4) missing cues.

User-defined cues show

adaptability to the setting, whereas

predefined (system generated) and

blended cues often represent

stereotypical, culture- and gender-

biased nonverbal acts.

Information,

Communication &

Society

Axelsson, Abelin,

Heldal, Schroeder,

Wideström

2001 Comparison of

pairs in two

situations: some try

to carry out a task

in a virtual

In the VE setting, one participant

was collaborating on an immersive

system and the other using a

desktop system. Their task was to

collaborate on a Rubik’s cube-type

44 participants in 22

dyads.

In the virtual setting, only 6 pairs

out of 22 completed the task within

the time limit, whereas in the real

setting, all pairs could complete the

task in the given time. Participants

CyberPsychology

& Behavior


environment and

the others in a real

environment.

puzzle. In the real setting,

participants collaborated face-to-

face with blocks of similar size as

in the VE. A time limit of 20

minutes was set for each

experimental run, first in VE and

then in the real world. After the

experiments, participants answered

questions on presence, co-

presence, their contribution to the

task, and collaboration.

reported a significantly stronger

feeling of presence in the

immersive VE than the desktop

system. Both participants

perceived the person in the

immersive VE to be more active

than the participant at the desktop.

In the real-world setting, there was

no difference between the

participants in perceived

contribution to the task.

Participants in the real-world

setting reported stronger feelings

of collaboration than participants

in VE setting.

Axelsson, Abelin

& Schroeder

2003 How people using

different languages

interact in a virtual

Participant observation of the

virtual world Active Worlds,

consisting of 50 hours of field

Users of a virtual

world Active Worlds.

The phases of a language

encounter situation are

introduction, responses and

New Media &

Society


environment. observations and notes as well as

chat logs. The observations were

made in the central and most

populated places of the

environment, in language-specific

areas, and in themed areas.

consequences. The reasons for the

introduction phase are 1) finding

out whether there are any fellow

speakers present, 2) having a

language game, or 3) disturbing the

ongoing conversation. The

responses to a language encounter

situation are either acceptance,

rejection, neutral or mixed. The

direct consequences of a language

encounter are that the language

introduced either survives or

disappears. Indirect consequences

are shifts in the atmosphere or in

topics of conversation.

Bailenson, Beall,

Blascovich

2002 Gaze and task

performance in a

virtual

Experimental study in which three

participants sat in physically

remote rooms with HMDs, entered

27 undergraduate

students.

Participants felt involved in the

game, truly interacting with the

other partners, and enjoyed the

The Journal of

Visualization and

Computer


environment. a common virtual room, and

played games of 20 questions.

Avatar behavior was measured

under three different conditions: 1)

low behavior condition, where

there was no avatar but participants

collaborated through voice, 2)

medium behavior condition, where

each participant could see others’

avatars and hear their voices, and

3) high behavior condition, where

participants’ head movements were

rendered. Each participant played

three separate games of 20

questions, one under each of the

three conditions. The three games

played under each condition

resulted in nine games in total.

experience more under the high

behavior condition than the other

conditions. Most head movements

(in a horizontal plane) occurred

under the high behavior condition.

The highest percentage of speaking

occurred under the low behavior

condition.

Animation


After each game, participants took

off their HMDs and responded to a

questionnaire.

Bailenson, Beall,

Loomis,

Blascovich & Turk

2005 Social influence of

gaze.

Experimental study in which a

presenter read a persuasive passage

to two participants using HMDs

with three gaze conditions: natural,

augmented or reduced. In the

augmented condition, the

presenter’s avatar directed his/her

gaze toward each of two

participants 100% of the time,

whereas in the reduced condition,

the presenter’s avatar gazed down

100% of the time. Both gaze

conditions were implemented by

transforming the veridical head

movements of the presenter.

72 psychology

students interacting

with one another in

groups of three: two

participants and a

presenter

(confederate).

Women were overall less

persuaded than men as a group.

Agreement in the augmented gaze

condition was significantly higher

than in the other gaze conditions

with female participants but not

with male participants. There was

no significant main effect of gaze

condition on how participants

remembered the persuasive

passage. Participants reported

lower scores of social presence in

the augmented condition than in

the other conditions. Overall,

participants did not detect that the

Human

Communication

Research


veridicality of the gaze had been

breached.

Bailenson,

Blascovich, Beall

& Loomis

2003 Proxemic behavior

study.

Two experiments using IVE with

HMD: 1) walking participant’s

avatar (distance) approached

virtual human and 2) virtual human

approached standing participant’s

avatar (avoidance). Gaze and

position were tracked.

Altogether 160

undergraduate

students.

Generally, similar distance

behavior was detected as in real

life. In study 1, greater distance

was kept from avatars than agents,

and this was enlarged if avatars

looked at the participants. In study

2, some avoidance of approaching

avatars was found, but less than

expected.

Personality and

Social Psychology

Bulletin

Bailenson,

Blascovich &

Guadagno

2008 How does

interaction differ if

individuals interact

with virtual

representations of

themselves or of

others in an IVE?

Experimental study with 2

(participant gender) X 3 (agent

identity: high-similarity self-

representation vs. low-similarity

self-representation vs. other

representation) factorial design.

Participants in similarity self-

64 university

students.

Participants in the other

representation condition remained

twice as distant from the agent as

participants in high-similarity self-

representation condition.

Additionally, the difference in

minimum distance between the

Journal of Applied

Social Psychology


representation condition were told

to walk around an animated agent

who represented him- or herself,

and in other representation

condition, they were told to walk

around of a virtual representation

of someone else. In high-similarity

self-representation condition, the

agents bore a photographic

resemblance to the participant,

whereas in low-similarity self-

representation condition, they did

not. In other representation

condition, the agent ostensibly

represented an unfamiliar stranger.

Participants’ location was recorded

during the experiment and they

were administered questions to

other representation condition and

the low-self-representation

condition was significant, but there

were no differences in distance

between the two similarity self-

representation conditions. The

main effect of gender to distance

was not significant. Participants

were more willing to commit

embarrassing acts in front of the

agent in both similarity self-

representation conditions than in

other representation condition but

there was no difference between

the high- and low-similarity self-

representation conditions and no

main effect of gender. Both

similarity self-representation


gauge liking for the virtual

representation and willingness to

commit embarrassing acts in the

presence of it.

conditions were also preferred to

the other representation condition

in terms of attraction and liking,

but there were no differences

between the two similarity self-

representation conditions.

Bailenson & Yee 2005 Social influence of

mimicry

(chameleon effect).

Two randomized conditions with

IVE and HMD: 1) agent mimicked

the participant; 2) agent acted as

previous participant. In both

conditions, the agent read a

persuasive message.

69 undergraduate

students.

Mimicking agents were perceived

to be more persuasive and rated

more positively than non-

mimickers, even though the

participants were aware that the

agent was a computer-directed

non-human. Moreover, few

participants noticed the mimicking.

Psychological

Science

Bailenson & Yee 2006 Effects of

similarity on task

performance

(different task

Longitudinal laboratory

experiment using collaborative VE

and HMD. Manipulation of facial

and behavioral similarity

Nine undergraduate

university students in

triads.

Performance was best in the facial

similarity condition, followed by

the mimicry and realistic

conditions. The more familiar the

Presence

https://www.researchgate.net/publication/275004526_Transformed_Social_Interaction_Augmented_Gaze_and_Social_Influence_in_Immersive_Virtual_Environments?el=1_x_8&enrichId=rgreq-d697aeed-7538-44ab-aa74-235fef931712&enrichSource=Y292ZXJQYWdlOzI1ODE4NzU4MjtBUzoyMDY0NjQ3MjQ1NDE0NDNAMTQyNjIzNjQzNDI3NQ==


types) and other

DVs over time.

(mimicry) of avatars to each other.

In a repeated problem-solving task,

presence, co-presence, entativity (

cohesion) and simulator sickness

were measured.

participants were with the system,

the less they looked at each other

and the less they reported

simulator sickness. Entativity

(cohesion) was high and increased

over time as in FTF interaction.

Bailenson, Yee,

Merget &

Schroeder

2006 How different

communication

modalities affect

verbal and non-

verbal self-

disclosure,

co-presence,

transmission and

identification of

emotions.

Experimental study with three

different conditions:

videoconference,

voice only and

“emotibox” (system that renders

the dimensions of facial

expressions abstractly in terms of

color, shape, and orientation on a

rectangular polygon).

30 university students

in dyads.

Verbal and non-verbal self-

disclosure was lowest in the

videoconference condition.

Nonverbal disclosure was highest

in the voice-only condition. Self-

reported co-presence and success

of transmission and identification

of emotions were lowest in the

emotibox condition. However,

people seemed to be able to

identify certain emotions even if

expressed in abstract fashion

Presence


(emotibox).

Bente,

Rüggenberg,

Krämer &

Eschenburg

2008 Influence of

different

communication

modalities on

interpersonal trust,

social presence,

communication

quality, nonverbal

behavior and visual

attention.

Laboratory experiment with 71

dyads unknown to each other

beforehand in two separate rooms

using PC interface. Five

conditions: 1) text, 2) audio, 3)

audio-video, 4) low fidelity avatar,

and 5) high fidelity avatar. Each

dyad conducted a job candidate

selection task. Tracking sensors

were used to record body

movements and questionnaires to

measure other DVs.

142 university

students in dyads.

The results indicated that text

mode was inferior to all other

communication modes in

predicting co-presence, emotional

closeness and affect-based trust.

Moreover, video and avatar modes

were rather similar with respect to

nonverbal activity and visual

attention to the other via the

computer’s communication

window.

Human

Communication

Research

Diehl & Prins 2008 Exploration of the

construction of

(individual)

cultural identity

and intercultural

The data consisted of observations

and interviews of Second Life

characters by the first author’s own

avatar via chat. Moreover, 29

habitants completed a survey on

Characters of Second

Life representing a

wide range of

demographic and

social categories.

Drawing on Activity Theory and

Heyward’s model of cultural

literacy, it was found that SL

interaction increased intercultural

literacy (e.g., through the use of

Language and

Intercultural

Communication


literacy in Second

Life.

their activities in the Second Life. many languages, cross-cultural

friendships and understanding).

Moreover, avatar appearance was

used in cultural identity

construction although only 76% of

the avatars were said to resemble

the real life appearance of those

they represented. Identity shifts

and literacy were seen more as

opportunities provided by the

virtual world than as causes of it.

Dotsch &

Wigboldus

2008 How implicit

prejudice is related

to automatic

proxemic behavior

and to skin

conductance level

(SCL).

Laboratory experiment utilizing

IVE. White Dutch participant

walked individually among 12

avatars (white-Dutch and

Moroccan-Dutch looking). Implicit

prejudice was measured by implicit

association test (IAT) and explicit

33 university

students.

Participants had no explicit

(conscious) prejudices toward

Moroccan-Dutch persons, but IAT

revealed more negative

associations. They kept a greater

personal distance from Moroccan-

looking than white avatars and

Journal of

Experimental

Social Psychology


prejudice with a questionnaire. their SCLs were higher while near

Moroccan-looking avatars. In

addition, SCL fully mediated the

relationship between implicit

discrimination and distance kept.

Feldon & Kafai 2008 How users engage

with their avatars

in a virtual world

and in real life.

A field study of the virtual world

Whyville. Online tracking data and

surveys were collected from all

participants. In addition, a subset

of the users were observed offline,

and their behavior was recorded

while they were using the virtual

world. These participants were also

interviewed offline.

595 children (8-18

years old) using

virtual world

Whyville. A subset of

the users (88

participants) were

recruited from an

elementary school

science class and

after-school clubs.

Avatar-related behavior (such as

customizing the appearance of the

avatar) was the most commonly

used behavior in the virtual world.

In second place was socializing,

such as sending messages and

chatting. Two primary types of

interactions centered on avatars

who were critiquing or admiring

others’ avatars and donating or

selling face parts to others’ avatars.

Regardless of differences in

participants’ profiles, the

Educational

Technology

Research &

Development


proportion of online activity

related to avatars did not vary.

Avatar customization was an

important part of this specific

virtual world.

Fiedler 2009 A study on the

effects of

strategically

irrelevant

communication,

media richness,

experience with

the

communication

medium, social

distance,

and collective

orientation on

An experiment with three

randomly assigned dyads was

conducted. The communication

media were a) communication via

Second Life voice, b)

communication via Skype chat,

and c) no communication. In the

trust-game experiment, the

amounts of money sent (proposer)

and returned (responder) were the

dependent variables and proxies

for cooperation.

304 university

students in dyads.

It was found that the collective

orientation predicted cooperation

both in sending and returning

money. The strategically irrelevant

communication had a positive

effect on cooperation only for the

sender since s/he was the only one

expecting reciprocation.

Communication media had no

clear effects on cooperation. Other

empirical results remained

somewhat mixed.

Schmalenbach

Business Review


cooperation.

Fiedler & Haruvy 2009 A study of

differences

between

laboratory and

field experiment

in trust-‐games.

The effects of

communication

and allocation

motives were also

studied.

The trust-game was the same as

that of Fiedler (2009) above,

except that in all conditions the

game was played in SL. Three

conditions were created: 1) 40

university students played in lab

without prior communication, 2)

136 university students played the

game in lab after 15 minutes of

game-irrelevant discussion, and 3)

216 volunteer SL residents (field

experiment) played the game after

similar discussion as in condition

2). Multiple independent variables

were measured.

176 university

students and 216 SL

residents in dyads.

The main finding was that

interaction through a virtual world

interface (cond. 2 & 3) increases

the amount sent relative to

laboratory results (cond. 1), but

that subjects recruited in the virtual

world (cond.3) give and return less

than the laboratory control group

with the same virtual world

interface (cond. 2). The analyses of

such allocation motives as trust,

cooperative orientation, obligation

and utilitarianism produced mixed

results. Hence, the authors

speculate that different aspects of

laboratory and field conditions

might have triggered the found

Journal of

Economic

Behavior &

Organization


differences.

Fox & Bailenson 2009 Does the exposure

to a sexualized

dress and amount

of eye gaze of a

female agent in an

immersive virtual

environment affect

sexist and rape-

supportive

attitudes?

A 2x2x2 (Dress x Gaze x Gender)

between-subjects design was

employed. Participants were

placed in an IVE with HMDs. The

female agents were selected

based on a pretest. 30

participants from a separate pool

from the main experiment

viewed 16 female agents and

rated them on several qualities.

Two agents that demonstrated

extreme scores as sexy and

suggestively dressed and two

that ranked as not sexy and

conservatively dressed were

selected for the experiment. In

the high gaze condition, an

83 participants

recruited from a

university.

Participants in the sexualized

condition indicated that the

agent was sexier and more

suggestively dressed than

participants in the nonsexualized

conditions. Moreover,

participants in the high gaze

condition indicated that the

agent saw and acknowledged

them more than participants in

the low gaze condition. There

was no main effect for dress on

rape myth acceptance,

benevolent sexism or hostile

sexism. Those in the suggestively

dressed high

gaze condition (“vamp”) and the

Sex Roles


algorithm determining the

agent’s head movements

assessed where the participant

was in the room and

appropriately oriented the

agent’s head toward the

participant, whereas in the low

gaze condition, a different

algorithm with random eye gaze

was used. The participants were

told to walk toward the agent in

the IVE and examine it carefully

to feel prepared to answer

questions about the agent’s

appearance and behavior later.

After the experiment,

participants answered two

questionnaires regarding the

conservatively dressed low gaze

condition (“virgin”) exhibited

significantly higher rape myth

acceptance than those in the

suggestively dressed low gaze

and conservatively dressed high

gaze conditions.


avatar appearance, gaze, sexism

against women and rape myth

acceptance.

Friedman, Garmiel

& Dinur

2009 How group

discussions in FTF

situations and in

VEs differ from

one another in

dynamics and

content.

Participants of four groups carried

out a conversation on same topic,

global warming. Two of the groups

interacted FTF and two of the

groups communicated via chat in

the virtual world Second Life. The

participants were instructed that

they had one hour to reach

consensus whether they were ready

or not to decrease their car usage to

address climate change. Before the

experiment, each subject filled in a

pre-experiment questionnaire and a

short questionnaire assessing their

view of global warming. In the

Four groups of 7–12

student participants,

altogether 36

subjects.

In all groups, there was an increase

in the concern for global warming

after the discussion. The number of

sentences was almost the same in

both conditions, but the sentences

in VE chat were much shorter and

the number of words per minute

smaller than in FTF condition. The

number of on-topic sentences per

speaker, normalized by group size,

was bigger in the FTF than in the

VE chat condition. The most

dominant speaker in the FTF

condition was also dominant in the

number of on-topic sentences, but

Presence


FTF condition, the participants

were located in a meeting room,

and in the VE condition they were

sitting in the same room using PCs

and collaborating via text chat in

Second Life. Both types of

sessions were recorded and

transcribed. After the experiment,

all participants filled in the same

questionnaire assessing their

opinions on global warming.

this was not the case in VE chat

condition. FTF groups had more

topics discussed in-depth, whereas

VE chat groups had more topics

discussed superficially.

Gillath, McCall,

Shaver &

Blascovich

2008 How participants

react in VE to

people in need.

Two experiments using IVE and

HMD when participants navigated

in VE. Study 1: emotional

reactions towards a blind man in an

accident. Study 2: proxemic

behavior towards a beggar vs. a

businessman on the street.

Altogether 107

university students.

Dispositionally compassionate

people reacted in a more

sympathetic way to the person in

need than the others (36%

emphatic) – similar to results for

real-life studies. Moreover,

dispositionally compassionate

Media Psychology


participants were more inclined to

look at and stay near the beggar

than the others.

Gong & Nass 2007 How people react

to inconsistencies

between human vs.

humanoid voice

and face and

whether humanoids

(i.e., computer-

generated artificial

agents ) are

considered to be a

different category

from natural

humans.

In two laboratory studies using

“talking faces” shown on a

computer screen, the four different

conditions (human face and voice,

humanoid face and voice =

consistent, human face and

humanoid voice, humanoid face

and human voice = inconsistent)

were tested for males and females.

In study 1, the task was self-

disclosure and in study 2

comprehension of a text read

aloud. The level of trust, negative

attitudes toward the human or

humanoid, used processing time

Altogether 160

undergraduate


In line with (social cognition)

literature, study 1 showed that

inconsistency of voice and face

(the stimulus seen and heard from

computer) led to lower levels of

trust and longer processing time of

responses. Moreover, female

participants showed more negative

attitudes toward, and were more

sensitive to, inconsistency than

males. Study 2 replicated the

processing time and females’

negative attitude effects regarding

inconsistency. Due to the

comprehension focus of study 2,

Human

Communication

Research


and in study 2 comprehension of

the message were the DVs

measured.

voice clarity overshadowed the

inconsistency effect, that is, the

human voice was preferred and it

was associated with more positive

judgments of the stimuli than the

humanoid voice, regardless of the

face. Overall, the studies indicated

that humans were perceived to be

qualitatively a different category

from humanoids.

Guadagno,

Blascovich,

Bailenson &

McCall

2007 Attitude change

after listening to a

persuasive

communication by

an in-group

member.

Laboratory experiment using IVE

and HMD. Study 1 investigated

gender and behavioral realism

effects on attitude change after a

persuasive communication. Study

2 extended study 1 by

manipulating the participants’

perceptions of agency.

Altogether 239

undergraduate


There was a greater attitude change

for same-gender virtual humans

(in-group) than for those of

different gender. Moreover,

behaviorally realistic in-group

(same gender) virtual humans were

perceived to be more influential,

but this applied only to male

Media Psychology


participants.

Hindmarsh, Heath

& Fraser

2006 How CVEs support

object-focused

collaboration.

Qualitative, ethnomethodological

study including a series of short

trials with novice users in a CVE.

Users’ task was to collaboratively

organize the layout of furniture in a

virtual room. The interaction of the

users was analyzed by

conversation analysis methods.

Six trials of two

participants and two

trials of three

participants (N=18).

Most participants

were students.

Due to the narrow field of view

available and the inflexibility of

avatars, referencing objects and

spaces became a topic in and of

itself within CVEs. The technical

limitations of CVE made it

difficult for participants to assess

what others in the environment

were able to see and how they

engaged in the activities. However,

during the course of trials,

participants became familiar with

some of the problems related to

inflexibility and the narrow view

and learned to manage them

through their talk.

The Sociological

Review

Hoyt, Aguiular, 2007 Attributional Laboratory experiments using Altogether 99 The studies confirmed the Journal of


Kaiser, Blascovich

& Lee

ambiguity effects

on well-being for

White vs. Latino

(socially

stigmatized

identity) in a

feedback-giving

situation.

IVET and HMD. In study 1,

participants performed an

employee-hiring task as leaders

with two white virtual subordinates

giving feedback. The participants’

virtually represented ethnicity was

varied (known by the participants)

and their well-being and

attributions to discrimination were

measured. Study 2 replicated study

1 but the participants were all

Latinos.

undergraduate

students, half White

and half Latino.

hypotheses, according to which 1)

a real or portrayed stigmatized

ethnicity (i.e. Latino) provoked

attribution of negative feedback to

discrimination and consequent

high-levels of well-being. The

reverse was true for positive

feedback, i.e., the stigma was

perceived to result in overtly

(normatively) positive evaluations

and post-feedback well-being was

thus rated lower. Attributions to

discrimination mediated the

relationship between stigma and

well-being.

Experimental

Social Psychology

Hoyt & Blascovich 2003 Comparisons of

groups under

transactional

Laboratory experiment with 2 X 3

between-subject factorial design

(leadership style: transformational

144 undergraduate


groups of three

Post-experimental manipulation

check confirmed clear differences

in the perception of leadership

Small Group

Research


leaders and

transformational

leaders based on

performance,

leadership

satisfaction, group

cohesiveness, trust,

and value

congruence when

operating in face-

to-face or virtual

environments.

or transactional and group setting:

face-to-face, immersive virtual

environment, or intercom [audio-

only]). Participants performed two

tasks collaboratively under a

confederate leader who had been

trained to exhibit behaviors

associated with either the

transactional or

transformational leadership style.

Participants also answered two

questionnaires.

including the

confederate group

leader.

style across leadership style

conditions, and these perceptions

did not differ as a function of

group setting. The followers of

transformational leaders produced

less quantity, greater quality, were

more satisfied, and had greater

cohesiveness than those led by

transactional leaders. Group

members were more satisfied with

their leader when interacting face-

to-face than in the two other

conditions, but there were no

differences in qualitative or

quantitative group performance or

group cohesiveness across settings.

Followers reported higher levels of

trust and value congruence when


led by a transformational as

opposed to transactional leader.

Trust mediated fully the

relationship between

transformational leadership style,

satisfaction and cohesiveness.

Hoyt & Blascovich 2007 How gender

stereotype

activation on

women leaders is

moderated by low

or high leadership

efficacy and how

stereotype

activation interacts

with perceptions of

leadership

performance, rated

Laboratory experiment using IVE

and HMD with between-subjects 2

X 2 design (leadership efficacy

[high or low] and stereotyped

activation [primed or not]).

Subjects were selected based on a

score they had got from a Self-

Efficacy for Leadership

prescreening. Before the

experiment, subjects in the

stereotype activation condition had

to examine a folder containing

53 female

undergraduate

students (study 1) and

72 female

undergraduate

students (study 2)

The findings show that the

stereotype activation serves to

increase high efficacy leaders’

perceived performance. High

efficacy leaders were also rated to

perform better in the stereotype

activation condition than in the

control condition. Additionally,

high-efficacy leaders evince

heightened identification with the

domain of leadership. Stereotype

activation has also a positive effect

Group Processes &

Intergroup

Relations


performance,

domain

identification and

well-being

responses.

stereotypical images of male

leaders and information on the

gender gap in top leadership roles,

whereas the subjects in the control

condition had to peruse a folder

containing material on the virtual

reality lab. The subjects ostensibly

participated in a leadership task

with two other participants (who

were in reality automated avatars).

They received instructions

regarding the task and were given

time to prepare for a recorded,

three-minute meeting with the

other ‘participants’. Their task was

to chair a selection committee

hiring a new employee, but the

meeting allowed only one-way

on the psychological well-being

(self-esteem and depressed affect)

of high efficacy leaders. Moreover,

stereotype activation on high- and

low-efficiency leaders’

identification with their domain as

well as their self-reported well-

being are mediated by the leaders’

perceived performance.


communication from the leader to

followers. After the meeting,

subjects completed the perceived

performance and domain

identification questionnaires. Study

2 was conducted with similar

methods and extended to well-

being responses (subjects filled out

a post-experiment questionnaire

measuring self-esteem and

depressed affect).

Hoyt, Blascovich

& Swinth

2003 Are social

facilitation and

social inhibition

exhibited in avatar,

agent or alone

conditions in

virtual

A 2 x 3 between-participants

experiment with varying task types

(novel or well-learned) and

audience types (alone, agent or

avatar). Before the experiment,

participants had a learning phase

consisting of a categorization task

39 undergraduate

students

Participants in the novel task

condition reported significantly

higher levels of task anxiety

compared to participants in the

well-learned task condition.

Participants in the avatar condition

reported significantly higher levels

Presence


environments? and a pattern recognition task in

VE with HMDs and with the help

of immediate audio feedback.

After learning the criteria,

participants were assigned to

different task conditions. In the

well-learned condition, the

participants performed the same

task they had just completed, and

in the novel task condition, they

performed another task.

Participants in the no-audience

condition performed the task alone,

participants in the agent-audience

condition were told that two

computer-controlled virtual

observers would join them in VE,

and participants in the avatar

of co-presence than participants in

the agent condition, as did the

participants in the well-learned

task condition compared to the

novel task condition. Participants

in the well-learned task condition

performed significantly better than

participants in the novel task

condition. Participants in the novel

task condition performed

significantly worse in the presence

of avatar observers compared to

when they performed in the

presence of the agent observers or

alone. Participants in the avatar

audience condition experienced

significantly greater co-presence

and performed significantly worse


condition were joined by two

assistants with HMDs observing

them in VE. Upon completion of

the experiment, participants

completed a short questionnaire.

in the novel task condition than

those in the agent condition.

Jackson, Zhao,

Qui, Kolenic,

Fitzgerald, Harold

& Von Eye

2008 Cultural

differences in

morality in real and

virtual worlds.

Cross-sectional survey

methodology.

600 US and 600

Chinese school pupils

(mean age about 12

years).

Chinese youth (especially girls)

emphasized more (conventionally)

moral off-line behaviors than US

counterparts. However, the

Chinese accepted more

questionable on-line behavior as

compared to their US counterparts.

Real world moral behavior

predicted VE moral behavior.

CyberPsychology

& Behavior

Jung & Lee 2009 Implicit and

explicit attitudes

and personality in

predicting

The experiment tested deception

by letting participants walk a

designated route in a VE. Lying

about following rules was the

60 undergraduate

students.

The results showed that

spontaneous deception was only

explained by implicit attitudes. The

explicit measures were argued not

Acta Psychologica


deceptive behavior. indicator of deception.

Manipulative personality (self-

report) and implicit and explicit

attitudes towards deception and

honesty and were measured as IVs.

to have explanatory power because

explicit favoritism of deception is

socially very undesirable.

Lantz 2001 Differences in

meetings held face-

to-face or via chat

or collaborative

virtual

environments

between

geographically

dispersed experts.

The participants held their regular

face-to-face meetings three times

through text-based chat and three

times in the virtual environment

Active Worlds. A questionnaire

was sent after these meetings.

4 distributed work

group members who

already had

experience of

continuous work

meetings face-to-face

Meetings via chat and

collaborative virtual environment

were perceived as more task-

oriented than face-to-face

meetings. Avatars were perceived

as enjoyable, and they helped turn-

taking in the meetings. The

participants perceived it as

important that the avatars were in

the same place during the meeting,

even though they could have used

the text-chat of the VE if located in

different parts of the environment.

Behaviour &

Information

Technology


Linebarger,

Janneck & Kessler

2005 How a virtual tool

that helps to

visualize and create

sub-groups affects

time, quality and

productivity of

simulated product

design task in a

CVE.

Experiment with 18 ad hoc groups

of six persons. Two in each group

used the VE through HMD and

four through PC. The task was for

groups to build a virtual roller

coaster from blocks. Three

experimental conditions were

tested using six groups for each: 1)

no sub-grouping support tool, 2)

manual use of it, and 3) automatic

version of the tool. All group

members were in the same room so

that comments could be heard.

108 undergraduate


groups of 6.

There were no differences in

completion time between the three

conditions, but quality and

productivity were superb when

group members could manually

use the sub-grouping tool. In fact,

without the tool, all groups failed

to complete the task, and in the

automatic sub-grouping condition

only one group succeeded.

Presence

Nakanishi 2004 Test of how

different features

of a VE called

FreeWalk 1)

affected different

Study 1 compared the VE as a

group discussion platform to FTF

and videoconferencing for seven

persons in three tasks (decision-

making, idea-shaping and free

University students:

1) 21 students in

groups of 7 and 2)

unknown number of

students in dyads.

In study 1, it was found that VE

initiated more turn-taking in

discussions than FTF or

videoconference. Moreover, the

length of utterances was most

International

Journal of Human-

Computer Studies


group discussion

topics and 2)

helped to overcome

awkward pauses in

the cross-cultural

discussion of pairs.

Note: This VE did

not use avatars but

placed the video

images of

participants’ faces

in 3D space.

discussion). FTF was videotaped

and used as a tracking tool for

tracking the movements and

speech of participants. In study 2,

an agent was developed for the VE

system. The agent (pictured as a

dog) detected awkward pauses in

discussions and presented

discussion-promoting questions to

both participants in a chat bubble.

Preferences and attitudes were

measured by means of a

questionnaire.

equal in VE and participants had

more non-task-related

communication than in other

conditions. Especially in the free

discussion condition, the

participants moved freely in the 3D

space. It was concluded that VEs

can support casual communication.

In study 2, the reactions to agent

intrusions of the American-

Japanese pairs’ discussion revealed

culture-specific effects. Americans

preferred safe discussion openings

by the agent (e.g. weather),

whereas the Japanese preferred

unsafe openings (e.g. religion).

Preferences also related directly to

perceptions of the other person and


stereotypical representation of the

other’s culture.

Nowak 2004 Study of the

influence of

anthropomorphism

and agency (human

vs. computer

operated virtual

image) on social

judgments in VE.

A 2 X 3 experimental set-up was

created: 1) participants were told

that they were interacting with

agent or avatar, and 2) they were

randomly given a virtual image of

their interaction partner that

represented three levels of

anthropomorphism (high, low, no

image). The task was 15 minutes

of voice interaction with

experimenter-controlled virtual

image in a desktop VE. Social

attraction, credibility and

uncertainty were measured.

134 university

students.

Contrary to hypotheses, the results

showed that low anthropomorphic

images were perceived as more

likable and credible than no image,

which was more credible and

likeable than the highly

anthropomorphic images. The

degree of agency had no effects on

any of the dependent variables and

no differences in uncertainty were

found.

Journal of

Computer-

Mediated

Communication

Nowak & Biocca 2003 Do people feel

more presence

A 2x3 between-participants

experiment with a variously

134 undergraduate

students.

The results showed no differences

in feelings of presence between

Presence


when interacting

with an avatar than

with an agent, and

is the sense of

presence lower

with partners

represented by less

anthropomorphic

images than with

partners

represented by

highly

anthropomorphic

images or no

images?

interactive participant (avatar

condition and agent condition) and

degree of anthropomorphism

(highly anthropomorphic image,

low anthropomorphic

image, and no image control).

avatars and agents. Existence of a

virtual image increased

telepresence. Participants

interacting with the less

anthropomorphic image reported

more co-presence and social

presence than those interacting

with partners represented by either

no image at all or by a highly

anthropomorphic image of the

other. This indicates that the more

anthropomorphic images set up

higher expectations that lead to

feelings of reduced presence when

these expectations are not met.

Nowak & Rauh 2005 Evaluation of

avatars in a

static context in

30 images were created from 3D

models which were divided into 4

image types: 10 male characters,

255 university

students

A test for a potential order effect

revealed that regardless of which

image participants saw first, the

Journal of

Computer-

Mediated


terms of their

androgyny,

anthropomorphism,

credibility,

homophily,

attraction, and

likelihood of

choosing the avatar

to represent

oneself.

10 female characters, 5 animal

characters and 5 objects. 4 human

images were purposefully children

and 4 were purposefully lower

rendering quality. All images

(including the objects) had

identifiable eyes and mouth to

ensure some level of consistency.

Measures were taken of

participants' gender, computer use,

computer efficacy, and the

perception of each presented

avatar's anthropomorphism,

androgyny, credibility, homophily,

attraction and their likelihood to be

chosen for an interaction.

first image shown was rated as

more androgynous and less

anthropomorphic than the

subsequent images, and less

homophily was felt towards the

first image than towards

subsequent images. Finally,

participants were less likely to

choose the first image to represent

themselves than subsequent

images. The most feminine avatar

was perceived as the most

attractive and the most masculine

avatar as the most

anthropomorphic. Human male

avatars and human female avatars

were rated as the most

anthropomorphic groups, followed

Communication


by animals and objects. Female

avatars were perceived as the most

attractive followed by men and

then by non-humans and objects.

Participants were more likely to

choose avatars that were human-

like and of the same gender as

themselves. The quality of the

image reduced the androgyny

perception.

Nowak & Rauh 2008 The study

examined the

reliance on visual

characteristics in

online avatar

perception and

especially the

relationships

In Step 1, respondents rated 30

static avatar representations on the

three dimensions in a survey. The

survey ratings were used to select

eight maximally differing avatars

for the experiment in Step 2. In the

second step, a different set of

participants rated their text chat

Altogether 485


Path analysis revealed that

perceived avatar androgyny

negatively predicted perceived

avatar anthropomorphism.

Perceived avatar

anthropomorphism, in turn,

positively predicted perceptions of

partner (represented by avatar in

Computers in

Human Behavior


between avatar

androgyny,

anthropomorphism

and credibility.

partners (represented visually by

randomly assigned avatars) on the

three dimensions.

chat experiment) credibility.

Overall, the experimental results

suggest that the visual cues of the

avatar were stronger predictors of

person perceptions than the

behavior (chat) of the person.

Peña, Hancock &

Merola

2009 Study of priming

effects in VEs.

That is, do

situational cues

activate

unconscious cue-

congruent thoughts

and deactivate cue-

incongruent

thoughts also in

VEs?

Two experiments were conducted

in desktop VEs. In experiment 1,

the priming was created by

dividing the participants into two

groups: those wearing black and

those wearing white cloaks.

Aggression and cohesion were

measured. In experiment 2, the

priming was done by randomly

dividing participants to use a) a

Ku-Klux-Klan (KKK) avatar, b) a

transparent avatar, or c) a doctor

Altogether 139

(51+88) university

students in groups

of three

(experiment 1) and

individually

(experiment 2).

After a three-person group decision

task (all having the same cloak),

the black-cloaked participants in

experiment 1 developed more

aggressive intentions and attitudes

but reported less cohesion than the

white-cloaked participants. In

experiment 2, persons with KKK

avatars wrote more aggressive and

less affiliative TAT stories than

persons with other avatars. The

results suggested that the automatic

Communication

Research


avatar. Participants wrote stories

based on Thematic Apperception

Test (TAT) pictures, which were

analyzed.

cognitive priming models are

applicable also in VEs.

Sallnäs

2005 How does

communication

mode affect

people’s

experience of

social presence,

presence, and

performance, and

how does it affect

their actual

collaboration in a

virtual

environment?

Two between-participants

experiments with three conditions:

VE text-chat, VE audio, and VE

video. The participants were

instructed to perform a decision-

making task in the VE. The time to

finish the task, number of words

used, and the number of words

used per second in

the dialogue between the

participants were measured.

Sixty subjects,

university students

and administrative

personnel.

Presence and social presence were

higher in audio and video

conditions than in the text

condition. No significant effect

was found in task performance, but

perceived performance was rated

highest in the video condition.

Significantly fewer words per

second were spoken in the text

condition than in other conditions.

In a web environment with audio

and video conditions, there was no

difference in presence, but

perceived presence was rated

Presence


higher in the video condition than

in the audio condition. There was

no difference in social presence or

performance.

Scheck,

Allmendinger &

Hamann

2008 Media richness

effects in a three-

person negotiation

task.

PC-based collaborative VE

experiment on multilateral and

multi-thematic negotiation.

Avatars with 1) text, 2) audio, and

3) audio + text conditions.

Measurement of objective

performance (time, quality) as well

as subjective satisfaction,

difficulty, concentration and

response speed.

Undergraduate

students in triads.

The performance quality was

similar in all conditions, but

solution was slower in condition 1

than in 2 or 3. No difference

between conditions 2 and 3 was

found. Moreover, higher ratings of

user satisfaction, lower difficulty

higher concentration and higher

response speed were reported for

conditions 2 and 3 as compared to

text condition.

Journal of Media

Psychology

Slater, Sadagic,

Usoh & Schroeder

2000 Comparison of

various aspects

(e.g. emerging

Experimental study that took place

over a two-week period. One

participant with HMD and two

Ten groups of three

people recruited by

an advertisement on

The immersed participant (with

HMD) was rated clearly as leader

and as most talkative after the

Presence


leadership and

group accord) of

small-group

behavior among

group members in

a virtual and a real

environment.

participants using workstation

displays met first in a shared VE

and carried out a collaborative

puzzle task of solving riddles. One

of the group members continually

monitored a particular one of the

other members in order to assess

social discomfort. After 15

minutes, the virtual session was

terminated, the participants

completed a questionnaire and

continued the same task in the real

world for another 15 minutes.

the university

campus.

virtual meeting, whereas after the

real meeting, each participant was

assigned approximately the same

leadership rating, and there were

no differences in perceived

talkativeness. Overall group scores

revealed a significant difference in

group accord, which was higher

after the real session than the

virtual session. However, this

difference could not be found at an

individual level. Women tended to

show higher accord scores than

men and the more riddles solved

the greater the accord was.

Talamo & Ligorio 2001 Construction of

identities in VE.

A community building a VE within

Active Worlds over eight months

for about four hours weekly both

38 students and 10

teachers from 7

schools in two

Identity construction was heavily

context-dependent, and identities

were not static but negotiated in

CyberPsychology

& Behavior


on- and off-line. Real-life identities

were known from the beginning.

Ethnographic and conversational

methodology was used. Identity

manipulation was possible through

changing avatar (fully), naming it,

and chat. Two types of data were

collected: 1) open responses to a

questionnaire, 2) recorded chat.

countries (age range

10 to 15 years)

interactions. Many strategies for

playing with different identities

and making them distinct were

found (most commonly trying

different avatars). Chat analysis

revealed different ways of

supporting one’s own and others’

identity construction (e.g. shifts

between real and virtual identity

talk).

Vasalou & Joinson 2009 The effects of

interactional

context on self-

presentation

through avatars.

First, the participants were divided

into one of three context groups

with a task of creating an avatar.

The contexts that were

hypothesized to create different

avatars were on-line 1) blogging

(supposedly neutral condition), 2)

dating and 3) detective gaming.

71 university

students.

The analysis of the experimental

data revealed no differences in

avatar attributes with one

exception: avatars for dating were

reported to be more attractive than

those for blogging. Moreover, the

participants created avatars that

were highly similar with their

Computers in

Human Behavior


The participants were led to

believe that their avatars would be

used later for the three respective

purposes. Second, after the

experiment, everyone was

interviewed and the data were

analyzed qualitatively.

features. No differences in

participants’ self-awareness across

the three groups were found. The

interviews revealed several

strategies of self-presentation but

these resulted in general tendency

to create self-representative

(realistic) avatars across contexts.

Vasalou, Joinson,

Bänziger, Goldie &

Pitt

2008 How users

negotiate their self-

presentation via an

avatar in social

media.

A thinking aloud method was

applied in two scenarios using

Yahoo! Avatars: 1) constructing an

avatar for a romantic date and 2)

making a postcard including the

avatar to send to a family member

regarding a nice surprise. The

participants’ speech and actions

were viewed and recorded by two

observers in an adjacent room.

20 volunteers from

university.

Three strategies or motives of self-

presentation via an avatar were

found independently of the two

scenarios. 1) An avatar was

customized to reflect a more or less

accurate self by using stable self

attributes or idealizing them to suit

social expectations. 2) Ambiguity

enabled playful avatar design in

order to engage the receiver in

International

Journal of Human-

Computer Studies


This data was used in qualitative

analysis.

reinvented interpretations of social

life. 3) Avatars were used as

proxies: they were designed to

embody and convey a literal or

symbolic message to the receiver.

Vinayagamoorthy,

Steed & Slater

2008 The impact of

character posture

on the

communication of

affect in IVE.

An experimental study of an IVE

where participants observed a

discussion of two virtual

characters. One character was

designed to be active using two

observable behavioral cues

(posture and facial expression) and

two emotional states (anger and

sadness), and the other character

was designed as passive,

portraying the same neutral state

throughout the experiment. The

verbal content of the discussion

49 male participants

recruited from a

university campus.

In angry conditions with postural

cues and facial expressions, the

participants were able to recognize

the emotional state of the active

character as anger, but this was not

the case in the sad condition.

Moreover, participants did not

perceive the affective state of the

passive character as being neutral,

even though the passive character

was designed to portray neutral

cues throughout the experiment.

The affective state of the passive

IEEE Transactions

on Visualization

and Computer

Graphics


was muffled in order to ensure that

the affective states could only be

inferred through the nonverbal

behavioral cues. The participant

was instructed to move toward the

characters and once s/he got close

enough, the active character

stopped the discussion, adopted a

neutral state and started an

experimenter-controlled discussion

with the participant while the

passive character remained a

listener. Two sets of between-

groups 2 (facial expression) X 2

(postural cues) factorial designs

were employed, as well as

interviews after the experiment.

character was instead perceived as

a function of the active character’s

behavioral animations. When

measuring personal distance, most

participants maintained smaller

interpersonal distance from the

active than the passive character.

Yee & Bailenson 2007 Does a change of Two between-subjects designs Altogether 82 In study 1, participants in the Human


one’s self-

representation

change one’s

behavior in a

virtual

environment?

with several conditions: study 1

avatars with attractive or

unattractive faces, and study 2

avatars of the same height, taller,

or shorter than the confederate.

undergraduate

students.

attractive condition walked closer

to the confederate and revealed

more information than

unattractively-represented

participants. In study 2,

participants represented as tall split

money more in their own favor

than those represented as short,

who were more likely to accept an

unfair offer than participants

represented as tall or of normal

height.

Communication

Research

Yee & Bailenson 2009 The relative

contribution of

self-perception and

priming to

behavioral changes

via digital self-

In study 1, participants interacted

with a confederate in a 2 (attractive

or unattractive avatar) X 2 (mirror

or playback condition) between-

subjects experiment. Participants

were assigned avatars with faces

73 undergraduate

students.

In the mirror condition,

participants in the attractive

condition were chosen more often

as a partner than participants in the

unattractive condition, and

participants in the unattractive

Media Psychology


representation that had been pretested for

attractiveness. In the mirror

condition, participants were

exposed to their avatar in a virtual

mirror whereas in playback

condition they saw a virtual

recording of the previous

participant in the same

attractiveness condition. In both

conditions, participants were

immersed in a virtual environment

by using HMD and asked to walk

closer to the confederate. The

confederate asked questions from

the participants to allow enough

time for social interaction. After

study 1, participants were asked to

pick two pictures from the photos

condition were significantly more

likely to increase their reported

height than participants in the

attractive condition. In the

playback condition, participants

in the attractive condition did not

have significantly different

reported height differences or

higher partner choice scores than

participants in the unattractive

condition. The means showed

that in the mirror condition,

participants in the attractive

condition walked closer to the

confederate than in the unattractive

condition, whereas in the playback

condition, the distance for

both attractive and unattractive


of nine people of opposite gender

in whom they were most interested

and who they thought would be

most likely be interested in them.

Participants were asked to tell their

height, but they were unaware that

their height was also automatically

measured in the IVE.

conditions was very similar.

Yee, Bailenson &

Duchenaut

2009 How avatar’s

appearance

changes user’s

online and offline

behavior.

In study 1, an automated script was

implemented in a multiplayer

online game (World of Warcraft,

WoW) which gathered information

on all individual characters, such

as their race, their location and

their current level within the game.

Screenshots of each race were

captured and the first four

randomly generated avatars for

In study 1 data was

collected from 76,843

players of WoW. In

study 2 the

participants were 40

undergraduate

students.

The findings of the study 1 show

that tall, attractive avatars are

expected to be the highest level in

the game, whereas the short,

attractive avatars are expected to

be the lowest level. Study 2

showed that the effect of height

was significant on the first split of

money both in virtual environment

and face-to-face trials but not

Communication

Research


every race and gender were

selected, resulting in 64 images of

WoW avatars. 22 undergraduate

students rated these images based

on their attractiveness. Height of

the race was calculated by

measuring the number of pixels of

their avatar’s height.

In study 2 participants were evenly

divided into short and tall

condition. Participants wore an

HMD and conducted a money

sharing task with a confederate in

IVE, where one designated a split

and the other would accept or

reject the split. After the

experiment, the participants

performed the same task again

significant on the second split or

on the final acceptance of money.


with the same person face-to-face.

Yee, Bailenson,

Urbanek, Chang &

Merget

2007 Do social norms of

gender and

interpersonal

distance (IPD) and

gaze transfer from

real life to VE?

Field experiment in Second Life.

Data gatherers carried a script that

saved information about avatars

(e.g. position, gaze direction).

From snapshots and script data

IPD and sum of gaze directions

were calculated for interacting

dyads.

835 dyads of Second

Life inhabitants

observed by

undergraduate

research assistants.

Norms were transferred to VE.

Male-male dyads maintained

greater IPD and maintained less

gaze-contact than female-female

dyads. However, mixed gender

dyads had the smallest IPD.

Decrease in IPD was compensated

for avoiding eye-contact. During

chat, the avatars were more likely

to look at each other than when not

chatting.

CyberPsychology

& Behavior

Abbreviations in the Appendix: DV = Dependent Variable; FTF = face-to-face; IV = Independent Variable; IVE = Immersive Virtual Environment; HMD = Head-Mounted Display


Biographies

Anu Sivunen, Ph.D. (Speech Communication), is Research Manager at Aalto

University, School of Science and Technology, Finland. Her research interests include global

virtual teams, collaborative virtual environments, group identity, and computer-mediated

communication. She has published articles of these issues in journals including IEEE

Transactions on Professional Communication and Group Decision and Negotiation.

Marko Hakonen, Ph.D. (Organizational Psychology), works as a researcher at Aalto

University, School of Science and Technology, Finland. He has studied virtual work and

virtual teams for over a decade. His specialties include research of social identification,

perceived justice and interpersonal trust in new ways of work.

Acknowledgement

The authors wish to thank the anonymous reviewers for their insightful comments as

well as Johanna Haapamäki and Petra Bosch-Sijtsema for their hard work on coding the

articles. This study was supported by the MIDE program, Aalto University School of Science

and Technology, Finland (http://mide.tkk.fi/en/).

Review of Virtual Environment Studies on Social and Group Phenomena

Documents