a study of group support systems and the intergroup setting

Available online at www.sciencedirect.com

45 (2008) 452–460www.elsevier.com/locate/dss

Decision Support Systems

A study of group support systems and the intergroup setting

John Lim ⁎, Xiaojia Guo

School of Computing, National University of Singapore, 3 Science Drive 2, Singapore 117543, Singapore

Available online 23 June 2007

Abstract

In this study, we explore the supporting role of group support systems (GSS) in organizational workgroups that involve groupsof distinct social identities. Intergroup processes become relevant and prevalent in such workgroups and trigger new issues.Anonymity is an important feature of GSS that concerns social influence in the intergroup setting. We report on the results of anexperiment consisting of 20 workgroups. The enforcement of common fate and its impact on social influence are also examined.Our findings provide preliminary indication of the usefulness of GSS in the intergroup setting of organizational workgroups.© 2007 Elsevier B.V. All rights reserved.

Keywords: Workgroup; Intergroup; Group support systems; Anonymity; Common fate

1. Introduction

The emergence and spread of teamworking and team-based organizations has been seen to be a principalcharacteristic of the world of work in the 1980s and1990s [46]. The characteristic is ever expanding withnew organizational forms (e.g., virtual and networkedorganizations) which arise in response to social andtechnological advances [13]. Different groups need towork together on certain projects (e.g., strategic planningfor the organization). Meetings of different groupsconducted electronically or otherwise, are consideredcommonplace. Intergroup processes become relevantand prevalent as groups of distinct social identities gatherand interact, which triggers issues that warrant separatetreatment from prior studies. In particular, the interven-tion of technologies with these issues and its con-sequences constitutes the research question in this paper.We will look specifically into group support systems

⁎ Corresponding author.E-mail addresses: [email protected] (J. Lim), [email protected]

(X. Guo).

0167-9236/$ - see front matter © 2007 Elsevier B.V. All rights reserved.doi:10.1016/j.dss.2007.06.007

(GSS) which have been pursued by researchers fordecades.

GSS are designed to facilitate interaction and fostercollaboration and decision making within groups. Theyaccomplish this purpose by providing group memberswith a set of features that aim to increase process gains(e.g., “learning” and “objective evaluation”) and reduceprocess losses (e.g., “conformance pressure” and“evaluation apprehension”) [34]. Although not alltheoretical treatments of GSS have agreed upon a singlelist of “essential features”, anonymity has been recog-nized by many researchers to play a key role inproducing effects (see [24] for a comprehensive review).While the feature of anonymity has been examined insingle-group contexts, there is relatively little or noknowledge of how it will affect intergroup interaction.How the anonymity feature of GSS may affect socialinfluence in an intergroup setting is an importantquestion that must be looked at [27].

This paper proposes a theoretical framework inter-relating three concepts: anonymity, intergroup, andinfluence. Hypotheses derived from this model weretested with an experiment. The paper is organized as

mailto:[email protected]

mailto:[email protected]

http://dx.doi.org/10.1016/j.dss.2007.06.007

453J. Lim, X. Guo / Decision Support Systems 45 (2008) 452–460

follows. Section 2 reviews the literature pertaining to theintergroup setting. Section 3 presents our research modeland hypotheses. The research method is described inSection 4. Section 5 reports the experimental results,followed by a discussion of the findings in Section 6.Section 7 concludes this study with future researchopportunities highlighted.

2. The intergroup setting

GSS have been defined as combining “communica-tion, computer, and decision technologies to supportproblem formulation and solution in group meetings”([9], p. 589). Much work has been done on GSSaddressing various aspects. These include four majorcategories of variables: contextual or independentvariables (e.g., characteristics of the GSS being used,of the group, task, environmental, and organizationalcontexts), intervening variables (e.g., session length,number of sessions, and presence and role of afacilitator), group adaptation processes (e.g., their levelof effort, their attitude toward the GSS, and participationpatterns), and outcomes (e.g., efficiency measures suchas calendar time to decision, effectiveness measuressuch as decision quality, usability of the system andmethods used, and subjective satisfaction measures) (see[12] for a comprehensive and integrative review).

However, an important phenomenon as regards smallgroup research that deserves attention is intergroupsetting. Most past GSS studies have examined the impactof GSS (or their features) on decision-making in a settingdefined by a small group isolated from all other socialgroups. This approach is appropriate when the field wasin an early stage of development. As the complexity ofGSS studies keeps increasing, it would be imperative tointroduce a more encompassing context. Consistent withthis view, the proposed study provides an intergroupenvironment in which two groups of distinct identitiesmake a decision together using GSS. Accordingly, weshall use the term “workgroup” to refer to the assemblyof all meeting participants, and “group” to refer to eachsmaller assembly defined by the relevant identity.

As far as intergroup literature is concerned, self-categorization theory (SCT) probably represents themost established line of research. SCT is a set of relatedassumptions and hypotheses about the functioning ofthe social self-concept — the concept of self based oncomparison with other people and relevant to socialinteraction. This theory grew out of the research onsocial identity. Social identity is defined as “theindividual's knowledge that he (she) belongs to certainsocial groups together with some emotional and value

significance to him (her) of the group membership”([44], p. 234). Much empirical evidence has attested tothe validity of SCT.

A central idea of SCT states that factors whichenhance the salience of ingroup-outgroup categoriza-tions tend to increase the perceived identity between selfand ingroup members (and the difference from outgroupmembers) and so depersonalize individual self-percep-tion on the stereotypical dimensions which define therelevant ingroup membership. Depersonalization refersto the process of “self-stereotyping” whereby peoplecome to perceive themselves more as the interchange-able exemplars of a social category than as uniquepersonalities defined by their individual differences fromothers. SCT argues that depersonalization of self-perception is the basic process underlying groupphenomena including group cohesiveness, subjectiveuncertainty, and social influence processes.

SCT provides an explanation for social influence thatemphasizes the social identities of the source and targetof influence [42–44]. The Social Identity Model ofDeindividuation Effects (SIDE) [38] builds on SCT andtries to extend it to provide a more detailed analysis ofthe effects of situational factors such as anonymity onsocial influence processes.

The SIDE model suggests that when social identity issalient – that is, when people define themselves as groupmembers rather than as individuals – the anonymity ofthe members would decrease attention regarding inter-personal differences and enhance the salience of thegroup and social identity. Such situations would result ina greater adherence to the group norms and greater groupinfluence. However, this model also points out that whenpersonal identity is salient, anonymity would notpromote normative responses, but rather it wouldencourage personal and individual responses [38–40].

According to SCT, group cohesion is produced andincreased by factors which lead to the formation andsalience of shared ingroup memberships. Supportiveevidence (see [28]) shows that common fate and explicitgroup membership is probably the most powerfuldeterminant of cohesion yet identified. Common fatehas been defined as representing “a coincidence ofoutcomes among two or more persons that arises becausethey have been subjected to the same external forces ordecision rules” ([3], p. 118). [26] and [4] saw commonfate (or “interdependence of fate”) to be a criticalprecondition for groups to become real, in a psycholog-ical sense.

Subjective validity [11,23], one's confidence in theobjective validity of one's opinions and beliefs, is adirect function of the extent to which similar people (in

454 J. Lim, X. Guo / Decision Support Systems 45 (2008) 452–460

relevant respects) in the same stimulus situation areperceived, expected, or believed to agree with one'sown response. Conversely, subjective uncertainty, acondition leading to social influence, is a direct functionof the extent to which similar others are not perceived orbelieved to respond similarly to oneself in the samestimulus situation. The negative feelings associated withuncertainty may be mitigated by conforming to themajority.

The direction of effective influence within the group(i.e., who successfully influences whom) is a function ofthe relative persuasiveness of the members, which isbased on the degree of relative consensual support for amember. In other words, other things being equal,majority opinion prevails. The perception of others as anappropriate reference group creates the shared expecta-tions of agreement necessary for the arousal ofuncertainty and mutual influence [1]. An individualwho holds minority view in his social category willexperience uncertainty because he disagrees with peoplecategorized as identical to him (i.e., the majority) (see[20]). Moreover, such uncertainty need not be lessenedeven if his view is shared by members of a differentsocial category, as they are not categorized as identicalto himself. If anything, sharing opinions with people ofa different social category may even increase theminority's uncertainty, thereby reinforcing the major-ity's influence.

3. Research model and hypotheses

Taking subjective uncertainty and conformity as thedependent variables and using the SCT conceptsdeliberated in the previous section, the research modelis put forward (Fig. 1).

Hypotheses 1 and 2 address the relationship betweencommon fate and salient aspects of intergroup relationswith respect to the minority. An individual perceives hisown image or esteem through his association, ormembership, with a social group. A group derives its

Fig. 1. Research model.

social identity through perceived differences with othergroups, as well as the perceived similarities within thegroup. [4] suggested that common fate is the dominantfactor in establishing group boundaries (i.e., differenti-ating between the ingroup and the outgroups). Membersin groups with high common fate perceive themselves tobe highly similar to each other in terms of goals andmembership [28]. In particular, they will actively seeksocial agreement, or subjectively correct perception,with other group members — people categorized assimilar to self [20]. Therefore, when a minority in a highcommon fate group disagrees with the majority, he orshe will experience subjective uncertainty [42]. Subjec-tive uncertainty refers to the lack of confidence in theobjective validity of one's beliefs, opinions, etc. [20]. Tomitigate uncertainty, he will in turn conform to themajority's views [27]. In other words, he or she willcomply with the ingroup's attitudes, opinions, beliefs,and behaviors.

Hypothesis 1. Greater minority uncertainty will beexhibited when common fate for a group is high thanlow.

Hypothesis 2. Greater minority conformity will beexhibited when common fate for a group is high thanlow.

Hypotheses 3 to 4 address the relationship betweenmode of GSS communication (anonymous versusidentified) and salient aspects of intergroup relationswith respect to the minority. Turner and Oakes ([44],p. 234) stated that part of an individual's self-conceptderives from his membership in social groups, whichencompasses the value and emotional significanceattached to it. Hence, people actively pursue member-ship and subjective social agreement in groups; when agroup member disagrees with his ingroup's majorityopenly (i.e., under identified communication), he risksbeing chastised and considered an outcast. This in turnleads him to experience subjective uncertainty. To avoidthis uncertainty, the minority conforms to the majority.However, anonymity in the intergroup context obscuresnot only the individual identities but also identities ofthe groups. Since group members are not able toassociate views with their originators, the endorsementfor social ingroup agreement becomes irrelevant orunnecessary. Therefore, when anonymous minoritiesvoice contrasting views, they experience less or nouncertainty. Subsequently, anonymity may insulategroup members from the conformance pressures thataccompany group interaction [2,31], resulting in lowerminority conformity.


Hypothesis 3. Greater minority uncertainty will beexhibited with identified communication than withanonymous communication.

Hypothesis 4. Greater minority conformity will beexhibited with identified communication than withanonymous communication.

Anonymity has been identified as one of the majordeterminants affecting decision processes and outcomesin GSS research (see [24] for a review). Experimentalstudies on GSS have generally found that anonymityleads to an increase in production and critical thought(e.g., [5,14,21,22,30]). It allows individual groupmembers to be less inhibited in their expression ofideas [10]. In addition, criticisms of others' contribu-tions will not be met with threat of direct repercussions[45]. Process losses, such as evaluation apprehension,member domination, conformance pressure and statuscompetition, are removed or reduced [34,37]. Therefore,anonymity will be particularly beneficial to the minority,encouraging him or her to voice out opposing views.Thus, anonymous minorities will experience highersatisfaction than identified minorities. Moreover, ano-nymity promotes task focus, encouraging members tofocus on ideas, regardless of who generated them [18],and providing an environment conducive to criticalrather than supportive behavior [22]. Subsequently,exploration of alternatives and surfacing of assumptionsincreases [17,18], which will, in turn, produce decisionswith high qualities. Thus, we derive the following twohypotheses:

Hypothesis 5. Lower minority satisfaction with deci-sion process will be experienced with identifiedcommunication than with anonymous communication.

Hypothesis 6. Lower decision quality will be producedwith identified communication than with anonymouscommunication.

4. Research method

4.1. Research design and subjects

A laboratory experiment with a 2×2 factorial designwas conducted to test the hypotheses. The independentvariables are common fate (high versus low) andmode ofGSS communication (anonymous versus identified),resulting in four conditions. Each condition was assignedfive workgroups.

One hundred and twenty undergraduate students froma large university served as subjects in the experiment.Each experimental session involved six subjects who

were randomly assigned to two groups (with threemembers each), being the “Advisory Committee” andthe “Board of Directors” (see the “TASK” section forexplanations).

A level-1 group support system was used for thisexperiment (see [9]). By definition, level-1 GSS arecommunication media aimed at improving the decisionprocess by removing common communication barriersand facilitating information exchange among members[9]. Its features include a public screen for instantaneousdisplay of ideas, anonymous input of ideas andelectronic message exchange between members.

4.2. Task

A strategic planning case from [47] was used. The taskrevolved around Tidewater College, which was describedto be situated in a rural mountain area and founded by thelocal church. With an annual enrolment of 450 students,Tidewater College provided an opportunity for the localresidents to obtain a college education, focusing on liberalarts and teacher preparation. However, the 1960s' babyboom brought an insurge of students, which greatlyoverextended the faculty teaching loads. Encouraged bysuch overwhelming enrolment, the former collegepresident invested heavily in new facilities to accommo-date the sudden inadequacies. Funding for the newfacilities was supported by the government, on thecondition that the college offered several new programs.Unfortunately, the college was soon saddled withmultipleproblems, including dwindling enrolment, financialproblems, bad reputation and low morale. The inabilityof the college to repay its debt, as well as its incompetenceto complete certain government programs, amplified itsdismaying predicament.

The subjects were informed that the previouspresident of the college had resigned in despair. Thenew president had assigned them, members of theAdvisory Committee and Directors' Board, to solvethese grappling problems together.

4.3. Experimental procedure

Upon arriving for the study, subjects were informed oftheir respective groups (i.e., Advisory Committee orBoard of Directors) by the experimenter, and led into thefirst stage of the study. This consisted of a ten-minutewarm-up discussion about a topic totally independentfrom the actual experimental discussion topic. Thepurpose of this warm-up discussion was to enable thesubjects to become more accustomed to the environment,and to each other. After the topic sheets were handed out,

Table 1ANOVA table for degree of uncertainty

Source Sum-of-squares df Mean-square F-ratio P

Anonymity 16.70 1 16.70 3.84 0.06Common fate 17.70 1 17.70 4.07 0.06Anonymity⁎

common fate10.50 1 10.50 2.42 0.13

Error 99.96 23 4.35


the subjects discussed the warm-up task and came upwitha workgroup consensus ranking. At the end of thisdiscussion, the subjects underwent a brief training sessionon how to use the system.

Subsequently, the experimenter handed out the de-scription of the problem. Before the subjects commencedon the actual discussion topic, the experimenter informedtheworkgroup that their ultimate goal was to come upwitha single coherent recommendation on how to solve theproblem. In addition, the workgroup was told that theywere given 45 minutes to come up with the solution. Forthe anonymous condition, subjects were logged into thesystem with similar-looking user identification names.They were also told not to reveal their identities to otherparties during the entire course of the discussion. Subjectsin the identified condition were logged into the systemwith identification names specifying distinctly the groupsthey belonged to. Common fate was manipulated using areward mechanism. Subjects in the low common fate cellwere informed that their performance would be evaluatedbased on the quality of their individual contribution to thefinal outcome. Subjects in the high common fate cell, onthe other hand, were informed that their performancewould be evaluated based on the quality of their group'scontribution to the final outcome. A number of studieshave shown that groups governed by such rewardstructures tend to work together more cooperatively andproductively than when individuals within groups arerewarded differentially (e.g., [48]).

The subjects were informed that they, as members ofthe Advisory Committee or the Board of Directors, haveto tackle the college's problems together with the othergroup. The workgroup then proceeded to solve theproblem with all members interacting via the groupsupport system. When the workgroup had reached asolution, they recorded it on the answer sheet provided.Then they were administered a questionnaire adaptedfrom [15] to measure the workgroup members'satisfaction levels. After that, the experimenter thankedthe subjects for their participation and dismissed themfrom the experiment.

4.4. Measurement of dependent variables

4.4.1. Determination of minoritySubjects were asked to what extent they perceived

themselves to be minorities of their ingroups. Theratings of their answers were summed up and the highestscore for each group was taken to be the minority's.Minority variables' (i.e., minority uncertainty, confor-mity and satisfaction with decision process) measure-ments were correspondingly derived.

4.4.2. Minority uncertaintyThe degree of minority's uncertainty was measured

using the following questions: (a) the degree ofconfidence with own views: not at all/to a large extent;(b) the degree of uncertainty during the discussion: notat all/to a large extent; (c) the degree of perceptions ofcorrectness of own views: not at all/to a large extent.These items were measured using a rating continuumwith five intervals.

4.4.3. Minority conformityThe degree of minority's conformity was measured

using the following two questions: (a) the frequency ofconforming to ingroup: not at all/to a large extent; (b) thedegree of conformity to ingroup: not at all/to a largeextent. These items were measured using a ratingcontinuum with five intervals.

4.4.4. Minority satisfaction with decision processQuestions from [15] were adapted to measure the

minority's level of satisfaction with the decision process.

4.4.5. Decision qualityThe solution on the answer sheets was evaluated by

two independent raters. Each solution was graded on itseffectiveness, feasibility, creativity, significance, andcompetence, similar to the five rating criteria in Leathers'Productivity Rating Instrument (PRI) [25], and rated on ascale of zero to ten. For decisions consisting of more thanone “action”, each action was rated using the ratingcriteria as stated above; subsequently, such ratings foreach workgroup were summed up and averaged.

5. Results

As we are only analyzing the data of the minority, theeffective sample size is relatively small. Hence, we shalluse p= 0.10 as the significance level for the data analysis.

5.1. Minority uncertainty

ANOVA showed a main effect due to anonymity(F=3.84; p=0.06) (see Table 1). Minorities under the

Table 2ANOVA table for degree of conformity



common fate7.83 1 7.83 2.84 0.11

Error 63.35 23 2.75

Table 4ANOVA table for decision quality



common fate0.02 1 0.02 0.06 0.80

Error 5.98 16 0.37

Table 5Summary of hypotheses versus experimental results

Hypothesis Supported?

1 Greater minority uncertainty will be exhibitedwhen common fate for a group is high than low.

Yes

2 Greater minority conformity will be exhibitedwhen common fate for a group is high than low.

No


anonymous condition experienced lower uncertainties(mean=6.54, S.D.=1.81) than their counterparts underthe identified condition (mean=7.71, S.D.=2.61).

In addition, there was also a main effect due tocommon fate (F=4.07; p=0.06) (see Table 1).Minorities in high common fate groups experiencedhigher uncertainties (mean=7.92, S.D.=2.50) thanminorities in low common fate groups (mean=6.53,S.D.=2.00).

5.2. Minority conformity

ANOVA showed a main effect due to anonymity(F=3.77; p=0.07) (see Table 2). Minorities under theanonymous condition conformed less to their ingroup(mean=5.77, S.D.=1.42) than their counterparts underthe identified condition (mean=6.86, S.D.=1.92).

5.3. Minority satisfaction with decision process

ANOVA showed a main effect due to anonymity(F=3.66; p=0.07) (see Table 3). Minorities under theanonymous condition were more satisfied (mean=16.62,S.D.=2.84) as compared tominorities under the identifiedcondition (mean=13.86, S.D.=5.29).

5.4. Decision quality

ANOVA indicated that anonymous workgroups(mean=8.08; S.D.=0.67) achieved higher decisionquality than identified workgroups (mean=7.40; S.D.=0.51) (F=6.34; p=0.02) (see Table 4).

Table 3ANOVA table for satisfaction with decision process



common fate9.96 1 9.96 0.54 0.47

Error 426.23 23 18.53

Table 5 summarizes the experimental results withrespect to the hypotheses proposed earlier.

6. Discussions and implications

6.1. Common fate

Experimental results indicated that minorities in highcommon fate groups experienced greater uncertainty thanthose in low common fate groups. This is consistent withour hypothesis. [20] suggested that it is fundamentallyimportant for individuals to be confident and certain aboutthe correctness and validity of their perceptions, attitudesand behaviors; accordingly, subjective uncertainty is apoor basis for functional conduct, and individuals need tobe assured that they understand and agree with others whoshare the same group memberships. Hence, the definingand prescriptive attributes of a group, or the groupstandard or norm, are defined from perceived intragroupsimilarities and intergroup differences. Internalized bygroup members through the process of self-categoriza-tion, this standard enhances self-perception, and reducessubjective uncertainty. In a group with high common fate,

3 Greater minority uncertainty will be exhibitedwith identified communication than withanonymous communication.

Yes

4 Greater minority conformity will be exhibitedwith identified communication than withanonymous communication.

Yes

5 Lower minority satisfaction with decision processwill be experienced with identified communicationthan with anonymous communication.

Yes

6 Lower decision quality will be produced withidentified communication than with anonymouscommunication.

Yes


group members perceive themselves to be very similar toothers in the same group. As a result, the strong bondingamong group members induces them to actively seekintragroup consensus; when a group member discoversthat he disagrees with other members of his own group(i.e., he belongs to the minority), subjective uncertaintyemerges.

ANOVA results indicated that there was nomain effecton conformity due to common fate. However, amarginally significant interaction on conformity waspresent (F=2.84; p=0.11). A further exploration wasconducted into the apparent moderating effect ofanonymity on the relationship between common fateand conformity. T-test results revealed that under identi-fied condition, minorities from high common fate groupsexhibited marginally greater conformity (mean=8.00;S.D.=1.58) than minorities from low common fategroups (mean=6.33; S.D.=1.80) (t=1.73; p=0.11).Thus, the hypothesis that minorities from high commonfate groups would exhibit greater conformity than theircounterparts in low common fate groups appears to besupported under identified communication. In groupswhere communication is identified and common fate ishigh, intragroup cohesion and cooperation become toppriorities and accentuate the pressure to conform. Whenminorities from these groups voice opposing views, theiridentities are known and they risk exclusion from thegroup. As a result, conformity to the majority's judgmentis advocated and any deviation from the group consensusis discouraged. This is consistent with [2], a study ongroup pressure, which reveals that subjects generallyconformed because of “social” reasons such as the fear ofgroup exclusion. On the other hand, the degree ofconformity exhibited by minorities from both high andlow common fate groups was insignificant whencommunication is anonymous. Anonymity obscures thegroups and provides a shield for minorities' identities; itassures minorities that their identities will not be knownand hence, deviation will not be punished. Consequently,the pressure to conform evaporates as maintenance ofgroup harmony becomes less of an issue.

6.2. Anonymity

Results showed that minorities under the identifiedcondition experienced more uncertainty and conformedmore to the ingroupmajority than their counterparts underthe anonymous condition. In addition, they (identifiedminorities) also experienced lower satisfaction withdecision process. Under identified communication, mem-bers' identities are known to the others. Therefore, when amember's opinions deviate from those of the ingroup

majority, he or she experiences subjective uncertainty. Asa result, to avoid being rebutted or being cast as the“outsider”, the minority member conforms to themajority's views.However, under anonymous conditions,the groups become obscure as the identities of groupmembers are not revealed. In this case, social identitypersonalizes to self identity. Since group members are notable to associate views with their originators, theendorsement for social ingroup agreement becomesirrelevant or unnecessary. Thus, a minority memberexperiences a lesser degree of uncertainty. Moreover,since the anonymity feature acts as a protective shield forthe deviant's identity, the fear of being singled out byothers is eliminated. The minority could then voiceunpopular and opposing views without fear of beingadmonished by his ingroup members; the pressure tocomply lessens and the minority conforms less underanonymous conditions. The findings on minority uncer-tainty and conformity are consistent with the SIDEmodel,which points out that when personal identity is salient,anonymity would not promote normative responses, butrather it would encourage personal and individualresponses. Correspondingly, minorities are less appre-hensive and more involved in the group problem-solvingprocess, resulting in higher satisfaction. This is consistentwith several GSS studies, which found that anonymityincreases satisfaction (e.g., [7,8,16,32,33,36]).

Decision quality was better for anonymous work-groups than for identified workgroups. Process losses,such as conformance pressure and evaluation apprehen-sion, lead to inadequate information search andevaluation, resulting in inferior decisions [35,41]. [35]proposed that anonymity reduces or eliminates evalua-tion apprehension and conformance pressure. Shymembers are encouraged to speak up and criticizewithout repercussions or reprisals, inadvertently encour-aging task participation and increasing process gains bycatching errors [35]. As a result, more creative andcritical suggestions are generated. Moreover, withoutthe knowledge of contributors' identities, the worth ofcontributions, and not the contributors, is appraised andjudged objectively [45]. It is only through extensive andcritical debate on the merits and demerits of a particularalternative that it can be appraised and analyzedthoroughly; complacency in exploring the feasibilityof an alternative may well result in an inferior solution.Hence, anonymity produces better decision quality.

7. Concluding remarks

This paper has presented a model as well as empiricalfindings concerning the impact of anonymity, an integral


component of GSS, on influence-related variablesincluding uncertainty and conformity. It provides astarting point for the examination of issues related to theintergroup setting, a much neglected aspect in GSSresearch, yet one which is increasingly important arisingfrom the emphasis on globalization and new organiza-tional forms. Teams can work together virtually (accord-ingly termed “virtual teams”), using GSS to workdistributed in time and space over the Internet [6].Distributed virtual work significantly reduces traditionalverbal and visual communication, forcing groups to relymore on the electronic communication provided by theGSS. The intergroup setting is prominent for virtual teamsas they are characterized by the combination ofgeographic dispersion of team members, disciplinaryheterogeneity and different organizational affiliations[29]. Therefore, the findings of this paper are likelyto be generalized to virtual teams with meticulousjustifications.

The paper, limited in certain aspects, calls for futureresearch in various directions. Firstly, unlike organiza-tional workgroups which usually take more than onemeeting to reach a decision, each workgroup in ourexperiment had only 45 minutes to reach a consensus.Such short period may not fully bring out the behaviorsof workgroups in organizations. Therefore, futureresearch of a longitudinal nature is warranted. Secondly,taking an exploratory step on the topic, this study usedthree-member groups; as the majority-versus-minoritydemarcation may become more pronounced in largergroups, group size should be a factor to be examined infuture research. Similarly, the number of groups inter-acting is another factor that deserves further investiga-tion. Thirdly, cultural factor concerning organizationalworkgroups is important to consider in light of culture'sconsequences for social influence (e.g., minority uncer-tainty and conformity) (see [19]). Lastly, this study hasfocused on the communication feature of GSS. Otherfeatures, especially those of level-2 and level-3 GSS [9],should also be looked at in future research.

References

[1] C.N. Alexander, L.G. Zucker, C.L. Brody, ExperimentalExpectations and Autokinetic Experience: Consistency Theoriesand Judgmental Convergence, Sociometry, 1970.

[2] S.E. Asch, Effects of group pressure upon the modification anddistortion of judgement, in: H. Guetzkow (Ed.), Groups,Leadership and Men, Carnegie Press, Pittsburgh, PA, 1951.

[3] M.B. Brewer, Superordinate goals versus superordinate identityas bases of intergroup cooperation, in: D. Capozzo, R. Brown(Eds.), Social Identity Process: Trends in Theory and Research,Sage, London, 2000.

[4] D.T. Campbell, Common fate, similarity, and other indices of thestatus of aggregates of persons as social entities, Behavioral Science3 (1958).

[5] T. Connolly, L.M. Jessup, J.S. Valacich, Idea generation in aGDSS: effects of anonymity and evaluative tone, ManagementScience 36 (6) (1990).

[6] A.R. Dennis, B.H. Wixom, Investigating the moderators of thegroup support systems use with meta-analysis, Journal ofManagement Information Systems 18 (3) (2002).

[7] A.R. Dennis, A.R. Heminger, J.F. Nunamaker, D.R. Vogel,Bringing automated support to large groups: the Burr–Brownexperience, Information and Management 18 (3) (1990).

[8] A.R. Dennis, C.K. Tyran, D.R. Vogel, J.F. Nunamaker, Anevaluation of electronic meeting support for strategic management,Proceedings of the 11th International Conference on InformationSystems, 1990.

[9] G. DeSanctis, R.B. Gallupe, A foundation for the study of groupdecision support systems, Management Science 33 (5) (1987).

[10] M. El-Shinnawy, A.S. Vinze, Technology, culture, and persua-siveness: a study of choice-shifts in group settings, InternationalJournal of Human-Computer Studies 47 (1997).

[11] L. Festinger, Informal social communication, PsychologicalReview57 (1950).

[12] J. Fjermestad, S.R. Hiltz, An assessment of group supportsystems experimental research: methodology and results, Journalof Management Information Systems 15 (3) (1999).

[13] J. Fulk, G. DeSanctis, Electronic communication and changingorganizational forms, Organization Science 6 (4) (1995).

[14] R.B. Gallupe, W.H. Cooper, S. Pollard, J. Cadsby, Electronicbrainstorming, anonymity and deviance, Working Paper, Queen'sUniversity, Kingston, Ontario, 1997.

[15] S.G. Green, T.D. Taber, The effects of three social decisionschemes on decision group processes, Organizational Behaviorand Human Performance 25 (1980).

[16] J.F. George, G.K. Easton, J.F. Nunamaker, G.B.Northcraft, A studyof collaborative group work with and without computer-basedsupport, Information Systems Research 1 (4) (1990).

[17] S.C. Hayne, R.E. Rice, Attribution accuracy when usinganonymity in group support systems, International Journal ofHuman-Computer Studies 47 (1997).

[18] S.C. Hayne, R. Rice, P. Licker, Social cues and anonymous groupinteraction using group support systems, Proceedings of the 27thHawaii International Conference on System Sciences, vol. 4, 1994.

[19] G. Hofstede, Culture's Consequences: International Differencesin Work-Related Values, Sage, Beverly Hills, CA, 1980.

[20] M. Hogg, D. Abrams, Towards a single-process uncertainty-reduction model of social motivation in groups, in: M. Hogg, D.Abrams (Eds.), Group Motivation: Social Psychological Per-spectives, Harvester Wheatsheaf, Hertfordshire, 1993.

[21] L.M. Jessup, D.A. Tansik, Decision making in an automatedenvironment: the effects of anonymity and proximity with agroup decision support system, Decision Sciences 22 (1) (1991).

[22] L.M. Jessup, T. Connolly, J. Galegher, The effects of anonymityon group process in an idea-generating task, ManagementInformation Systems Quarterly 14 (3) (1990).

[23] H.H. Kelly, Attribution theory in social psychology, in: D. Levine(Ed.), Nebraska Symposium on Motivation, University ofNebraska Press, Lincoln, 1967.

[24] E.E. Klein, C.C. Clark, P.J. Herskovitz, Philosophical dimensionsof anonymity in group support systems: ethical implications ofsocial psychological consequences, Computers inHumanBehavior19 (2003).


[25] D.G. Leathers, Quality of group communication as a determinantof group product, Speech Monographs 39 (3) (1972).

[26] K. Lewin, Resolving Social Conflicts: Selected Papers on GroupDynamics, Harper, Oxford, England, 1948.

[27] L.H. Lim, Toward globalization/regionalization: the intergroupdimension of meetings technology, Proceedings of the 1st Asia-Pacific DSI Conference, III, 1996, Hong Kong.

[28] A.J. Lott, B.E. Lott, Group cohesiveness as interpersonal attraction:a review of relationships with antecedent and consequent variables,Psychological Bulletin 64 (1965).

[29] A. Majchrzak, R.E. Rice, A. Malhotra, N. King, S. Ba,Technology adaptation: the case of a computer-supported inter-organizational virtual team, Management Information SystemsQuarterly 24 (4) (2000).

[30] P. McLeod, An assessment of the experimental literature onelectronic support of group work: results of a meta-analysis,Human-Computer Interaction 7 (1992).

[31] S. Milgram, The Individual in a Social World, Addison-Wesley,Reading, MA, 1977.

[32] J.F. Nunamaker, L.M. Applegate, B.R. Konsynski, Facilitatinggroup creativity with GDSS, Journal of Management InformationSystems 3 (4) (1987).

[33] J.F. Nunamaker, L.M. Applegate, B.R. Konsynski, Computer-aided deliberation: model management and group decisionsupport, Journal of Operations Research 36 (6) (1988).

[34] J.F. Nunamaker, A.R. Dennis, J.S. Valacich, D.R. Vogel, J.F.George, Electronic meeting systems to support group work,Communications of the ACM 34 (7) (1991).

[35] J.F. Nunamaker, A.R. Dennis, J.S. Valacich, D.R. Vogel, J.F.George, Group support systems research: experience from the laband field, in: L.M. Jessup, J.S. Valacich (Eds.), Group SupportSystems: New Perspectives, Macmillan, New York, 1993.

[36] J.F. Nunamaker, D.R. Vogel, A. Heminger, B. Martz, R.Grohowski, C. McGoff, Experiences at IBM with group supportsystems: a field study, Decision Support Systems 5 (2) (1989).

[37] A. Pinsonneault, K.L. Kraemer, The effects of electronic meetingson group processes and outcomes: an assessment of the empiricalresearch, European Journal of Operational Research 46 (2) (1990).

[38] S. Reicher, R. Spears, T. Postmes, A social identity model ofdeindividuation phenomena, EuropeanReviewof Social Psychology6 (1995).

[39] R. Spears, M. Lea, Social influence and the influence of the“social” in computer mediated communication, in: M. Lea (Ed.),Contexts of Computer-Mediated Communication, HarvesterWheatsheaf, Londres, 1992.

[40] R. Spears, M. Lea, Panacea or panopticon? The hidden power incomputer-mediated communication, Communication Research21 (1994).

[41] I.D. Steiner, Group Process and Productivity, Academic Press,New York, 1972.

[42] J.C. Turner, Rediscovering the Social Group: A Self-Categori-zation Theory, Basil Blackwell, Oxford, UK, 1987.

[43] J.C. Turner, Social Influence, Open University Press, Bucking-ham, UK, 1991.

[44] J.C. Turner, P.J. Oakes, Self-categorization theory and socialinfluence, in: P.B. Paulus (Ed.), Psychology of Group Influence,Lawrence Erlbaum, Hillsdale, NJ, 1989.

[45] J.S. Valacich, L.M. Jessup, A. Dennis, J.F. Nunamaker, Aconceptual framework of anonymity in group support systems,Proceedings of the 25th Hawaii International Conference onSystem Sciences, vol. 4, 1992.

[46] M.A. West (Ed.), Handbook of Work Group Psychology, Wiley,Chichester, UK, 1996.

[47] D.A. Whetten, K.S. Cameron, Developing Management Skills,Scott, Foresman and Company, Glenview, Illinois, 1984.

[48] S. Worchel, H. Rothgerber, E.A. Day, D. Hart, J. Butemeyer,Social identity and individual productivity within groups, BritishJournal of Social Psychology 37 (1998).

Dr. John Lim is Associate Professor in theSchool of Computing at the National Uni-versity of Singapore. Concurrently, he headsthe Information Systems Research Lab. Dr.Lim graduated with First Class Honors inElectrical Engineering and an M.Sc. in MISfrom the National University of Singapore,and a Ph.D. from the University of BritishColumbia. His current research interestsinclude e-commerce, collaborative technol-
ogy, negotiation support, IT and education,
and IS implementation. He has published in MIS and related journalsincluding Journal of Management Information Systems, Journal ofGlobal Information Management, Decision Support Systems, Interna-tional Journal of Human Computer Studies, Organizational Behaviorand Human Decision Processes, Behaviour and Information Technol-ogy, International Journal of Web-based Learning and TeachingTechnologies, Journal of Database Management, and Small GroupResearch.

Xiaojia Guo is currently a Ph.D. candidate inthe Department of Information Systems,School of Computing, National University ofSingapore. She has received Bachelor ofComputing (First Class Honors) in June2004. Her primary research interests includenegotiation support systems and their impactupon the process and outcomes of inter-teamnegotiations and group negotiations. Herpapers have appeared in the proceedings ofHCI International Conference, Pacific Asia
Conference on Information Systems, Grouption Conference, and IRMA International Decision and Negotia
Conference. She has also published in the Information and SoftwareTechnology journal.

a study of group support systems and the intergroup setting

Documents