The Psychological Dynamics and Social Climate of Text-Based Virtual Reality
Dennis A. Johnston
MUDs, though argued by a number of Computer Mediated Communications (CMC) researchers to be the most indicative of emergent cyberculture (Bruckman, 1992; Carlstrom, 1992; Roush, 1993; Cherney, 1994, Reid, 1994, Snell, 1994, Sempsey, 1997), have to date received little attention from CMC researchers and even less from main stream social psychologists, yet their increasing utilization as educational, mutual support and work environments warrants deeper investigation. As with face to face groups, MUD groups come in a variety of types and sizes, as do the types of groups that inhabit them. The matter of what formations definitively constitute a coherent group is as debatable for virtual spaces as it is with face to face situations. Certainly, a group must consist of several individuals, and ideally there should exist some degree of group identity. That is, a group should be able to identify that it is indeed a group. Generally, in the field of group dynamics a clear distinction is made between a group and a mere gathering of people in a common space, such as is found in such situations as people riding on a bus together, walking down the same street, or gathered in a hotel lobby. According to Leland Bradford (1978), a group identity can be said to have been formed when the following criteria are met: patterns of interaction are proven effective; differences in perception about task, communication, and procedures are clarified; relationships to other persons and groups are delineated; standards for participation are set; methods of work that elicit rather than inhibit contributions are established; a respected "place" for each person is secured; and, trust is established among members. To this extent then, MUD groups might be expected to conform to these criteria as much as any face to face group.
Certain key differences between MUD and face to face groups should be noted. In face to face groups, individuals often meet and gather due as much to geographical proximity as common interests. However, no such geographical restrictions apply to MUD membership. MUD group membership is strictly voluntary and it follows that common interest or common personal temperament might exert a greater influence upon group formation and maintenance. In addition, due to the nature of the medium there is a varying amount of lag time between member's responses. Therefore, they will often indulge in, and be able to keep track of, several separate conversations at a time (often with individuals from outside the present group) without the confusion and noise that would normally result from such activity in the real world (Carlstrom, 1992; Curtis, 1992). These differences are noteworthy in that formation processes and group boundaries in MUDs may be qualitatively unique as compared to face to face groups.
The variables examined within this study fall under the rubric of Social Climate as measured by form R of the Group Environment Scale (GES). Moos, (1994) describes social climate as the "...'personality' of a setting or environment, such as a family, a workplace, or a classroom." Additionally, Moos stresses that "... each setting has a unique 'personality' that gives it unity and coherence." Social climate can have a strong influence on individuals within a particular setting, effecting such variables as morale, well-being, achievement, impulse control and so on. Interestingly, it has been demonstrated that group members' sociodemographic and other personality attributes are only minimally related to their perceptions of the group social climate (Moos, 1994). Measurement of these environmental characteristics can help researchers understand the overall behavior of groups, particularly as they relate to outcomes on both the aggregate and individual level. In addition, such measurements make it possible to compare groups and examine the determinants of group differences.
It appears to be generally accepted that individuals are less inhibited when interacting in MUDs than they might be in face to face circumstances (Bruckman, 1992; Carlstrom, 1992; Rosenberg, 1992; Roush, 1993; Penkoff, 1994; Reid, 1994), which is consistent with findings for computer mediated communication (CMC) in general (Kiesler et al., 1984; Hiltz, et al., 1986; Daly, 1993). It has been postulated that this phenomenon may be attributable to the lack of regulating social context cues (e.g., body language, tone of voice, etc..) and that the lack of such cues are read as an obscuring of the boundaries which delineate the forms of behavior which are deemed acceptable or unacceptable (Kiesler et al, 1984; Rosenberg, 1992; Reid, 1994). It should be noted however, that disinhibition is not necessarily always negative in these settings and that there appears to be just as much positive disinhibition as negative (e.g. displays of affection, mutual support, volunteerism, etc..). In fact, people may tend to be friendlier towards one another in MUDs than in face to face groups. Resnick and Bruckman (1993) claim, through their interviews with various participants of the MediaMoo project, that they found that participants reported the most salient aspects of virtual community as, "ease of collaboration, availability of technical assistance from peers, playfulness, availability of an audience for completed work, and community spirit." A number of studies have found that CMC based groups exhibit greater equality of participation than do face to face groups (Hiltz et al., 1986; Siegal, Dubrovsky, Kiesler & McGuire, 1986; Dubrovsky et al. 1991; Weisband, 1992). In addition to overall equality of participation, it has also been reported that the difference between high and low status member participation is diminished in CMC groups (Dubrovsky et al. 1991). However, Adrianson and Hjelmquist (1991) found no difference between CMC and face to face groups on participation, but did find that the levels of participation equality were greatly influenced by experience with CMC. That is, experienced group members participated more equally in computer based groups than did inexperienced members. Likewise, Straus (1996) reported that although CMC groups do show greater equality of participation, there still exist differences among CMC group member participation rates associated with individual differences in extraversion. Findings suggest that there is less normative interpersonal or social pressure in CMC based groups. McGuire et al. (1987), Dubrovsky et al. (1991) and Weisband (1992) all reported that CMC groups took longer to reach consensus. These findings indirectly indicate reduced levels of normative pressure in CMC groups. In addition, Smilowitz, Compton and Flint (1988) conducted an experiment wherein computer based groups were exposed to an Asch (1956) type conformity study, and the results compared to those found by Asch. The authors came to the conclusion that there is less conformity in CMC environments. Daly (1993), in a comparison of the performance of face to face groups and CMC groups on a task involving collective problem recognition and solution found greater influence of people proposing plausible hypotheses in face to face groups. There is one major caveat to these prior findings. Many of these studies were conducted in areas of CMC outside of the more environmentally complex domain of MUDs and as such may not be wholly applicable to the behaviors exhibited within them. Previous studies that directly pertain to the psycho/social aspects of MUDs have tended to be observational in nature and thus leave open to question the degree of difference to be found between MUD based and face to face groups. Due to the paucity of research in this area, the purpose of the present study is to compare MUD groups to face to face groups on various dimensions of social climate.
SUBJECTSFor a group to be considered for inclusion, it was decided that at least 75% of the respondents from any one particular group must have returned completed surveys. This is in accordance with the recommended criteria for group survey inclusion given in the user's guide of the GES (Moos, 1994). Thus, for example if a group consisted of 4 individuals, 3 of them would have to return completed surveys in order for the group to be included in the overall sample. A total of 31 groups met this criteria and were used in the present study.
SETTING AND MATERIALSThe MUD sites used in the present investigation were selected according to general accessibility from the random MUD link generator located at the comprehensive MUD Connector web site (http://www.mudconnect.com/mud.html). This site contains a comprehensive database of MUD addresses and offers a program which randomly selects and connects users to one of several hundred MUD sites. A complete list of the MUDs surveyed, along with their internet addresses, code bases, themes, degree of player killing allowed, survey compliance rates and sizes of the subgroups surveyed are available in Table 1.
TABLE 1: MUDs SURVEYED
Form R of Moo's (1994) Group Environment Scale was used to assess the social climate of various subgroups encountered within each MUD site. The GES has a long and well documented history as a valid measure of social climate (Finney & Moos, 1984; Moos, 1986, 1993; Toro, Rappaport & Seidman, 1987; Young & Williams, 1989). The GES consists of 90 true/false items composing 10 subscales that measure such constructs as: Cohesion, Support, Expressiveness, Independence, Task Orientation, Self-Discovery, Anger, Order, Leader Control and Innovation (Moos, 1994). For a complete definition of these constructs see Appendix A.
Administration of the GES results in a group profile which can then be compared to norms established for other groups. Normative data have been obtained using face to face social recreational groups such as chess, bridge, book, and cooking clubs; church groups, softball, canoeing and backpacking groups; and boys' and girls' sports teams (Moos, 1994). Since this study concerned itself with MUDs which are social-recreational in nature, the social-recreational data for face to face groups was used as the baseline for comparison to MUD groups.
PROCEDURETypically, upon logging into a particular MUD one can encounter any number of sub-groups gathered and engaged in various activities - much like if one were to step into a busy office building with its numerous floors, subdepartments, workgroups, etc... Each of the subgroups that were contacted were invited to participate in the study. Those choosing to participate were sent a copy of the survey and group members were asked to complete and email it back to the researcher at the end of their meeting or activity. Names of the participants were not gathered, although e-mail addresses were collected for purposes of follow-up requests for completed surveys (if necessary). In addition to the survey questions, an attempt was made to gather sample demography including the age and sex of the respondents. Participants were also afforded the opportunity to include any additional comments they considered relevant to the study. Upon receipt of complete surveys, data for each respondent were tallied and entered into a database. Respondent's scores were combined with other respondents' belonging to the same group. A oneway analysis of variance procedure using matrix input was employed to assess group differences on each of the subscales of the GES. The Statistical Package for the Social Sciences (SPSS) was used to conduct the analyses. SPSS provides strategies for users to conduct powerful inferential analyses without the use of raw data elements. Using the matrix-input functionality in SPSS, ANOVA's were conducted by supplying means, sample sizes, pooled variance estimates, and degrees of freedom for each group. Because group comparisons were made across multiple dependant measures the bonferroni procedure was used to correct for potential inflation of type one error.
PARTICIPANT DEMOGRAPHICSA total of 31 separate groups returned completed surveys (128 respondents in all). Group sizes ranged from 3 to 10 members with a mean group size of 4. Among the respondents, 30.5% of the group members are female and 67.2% are male, with gender being unreported in 2.3% of the cases. The mean age of the respondents was 22.9 years, with males averaging 21.6 years and females averaging slightly higher at 23.6 years of age.
STATISTICAL RESULTSSignificant differences between face to face and MUD groups were found on 6 of 10 GES subscales. A comparison of the means and standard deviations between the MUD and GES Social Recreational groups is available in Table 2.
TABLE 2: COMPARISON OF MEANS & STANDARD DEVIATIONS FOR
MUD Groups Social-Recreational Groups
Specifically, Mud groups scored significantly higher than face to face groups in the areas of Expressiveness, F (1,91) = 19.44, p < .01, Innovation, F(1,91) = 37.23, p < .01, Independence, F(1,91) = 10.08, p < .03, and Self-Discovery, F(1,91) = 21.81, p < .01. In addition, MUD groups scored significantly lower than face to face groups in terms of Order and Organization, F(1,91) = 11.59, p < .01, and Leader Control, F(1,91) = 26.8, p < .01.
EXPRESSIVENESSMUD groups allow for a greater degree of self expression than do face to face groups. These findings seem to be consistent with the research pertaining to CMC in general suggesting that people tend to be less inhibited on-line than in face to face situations. Though the potential explanation for this, as mentioned previously, is couched in the notion that lack of social context cues contributes to disinhibition, it may also be that such expression can plausibly be attributed to the membership regulation and selection processes among these groups. Due to the degree of control that computer mediated environments offer (one can always disconnect with the push of a button) it should be much easier for group members to relinquish membership and go elsewhere should they find the group to their disliking. Also, many group members (leaders as well as common members) have the ability to oust members that do not "fit in". Numerous respondents made statements in support of this contention:
...if people disagree too much they soon feel they can't work properly and so make appropriate arrangements. (Respondent #38)
INNOVATIONMUD groups evidence a significantly higher degree of diversity and change in their internal functions and activities than do face to face groups (Innovation). It could be argued that many of the hidden agendas found in face to face groups that lead to group dysfunction are largely absent from virtual groups due to the evolutionary dynamics of self selection. That is, core group members are more willing to accept change in their internal functions and activities because such shifts are perceived as posing no real threat to individual member roles. In addition, with a computer dependent environment it would seem essential that new challenges be created and new spaces built within the virtual framework of the existing simulation. Otherwise, members would quickly become bored with the place after having explored all that it has to offer (outside of the social interactional aspects). Indeed, Richard Bartle (1996) claimed that without a good balance of builders (Innovators) and other players, MUD populations tend to drop rapidly.
INDEPENDENCEThe degree to which MUD groups encourage independent action among their members is significantly higher than for face to face groups. This is understandable in light of the above since any restrictions upon individuals that are deemed too confining by individual members would again most likely result in changes in group membership. Indeed, one could argue that one of the most attractive aspects of virtual reality is its malleability, both in terms of the environment and social structure. Thus, people may be drawn to virtual environments because they offer a greater degree of freedom than the more restrictive environments found in the "real" world. While this is likely to be the case, there does appear to be some degree of limitation of independence within MUD groups as evidenced by the respondents' commentary:
You do as you please, and if your clan, or the admins do not like it, they will deal with it as they see fit. (Respondent #19)
SELF DISCOVERYMUD groups were found to be significantly more open and encouraging than face to face groups in matters concerning personal problems. With such high levels of independent action and self expression, this result is not surprising. In addition, as pointed out previously, the degree of anonymity that virtual groups offer probably promotes a sense of invulnerability in terms of the repercussions such self disclosure might have upon their "real" lives. Much as the environment of a therapist's office is conducive to self disclosure due to its sanctioned status as "separate from" the rest of the world, the virtual world may encourage disinhibition for this same reason. The combined expectations of group members stemming from the degree of invulnerability perceived in MUDs probably fosters an atmosphere, and perhaps even the expectation, that group members should be permitted a higher degree of expression concerning personal matters.
ORDER AND ORGANIZATIONThe formality and structure of MUD groups and the explicitness of their rules and sanctions are significantly lower than for face to face groups. Since MUDs are highly self selective, the need for rigid rules and sanctions is probably lower since those dissatisfied with the inner workings of the group can easily go elsewhere. Additionally, rules and sanctions need not be made very explicit since tacit rules function to weed out non-responsive members. According to the respondents' commentary, enforcement of the few rules that exist is only occasionally conducted. Most members simply agree to comply by their sustained membership. Leader Control There are significantly lower levels of leader control among MUD groups than in face to face groups. Given the highly voluntary nature of group membership, a leader perceived as exceedingly controlling would soon find that they no longer had a group to lead. Likewise, the need for high levels of leadership control is diminished in such environments, since members seemingly have a clear conception of the group function and their places within it. Additionally, control does not seem to be concentrated in the hands of the highest ranking leaders, but is rather dispersed among the membership according to levels attained in the overall hierarchy. Thus, members can act to enforce rules and sanctions for lower level members without the need to appeal to the authority of higher ranking individuals. This chain of voluntary responsibility and control tends to make it appear as though no one in particular is in control. Indeed, many respondents claimed that there were no leaders as such within their groups and that all members are at times leaders as well as followers:
To some extent, the Elders are just figureheads, because even when we do make/change policies, we seek the approval of the majority of Clan members. (Respondent #9)
CONCLUSIONS AND IMPLICATIONS
Although these findings are the result of only one study, they do indicate the possibility that MUDs may foster a highly characteristic social climate that can be said to be consistent to a degree with other forms of CMC. This bodes well for the potential to synthesize findings from research conducted in synchronous and asynchronous areas of on-line interaction and lends itself to the formation of a more coherent body of CMC research in general. The current authors believe that the customization of real-world survey instruments (including standardized psycho-social measurements) to better suit and capture the unique nature of virtual environments may be a starting point for further exploration of the psychosocial characteristics of interpersonal communication on-line. A question arises concerning how much such instruments might be missing due to their being so firmly grounded in research, practices and terminology based upon group behavior within face to face settings. For example, with the ability for virtual group members to participate simultaneously in several groups at once via channels that are not always obvious to observers (e.g. chat channels, pages, etc.), how might constructs such as group cohesion, participation, and presence qualitatively differ for such groups? We're not suggesting that standard research practices and well grounded constructs be abandoned or that all such problem areas can be easily overcome, but that preliminary exploration needs to be considered in light of the differences and challenges that on-line groups pose to the study of human behavior and existing methodologies.
Findings from the present investigation suggesting that online group characteristics differ from face to face groups highlight the need for further inquiry with regard to MUDs and other related CMC environments. Though this study indicates some more specific areas in which face to face and MUD groups may differ than was uncovered in previous research, the causal factors that may contribute to such differences are yet to be determined. Additionally, although just a preliminary finding, the combination of GES scores and respondent feedback suggests that MUD groups exhibit a number of phase four group developmental typological traits such as shared leadership and task delegation, an open and participatory communication and decision making structure, encouragement of innovation, acceptance and integration of subgroup formations/coalitions, and matching between role assignments and member abilities, which are all indications of well functioning and effective work groups (Wheelan, 1994). This implies that MUD environments may be highly suited to various work group related activities, an issue that should gain increasing attention as more people begin to use the Internet to telecommute to work and for other group related endeavors such as educational and mutual support groups.
GES ConstructsCohesion: The degree of group members' involvement in and commitment to the group and the concern and friendship
they show for one another.
Support: The perceived amount of help, concern, and friendship that group leaders show.
Expressiveness: The degree to which freedom of action and expression of feelings are encouraged.
Independence: The degree to which the group will encourage and tolerate independent action and expression of feelings
Task Orientation: The degree of emphasis on completing concrete practical tasks and on decision making and training.
Self-Discovery: The degree to which groups encourage members' discussions of personal problems.
Anger/Aggression: The degree to which open expression of anger and disagreement is displayed.
Order/Organization: The formality and structure of the group and the explicitness of rules and sanctions.
Leader Control: The extent to which the leaders direct the groups, make decisions and enforce rules.
Innovation: The degree to which the groups promote diversity and change in their own functions and activities.