It appears to be generally accepted that individuals are less inhibited when interacting in MU*s than they might be in real life circumstances (Reid, 1991; Bruckman, 1992; Carlstrom, 1992; Rosenberg, 1992; Curtis, 1993; Roush, 1993; Elmer-Dewitt, 1993; Penkoff, 1994; Snell, 1994), and there is evidence to suggest that such disinhibition is characteristic of CMC in general (Kiesler, Siegal & McGuire, 1984; Turkle, 1984; Serpentelli, 1993; Odegard, 1993; Bordia, 1994; Brown, 1995). 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, Siegal & McGuire, 1984; Reid, 1991; Rosenberg, 1992). However, other researchers have attributed this phenomenon to such causes as real life anonymity (Cutler, 1994; Barrett & Wallace, 1994; Escobar, 1994) and immunity from physical reprisal (Serpentelli, 1993, Rheingold, 1993). All of these conclusions are based upon the concept of limited bandwidth as contributing to such behavior. However, evidence is beginning to accrue from observations of experimental high bandwidth systems suggesting that even with full audio/visual interface (which one would assume to mitigate the dramaturgical shortcomings and degree of anonymity) disinhibition in CMC remains quite high (Marshall, 1993).
It should be noted however, that disinhibition is not necessarily always negative in these settings (e.g. flaming, player killing, spamming, etc..). There appears to be just as much "positive" disinhibition (e.g. newbie mentoring, tinysex, whuggles, etc..). However, Reid argues that such disinhibition does not necessarily indicate some sort of "flaw" in MU* culture or that MU* environments are "lawless and socially chaotic":
...being disinhibited is not the same as being uninhibited. MUD players experience a lowering of social inhibitions; they do not experience the annihilation of them. The social environments found on MUDs are not chaotic, or even anarchic. There is indeed no moment on a MUD in which players are not enmeshed within a web of social rules and expectations. ....However, these webs of meaning and control are not as immediately apparent on MUDs as they might be in actual life. Substitutes for the contexts and atmospheres that we rely on to regulate and define our behavior may have been developed on Muds, but it takes time for players to learn to recognize and to adopt these substitutes (Reid, 1994).
Another finding, related to the issue of disinhibition, is that men and women appear to behave differently from one another within MU* environments. In a study conducted by Diane Penkoff of Purdue University it was demonstrated that both the types and frequency of graphic accents (emoticons & articons) correlated with player gender, with men being less prone to utilize such devices than women (Penkoff, 1994). However, it did not appear that there was a significant difference in the rate or degree of flaming. She proposes that this tendency for women to be more challenging in CMC groups as compared to face-to-face interactions may be due to women feeling more at ease in CMC because of the relative anonymity, or that it may be that the women engaged in CMC might already be involved in male-dominated endeavors such as high tech. organizations or academia (Penkoff, 1994). It should be noted that Penkoff's study, though informative, was conducted using data from newsgroups which are asynchronous forms of CMC. Since MU* interaction is essentially synchronous, these findings have to be considered with this distinction in mind.
Lynn Cherney's study of MU* interaction, when considered in conjunction with Penkoff's findings, seems to support the notion that indeed gender differences do persist across both synchronous and asynchronous forms of CMC. Cherney found that women are 4 times as likely to initiate whuggles than are men, that men tend to use more physically violent imagery, and that women are overall more affectionate towards other characters (Cherney, 1994). However, these findings appear to partially contradict Penkoff's in that the women in Cherney's study did demonstrate a higher degree of affection.
One factor that complicates gender related research in such environments is that, due to the anonymity of the medium, one can never be certain of the sex of a respondent. Researcher Amy Bruckman has studied the phenomenon of gender swapping in MU*s - wherein individuals adopt an identity of the opposite sex. She has found that this is a fairly common practice, though it is not necessarily indicative of real-life homosexuality or transvestism (Bruckman, 1993). Gender switching has been documented by other researchers as well (Curtis, 1991; Serpentelli, 1992; Rheingold, 1993; Kennedy, 1994; Penkoff, 1994; Reid, 1994). An interesting finding is that the vast majority of those adopting opposite genders seem to be men; women rarely report such behavior (Serpentelli, 1992). Of the women that did admit to interacting as opposite gender characters, many of them expressed shock at what was perceived as an abrupt drop in the degree and frequency of helpful gestures and general civility offered by other "male" characters. Conversely, men consistently reported that their female characters received more assistance and generally prospered better than did their alternate male characters (Serpentelli, 1992; Bruckman, 1993; Rheingold, 1993).
It should be noted however, that information gathered through on-line interviews (such as Sepentelli's) may not be wholly reliable due to the very fact that one can never be absolutely certain of the true gender of the respondent. Indeed, any self-reports stemming from such an inherently fantasy laden construct must be accepted with a degree of skepticism. Again, exact demographics for the Internet are hard to come by, but it is estimated that 75% of the current cyberspace population is male (Cassidy, 1995) and Bruckman's survey of the MediaMOO revealed that 84% of the respondents were male, 16% were female (Bruckman, 1992).
Sociologist Jay Chaskes of Rowan college has been conducting research pertaining to the "reinvention of self" in multi-user dimensions. Though his work is still in the pilot stage, he has found that it is common practice for individuals to assume several identities, which are employed for different purposes throughout various virtual worlds. As a disembodied entity, one may "appear" as whatever one wishes (Chaskes, 1995). This phenomenon of adopting different character names and personas for use in various settings has also been documented by other researchers (Curtis, 1991; Bruckman, 1992; Dibbell, 1993; Reid, 1994; Rheingold, 1994).
Another related topic, also explored by Bruckman, is that of MUD addiction - wherein people so prefer their on-line lives that they actually begin to abandon their "real" lives, spending up to 18 hours per day living in virtual reality (Kelly & Rheingold, 1993; Bruckman, 1994). There have even been reports of Internet related deaths, such as cardiac arrests resulting from sleep deprivation and lack of personal maintenance or suicides due to Net related stress (Elmer-Dewitt,1993; Rheingold, 1994). Such obsessive behavior has become quite a problem on some college campuses where computer facility administrators have had to literally "pull the plug" on some addicted students (Cassidy, 1995). In fact, MUDing has been outlawed on the continent of Australia because of the excessive satellite link-up costs associated with such activities (Rheingold, 1994). Little is known about this new form of addiction, but as mentioned previously a preliminary draft of the DSM-V includes a section delineating some of these types of "Cyber Disorders" (Zenhausen, 1995).
People may be friendlier towards one another in cyberspace. Mitchel Resnick et al of MIT claims that, through their interviews with various participants of the MediaMoo project, he and his colleagues 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." (Resnick, 1994). Howard Rheingold reports that "Your chances of making friends (in cyberspace) are magnified by orders of magnitude over the old methods of finding a peer group." (Rheingold, 1994). In concurrence, a recent survey of public school teachers who utilize the Internet as a classroom resource revealed that, except for its function as an information archive, "the rewarding social atmosphere" was viewed as the most outstanding feature of Net access (Kinslow & Sempsey, 1995). Likewise, studies conducted in educational settings where INTERNET access is available have revealed that children are less critical of differences via CMC than in face to face groups (Sauer, 1994; Simms & Simms, 1994). Additionally, a study by Masinter & Ostram on the effectiveness of multi-user information retrieval systems revealed that peer collaboration is higher in on-line information access than it is in "physical" libraries (Masinter & Ostram, 1993).
Yet a study of electronic bulletin boards conducted at Temple University revealed that Internet groups seem to form, interact and go through distinct stages just as do "real-life" groups (Bordia, 1994), implying that virtual groups may only seem to be more cooperative, friendly, etc.. Likewise, Kiesler et al, in one of the first studies ever conducted on Net related psychological phenomena, concluded that not only did computer mediated communication impede a group's ability to reach a consensus, but that it was practically devoid of the types of nonverbal social cues necessary to sustain the type of social climate Resnick claims is inherent in the medium (Kiesler, 1984). Another researcher, Sara Brown also of Temple University, has come to mixed conclusions regarding CMC. She reports that groups were slower to develop relationships and commit to a common purpose via CMC than in face to face groups. The groups shared general information more freely, however they avoided the sharing of sensitive and personal information, and they found it more difficult to resolve conflicts (Brown, 1995). Finally, Dunlop and Kling have determined that the reduction in computerized BBS's of social cues that support "chat modes" sometimes foster special senses of personal trust and intimacy. That is, although they are less personal, they can also reduce social distance (Dunlop & Kling, 1991).
One of the most evident aspects of MU* related research is that there seems to be so little of it. In recent years there have been a number of studies conducted on CMC, most notably on asynchronous forms, but it is unclear as to what extent such research can be extrapolated to synchronous CMC (Metz, 1994). Furthermore, much of the CMC research has been conducted via relatively autonomous methodologies and with little emphasis on reconciling the various (and sometimes conflicting) results from previous studies. In an attempt to address this issue, Metz has advocated that future studies of CMC be conducted from within a CMC context rather than treating CMC as part of a broader set of communication (Metz, 1994). I concur with Metz in that if one is to begin to grasp the gestalt of CMC, then there needs to be established a cohesive theoretical framework. Though I am uncertain of what such a framework might consist.
The viability of studying MU*s from a CMC theoretical perspective, should one develop, may prove to be limited due to the extent to which MU* constructs can differ. One might be able to construe some axiomatic principles by which all such systems operate, but I doubt that these would be of any particular utility. As an analogy, most anthropologists can claim with certainty that all human cultures (in the real world) have one form or another of marriage; but this rule sheds little light upon the intricacies of any particular culture.
Though common themes have begun to emerge from the body of research pertaining to MU*s, many issues seem to have received only perfunctory consideration and some of the findings seem to be conflictual. For example, the issue of disinhibition - although fairly universally accepted - needs to be explored in greater detail. What is the underlying cause of this phenomenon? The major hypotheses, as presented earlier, all appear to be plausible, but to what extent do the factors of anonymity, dramaturgical weakness and fearlessness contribute to disinhibition; why do they have such an effect (if this is the case), and what other effects can be attributed to them? If anonymity and dramaturgical weakness are causal factors then should not disinhibition decrease with the extension of bandwidth? This does not appear to be so (Marshal, 1993). Also, if MU* environments are demographically skewed towards predominantly college aged men (Bruckman, 1992), than could this not also be construed as a causal factor in such behavior? As stated at the beginning of this paper, one needs to consider MU*s in light of the fact that users are situated in both the "real world" and the "virtual world", especially since MU populations appear to be comprised of an unrepresentative subset of the real-world population.
With respect to the issue of gender differences in MU* behavior, again it seems that too little data exists to even claim with certainty that such phenomena are more than mere anomalies. I would imagine that one way to establish the validity of such claims would be to form an all female experimental group, give them exclusive access to a MU*, and then compare the findings to another all male group occupying an identical environment. This might serve to shed some light on the causal factors contributing to MU* related disinhibition as well. However, even though computer mediated environments seem tailor made for experimental observations, there seems to be a general bias on the part of MU* researchers towards non-experimental research designs; all of the MU* related research I could find was ethnographic and narrative in nature.
The two phenomena of gender swapping and multiple identities seem to be closely related. In such fantasy based - socially constructed environments one should expect that individuals will avail themselves of the opportunity to "try on" different persona, particularly in light of the fact that there seems to be little chance that such behavior will have any "real world" repercussions. Yet, when one considers the associated phenomenon of MU* addiction, it appears that there can indeed be real world consequences to on-line interaction, just as it has been argued that there are real-world consequences to televised reality (Bagdikian, 1983). With this interplay of actual and virtual reality in mind, finding answers to such questions as: 'Why do individuals become so attached to their virtual selves?', 'Why do women seem less prone to swap gender?', or 'Why do different virtual environments give rise to different personal manifestations?' become problematic, since it seems unclear where the source of such answers may lie (i.e. in the real world, the virtual environment, or in some combination).
Finally, the issue of whether people are friendlier in MU*s seems disputable. True, Resnick does report that his interviewees claimed a friendlier atmosphere exists in MU*s than in face to face groups, but this is only one study. Dunlop and Kling's findings seem to suggest that people are friendlier because there is more distance between them (Dunlop & Kling, 1991). The remainder of studies that support this contention were conducted in CMC areas outside of MU*s, and as such are only tenuously acceptable. Besides, I discovered just as many CMC studies which contradict or problematize these findings (Kiesler, 1984; Bordia, 1994; Brown, 1995). On the other hand, as was pointed out by Reid, it may not be fair to judge the quality of MU* interactions without first attempting to become familiar with the overall social dynamics of a given situation; what may seem unfriendly or chaotic to the uninitiated may be more a function misinterpretation than of intent (Reid, 1994). The perception of friendliness may also be confounded by the temperament of the individuals under study as it relates to computer environments in general. For example, it seems plausible that those who are somewhat technophobic might find any form of CMC to be unfriendly, whereas individuals that embrace technology might perceive the same environment as warm and inviting. In terms of specific research issues raised by this review, one could select any of the features that I've highlighted and pursue it as an isolated topic. Its all very new territory and even replications of the existing studies would be beneficial. However, such an approach does not appeal to me personally because at this juncture I feel that a better grasp of the gestalt of MU* phenomena is in order.
It does seem clear that one strong implication of this review is that more research needs to be done concerning synchronous forms of CMC, of which MU*s are a major constituent, and that future CMC studies should be conducted with this synchronous/asynchronous distinction in mind. Furthermore, steps should be taken to establish some meta-theoretical framework through which researchers from different disciplines might be able to better benefit and learn from each other's work in this area. This implies that a hybrid discipline - Cyberology - may be in order.