The Social Identity Model of Deindividuation Effects (SIDE) proposes that depersonalization of self and others is responsible for the effects of visual anonymity on group behavior. The authors investigated these mediating processes by assessing the effects of group-based self-categorization and stereotyping of others on group attraction within visually anonymous or video-identifiable groups communicating via computer.
Structural equation modeling showed that visual anonymity increased group-based self-categorization, which directly increased attraction to the group and indirectly increased group attraction by enhancing group-based stereotyping of others. Visual anonymity had no effect on self- categorization in terms of a wider social category (nationality).
Predictions derived from alternative perspectives that visual anonymity decreases group attraction by increasing impersonal task focus or by attenuating evaluation concerns were not supported.
Anonymity has been implicated in research on deindividuation, social facilitation, and brainstorming, decision-making, group size and crowd behavior. Yet its conceptual status and the processes by which anonymity achieves its effects are still far from clear. This is nowhere more true than with regard to the deindividuation literature. Deindividuation has traditionally been defined as a state of reduced self awareness, or even “loss” of self, often associated with immersion in the group or crowd (Festinger, Pepitone & Newcomb, 1952; Diener, 1980; Prentice-Dunn & Rogers, 1989; Zimbardo, 1969). Early formulations saw anonymity as a central antecedent or input to this state (e.g., Festinger et al., 1952; Zimbardo, 1969).
Contemporary accounts see anonymity as somewhat less central to deindividuation in theoretical terms, focusing on the arousal and external focus of attention associated with group immersion (Prentice-Dunn & Rogers, 1982, 1989). Nevertheless, a rather unsatisfactory feature of virtually all research on this topic is that the underlying processes remain unclear (Diener, 1979, 1980; Postmes & Spears, 1998). Mediation analyses that could provide the necessary evidence that changes in self-awareness mediate deindividuation effects has been largely absent or only partially successful in confirming hypothesized effects (Matheson & Zanna, 1988; Prentice-Dunn & Rogers, 1982).
The explanatory role of anonymity has recently returned to the foreground in research proposing an alternative mechanism for explaining deindividuation effects in terms of social identity processes (Lea & Spears, 1991; Postmes & Spears, 1998; Reicher, Spears & Postmes, 1995; Spears & Lea, 1992, 1994). Here too however, evidence that social identity processes mediate the effects of anonymity on groups has so far been lacking. The main purpose of the present research is to address this shortcoming, and to arbitrate between the different theoretical approaches to deindividuation in the process. Specifically, we evaluate the social identity explanation of deindividuation effects, that argues that anonymity within a salient group promotes categorization of self and others in terms of the group, thereby enhancing group behavior (Reicher et al., 1995; Spears & Lea, 1992, 1994). The second, related purpose, is to test between competing predictions about the negative or positive effects of anonymity on group attraction, derived from traditional group cohesiveness and social identity approaches respectively (see Hogg, 1993 for a comparative review). In both cases we assess the alternative processes proposed to mediate the effects of anonymity in the group.
Anonymity, deindividuation theory and SIDE
Visual anonymity within groups has generally been considered to have negative consequences, resulting, for example, in disinhibited and aggressive behavior (e.g., Singer, Brush & Lublin, 1965; Zimbardo, 1969). Processes held to be responsible for these effects include reduced objective self-awareness (Diener, 1979), reduced private self-awareness (Prentice-Dunn & Rogers, 1982, 1989), and reduced public self-awareness, or evaluation concern (Festinger, Pepitone & Newcomb, 1952; Prentice-Dunn & Rogers, 1982, 1989).
Anonymity and deindividuation have received additional attention in recent years from research into the effects of new communication technologies such as computer-mediated communication (CMC) in which interactants can be physically isolated and hence visually anonymous. Here, anonymity has been argued to increase disinhibited, hostile behavior, and extreme decision-making, and to reduce attraction within computer-mediated groups. These effects are held to result from the relative lack of interpersonal cues resulting in a state of deindividuation, reduced evaluation concern, and a more impersonal and task-oriented attentional focus (Kiesler, Siegel, & McGuire, 1984; Siegel, Dubrovsky, Kiesler & McGuire, 1986; Sproull & Kiesler, 1991; Walther, 1992, 1997).
Whether anonymity produces generic effects within groups is open to question, however. The outcomes predicted by deindividuation theory have not always been observed; decreased aggression, and even increased affection have sometimes been expressed under anonymous conditions (e.g., Gergen, Gergen, & Barton, 1973; Johnson & Downing, 1979). Likewise, in some CMC studies, greater self-awareness, more cautious decision-making, increased attraction, and lack of disinhibition have all been observed under conditions of anonymity (Hiltz, Turoff and Johnson, 1989; Lea & Spears, 1991, 1992 Study 2; Matheson & Zanna, 1988). In a recent study, Coleman, Paternite and Sherman (1999) found that anonymous CMC groups reported more submergence in the group, and reduced perceptions of individuality, in accordance with a state of deindividuation. However, there was no evidence of a corresponding increase in the negativity of group discussions.
One attempt to reconcile these conflicting observations has departed from the attempt to define generic anonymity effects and instead considers the effects of anonymity in relation to the specific social group context that is salient. The Social Identity Model of Deindividuation Effects (SIDE) aims to explain cognitive and strategic effects of visibility and anonymity in both intragroup and intergroup contexts within a single theoretical framework (for overviews, see Reicher et al., 1995; Spears & Lea, 1994; Spears, Postmes & Lea, in press). The SIDE model developed out of social identity theory (Tajfel & Turner, 1985) and self-categorization theory (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987), and further specifies the effects of situational factors on the operation of processes proposed by these theories. Underlying this approach is a conceptualization of self-construal as flexible and situation-specific. It comprises a range of self-categories that define people in certain contexts as unique individuals—their personal identity; but in other contexts in terms of membership of specific social groups (such as an activity group) and wider social categories (such as nationality or gender)—their social identities. A person’s behavior in any situation can be placed along a continuum ranging from entirely personal (conforming to personal standards) to entirely group-based (conforming to salient group norms and standards). According to SIDE, visual anonymity reduces the communication of interpersonal cues within the group, allowing certain social group and category information which is less dependent on visual cues for its communication to become more salient. This has the effect of shifting perceptions of self and others from the personal to the group level, thus encouraging behavior that is normative for the salient group. In contrast to deindividuation theories, SIDE proposes that anonymity promotes a shift in the kind of self-awareness from the personal to the group self, rather than a loss of self-awareness. Similarly, perceptions of others shift from being primarily interpersonal to being group-based perceptions (stereotyping), under anonymity, rather than there being a loss of attention to others.
A number of studies have been carried out that support this general prediction. More normative behavior has been observed in face-to-face groups, in which the visual anonymity of group members was achieved by means of masks and overalls, as in classic deindividuation studies (Reicher, 1984), and also in computer-mediated groups in which interactants were physically isolated and hence visually anonymous (Lea & Spears, 1991; Postmes, Spears, & Lea, 1998; Spears, Lea & Postmes, in press). Moreover, where it was measured in these studies, little evidence was found of variations in amount of (private) self-awareness.
Anonymity and group attraction: Competing perspectives on group cohesiveness
We turn now to consider how anonymity affects group attraction according to the different theoretical perspectives on group cohesiveness. Most traditional and contemporary formulations of group cohesiveness argue that attraction to the group is based upon interdependent interaction towards the satisfaction of mutual goals, and/or the perception of interpersonal similarities among individuals, both of which produce interpersonal attraction. According to this interdependence approach, group attraction can be viewed as the aggregate of interpersonal bonds of attraction or interdependence (e.g., Cartwright, 1968; Lott & Lott, 1965; Mudrack, 1989). In contrast, the social identity approach regards group attraction as a conceptually distinct process involving depersonalized perceptions of self and others in terms of a common group categorization (Hogg, 1993; Turner, 1982; Turner et al., 1987).
Depersonalization does not necessarily mean that interactions with others become impersonal, or task-focused, however. Rather it simply means that perception and behavior become stereotyped in terms of the salient group. We argue that anonymity encourages depersonalized perceptions of self to occur because it reduces the (inter)personal basis for social comparison, self-awareness and self-presentation. As a result the self tends to be perceived and presented less as a unique individual and more in terms of its similarity to the perceived prototypical attributes of the salient social group (i.e., self-stereotyping). The anonymity of others means that they also tend to be perceived as interchangeable representatives of the group rather than as unique individuals (Turner et al., 1987). Under anonymity, perceptions of individual differences among interactants are reduced, leading to less individuated impressions and a less interpersonal basis for interaction (Coleman et al., 1999; Reicher, Spears & Postmes, 1995; Spears & Lea, 1992). In short, SIDE proposes that depersonalized perceptions of self and others increase attraction towards group members and that this process is stimulated by the dearth of individuating cues in visually anonymous interactions (Lea & Spears, 1995).
In contrast to this social identity route to attraction, a number of other perspectives propose the opposite, namely that anonymity decreases attraction to group members. Once again, recent research into computer-mediated groups has focused attention on this issue. These accounts implicitly invoke a traditional interdependence explanation of group attraction that focuses on the development and maintenance of interpersonal bonds as the basis of group cohesion and attraction. They imply that conditions that prevent or retard the formation of such bonds, such as visual anonymity, will reduce attraction to the group. Because anonymity removes interpersonal cues, it decreases attention to others, reduces concerns about being positively evaluated by others, and creates an impersonal task-oriented focus for group interaction. This impersonal focus reduces politeness and tolerance, promotes conflict and hostile behavior and impedes the development of attraction and interpersonal relations (Jessup, Connolly & Tanisk, 1990; Kiesler et al., 1984; Walther, 1992). A related suggestion is that impersonal task focus may be a general norm affecting temporary or short term groups (i.e., as in most experimental settings) that is amplified by anonymity (Walther, 1997).
However, while reduced attraction has been observed in anonymous computer groups (Kiesler, Zubrow, Moses & Geller, 1985) other studies have found no effect (Coleman et al., 1999; Walther & Burgoon, 1992), or even increased attraction (Lea & Spears, 1992, study 2; Walther, 1995). Furthermore, the evidence for the role of evaluation concern or impersonal task focus is weak. Kiesler et al. (1985) found less responsiveness to others and greater adherence to the task accompanied by significantly less positive interpersonal evaluations in anonymous CMC, compared to face-to-face interaction. However, the crucial mediation analysis that would establish task focus and reduced attention to others as causal processes was not reported. Walther (1997) observed lowered attraction in short-term groups interacting in anonymous CMC under high group salience conditions than within groups interacting for longer or under individual salience conditions; however, there was no corresponding effect on task-orientation. Moreover, the high group salience conditions were significantly less task-oriented than the low group salience conditions, contrary to the prediction that reduced attraction responses were guided by a generic impersonal/task-focus norm. Other studies have also found significantly less impersonal task focus in anonymous CMC interactions (Walther, 1995; Walther & Burgoon, 1992).
Overview and hypotheses
To summarize, the SIDE model makes specific predictions regarding the effects of anonymity and the underlying processes that contrast with traditional group theories. These different perspectives were tested in a study in which physically isolated participants discussed a number of issues in small groups using text-based computer conferencing under visually anonymous or video-mediated (i.e., identifiable) conditions. On the basis of the SIDE model, we predicted that visual anonymity would lead to greater group attraction. In terms of the mediating process, we predicted that anonymity would increase the tendency for interactants to categorize the self in terms of the group, which in turn should increase group attraction. Second, self-categorization should enhance the tendency to see others in group terms (stereotyping), which should independently increase group attraction. In sum, the effect of anonymity on attraction may follow two routes, both of which crucially depend on anonymity enhancing self-categorization. The predicted shifts towards group-based perceptions of self and others as a consequence of anonymity are not predicted by deindividuation theory, which in addition, would generally predict a negative relation between anonymity and group attraction.
A number of alternative mediation paths were hypothesized from consideration of traditional deindividuation theory and the competing interdependence approach to group attraction described above. Both predict that visual anonymity should decrease group attraction by reducing evaluation concern and by increasing perceptions of an impersonal task focus, concomitant with a loss of attention to self and others (Festinger et al., 1952; Kiesler et al., 1984; Kiesler et al., 1985; Kiesler & Sproull, 1992; Martens & Lander, 1972; Prentice-Dunn & Rogers, 1989; Walther, 1992).
We also explored task focus and evaluation concern in relation to the SIDE model, and specifically their effects in relation to the central self-categorization variable. There were several possibilities here. If visual anonymity increases task focus because it is perceived to be normative for short-term groups (Walther, 1997) then the effect should be mediated by the extent to which participants see themselves as members of a group. Self-categorization might increase participants’ evaluation concern by increasing their focus on the group-based aspects of self-identity. On the other hand, self-categorization might decrease evaluation concern—because it reduces their focus on individual aspects of the self and encourages feelings of acceptance by the group.
Finally, another purpose of the present study was to assess effects of anonymity on self-categorization in terms of a wider social category (such as nationality) that was made salient during interaction, independently from anonymity effects on local group self-categorization. The effects of local group identification and social category identification on normative behavior have been confounded in previous research on SIDE in which the salience of both have been simultaneously raised (Lea & Spears, 1991; Postmes et al., 1998). Visual anonymity may enhance perceptions of self in terms of a wider social category. However, this effect may be small compared to the effect on local group self-categorization because participants have prior knowledge of a pre-existing social category and may therefore already hold a relatively clear and stable definition of the category and their relation to it. By contrast, the local group identity is constructed more immediately in the interaction situation. Supporting this, a recent meta-analysis of 60 independent studies of deindividuation found that group behaviour was explained more adequately by reference to situation-specific rather than general social norms (Postmes & Spears, 1998). Assuming that self-categorization was the underlying process responsible for these effects, the meta-analysis also suggests that local group self-categorization should have a stronger influence on group attraction than self-categorization with a wider social category. It was therefore predicted that there would be no additional independent effect of visual anonymity on attraction that was mediated by self-categorization with a salient social category (nationality).
Another possibility is that visual anonymity increases self-categorization with a social category only under conditions where social category salience is already high. For example, in the case of nationality, visual anonymity might become more influential when interacting with others who have a different nationality from oneself. Under these conditions, differences in nationality among interactants should increase self-perception and self-presentation in line with one’s own nationality while visual anonymity should reduce the (inter)personal basis for social comparison, self-awareness and self-presentation.