This paper argues that to achieve social presence in a distributed environment, it is not necessary to emulate face-to-face conditions of increased cues to the interpersonal. Rather, it is argued, that a sense of belongingness to the group, or perceptual immersion in the group, can be realised through the creation of a shared social identity between group members. From this perspective, social presence is a function of the cognitive representation of the group by group members and not the interpersonal bonds between group members. Furthermore, specific design features and characteristics of the distributed learning environment can be utilised to achieve and maintain this shared group identity. This approach, encapsulated by the SIDE model, is discussed and supported by two case studies of distributed students, each consisting of ten groups, collaborating through computer-mediated communication for a period of five weeks on group projects.
The concept of presence
Presence, although broadly deﬁned as the sense of ‘being there’ in a mediated environment, can be conceptualized in a number of ways. Indeed, presence can be described as a multidimensional concept, with Lombard and Ditton (1997) identifying a set of at least six diﬀerent conceptualizations found in the literature (e.g. presence as realism, immersion, transportation etc.). These multidimensional conceptualizations, however, can be grouped into two broad categories—social and physical presence (Ijsselstein et al. 2000). The physical category refers to the sense of being physically located somewhere and implies that the medium appears to be invisible; whereas the social category refers to being and communicating with someone, with the implication that the medium appears to be transformed into a social entity. Importantly, the antecedents necessary to establish these two forms of presence are not necessarily the same aspects of communication. It is possible that one can experience physical presence without a corresponding level of social presence, and conversely one can experience social presence in the absence of physical presence. It is suggested that although the distinction is often made between physical and social presence, they are nonetheless usually treated as synonymous, or at least closely inter- dependent. While there are undoubtedly commonalities between social and physical presence, it is argued that their division is a useful one in order to best decide on the appropriate form of presence that can inform system design and result in the desired outcomes from the use of the mediated technology. Moreover, it is suggested that the partitioning of the physical from the social enables a re- conceptualization of social presence that has signiﬁcant consequences for the representation of groups in computer- mediated or virtual environments.
One of the earliest and most inﬂuential theories concerning the issue of presence was the social presence model developed by the Communication Studies Group at University College, London (Short et al. 1976). Although the model was not applied directly to computer-mediated communication, but to telephone, audio and video links, it has had a signiﬁcant impact on many current approaches to presence. The model suggests that the critical factor in a communication medium is the ‘social presence’, deﬁned as a subjective quality of the medium. Social presence is conceived as the degree of interpersonal contact that a medium allows and is closely related to issues such as ‘intimacy’ and ‘immediacy’. In a series of initial studies, diﬀerent communication media were ranked in respect of their ‘social presence’. Face to face communication ranked highest, with a decrease in ‘presence’ of television, telephone and lastly business letters. Text only computer- mediated communication would occupy a relatively low ranking in this context, somewhere between business letters and the telephone (Spears and Lea 1992), whereas virtual reality environments would occupy a relatively high ranking.
Social presence has been typically characterized in research studies by dimensions such as cold – warm, unsociable – sociable and impersonal – personal (Short et al. 1976). Furthermore, in early studies, the impersonal – personal dimension was positively correlated with items designed to tap social presence. It is therefore unsurprising that media low in social presence were seen as less ‘social’. Moreover, the model predicts that media high in social presence are more conducive to social inﬂuence (the process by which attitudes and behaviours are shaped by others), whereas those low in social presence correspond with less social inﬂuence. However, studies testing this prediction found that inﬂuence and attitude change followed a pattern directly contrary to that hypothesized (Short 1974). Social inﬂuence was in fact seen to be strong under conditions where social presence was low. Although, this undermines the viability of this model, it has nonetheless been applied in more recent analyses of computer-mediated communication (Rice and Love 1987, Kiesler and Sproull 1992, Rice 1992, Monk and Watts 2000).
The basic tenet of the social presence model is that social presence, and therefore social inﬂuence in computer- mediated environments is restricted to the extent that interpersonal contact is constrained within the environ- ment, i.e. the social is equated with the interpersonal. This supposition has subsequently been the basis for further theories of social inﬂuence in mediated environments, e.g. the cuelessness model (Rutter 1984, 1987) and the reduced social cues approach (Kiesler et al. 1984, Kiesler 1986). The assumption is that, as (text-based) computer-mediated environments do not allow for communication of non- verbal cues such as gestures or facial expressions, which impact on face-to-face interpersonal communication, these media are less social and therefore enable less social presence. Indeed, many commentators have found low degrees of social presence in computer-mediated environments (e.g. Choon-Ling 2002). This, somewhat intuitive, position leads to the conclusion that for computer- mediated, or virtual environments, to aﬀord social pre- sence, one must maximize the number of visual and audio cues, thus attempting to emulate face-to-face communication. Indeed, Greenberg (1998: 246) stated that ‘Electronic virtual workspaces must emulate the aﬀordances of physical workspaces if they are to support a group’s natural way of working together’. Similarly, Lombard and Ditton (1997) stated that the number of sensory output channels are an important factor in generating a sense of presence, and Ijsselstein et al. (2000) suggested that as technology increasingly conveys non-verbal communication cues, social presence will increase. Implicit in these attempts to emulate face-to-face communication is the equation of physical and social presence, i.e. face-to-face communication ensures physical presence, encompassing interpersonal interaction, and thus social presence.
Despite the above, it would be churlish to assert that all presence commentators subscribe to the deterministic view that the use of technologies that aﬀord a closer emulation of face-to-face communication, will necessarily result in great- er social presence. For example, in discussion of virtual learning environments, Mantovani and Castelnuovo (2003) concluded that although a sense of presence makes learning experiences engaging and relevant, it is not always necessary to use the most highly technical solution. Furthermore, Cottone and Mantovani (2003) considered the overriding importance of social context and ‘common ground’ in virtual learning environments. Nevertheless, the tide of research into presence continues to attempt to emulate face- to-face interactions and thus sustain the belief that to achieve social presence one must also have physical presence. This, it is argued below, occurs due to the assumption of equity between the social and the inter- personal, an equation which may not always be appropriate.
Presence and SIDE
The SIDE (Social Identity model of DEindividuation Eﬀects) model (Reicher et al. 1995, Spears and Lea 1992, 1994) is critical of the assumption that interpersonal interaction is necessary for social presence to occur. Indeed, the lack of non-verbal cues in computer-mediated environments may in fact increase, rather than decrease social presence in group contexts.
SIDE is developed from Social Identity theory (Tajfel 1978, Tajfel and Turner 1986) and the closely related work of Self Categorisation theory (Turner 1982, Turner et al. 1987). According to this theoretical perspective, indivi- duals have multiple layers of ‘self’, including not only a personal identity, but also a range of possible social identities. Moreover, each social identity provides information about the social group, what is typical for that group and the behavioural norms associated with it. For example, various characteristics are associated with groups such as sports teams, work groups and gender aﬃliation. Each group situation presents a diﬀerent social identity with associated norms and at any one time either a particular social identity or the personal identity can be salient. Furthermore, the social context is a crucial factor in determining the salience of social categories. Consequently, an individual is more likely to act in accordance with the salient identity of that moment. From this perspective, the group exists within the individual as a cognitive representa- tion, rather than the individual existing within the external group. The cognitive representation of the group, therefore, contains information about the group and its associated norms.
Given the above, the social identity approach argues that an absence of social cues to interpersonal contact does not necessarily imply an absence of social cues per se (Spears and Lea 1992). The communication of social category information may not be as sensitive to the information richness of a virtual environment, as is the transfer of interpersonal information. Cues as to the membership of social categories can in fact be very easily communicated. For example, the textual information provided in message headers can provide category information such as gender and ethnicity (often from one’s name), or organisation aﬃliation and interests. The virtual group will also have a purpose to their communication and this shared purpose can form the basis of a shared social identity. The salience of category cues underlying social identiﬁcation is there- fore, to some extent, independent of interpersonal cues that are absent from many computer-mediated environments. Rather than the group being reduced to a set of interpersonal connections between group members, the group exists within the individual as a cognitive representation. Thus, a feeling of belongingness to a group, or identiﬁcation with a particular group, can occur in environments with relatively few sensory channels. This identiﬁcation with a group results in a perceptual immersion within the group, enabling the medium to become a social entity and thus conveying social presence.
The SIDE model argues that in situations where the transfer of personal, or individuating information is limited, this can increase the salience of a relevant social identity. Factors such as the lack of cues within virtual environments can, therefore, reinforce group salience and thus social presence. In this way, individuals need not be physically co-present or exchange interpersonal informa- tion in order to feel part of a group, or for the group to have real inﬂuence on the behaviour of each individual. Instead, they simply need to categorize themselves in terms of the social (group) identity rather than their personal identity. In terms of self-categorization theory, instead of deﬁning the situation in interpersonal terms (‘me’ vs. ‘other’), the self and communication partners are more likely to be included in a shared social category (‘we’), leading to a focus on shared similarity rather than diﬀerence. In other words, the self and others are depersonalized into categorically interchangeable group members rather than viewed as distinct individuals and thus the shared social identity has an increased inﬂuence on behaviour. However, this process will only result in increased social presence if the social identity remains salient. A salient personal identity could in fact undermine the shared group identity.
The SIDE model has been empirically tested in various contexts, including computer-conferencing and video-conferencing. For example, Lea et al. (2001) investigated the reinforcement of normative behaviour in computer- mediated environments under anonymous conditions. Of speciﬁc interest were the eﬀects of group based identiﬁcation on attraction within dispersed groups interacting under visual anonymity or video identiﬁable conditions. In the visually anonymous condition, communication was text-based, and in the visually identiﬁable condition this was supplemented by two-way real-time silent video. Under anonymous conditions, identiﬁcation in terms of the group and attraction to the group were increased. Through the use of path analysis, it was shown that visual anonymity increased group based identiﬁcation, which in turn increased attraction to the group. Therefore, that group members felt a greater belongingness to the group under conditions of reduced visual cues has important implications for designing ‘presence’ in virtual environments. If the intended result of social presence is to confer on the group greater capacity to communicate and collaborate, then the group will work more productively to the extent that group members identify with the group, thus making the group more cohesive. The group will then have greater inﬂuence over its members. Investigations of the SIDE model have demonstrated that greater adherence to group norms, and thus social inﬂuence, occur in computer-mediated groups in which a shared social identity is salient (e.g. Lea and Spears 1991, Postmes et al. 2000). Furthermore, it is suggested that for an individual group member to be motivated to work for the group, their salient identity should be congruent with the group identity (Haslam, 2001). In other words, it is social identiﬁcation with the group, or a feeling of social presence, that is the motivation for group behaviour.
In summary, if social aspects of communication are reduced simply to the aggregation of interpersonal bonds between group members, then to achieve social presence it will be necessary to increase cues to the interpersonal and thus emulate face-to-face communication. However, if the social is equated with a shared social identiﬁcation, which embodies the norms of the group, then social presence can exist in environments with few cues to the interpersonal. Indeed, environments rich in interpersonal information may, in fact, undermine group identity and result in process losses for the collaborating group (Lea et al. 2002).
Presence in distributed group environments
From a traditional approach, the mechanism by which social presence and group cohesion would be encouraged within a distributed group environment, would be to maximize the interpersonal communication between individuals, thus concentrating on the personal identity of the group members. To this end, it is often suggested to maximize the visual cues available through the use of technologies such as video conferencing or virtual reality. Alternatively, in less sophisticated environments, such as predominantly text-based computer-mediated systems, the display of group member biographical information, including photographs, is encouraged, thereby personalizing the virtual environment. Furthermore, an initial task of distributed groups is often the ‘ice-breaker’, whereby group members attempt to get to know each other before they begin working together.
From a social identity perspective, however, group members coming together to communicate and collaborate in a distributed virtual environment will bring to the collaboration a variety of possible identities. These could include identities relating to their gender, organisation aﬃliation, nationality and work interests, as well as a personal identity and the identity relating to the virtual collaborating group. Rather than focusing on the personal identity, we suggest that to facilitate social presence, the shared social identity relating to the virtual group should be made salient. In this way, the goals, priorities and norms of the collaborating group will become those to which group members are likely to adhere to, rather than those of conﬂicting identities. Moreover, the lack of face-to-face cues to the interpersonal will not be a barrier to the formation of a cohesive group with a strong sense of social presence.
There are a number of ways in which the group identity of the collaborating group can be made salient. For example, rather than the traditional ‘ice-breaking’ task, the ﬁrst meeting of the group could take place under anonymous conditions. That is, all personal information about the group members would be concealed, in order that the only cues to category membership are those concerning the current collaborating group. Even one’s e-mail address can display cues to a variety of identities, including gender, nationality and organisation aﬃliation. The removal of these from the initial group communications will serve to focus attention towards the goals and norms of the collaborating group. Indeed, in a study by Rogers (2000) groups performed a collaborative task, communicating via a text based computer-mediated environment, and were either anonymous or identiﬁable by their full name. It was demonstrated that group members in the anonymous condition reported a higher degree of identiﬁcation, or belongingness to the collaborating group, than those who were identiﬁable to others by their name. In the anonymous condition, therefore, the collaborating group was salient and the self and others were depersonalised into categori- cally interchangeable group members rather than viewed as distinct individuals.
It is, of course, impractical to maintain anonymity among group members for any length of time. It is also not suggested that interpersonal communication or bonds between group members should be discouraged per se. These are natural developments in group and indeed, human communication. Moreover, it is not the case that groups should consist of homogenous group members, showing little or no intragroup diﬀerentiation. Group member role diﬀerentiation is an integral part of colla- borative work and is essential for proper group functioning. However, the focus throughout the develop- ment of the group should be on the shared group identity that bonds the group together, thus ensuring that each group member holds salient in their mind the cognitive representation of the group. One technique to establish the salience of the group is through appropriate intergroup comparisons. This is based on the social identity approach that considers the analysis of groups incomplete without recognition of the fact that the ingroup cannot exist without outgroups and that the ingroup is deﬁned in terms of its relationship to the outgroup (Tajfel 1978, Hogg and Abrams 1988). The comparison also helps to undermine alternative categories of identiﬁcation, through a procedure of functional antagonism (Turner 1985). Further proce- dures to increase group salience include instructions given to the group as a whole, rather than to individual group members, and instead of personalising the virtual environ- ment, it is suggested that the environment be ‘collectivised’, reﬂecting the identity of the group and not simply the individuals that make up the group (see Lea et al. 2002). Furthermore, it is suggested that rather than prescribing roles to group members and restrictive working procedures to the group, the group should be allowed to manage itself, thus enabling the shared group identity to emerge and develop. Group members should then feel a greater stake in the group, are more likely to identify with that group and thus feel an increased sense of social presence. Moreover, behaviour follows from group identiﬁcation rather than from prescribed roles.
Two case studies, or ﬁeld trials, of this SIDE approach to social presence were carried out with collaborating under-graduate students from the University of Manchester and the University of Amsterdam.