Short Communication A social identity approach to trust: Interpersonal perception, group membership and trusting behaviour

Trusting behaviour involves relinquishing control over outcomes valuable to the self. Previous research suggests that interpersonal perceptions of trustworthiness are closely related to this behaviour. The present research suggests that the more proximal determinant of trusting behaviour is the expectation that the other will reciprocate. Based on the Social Identity model of Deindividuation Effects (SIDE) model, reciprocity expectations may be created by interpersonal perceptions of trustworthiness or a shared group membership. To investigate this, group membership and individual identifiability were experimentally manipulated (N1⁄4 139): When individuals were not identifiable, trusting behaviour was based on expectations of reciprocity inferred from group membership, not on perceived trustworthiness. In contrast, personal identifiability fostered perceptions of trustworthiness for both inand out-group members. In this case interpersonal trustworthiness enhanced expectations of reciprocity, which in turn increased trusting behaviour. Copyright # 2005 John Wiley & Sons, Ltd. There has been considerable debate about what trust is, and how it is influenced (for an overview, see Cook, 2001). Trust has been treated as a more or less static interpersonal difference construct (Granovetter, 1985; Rotter, 1967). Another perspective emphasizes the influence of contextual factors, in which trust is seen as a cognitive process associated with the confidence in another’s goals or purposes, or the perceived sincerity of another’s word (Hosmer, 1995; Lewicki & Bunker, 1995; Mellinger, 1956). In this view, the level of trust is specific to the relationship and to the contextual factors which enhance or inhibit the development and maintenance of trust (Lewicki & Bunker, 1995). In these approaches, definitions of trust are operationally if not conceptually confined to perceptions of one regarding the other: that is, trust is defined as perceived trustworthiness. This, we argue, is a rather restrictive conceptualization, and does not clearly differentiate trust from other Received 9 July 2004 Copyright # 2005 John Wiley & Sons, Ltd. Accepted 2 January 2005 *Correspondence to: Martin Tanis, Free University, De Boelelaan 1081, 1081 HV, Amsterdam, The Netherlands. E-mail: [email protected] Contract/grant sponsor: ESRC; contract/grant number: RES-000-27-0050. ‘relational goods’, such as respect, attraction, support, or endorsement. Trust of this kind is undoubtedly valuable in interpersonal relations, and would appear to be closely bound up with the positivity of the relationship between two or more individuals. However, there is a rather different aspect of trust which is not quite captured by this perceptual or relational conceptualization. This feature has to do with the behavioural consequences of trust, in that trust also entails relinquishing some degree of control or power to the other. For example one could trust another to complete a particular task, or give them the keys to one’s house and trust them to put the rubbish out for you. There is a subtle difference between perceptions of trustworthiness and trusting behaviour. Unlike perceptions of trustworthiness, trusting behaviour involves relinquishing power over outcomes valuable to the self (cf. Messick & Kramer, 2001). One key difference between the two, then, is that one is a passive and potentially inconsequential interpersonal evaluation, whereas the other requires an active investment of self-relevant outcomes. Moreover, whereas the perception of trustworthiness is an assessment of someone’s character, trusting behaviour involves expectations about the other’s actions. It could be argued that this distinction is one of degree. Indeed, the perception of trustworthiness is likely to affect trusting behaviour to a certain extent. However, we argue that the interpersonal evaluation of someone as ‘trustworthy’ is not a necessary condition for trusting behaviour to follow, and that there is a qualitative difference between perceiving someone as trustworthy and expecting him or her to reciprocate. Indeed, we believe that perceivers are well aware that a target’s behaviour is (sometimes) dictated by factors other than their personality. Of course, the two are likely to correspond to each other in most cases, but exceptions exist. For example we might rely on an untrustworthy person to do something for the community (i.e. to reciprocate) because it is in their self-interest to do so (a case where people’s actions are for the greater good but not motivated by their personality). Also, we might expect reciprocity simply because rules or laws oblige people to behave in a certain way (a case where compliance is assumed). These examples illustrate two things. One is that trusting behaviour and perceptions of trustworthiness do not always coincide. Another is that the expectation of reciprocity appears to determine whether people will behave in trusting ways. Although reciprocity expectations can be created by interpersonal perceptions of trustworthiness (and indeed many have treated them as synonymous), we know from research that they can also be created by higher order perceptions of similarity and interchangeability, such as those induced by shared social group membership (Brewer, 1981; Yamagishi & Kiyonari, 2000). Shared group memberships (when salient) reduce the relevance of interpersonal distinctions (Turner, Hogg, Oakes, Reicher, & Wetherell, 1987) and produce the possibility for individuals to engage in co-action and collective action (Reicher, 1996). So, in certain group contexts, trust is not so much based on the economic calculation of what happens if the other individual preserves or violates the trust (so called calculus-based trust) but is based on common membership of a salient social group—i.e. identification-based trust (Kramer & Wei, 1999; Lewicki & Bunker, 1996). INDIVIDUAL IDENTIFIABILITY, GROUP MEMBERSHIP AND TRUST In order to examine what affects trusting behaviour, the present study examines the two factors underlying reciprocity expectations based on predictions derived from the Social Identity model of Deindividuation Effects (SIDE, Reicher, Spears, & Postmes, 1995; Spears & Lea, 1992). More specifically we manipulate information about the target by providing so-called cues to identity. It is important here to differentiate between cues that make salient aspects of personal identity and cues to 414 Martin Tanis and Tom Postmes Copyright # 2005 John Wiley & Sons, Ltd. Eur. J. Soc. Psychol. 35, 413–424 (2005) social identity (or group membership, see Tanis & Postmes, 2003). Both kinds of information are believed to play an important role in the perception people form of each other. In contexts in which cues are informative about personal identity, even relatively minimal cues such as portrait pictures and first names have been shown to reduce ambiguity and to result in more positive interpersonal impressions (Tanis & Postmes, 2003). Not ‘knowing’ one’s interaction partner might increase uncertainty or even apprehensions about the other, and provide a less firm basis for trusting the other. This idea resonates with the general assumption that ‘trust needs touch’ (Handy, 1995), i.e. that interpersonal contact is important (if not vital) in order to trust someone. Therefore, our first hypothesis would be that cues to personal identity increase perceived trustworthiness, but that group membership does not. However, people may also engage in trusting behaviour simply because they expect reciprocity from ingroupers. Past research on the SIDE model has demonstrated that when cues to shared social identity are available (and when this identity is salient), the inability to tell group members apart may accentuate the perceptual unity of the group, and thereby enhance group members’ feelings of attraction and identification to the group (e.g. Lea, Spears, & de Groot, 2001; T. Postmes, R. Spears, T. Lee, & R. J. Novak, in press; Postmes, Spears, Sakhel, & de Groot, 2001; Sassenberg & Postmes, 2002). Indeed, Sassenberg and Postmes (2002) showed that an inability to individuate ingroup others reduces interpersonal attraction, but at the same time increases a sense of shared identity. Conversely, information about idiosyncratic characteristics of group members stresses the unique individuality of each of them, thereby drawing attention away form the person’s identity as an ingroup or outgroup member. Extending this to the realm of trust, our second hypothesis would be that a person’s group membership would only affect expectations of reciprocity and trusting behaviour when individuation is not possible: In the absence of cues to personal identity, there may be an accentuated expectation of reciprocity based on shared ingroup membership. This leads us to our third hypothesis which is that trusting behaviour is mediated by the expectancy of reciprocity. As mentioned, the expression of trust in behaviour means that one invests trust in that person by giving the other power over outcomes valuable to the self. Therefore, trusting behaviour will only be displayed when people think that others will not take advantage of the situation and when reciprocity is expected. Reciprocity can be expected either on the basis of group membership, or on the basis of individuating cues to personal identity. When the other is an ingroup member, reciprocity would be expected from the other irrespective of whether they are individuated or not (i.e. irrespective of whether there is a basis for interpersonal trustworthiness or not). But when the other is an outgroup member, reciprocity can only be based on an interpersonal evaluation of the other’s trustworthiness— in other words an evaluation that depends on one’s capacity to individuate the other. Therefore, our fourth hypothesis is that when one’s partner is from the outgroup, expected reciprocity and trusting behaviour will be based on perceived interpersonal trustworthiness, when cues to personal identity are available.