مسابقه، تنوع و رفتار حرفه ای و اجتماعی در یک تقسیم جامعه

This paper examines the impact of racial identity on behavior in trust games played by public high school students in South Africa. There is a systematic pattern of distrust towards Black partners, even by Black proposers, partially attributable to mistaken expectations. Non-Black proposers are significantly less likely to engage in a strategic interaction at all when paired with a Black partner, while Black proposers engage in exchange but at lower levels than when paired with non-Blacks. However, greater racial diversity in public schools promotes pro-social behavior towards Black partners.

A rich tradition of social psychology literature affirms that group identity or affiliation matters for outcomes in society, particularly since individuals tend to favor members of their own group over outsiders. The groundbreaking work in these minimal group experiments is due to Tajfel et al. (1971) who demonstrated that the simple categorization of individuals into groups, on the basis of some trivial criteria, such as the tendency to over- or under-estimate the number of dots on a screen, or a preference for the artistic work of Kandinsky over Klee, was sufficient to induce a favorable bias in behavior by subjects towards in-group members. This stood in stark contrast to previous work in this area, which had attributed in-group favoritism to perceived similarities in attitudes and beliefs among in-group members relative to outgroup members (Byrne, 1969), inter-group conflict over resources (Sherif and Sherif, 1961), or a common identity forged through a common shared fate (Rabbie and Horwitz, 1969). The minimal group experimental results were attributed to the fact that to the extent that the self-identity of individuals was at least partly derived from the social identity of the groups to which they belonged, individuals would treat in-group members positively in order to enhance the positive image of the group, thereby enhancing their own self-identity. This is not conceptually different from the view articulated in the economics literature by Akerlof and Kranton (2002) that an individual gains utility when her actions and those of others enhance her own self-image. (Billig, 1973, Messick and Mackie, 1989 and Hogg and Abrams, 1988). In the event that the social identity of the group became unfavorable, individuals would attempt to leave the group (physically or psychologically through disassociation) and join a better group (Turner et al., 1987). Subsequent work in the minimal group tradition (MGE), while cautioning against the generalizability of these results to real world groups (Yamagishi et al., 1999), has tried to elucidate the mechanisms which generate the observed in-group bias. One important mechanism is that group membership or identity must be public knowledge (Hoff and Pandey, 2004 and Yamagishi et al., 1999). Secondly, identity per se may be insufficient to produce an in-group bias unless it comes with an expectation of in-group reciprocity ( Yamagishi and Kiyonari, 2000). 1 There is some evidence that when one controls for the expectations of proposers concerning reciprocity from in-group members, the in-group bias disappears, suggesting that it is an expectation of (generalized) reciprocity from in-group members that drives the results ( Yamagishi and Kiyonari, 2000 and Yamagishi et al., 1999). 2 This paper adds to the substantial insider–outsider literature by examining the impact of observable racial identity on behavior in trust games played by public high school students in South Africa, a country with a well-documented history of legislated discrimination. To the extent that the expectations and social meanings created by apartheid persist even in the post-apartheid era, one might expect the racial identity of participants to affect trust relationships, particularly those involving Black South Africans, as it was this group that was most severely marginalized by apartheid institutions. Yet, these effects have been little studied in South Africa. In an important study, one of the first of its kind in South Africa,3Ashraf et al. (2006) find that Black proposers make significantly lower offers in a trust game, supporting previous work suggesting that members of previously disadvantaged groups in a society may be less trusting (Alesina and La Ferrara, 2002). The work reported here extends this earlier work by examining the impact of the racial identity of both the proposer and the responder on individual behavior in this strategic setting. Moreover, the students who participated in these games form part of the first generation of South African students who have not only had the opportunity to participate in a more integrated schooling environment, but have also spent much of their lives living in the “new” South Africa, where attempts to redress the devastating effects of racial segregation under apartheid have been made. This provides an opportunity to assess the extent to which increased racial integration, measured here through racial diversity within the schools attended by these students, might affect inter-racial co-operation.

The results presented here suggest that, at least for this sample of South African students, racial identity remains a salient cue for decision making in a strategic setting characterized by limited information. While the design of the game means that behavior cannot definitively attribute behavior to trust, the game does allow the opportunity to examine the impact that racial identity has on pro-social decision making in a strategic setting characterized by limited information. The emphasis in this paper is on behavior towards Black partners, since this was the group most severely marginalized by the apartheid regime. As such, if it is the case that proposers were motivated by altruism or the need to be generous to partners from racial groups with poorer socio-economic status on average, then one might expect that these motivations should have biased offers in favor of Black partners. The fact that Black responders were clearly disfavored in terms of offers made only confirms the extent to which this group might be stigmatized. The significantly lower offers made to Black responders are at least partially attributable to an expectation by proposers that Black responders would remit less. While such expectations may arise from the persistent socio-economic differences between race groups, with Black South Africans being significantly poorer than others on average, Black responders did not conform to this expectation in these games. Their behavior sends a clear signal that in spite of their poorer socio-economic backgrounds on average, they exhibit similar levels of pro-social behavior in their responses as members of other race groups. But perhaps more important is the result that proposers engage with Black partners in different ways, with non-Black proposers being significantly less likely to engage in an exchange at all, while Black partners engage but at significantly lower levels than when paired with non-Black partners. These differences in behavior are important. By refusing to engage in any exchange at all, non-Black proposers are unable to gain additional information about the pro-social proclivities of Black partners, and thus, racial stereotypes remain entrenched. This behavior of non-Black proposers paired with Black partners is consistent with Loury's (2002) notion of racial stigma, the argument being that when subjects are racially stigmatized, observers are less likely to engage in the critical experimentation required to reveal flaws in their beliefs concerning these individuals. In contrast, Black proposers appear more willing to undertake this experimentation process, making low offers to Black responders, even though they may expect that these offers might not pay off. In so doing, these proposers are able to acquire important information that allows them to update their expectations and stereotypes concerning the pro-social traits of Black partners. The systematic pattern of low offers towards Black partners might be characterized as “discrimination in contract” (Loury, 2002, p. 95)22 and while such discrimination may be morally objectionable, recourse to third parties such as the State to prevent such discrimination from occurring in real world contexts is certainly possible. However, these results also shed light on the more invidious problem of “discrimination in contact” (Loury, 2000, p. 95), or “race-mediated social relations” (Loury, 2002, p. 100). Pro-social behavior, particularly towards Blacks, is positively associated with increasing racial diversity in the school environment, which reduces “discrimination in contact”. Additionally, the racial composition of local networks such as peer groups may also be important, in that individuals may continue to practice discrimination in contact in terms of their close associates, despite attending racially diverse schools. The results presented here cannot speak to this. However, to the extent that attendance at a more racially mixed high school increases the incidence of mixed race friendships, as documented by Quillian and Campbell (2003), these two factors should work together over time to enhance inter-racial trust and co-operation.