opportunity to stereotype an outgroup member when alone, in the presence of an audience, or in the presence of ingroup members ( Wilder & Shapiro, 1991). The authors anticipated that stereotyping would be increased in the presence of ingroup members, where categorization would make the ingroup distinction salient and would thus reduce the possibility of personalization. As predicted, stereotyping of the outgroup and identification with the ingroup were increased in the presence of the ingroup, relative to the other two conditions. This finding suggests that, as would be expected by personalization theory, identification with the ingroup is positively associated with stereotyping.
Unfortunately, cross-cutting categories may yield negative as well as positive effects. Although cross-cutting categories decrease negative evaluations of groups that are ingroups on one dimension and outgroups on the other (e.g., people from one's own sex but another race), the judgments made about the double outgroups (e.g., people of the other sex and another race) are the most negative ( Hewstone , et al., 1993; Islam & Hewstone, 1993; Vanbeselaere, 1991). In some studies, derogation directed toward outgroup/outgroup members has been even more pronounced than single-category outgroup derogation ( Islam & Hewstone, 1993; Vanbeselaere, 1991).
Samuel Gaertner and John Dovidio believe that intergroup bias can be mitigated by transforming the cognitive representations of ingroup and outgroup members from the perception of two groups to perceptions of a single superordinate group ( Gaertner, Mann, Dovidio, Murrell , & Pomare, 1990; Gaertner, Mann, Murrell, & Dovidio, 1989). If group members can recategorize themselves as belonging to a superordinate group, then attitudes should become more positive toward members of the former outgroup. This effect is predicted on the basis of the social identity theory tenet that ingroup favoritism is one way of using groups to maintain or enhance the individual's self-esteem ( Tajfel, 1978, 1982), and on Sherif's findings regarding the reduction of negative intergroup relations through identification with a superordinate goal ( Sherif et al., 1961).
In a study demonstrating the benefits of recategorization, subjects from previously formed pairs of three-person groups were induced to view themselves as one 6-person group, as two 3-person groups, or as six individuals by treatments such as segregation of seating, color coding of groups or individuals, naming of groups or individuals, and the nature of the interdependence among the individuals ( Gaertner et