When does activating diversity alleviate, when does it increase intergroup bias? An ingroup projection perspective.

The question how intergroup bias can be alleviated is of much theoretical and practical interest. Whereas diversity training and the multiculturalism ideology are two approaches prominent in practice, most theoretical models on reducing intergroup bias are based on social-identity theory and self-categorization theory. This social-identity perspective assumes that similar processes lead to intergroup bias in very different intergroup contexts if people identify with the respective social groups. A recent prominent model based on these theories is the ingroup-projection model. As this model assumes, an ingroup's norms and standards are applied to outgroups included in a common superordinate category (this is called ingroup projection). Intergroup bias results because the outgroup fulfils these norms and standards less than the ingroup. Importantly, if the diversity of the superordinate category is induced as the norm, ingroup projection and thus intergroup bias should be reduced. The present research delineates and tests how general this process is. We propose that ingroup prototypicality is not only an outcome variable, as the ingroup-projection model originally assumes, but can also be an important moderator. We hypothesize that for members considering their ingroup highly prototypical ("pars pro toto", large majorities), the superordinate group's diversity may question their ingroup's position and thus elicit threat and intergroup bias. In contrast, for members who consider their group as less prototypical (one among several, or "una inter pares" groups), activating diversity should, as originally assumed in the ingroup-projection model, reduce intergroup bias. Three experiments (total N = 345) supported these predictions in the contexts of groups defined by gender or nationality. Taken together, the ingroup-projection model can explain under which conditions activating superordinate-category diversity induces tolerance, and when it may backfire. We discuss in how far the ingroup-projection model can integrate conflicting findings on the multiculturalism ideology