The broad goal of this research is to further our understanding of the processes that lead to the formation, maintenance and modification of stereotypic beliefs. Stereotypic beliefs are complex, and are learned from the culture at large, as well as by experiences with individual group members. Thus, stereotypes may be changed by a shift in cultural norms or values, and it is also assumed that they can be changed by experiences with individual group members. This last assumption-central to most forms of the Contact Hypothesis-has proved troublesome. Favorable experiences with outgroup members frequently fail to generalize to the group as a whole, as noted half a century ago by Lewin and Grabbe. One possible reason for this failure of generalization is the poor fit between the group members who disconfirm the stereotype and the stereotype as a whole (Rothbart and John, 1985). If goodness-of-fit, or typicality, is a critical condition for generalization to occur, then how well do our images of groups fit the characteristics of group members? Most research suggests that our images of groups are too extreme and too simple, and the extremity of our images are strongly influenced by the labels we apply to groups. The specific goals of this research, then, are to understand how the use of category labels, and the imposition of category boundaries, influences our perception of the group as a whole as well as the individuals who make up the group. Four general classes of studies are proposed, all concerned with the nature and consequences of the imposition of category labels and category boundaries. The first examines how strong or potent category labels may serve to inhibit the attribution of characteristics not directly implied by the label. These studies may explain why an individual's common or mundane attributes will dissociate that individual from category membership. While it may be desirable to """"""""release"""""""" an individual from a strong (and potentially destructive) label, the same process may make it difficult to temper our unrealistic images of the category as a whole. The second examines how the strength of category labels may increase the perceived differences between groups by both accentuating between group differences, and minimizing within group differences. The third set of studies examines the fundamental importance of ingroup-outgroup category boundaries and examines the pattern of inferences that exist between self, other, and ingroup, and also attempts to clarify the necessary conditions for establishing ingroup bonds. The fourth study involves completing the analysis of already collected longitudinal data designed to assess the magnitude, and correlates, of stereotype change over a four year period.
