The goal of this research is to understand how complex social beliefs may be modified by relevant information. This research addresses two general questions: 1) In what ways do stereotypic beliefs differ, one from the other, in their susceptibility to confirming and disconfirming information?, and 2) How is information about individual group members, whose characteristics confirm or disconfirm the stereotype of the group, incorporated into, or isolated from, the stereotype? Both questions deal with the same general issue of how concepts or categories are modified by experiences with exemplars of those categories. The first study investigates stability and change in stereotypes in a longitudinal design. A central focus of this study is on the disconfirmability of people's beliefs about the characteristics of individuals and groups, as expressed in the attribution of traits. Previous work by Rothbart and Park (in press) examined a number of dimensions along which traits differ in their susceptibility to confirming and disconfirming evidence, and a large number of trait terms used to characterize groups were scaled on these dimensions. We have begun a longitudinal study of stereotype change to determine whether the judged """"""""disconfirmability"""""""" of a trait can be used to predict changes in the content of a stereotype, and propose to continue this research for the next two year period. The second question concerns an important issue of long-standing interest to cognitive and social psychology: When do disconfirming exemplars of a category modify the nature of the category itself, and when are those exemplars functionally isolated from the category? We propose to test a preliminary theoretical model (Rothbart and John, in press) based on the nature of the social categories that become associated with particular individuals. We propose to conduct six basic experiments designed to test a number of implications of the model. Two major implications include: 1) Members of a group who possess attributes that are incompatible with the category are unlikely to be stored as instances of the category, and 2) As a group member become increasingly """"""""individuated"""""""" by additional information, that member becomes increasingly associated with subordinate categories, again leading to functional isolation from the stereotypic category. This research has importance not only for understanding the modifiability of socially destructive stereotypes, but also for understanding basic processes used for protecting existing belief systems (e.g., compartmentalization and differentiation).
