| Abstract: | Research relevant to Rothbart and John's (1985) Journal of Social Issues, 41, 81–104] model of stereotype change is examined. Contrary to predictions from the contact hypothesis, the attributes of category members frequently fail to generalize to the category as a whole. To account for this lack of generalization, Rothbart and John proposed that judgments about the attributes of a category are based, in part, on the attributes of the members most strongly activated by the category label. Embedded in this simple assumption, however, is the idea that as a category member becomes increasingly disconfirming of the stereotype, it decreases in its likelihood of being activated by the category. Category members who are too strongly disconfirming of the category are, in effect, not thought of as category members — a view that is consistent with prototype models of category structure. Current evidence for two critical assumptions is examined: (a) at the level of judgment, greater weight will he given to the attributes of good-than poor-fitting members of the category, and (b) typical category members are retrieved from memory more easily than atypical category members. In addition, evidence relevant to two implications of the model is also examined: (a) moderately disconfirming exemplars are more likely to change the stereotype than are strongly disconfirming exemplars, and (b) stereotypes should show considerable stability over time, given the tendency to “functionally isolate” highly disconfirming exemplars. The cultural images of groups tend to be both more extreme and more homogeneous than is warranted by reality. The implications of this discrepancy for understanding resistance to change through contact with group members are discussed. |
