Stereotypes are often presumed to exaggerate group differences, but empirical evidence is mixed. We suggest exaggeration is moderated by the accessibility of specific stereotype content. In particular, because the most accessible stereotype contents are attributes perceived to differ between groups, those attributes are most likely to exaggerate actual group differences due to regression to the mean. We tested this hypothesis using a highly accessible gender stereotype: that women are more socially sensitive than men. We confirmed that the most accessible stereotype content involves attributes perceived to differ between groups (pretest), and that these stereotypes contain some accuracy but significantly exaggerate actual gender differences (Experiment 1). We observe less exaggeration when judging less accessible stereotype content (Experiment 2), or when judging individual men and women (Experiment 3). Considering the accessibility of specific stereotype content may explain when stereotypes exaggerate actual group differences and when they do not.
All Science Journal Classification (ASJC) codes