Three studies tested the hypothesis that people assume that the identities of other people are tied more closely to their distinctive than to their nondistinctive traits. In Studies 1 and 3, subjects predicted the preferences of a target person who was a member of both a statistically distinctive and a statistically nondistinctive category (e.g., sky diver and tennis player). In Study 2, subjects judged the degree of interpersonal similarity between pairs of people sharing distinctive as opposed to nondistinctive category memberships. Consistent with the hypothesis, subjects linked targets with their more distinctive traits and assumed targets would be more similar to people who shared their distinctive traits than to people who shared their nondistinctive traits. The implications of this distinctive ness effect for an understanding of stereotyping are explored.