Using the Katz-Braly checklist subjects (N = 65) assigned five traits to a national group and estimated the percentage of group members who had those traits. This was either an in-group (Australians) or an out-group (Americans), and subjects either judged that group alone (one-group conditions) or also estimated the percentage of people from the other nation (the United States or Australia, respectively) who had those same traits (two-group conditions). Across one-group conditions there was a significant out-group homogeneity effect with traits being seen to apply to more Americans than Australians, but there was no such effect across the two-group conditions. These findings were predicted on the basis of self-categorization theory's analysis of the role of comparative context in determining level of social categorization. Across two-group conditions non-stereotypic traits were also applied to fewer in-group than out-group members. This result suggests that trait favourableness is an important normative—motivational determinant of perceived homogeneity. A second experiment (N = 297) confirmed this point through an additional manipulation of the favourableness of checklist traits. This study also replicated the effect for comparative context. Implications for the analysis of social categorization, perceived group homogeneity and stereotyping are discussed.