Three studies tested a self-categorization theory explanation for the third-person effect. In Study 1 (N= 49) undergraduate students judged the influence of the National Enquirer, Wall Street Journal, and TV show Friends on themselves, relative to low- and high-status outgroup members, and other undergraduate students. The profile of first- and third-person perceptions was largely consistent with predictions, and the size of the third-person effect decreased as perceived similarity to target others increased—but only for media that were normative for comparison others. Study 2 (N= 49) provided evidence for this process with different media and showed that the profile of first- and third-person perceptions matched closely with perceived norms of media consumption—but not the social desirability of those media. Study 3 (N= 64) showed that the third-person effect for the same media and target other shifts with the frame of reference in which the judgment is made. Taken together, the findings are consistent with self-categorization theory and difficult to reconcile with other explanations.