People possess idiosyncratic, self-serving definitions of traits and abilities. This observation was supported by 6 studies in which people articulated the performances along behavioral criteria (e.g., math Scholastic Achievement Test score) necessary to "qualify" for relevant traits (e.g., math ability) or made judgments about performances attained by other people. When making judgments of others, high-performing Ss tended to rate target performances less favorably than did low-performing Ss, with these disagreements most pronounced when the target's performance was low. These disagreements were mediated neither by perceptions of trait importance nor by differing beliefs about the distribution of performances along the behavioral metrics. Discussion centers on avenues for further study and on similarities and differences between these studies and classic work on attitudinal judgment.