For decades, a recurring question in person perception research has been whether people’s perceptions of others’ personality traits are related to how they see themselves on these traits. Indeed, evidence for such “assumed similarity” effects has been found repeatedly, at least for certain characteristics. However, recent research suggests that these findings may be an artifact of individual differences in how positively or negatively perceivers see others in general, irrespective of trait-specific content. Overcoming the limitations of prior studies, the present work provides a critical test of trait-specificity versus global positivity as sources of assumed similarity in personality judgments. In two large studies (Ns = 2,287 and 3,563) with preregistered hypotheses and analyses, perceivers rated 10 targets (strangers) each on the honesty–humility, emotionality, extraversion, agreeableness, conscientiousness, and openness to experience; HEXACO (Study 1) and Big Five (Study 2) dimensions to capture their perceptions of the “average other” (i.e., perceiver effects). We then computed “positivity-corrected” assumed similarity effects using trait-based and profile-based approaches. Although controlling for global positivity considerably reduced the strength of assumed similarity, perceiver effects were still positively related to self-reports. As predicted, these assumed similarity effects occurred foremostly for traits strongly linked to values. Specifically, in Study 1, positivity-corrected assumed similarity was observed only for honesty–humility and openness to experience, albeit meaningful effects merely occurred on one of the two self-report measures. In Study 2, traits’ value-relatedness remained a unique moderator of assumed similarity after accounting for traits’ positivity (i.e., social desirability). These findings demonstrate that assumed similarity is indeed, to some extent, trait-specific.