When Self-Descriptions Contradict Behavior: Actions do Speak Louder than Words

Subjects viewed two videotapes, one depicting a stimulus person's self-description and the other depicting that person's behavior in a conversation, according to a four-way factorial design personality descriptor used in the self-description ("introvert" or "extravert") × type of behavior displayed during the conversation (introverted or extraverted) × order of presentation (self-description seen first or conversation seen first) × stimulus person (one of two actresses). After viewing the stimuli, subjects rated the stimulus person on several personality dimensions related to introversion-extraversion and made attributions about the cause of her conversation behavior. Results showed a clear superiority of behavioral evidence over self-description in impact on the personality ratings. Although most of the personality dimensions showed significant effects of both the self-description and the behavior, the latter accounted for much more of the variance in these ratings. In contrast to previous findings, no order-of-presentation effects were found. Subjects tended to attribute the stimulus person's behavior to her personality except when she described herself as an extravert and behaved as an introvert; in that case, subjects tended to attribute more causality to situational factors. A subsequent study in which subjects read modified transcripts of the videotapes, within the same experimental design, yielded personality-rating results that were virtually identical to those of the initial study. Neither order effects nor attribution effects were found in this follow-up. Several differences between previous research and the current method of examining effects of inconsistent personality information are noted. The results are discussed in terms of strategies of information integration and errors in information processing.