Set/reset: use and disuse of concepts in impression formation.

In three experiments, impressions of an ambiguously described stimulus person were assimilated toward the implications of primed concepts when performance of the priming task was interrupted, but were contrasted with these implications when performance of the priming task was allowed to continue to completion. In addition, when the primed concepts were evaluatively consistent (Experiment 1), assimilation and contrast were observed on both prime-related and prime-unrelated dimensions. When the primed concepts were evaluatively inconsistent (Experiment 2), however, these shifts in impression were observed only on dimensions directly related to the primed concepts. When no concepts descriptively relevant to the stimulus information were primed (Experiment 3), the assimilation and contrast were relative to the favorableness of a primed general evaluative person concept. Taken together, these results suggest that a concept may be accessible to an individual and may be relevant to target information, yet not be used to encode that information; that assimilation and contrast may occur for reasons other than the discrepancy between the target and the contextual stimuli on the dimension of judgment; and that individuals may use the evaluative implications of their person representation as a cue in deciding which of several equally applicable, equally accessible descriptive concepts to use in interpreting information about a person.