Stereotype-based expectancy and social judgment: Rethinking from a Bayesian perspective

Abstract

To examine effect of prior stereotypical expectancy on social judgment from a Bayesian perspective, undergraduate subjects (N=204) were asked to infer a target person's attitude toward anatomic power problem. Half of them were told in advance that he was a member of Liberal Democratic Party (pro-expectancycondition), and the other half were told that he was a memberof Japanese Socialist Party (con-expectancycondition). Then subjects were given a series of hisprevious relevant utterances, which had either high orlow diagnostic values for the inference ofhis attitude. (a) "Labeling effect" occurred. That is, despitebeing given identical utterances, subjects given L. D. P. label estimated the target's attitude to be morefavorable toward the atomicpower than subjects given J. S. P. label. (b) This effect emerged mainly when subjects were givenlow diagnostic utterances. (c) Subjects given high diagnostic utterances inadequately underusedthe base-rate information (prior expectancy) compared with the Bayesian normative value. (d)Utterances congruent with prior expectancy were better recalled than utterances incongruent withprior expectancy. © 1986, The Japanese Psychological Association. All rights reserved.