Key points are not available for this paper at this time.
A stereotype of a group can be expressed by the estimated percentage of members that possess certain personality attributes (Brigham, 1971). In a multi-group design, the properties of the percentage measure were examined, and three common assumptions about stereotypes were supported. First, there was high consensus among American and Italian raters regarding the attributes of Americans, Italians, English, and Germans. Second, the perceived typicality of a trait depended largely on contrasts with other traits attributed to the same target group. Contrasts between attributions of the same trait to different groups were largely irrelevant. Third, most stereotype judgments revealed consistency biases. Compared with a Bayesian model of probability estimation, raters exaggerated the similarities between trait attributions (the percentage measure), social categorizations (percentage of people that belong to a group given they possess the trait), and Likert-scaled typicality ratings. Raters underestimated the effects of the traits' global base rates on the typicality ratings.
Joachim I. Krueger (Fri,) studied this question.