In the present research we examined predictions derived from the following three theoretical approaches to stereotyping: complexity-extremity theory, assumed characteristics theory, and expectancy-violation theory. In order to assess these predictions, we manipulated the race, personal appearance, and dialect style of target job applicants. White judges rated these applicants on a set of characteristics relevant to hiring decisions. Results were consistent with all three theories. Specifically, the range of judges' evaluations of black applicants was larger than the range of their evaluations of white applicants; the effects of personal appearance and dialect style were larger than the effects of race; and black applicants, on average, received more favorable ratings than white applicants. We present a model integrating all three perspectives, and we demonstrate its usefulness for explaining our results and for understanding past research on stereotypes.Three theoretical perspectives within social psychology address how stereotypes influence perceptions of individual members of in-groups and out-groups: complexity-extremity theory, assumed characteristics theory, and expectancy-violation theory. Each theory proposes that different processes underlie the impact of background information on evaluations of in-group and out-group members. When considered separately, these theories generate conflicting predictions. Rarely, however, have these theoretical approaches been directly compared with one another. Therefore, in this article we will examine these theories experimentally by investigating how the race, personal appearance, and dialect style of job applicants affects white observers' evaluations.First, we review the basic ideas of each theory to illustrate how they could lead to opposing predictions. For each theory, we generate one set of predictions by assuming that the theory describes the only influences on evaluations. These predictions are useful because they clearly reflect the ideas of each theory. However, researchers from each of these theoretical approaches generally acknowledge the possibility of other influences. Consequently, we also develop a second set of predictions for each theory. These predictions are more realistic and more complex because they reflect how each theory might function in the con-This project was based on work supported by a National Science Foundation Graduate Fellowship provided to Lee Jussim and by a Spencer Foundation Grant provided to Lerita M. Coleman.We wish to acknowledge Lisa Gibson for her help in collecting the data and

No abstract