Three types of supervisor-subordinate similarity were identified: (a) perceived similarity, perceptions of how similar the supervisor and subordinate are; (b) perceptual congruence, similarity of perceptions about behaviors important in receiving a high merit pay raise; and (c) actual similarity of individual characteristics. The relation(s) among the types of similarity and of each type with various employee outcomes were examined. Results supported the distinctions among types. Each type was related to subordinate performance. Results were less consistent for job satisfaction and pay ratings. Perceived similarity yielded the strongest relation with the dependent variables. The findings also suggest that similarity affects evaluations not only through bias, but also partly because of differences in supervisor-subordinate interactions. Numerous studies have identified factors that affect performance ratings. One such factor is supervisor-subordinate similarity (Landy & Farr, 1980). The present study explores possible effects of three types of similarity: (a) perceived similarity between the evaluator and another person, (b) similarity of supervisor and subordinate perceptions about aspects of the work environment, and (c) actual or demographic similarity. In research on perceived similarity it has generally been assumed that a person who is perceived as similar to the evaluator is more attractive, so that decisions and evaluations regarding that person are biased positively (Byrne, 196 l; Byrne, Young, & Griffitt, 1966). Experimental manipulations of similarity have generally supported this assumption. Persons seen as similar in attitudes and background were treated and evaluated more favorably than were those seen as dissimilar (Baskett, 1973; Golightly, Huffman, & Byrne, 1972; Griffitt & Jackson, 1970; Peters & Terborg, 1975; Rand & Wexley, 1975; Wexley & Nemeroff, 1974). However, studies conducted in the field suggested that the biasing effects of similarity might be less robust and more subject to individual differences than the aforementioned results indicate. Pulakos and Wexley (1983) found that perceived similarity between managers and their subordinates led both to give higher performance ratings to the other, but studies involving college and job-applicant interviews found sizable individual differences in the effects of perceived similarity (Frank & Hackman, 1975; Sydiaha, 1962). Finally, Dalessio and Imada (1984) found that the decision made by most interviewers re