Models of job performance ratings: An examination of ratee race, ratee gender, and rater level effects

This research extends the work of Hunter (1983) and Borman, White, Pulakos, and Oppler (1991) on models of supervisory performance ratings. A rating model that included measures of cognitive ability, practical intelligence, job knowledge, task proficiency, achievement orientation, and performance ratings was proposed and evaluated. A major purpose of this research was to evaluate the fit of the proposed model for different race and gender subgroups as well as for peer and supervisor raters. The models were tested on a sample of 456 professionals in a large government agency. Overall, results of the LISREL analyses showed no statistically significant differences in the fit of the model for the different race or gender subgroups examined. Significant differences in the fit of the model and variance accounted for in the peer versus supervisor groups were observed.