Racial or sexual discrimination in performance appraisals could potentially contribute to continuing race and gender gaps in pay in the federal service if (1) women and minorities receive lower performance ratings than comparable white males and/or (2) high performance ratings increase salaries and promotion probabilities less for women and minorities than for white men – what Castilla (2008) calls evaluation bias and rewards bias. Using a one percent sample of federal government personnel records, we test these two possibilities using OLS and logit analyses that control for individual and job-related factors. We find mixed evidence of evaluation bias: women receive equal or higher performance ratings than men and white women consistently receive higher ratings than white men, while black men tend to receive lower ratings than comparable white men. We find no evidence of rewards bias: outstanding ratings increase salaries and promotion chances as much for minority men as for white men, and increase them more for Hispanic, Asian and black women.

□ Key Words: Performance Appraisals, Women and Minorities, Salary Increase, Promotion, Federal Civil Service