Evaluation Bias or Reward Bias? Performance Appraisals and Gender and Racial Disparities in Federal Career Success

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 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 little evidence of evaluation bias: women receive equal or higher performance ratings than men and white women consistently receive higher ratings than white men, though 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.