Gender-based differential prediction of job performance in employment-oriented personality measures has been left virtually unexamined. The use of 3 personality composites from the U.S. Army's instrument to predict 5 dimensions of job performance across 9 military jobs was investigated. Differential prediction, occurring in one third of the cases, was predominantly in the form of overprediction of female performances (i.e., higher male intercepts); slope differences were not found at above-chance levels. Female performance on the Effort and Leadership dimension was overpredicted in 90% of all predictor-criterion-job combinations, suggesting the measurement of tiffs performance dimension as the source of the differential prediction rather than bias in the personality measures. Findings of overprediction of female performance parallel those of research investigating differential prediction by race in the ability domain.