There has been increasing interest in gender disparities in clinical decisionmaking. Few studies have examined this issue in nationally representative samples or focussed on primary care. In addition, few of the studies have examined the role of physician gender. The 1987 National Medical Expenditure Survey was used to examine the relationship between physician gender and screening deficiency in women for three gender-sensitive tests (breast examinations, Papanicolaou tests, and mammograms) and one gender-neutral test (blood pressure checks). Women reporting a female physician as their usual provider compared with those reporting a male physician were less likely to be deficient for Papanicolaou tests and mammograms. There was a smaller, but nonsignificant similar trend for breast examinations. No gender bias was evident for blood pressure checks. These results persisted after multivariate adjustment for patient age, race, education, income, insurance status, subjective health status, other health behaviors, and attitude toward health care and health insurance. The results confirm the existence of physician gender bias in clinical decision making and represent one area for quality improvement.