Significant disparities in health care based on patient socioeconomic status have been documented. The extent to which physician behavior accounts for these differences is not known. We examined the impact of patient socioeconomic status, measured by years of education, on physician behavior assessed by direct observation of office visits, chart audits, and patient reports among 138 family physicians in 84 practices. Outcomes included time use measured with the Davis Observation Code, delivery of preventive services recommended by the US Preventive Services Task Force, satisfaction assessed with the MOS 9-item Visit Rating Scale, and delivery of attributes of primary care measured by the Components of Primary Care Index. After controlling for patient characteristics among 2538 visits by adult patients who returned questionnaires, a high school education or less was associated with slightly greater proportion of time spent on physical examination and providing nutrition counseling, and less time on patient questions, assessing patient health knowledge, negotiation, and exercise counseling. This indicates that physicians adopt a more directive style with less educated patients. Screening tests were provided at lower rates among less educated patients, but there were no differences in rates of health habit counseling or immunization services delivered and no differences in delivery of patient-reported components of primary care. Less educated patients had similar overall visit satisfaction, but were slightly less likely to have their expectations met. These show that patients' education has relatively small, but potentially important, effects on the outpatient delivery of primary care.