Despite concern over the concentration on science in pre-medical education and the selection of medical students in the United States, no major changes have been made. The present study compares attrition rates and clinical competence levels of medical students who vary in pre-medical training, age, and sex, in order to determine the risk attached to accepting applicants with different background characteristics. No differences were found in levels of clinical competence, although the groups varied in average science scores on the medical college admission tests (MCATs). Students with non-science undergraduate majors had lower scores on this test than science majors. Younger women had relatively high attrition rates, but high performance on the MCAT science sub-test and in medical school science course. It was concluded that all groups had adequate science preparation for medical school, and there was no greater risk of producing doctors with low levels of clinical competence associated with accepting students from groups with lower science performance.