This paper explores the results of an experimental program at Tufts University designed to affect medical students' career choices. Students were located in a rural hospital for the entire third year of their medical training. It was hypothesized that the program would result in more positive orientations toward practice in rural settings and toward primary care. Longitudinal survey data on students exposed to alternative training settings and those remaining in the university were analyzed using a multivariate test of significance. The results indicate no treatment effects. Some of the possible reasons for the negative results, and the policy implications of those results, are discussed. (Author)