Two classes of entering freshman medical students at the University of Utah indicated career preferences prior to the start of preclinical classwork. A group of interested students applied to a special family practice curriculum track. From this group, 20 were randomly selected each year to participate in the track program. The number was limited because of the need for a small student/faculty ratio. Students participating in the program were found to select residencies in family practice at a significantly higher rate than their classmates who expressed an interest but were not selected to participate. Students who stated a preference for family practice as freshmen were also found to select residencies in family practice at a significantly higher rate than freshman not preferring it. Nevertheless, most senior students matched with residencies different from their freshman preferences.