BACKGROUND: There have been no prior studies of the political self-characterization of U.S. physicians-intraining, and little is known about physicians' political leanings or the critical relationship between medical issues and political orientations of physicians and physicians-in-training. METHODS: All medical students in the class of 2003 at 16 nationally representative U.S. schools were eligible to complete three questionnaire administrations (at freshman orientation, entrance to wards, and senior year). RESULTS: Among these medical students, 5% self-characterized as politically very conservative, 21% conservative, 33% moderate, 31% liberal, and 9% as very liberal."Being male, white, Protestant, intending to specialize in Surgery or anesthesiology/pathology/radiology, or currently or previously being married significantly (P = .001) increased the likelihood that a student self-identified as very conservative or conservative. Disagreement or strong disagreement with the statements, "I'm glad I chose to become a physician" and "Access to care is a fundamental human right," were also both associated with being very conservative or conservative. Being more liberal was reported by blacks and Hispanics; those intending to become ob-gyns, psychiatrists, and pediatric subspecialists; and atheists, Jews, and adherents of eastern religions. CONCLUSIONS: U.S. medical students are considerably more likely to be liberal than conservative and are more likely to be liberal than are other young U.S. adults. Future U.S. physicians may be more receptive to liberal messages than conservative ones, and their political orientation may profoundly affect their health system attitudes.