This research project investigates the value of using simulated v. real patients in teaching interviewing skills to third- and fourth-year medical students on a clerkship in family and community medicine. Sixty-four medical students (38 males, 28 females) were asked to make a videotaped patient interview at the beginning of the clerkship. Forty-one of the students interviewed trained, simulated patients and 22 interviewed real patients at their clerkship site. Students received feedback and faculty teaching of interviewing skills after the first interview. All students made another videotaped interview with a simulated patient at the conclusion of the clerkship. All interviews, pre- and post-clerkship, were scored for interviewing skills, focal areas, and nonverbal language. Multivariate and univariate analyses of pre- and post-interviews found simulated patients to be of most value in teaching medical students verbal interviewing skills and real patients to be of most value in teaching the focal content areas of the interview. Recommendations were made to include use of both simulated and real patients in the teaching of medical interviewing.