PURPOSE This randomized controlled study compared the interviewing skills of first-year medical students receiving feedback primarily from standardized patients (SPs) with those of students receiving feedback primarily from faculty. METHOD All 154 first-year students at the University of Oklahoma College of Medicine in 1993-94 were video-taped to assess baseline and post-instruction interviewing skills. All the students, randomized to one of three study groups, attended two four-hour workshops on interviewing skills. Instruction in the groups was as similar as possible except in the matter of who provided feedback. Two rating systems were used to rate the videotaped interviews for performances of targeted skills. RESULTS Complete, usable data were available for 120 (78%) of the students. Skill ratings using the Arizona Clinical Interview Rating Scale were significantly higher for the "types of questions used" and "use of empathy" items in the SP-led feedback group. No significant difference in ratings was detected among the groups as measured by the Rotor Interactional Analysis System. CONCLUSION The SPs were at least as effective as the faculty in effecting behavioral changes in the first-year medical students' interviewing skills.