Programs were randomly assigned to have their faculty participate in DOC training or not. Faculty from both groups attended a meeting where they watched 9 videotapes of residents interacting with patients. Residents and patients in the videotapes used scripts. All videotapes showed residents making some errors in history taking, physical examination, or counseling. Some videotapes showed residents making many errors, while others showed residents making only a few errors. After viewing each videotape, faculty noted the errors and rated resident competence. Next, faculty in both groups received a notebook that provided information about ways to evaluate trainees' clinical skills. Then the group assigned to DOC training attended a 4-day workshop while the other group returned to their workplaces. The workshop included lectures and 1 day devoted to DOC training, which involved videotaped evaluation exercises and practice sessions. Eight months later, 37 of the original 40 faculty members rated a second set of 9 videotapes.