In the study reported here, the authors evaluated the effectiveness of a teaching skills program for residents. Twenty-two residents in obstetrics and gynecology, medicine, and family medicine were randomly assigned to control and experimental groups. The experimental groups received instruction and feedback about teaching skills during their first and second postgraduate years. Both groups were evaluated at three times: in the first year before the instructional program, in the second year during the instructional program, and in the third year six months after instruction. During each study phase, videotapes were made of each resident teaching a student in the context of a case presentation. Trained raters evaluated eight teaching skills exhibited in the videotapes. In general, there was greater increase and less decline in the scores of the experimental than the control groups. These data suggest that teaching skills can be improved by instruction and that without support residents' teaching skills do not improve in relation to their clinical competence.