This report describes and evaluates a 42-station objective structured clinical examination (OSCE) administered to 29 pediatric residents and six medical students. In half of the stations, residents spent 5 minutes performing a clearly defined clinical task while being rated by an observer. In the other half of the stations, they answered questions based on the data just gathered. There were six interviews with real or simulated patients, four physical examinations, six laboratory tests or procedures, and one chart review. Eight rest stops were provided. The results of the OSCE were compared with those of resident performance ratings and the Pediatric Board's in-training examination. The OSCE scores could clearly separate the students from the residents and each class of residents from all others (construct validity). The in-training examination could not separate first-year post-graduate level and second-year postgraduate level residents. Resident performance ratings could distinguish only first-year postgraduate level from third-year postgraduate level residents. Residents uniformly agreed that the OSCE measured important clinical objectives attesting to its content validity. Reliability for the OSCE was calculated at the 0.8 to 0.83 range. It is concluded that valid and reliable clinical examinations in pediatrics are feasible, practical, and highly desirable.