Validation of otoscopic diagnostic accuracy is an important aspect of medical education and is necessary to substantiate clinical research observations. In addition, otoscopic accuracy is a prerequisite to optimal patient care. We describe an otoscopic validation program at the Children's Hospital of Pittsburgh (Pa) in which 27 physicians and three nurse practitioners participated from December 1980 to March 1990. The lowest acceptable limits for sensitivity and specificity regarding the diagnosis of middle-ear effusion were arbitrarily set at 80% and 70%, respectively. The 30 participating clinicians examined a total of 4147 ears. Mean sensitivity and mean specificity for the group as a whole were 87% and 74%, respectively. The chief benefit of the program was an improvement in diagnostic skills for some individuals. The chief limitation was the large amount of time required. We conclude that a formal validation program can document achievement of a reasonable standard of otoscopic diagnostic accuracy.