We compared two educational interventions (didactic lecture with handouts vs guidelines placed in patient charts) and assessed their impact on knowledge of early developmental services and identification of children eligible for these services. Forty-nine pediatric residents participated in preintervention and postintervention tests of knowledge of such services. Mean scores for all residents were 54% (pre-test) and 60% (post-test). First-year residents scored significantly lower than 2nd- and 3rd-year residents on pre- (p=.0001) and post-tests (p<.02). From our convenience sample of 1204 patients, ages newborn to 6 years, seen in the resident continuity clinic during the study period, residents identified children eligible for early developmental services. The mean percentage of identifying eligible children was 13% (range, 0-40%). We did not detect a statistically significant impact of either educational strategy. We conclude, therefore, that more comprehensive strategies are needed to change physician knowledge of and behavior toward early developmental services.