To determine if experienced primary care physicians are more likely to reach correct decisions on the telephone than their less experienced colleagues, we asked 31 first-year and 29 third-year residents, 21 faculty, and 36 private practitioners in pediatrics and family practice to evaluate three pediatric patients via a telephone interview with a simulated mother and to decide whether each patient needed to be seen that evening. Compared with first-year residents, the third-year residents, faculty and private practitioners decided less frequently to see children who were not severely ill (P less than .05) or injured (P less than .01); however, less than half obtained histories considered adequate to rule out potential serious illnesses. Faculty did better than either residents or private practitioners in managing a severely dehydrated child; 100% of the faculty, but less than 60% of the residents or private practitioners, chose to see the patient promptly (P less than .001). More than one third of all residents and private practitioners reached inappropriate management decisions despite obtaining information that should have altered their decisions. In these simulations, experience in private practice was not associated with improved telephone management of very sick children. Faculty physicians appeared to be better able to identify severely ill children without inappropriately evaluating those who were less ill. In all three simulations, attainment of the correct decision appeared to be determined not by the number or type of questions asked, but rather by the physician's interpretation of the information collected.