In order to evaluate the impact of simulation complexity on resuscitation skill performance, 15 pediatric residents were observed in bag-valve-mask ventilation on an infant manikin. Performance was evaluated in a simple exercise in which ventilation was performed as a single isolated task. For each subject, this was compared to performance when a cognitive task involving clinical problem solving was simultaneously presented. Performance was expressed as frequency, tidal volume, and estimated hypothetical alveolar Pco2 in each resuscitation. A correlation was observed (P less than 0.05) between simple and complex resuscitation simulations for each measure of performance. The only subject with inadequate performance in the complex exercise was also identified as deficient in the simple exercise. These observations suggest that an adequate evaluation of performance of some resuscitation skills may be obtained in simple, efficient simulation exercises.