Abstract Effectiveness of CPR performance on a manikin was evaluated immediately after training in public CPR classes by trained independent observers using validated measures and procedures. An instrumented manikin was used to assess critical skills thought to be related to survival following out-of-hospital cardiac arrest (compressions and ventilations), applying standards of the American Heart Association. The 226 subjects were enrolled in CPR classes offered to the public by the American Red Cross and the American Heart Association. Fifty percent of subjects performed 2% or fewer compressions correctly (the most common error being insufficient depth), and 50% performed 10% or fewer of ventilations correctly (the most common error being insufficient volume). Sixty-five percent failed to achieve a compression rate of 80 to 100/min. Forty-five percent of subjects failed to open the airway prior to a breathing check, 50% failed to adequately assess breathing, and 53% did not perform an adequate pulse check. Nearly half of all subjects made at least four errors in assessment and sequencing of skills. According to published criteria, trainee performance of CPR is poor. Failure in critical skills may contribute to poor survival rate following out-of-hospital cardiac arrest. CPR training programs must be developed with attention to learner outcomes.