Recent reports in the field of medicine have recommended the use of teamwork training to reduce the number of injuries and fatalities from human error. Teamwork training in the field of medicine appears promising, but few empirical evaluations of such programs have confirmed their effectiveness. Existing teamwork training studies have tended to use a traditional, lecture approach to training, with positive but modest results upon teamwork attitudes and behaviors. The current study developed and evaluated a more active teamwork training protocol for trauma resuscitation teams. The training protocol supplemented several medical and non-medical role plays with a lecture and guided discussion for feedback. Forty-one residents participated in the training on one of two days (groups) and completed evaluation measures prior to and immediately following the training program. The training was evaluated with measures of trainee reactions, attitudes towards teamwork, and responses to a situational judgment test (SJT). Analyses compared item and scale scores between pre-training scores and post-training scores. T-tests generally found higher means for post-training behavioral responses than pre-training responses. However, mean comparisons with teamwork attitudes and learning goal orientation did not yield significant differences. An item analysis of the SJT responses (using chi-square) indicated significant response shifts in many items that correspond to the teamwork training content. In summary, results indicated that teamwork training on behavioral choices, but little effect on the self-reported attitudes of trainees.