The impact of group discussion on the decision-making effectiveness of medical teams was examined. Three-person teams of physicians diagnosed 2 hypothetical medical cases. Some of the information about each case was given to all team members prior to discussion (shared information), whereas the rest was divided among them (unshared information). Compared with unshared information, shared information was more likely to be pooled during discussion and was pooled earlier. In addition, team leaders were consistently more likely than other members to ask questions and to repeat shared information and, over time, also became more likely than others to repeat unshared information. Finally, pooling unshared (but not shared) information improved the overall accuracy of the team diagnoses, whereas repeating both shared and unshared information affected bias (but not accuracy) in the diagnoses.