This project examined the effects of two team selection methods (self-selected and instructorformed based on matched academic performance) on team and individual student performance and on self-reported attitudes and team behaviors in a freshman-level core-required introductory course. The data included mid and end-of-semester self-reports. Matched-performance groups had significantly higher grades on several performance measures, with a larger effect on the team grades than on the individual grades; however, overall the effect sizes were small. Although there were no group differences for most self-reported items, a key finding was that self-selected teams were significantly more likely to already have friends on their team, and a significant correlation showed that already having friends on a team was negatively correlated with many of the performance measures. In contrast, members of both types of teams reported equally high likelihood to make new friends, which was positively correlated with performance. Understanding the impact of different approaches to team formation may guide instructors and lead to more well-functioning teams, higher student learning, and greater student satisfaction.