The authors assign semester- or quarter-long team-based projects in several Computer Science and Finance courses. This paper reports on our experience in designing, managing, and evaluating such projects. In particular, we discuss the effects of team size and of various peer evaluation schemes on team performance and student learning. We report statistical measures of the students' peer evaluations--do they always rate each other strongly or weakly? What are the means and ranges? More importantly, we discuss why we introduced these peer evaluations, and what effect they have had on student commitment and performance. We discuss a small number of cases where student participation was poor, and relate this to the peer evaluation process. (Contains 6 figures and 1 table.)