Although group work is commonly used in university-level instruction, social loafing, domineering team members, poor attenders, and inequitable distribution of marks have been identified as obstacles to team-based learning. Peer evaluation has been proposed as one vehicle to address these issues. For use in grading, peer evaluations are often anonymous; however, as tools to address team functioning, they should not be conducted anonymously, but rather with the results discussed as feedback. It has been suggested, however, that non-confidential peer evaluations will artificially elevate students' marks. In this study, we investigated the impact of peer evaluation confidentiality on students' marks. Without a weighted correction, confidential evaluations significantly dropped students' marks while non-confidential evaluations raised them. Implications for practice are discussed.