Peer Ratings in Cooperative Learning Teams

Synopsis A universal concern about cooperative learning is the possible existence of “hitchhikers, ” team members who fail to fulfill their team responsibilities but get the same high grade as their more responsible teammates. A common way to minimize hitchhiking is to use peer ratings to assess individual performance of team members and to adjust the team project grade for individual team members based on their average ratings. Peer ratings have potential drawbacks, however. Common concerns are that team members will agree to give one another identically high ratings, or give ratings based on gender or racial prejudice, or inflate their own ratings if self-ratings are collected. Some instructors also worry that many students will resent having their grades affected by their teammates’ ratings. The objective of this study was to examine the validity of these concerns. A peer rating system developed at the Royal Melbourne Institute of Technology was modified and used in two sophomore-level chemical engineering courses. The students completed their homework in instructor-fo rmed teams in each course, and an average homework grade was computed for each team. At the end of each course the students confidentially rated how well they and each of their teammates fulfilled their team responsibilities, taking the ratings from a prescribed list of nine terms ranging from “excellent” to “no show.” The instructor assigned numerical values to each rating and computed a weighting factor for each student as the student’s individual average rating divided by the team average. The student’s final homework grade was the product of the weighting factor and the team project grade. Correlations were computed between peer ratings and test grades, peer ratings and self-ratings, ratings given to teammates of the same sex and of the opposite sex, and ratings given to teammates of the same ethnic background and of different ethnic backgrounds. Peer ratings correlated significantly with test grades, indicating that the more responsible students tended to be those who did best academically and/or that the academically stronger students were perceived as contributing most to the team effort. Self-ratings were remarkably consistent with peer ratings. Students rarely rated themselves higher than the rest of their teammates rated them; in fact, more (although still relatively few) gave themselves ratings lower than any they received from teammates. The incidence of identical ratings for all members of a team was also relatively low, on the order of 5–10% of all teams. No evidence of gender bias appeared in the data. Nonminority students gave lower ratings to minority students than to other non-minority students; racial prejudice could account in part for this result, but other explanations are equally likely or more so. Roughly 7% of the students were revealed as possible hitchhikers (as evidenced by their receiving less than satisfactory peer ratings from their teammates), but complaints about the system were almost non-existent. Most of the concerns frequently raised about peer ratings in cooperative learning were thus not borne out by the results of this study. Much additional research will be needed before the concerns can be definitively set aside.