Group matching for peer mentorship in small groups

Abstract

The problem of assigning learners to groups is essential in collaborative venture like group peer mentorship. Usually, in group peer mentorship, the group of peers is focused on one topic or question, and engage in discussion, argumentation, and providing criticism and constructive feedback, related to the individual contributions made by the participants. Therefore, the problem here is how to group peers that provide feedback to each other to ensure that all peers benefit from the venture, and everybody has something to offer. This problem is akin to the group formation problem in the area of computer supported collaborative learning (CSCL). Many group formation algorithms exist in the area of CSCL, but they have problem with limited and fixed constraints, evaluation strategies, their initial grouping mechanism and the goal of the grouping strategy. Therefore, we proposed a grouping algorithm based on some constraints and the principles of the Hungarian algorithm, to achieve a diversified grouping of peers for every mentorship session. Although this algorithm had been used manually in some of our previous small scale studies, in this paper, we evaluated the algorithm using 1080 system generated data, and compared the performance of our algorithm with three other algorithms in CSCL.