Motivating Students in Collaborative Activities With Game-Theoretic Group Recommendations

Abstract

Recommending educational resources to groups of students is a common task in collaborative learning contexts. However, differences in within-group motivational factors might lead to conflicts in students' intention to use the resources. Previous methods fail to achieve high goodness of recommendation for the majority of students in heterogeneous groups. This study demonstrates a game-theoretic solution for recommending educational resources to homogeneous and heterogeneous groups. The group members are the players, the resources comprise the set of possible actions, and selecting those items that will maximize all students' motivation in the collaborative activity is a problem of finding the Nash Equilibrium (NE). In case the NE is Pareto efficient, none of the players can get more payoff (motivation) without decreasing the payoff of any other player, indicating an optimal benefit for the group as a whole. The suggested approach was empirically evaluated in a controlled experiment with a real dataset. The relevance of each delivered item to its corresponding students was explored both from the perspective of the group and its the individual students. The accuracy of the predicted group/individual motivation, the goodness of the ranked list of recommendations, and the problem-solving performance for the treatment group were significantly higher compared to the control groups. Limitations of the approach, as well as future work plans conclude the paper.