Antecedents of student team formation in higher education

Students often form project teams by themselves rather than impose pre-defined teams because it leads to better study outcomes. However, few scholars have investigated the mechanisms driving the formation of self-organized groups. We examine three underlying and interrelated antecedents of project team formation: friendships, familiarity with fellow team members due to prior co-working experiences, and similarity of team members regarding gender, academic achievement, and preferred collaborators. The data encompasses 70 first-year university students asked to self-form teams to carry out a semester project. We utilize Exponential Random Partition Models. This new method allows us to model how friendships, familiarity, and similarity affect self-organized group compositions while accounting for each factor's relative importance. The results show that friends are more likely to end up in the same team than non-friends. Similarity is also an important antecedent for team formation: Students with the same gender and similar grades are more likely to end up in the same team. Familiarity and preferred collaboration showed negligible effects. Our findings emphasize that educators must consider academic and non-academic factors when allowing students to self-organize in teams, depending on what goal educators strive to achieve with team projects.