An Examination of Group Dynamics from an Active Classroom Setting using K-Means Clustering

The use of group work to maximize the learning outcome in the physics classroom is a well-established idea. However, one challenge that instructors often face in implementing collaborative learning is the creation of effective student teams. The complex nature of student groups as collective learning entities is often mediated by the demographics and experiences of the different group members. In this study we apply K-Means clustering to peer survey data, collected using the Comprehensive Assessment of Team Member Effectiveness system, and institutional census data to identify features of well-functioning groups in an active learning classroom. Initial results have revealed interesting patterns; for example, the groups with the lowest levels of effectiveness were not the groups with the lowest levels of prior physics knowledge. When replicated on a larger scale, these results could inform the development of automated systems for effective and equitable team formation in active learning classrooms.