Using role theory in monitoring web group learning systems

Role theory has been proposed to explain group teamwork. Thus, it may also be valid to explain group learning performance. However, teachers in both conventional classrooms and web learning systems find it difficult to figure out what role a student played in a group and what relationship exists between roles and group performance. In a web learning system, interactions among group members can be recorded in a database. Computer tools can be developed to do the tasks for teachers. In this paper we develop a tool to capture the roles that a student plays in her/his learning group. Then, tools using machine learning techniques are built to find the relationship between existence of roles and group performance. A tool was then built to predict the group performance based on the relationship captured. An experimental result is shown that demonstrates that role theory is effective to predict group performance.