A method to form learners groups in computer-supported collaborative learning systems

Nowadays, most students take part in collaborative learning activities, which consist of carrying out academic tasks in groups. Computer-Supported Collaborative Learning (CSCL) systems offer tools to support these collective activities. The method used to form the learners groups can be a key element in achieving a successful collaboration. This paper proposes a method of group formation using indicators that analyze previous collaborative activities of the learners. This method is based on data depth, a statistical tool that allows the ordering of multivariate data. In the process, the data depth of the analysis indicators of each learner is calculated, providing a measure that compares the values of the indicators of each learner with those of other learners. Thus, the method allows us to group learners whose analysis indicators register similar or different values. In this way, a flexible approach for forming homogeneous or heterogeneous groups is offered. We develop a software tool for this method, which we use in a case study to form groups of learners who work on programming tasks using a CSCL system.