Using data mining for discovering relationships between collaboration skills and group roles

Abstract

Computer-Supported Collaborative Learning systems provide communication, coordination and collaboration tools that ease group dynamic regardless space-time location of group members. However, forming groups and having technology to support group tasks is not enough to guarantee students collaboration and the reaching of learning goals. Effective collaboration supposes the manifestation of specific roles by group members. Considering that group roles are conditioned (among others factors) by collaboration skills that students are able to manifest, it is necessary to discover non-explicit relationships between group roles and collaboration skills. In order to stablish this relationship data mining, in particular association rules, was applied to a set of interactions registered during online collaboration sessions where universitary students participated. Through associaton rules it was possible to discover relationships of Conversation and Active Learning collaboration skills with Monitor Evaluator, Coordinator, Resource Invesigator and Specialist group roles. The discoverd knowledge might be used for automatic recognition of student roles based on collaboration skills that students manifest in their groups. Furthermore, the discovered association rules could be used for group formation considering if group members have the skills related to the necessary roles for an adequate group dynamic.