Theory-Driven Group Formation through Ontologies Seiji Isotani and Riichiro Mizoguchi The Institute of Scientific and Industrial Research, Osaka University, 8-1 Mihogaoka, Ibaraki, Osaka, 567-0047, Japan isotani@acm.org, miz@ei.sanken.osaka-u.ac.jp Abstract. Group formation plays a critical role in collaborative learning (CL). It affects the acceptance of group activities by learners and the success of the collaborative learning process. Nevertheless, proposing an effective and pedagogically sound group formation is a very complex issue due to the multiple factors that influence group arrangement. The main goal of this paper is to present an ontology that works as a framework based on learning theories that facilitates group formation and CL design. To validate the usefulness and effectiveness of this ontology we present a method to use it and the results of an experiment carried out with four instructors and twenty participants. The results suggest that our ontology can be used adequately and the concepts represented on it can positively affect the performance of individuals during group learning. Keywords: Group formation, ontological engineering, collaborative learning. 1 Introduction Collaborative learning (CL) has a long history in Education [14]. According to [13], over the past decades the numbers of technologies that enable people to learn collaboratively have increased considerably. In CL, group formation plays a critical role that affects the acceptance of group activities and the success of the learning process. Some researchers claim that an inadequate group formation has been the main reason for many unsuccessful applications that rely on CL [5;6]. Nevertheless, according to [17], only a few CSCL systems provide the functionality for group formation. The large majority focuses on techniques for sharing resources or on improvements of group performance (which does not guarantee an improvement of learning [3]). The policy used by conventional methods concerns situation- independent CL activities where the idea of groups composed by heterogeneous participants is always the best solution. Such policy (lower-level policy) is applicable to any situation without regulation of the group. While it has satisfactorily facilitated the use of group formation in CSCL systems [12], the lower-level policy has difficulties in supporting well-structured groups where each learner has a defined role and learning goal. This limitation may impair the chances of successful learning and complicates the analysis of the CL processes. To overcome this problem our work deals with a higher-level policy that can be put on top of the lower-level policy to further increase the benefits of CL by bringing structure and context into the group. Thus, the main problem we are addressing is