An intervention detection framework for collaborative projects

Capstone projects are increasingly being used in Higher Education to emulate real world work environments and to provide valuable experiential learning. There these projects are run within a live commercial environment and with the close cooperation of a client mentor, they have proven to provide enhanced benefits to students. In such cases a vast amount of planning and cajoling may have gone into securing the cooperation of live clients and tutors are therefore eager to ensure that the projects they have provided are successfully executed in order to secure the cooperation of the client organisation in future years. With this in mind, this paper describes the proposal for a framework to monitor and detect dysfunctional behaviour within project groups using CSCL (Computer Supported Collaborative Learning), in order to allow timely interventions to be made to correct the collaboration process. The proposed intervention detection framework comprises of four dimensions: 1) individual temporal dynamic, 2) group dynamic, 3) sentiment analysis of discussions, and 4) analysis of content volume.