Towards Supporting 'Learning To Learn Together' in the Metafora platform

Computer-Supported Collaborative Learning (CSCL) has been demonstrated to improve student interaction in complex collaborative learning scenarios. When orchestrated appropriately, it also provides opportunities for learning high-level social learning skills, or “learning to learn together” (L2L2), but these opportunities are often only dealt with implicitly. This paper presents work towards an intelligent system that can scaffold L2L2 across many domains by (a) offering carefully-designed message templates that encourage peers to communicate with their groups about their learning process, (b) analyzing student work and recommending a specific set of these message templates that are pertinent to their moment-by-moment interaction. We present methods by which the system can use automated analysis techniques to recognize opportunities where students might benefit from these messages, and either send the message directly or prioritize message templates for students’ use.