Ranking learner collaboration according to their interactions

Collaboration is supposed to be easily implemented in Learning management systems (LMS). Usually the basic functionalities in that respect support grouping students and providing communication features so that they are able to communicate with each other. However, related collaborative learning and CSCL studies and developments, which have been investigating how to manage, promote, analyze and evaluate collaborative features for decades conclude that there is no easy way, and much less standards-based approaches to support effective collaboration. The mere use of a typical set of communication services (such as forums, chat, etc.) does not guarantee collaborative learning. Further, managing collaborative settings in those LMS approaches is usually a time consuming task, especially considering that a frequent and regular analysis of the group´s collaboration process is advisable when following and managing the collaborative processes. To improve collaborative learning in those situations we provide tutors and learners with timely information on learners´ collaboration in a domain independent way so that the model can be transferred to other domains and educational environments. After setting a collaborative experience in an open and standards-based LMS, we have analyzed, through various data mining techniques, the learners´ interaction in forums during three consecutive academic years. From that analysis we have built a metric with statistical indicators to rank learners´ according to their collaboration. We have shown that this rank helps learners and tutors to evaluate the collaborative work and identify possible problems as they arise.