Boolean prediction of final grades based on weekly and cumulative activities

The authors' programme of work focuses on designing online collaborative learning systems, courses and assessments as well as developing a range of data analytics to understand how social learning takes place in these systems. In this paper the results of an analysis of an online course, designed by the authors and delivered on Coursera is presented. As part of this research, learners were guided into our collaborative learning system to give and receive feedback on each other's developing work. Using only data gathered from learners from the learning system about how they interact with each other, it is shown that combined activity is the best predictor of success in the course. This differentiates this research apart from many clickstream analysis studies. The results indicate the potential of predicting which of the two fail/pass bands students fall into, and analyse the predictions throughout the duration of the term.