Openness, scalability, and reachability are intrinsic features of MOOCs. However, research studies in MOOCs indicated low participation from some cultural clusters, mostly from less privileged strata of the world's population. The impeding factors are not only related to individual student characteristics, but also are related to structure and curriculum design. This proposed PhD thesis will address this stratification, performance and achievement gap. In a cross-module analysis on Open University MOOCs at FutureLearn, temporal predictive modelling was used to explore learners' background, regional belonging and behavioral patterns that contribute towards engagement, and overall performance. Later on, clustering and temporal process mining will be employed to observe end-to-end processes of learning. Behavioral traces and learning trajectories for different clusters of learners will be explored in a variety of MOOC Learning Designs (LD). The research findings aim to provide useful actionable insights on how adaptations in LD can make MOOCs more inclusive and diverse.

Keywords

cultural clusters · educational data mining · educational process mining · learning design · massive open online courses