Predictive models of student success in Massive Open Online Courses (MOOCs)are a critical component of effective content personalization and adaptiveinterventions. In this article we review the state of the art in predictivemodels of student success in MOOCs and present a dual categorization of MOOCresearch according to both predictors (features) and prediction (outcomes). Wecritically survey work across each category, providing data on the data source,feature extraction from raw data, statistical modeling, model evaluation,prediction architecture, experimental subpopulations, and prediction outcome.Such a review is particularly useful given the rapid expansion of predictivemodeling research in MOOCs since the emergence of major MOOC platforms in 2012.This survey reveals several key methodological gaps, which include extensivefiltering of experimental subpopulations, ineffective student model evaluation,and the use of experimental data which would be unavailable for real-worldstudent success prediction and intervention, which is the ultimate goal of suchmodels. Finally, we highlight opportunities for future research, which includetemporal modeling and research bridging predictive and explanatory studentmodels.
展开▼
