Prediction in MOOCs: A Review and Future Research Directions

This paper surveys the state of the art on prediction in MOOCs through a systematic literature review (SLR). The main objectives are: first, to identify the characteristics of the MOOCs used for prediction, second, to describe the prediction outcomes, third, to classify the prediction features, fourth, to determine the techniques used to predict the variables, and, fifth, to identify the metrics used to evaluate the predictive models. Results show there is strong interest in predicting dropouts in MOOCs. A variety of predictive models are used, though regression and support vector machines stand out. There is also wide variety in the choice of prediction features, but clickstream data about platform use stands out. Future research should focus on developing and applying predictive models that can be used in more heterogeneous contexts (in terms of platforms, thematic areas, and course durations), on predicting new outcomes and making connections among them (e.g., predicting learners' expectancies), on enhancing the predictive power of current models by improving algorithms or adding novel higher-order features (e.g., efficiency, constancy, etc.).

I. Introduction

MASSIVE Open Online Courses (commonly known by the acronym MOOCs) are open courses designed to provide educational content to a large number of participants through an online platform, and with free access [1]. The term was coined in 2008, and has become more and more popular since 2012, creating a new paradigm in education. MOOCs originally appeared to enable learners all over the world to gain access to introductory courses from universities. However, currently there are MOOCs about many different thematic areas and levels, and as Christensen et al. [2] have shown, the target learners are not only undergraduate students anymore: Anyone with an interest can take one of these courses.