Exploration of educational empirical research based on big data from MOOCs platforms

Abstract

We reviewed numerous course reports and research papers generated through massive open online courses (MOOCs) and generally found two phases of MOOCs research, with a turning point in 2012, the so-called year of the MOOCs. Among these sources, research analyzing educational big data of MOOCs can be sorted into three groups: (1) Learning pattern and learner categories; (2) learners completing or dropping out of courses; (3) specific learning behaviors and different functions of the MOOCs platform. However, we find limitations in the existing data. This insufficiency is partly derived from the lack of conventional educational and psychological theoretical frameworks in MOOCs research based on big data from online platforms. We point out that researchers could use MOOCs platforms as valid data resources to further verify and to extend conventional educational research theories and frameworks, such as learning theories, pedagogy and educational policy. We also analyze different data types that MOOCs platforms offer and their roles in educational data mining (EDM). MOOCs data include user logs, learning paths and performance results. It is important in future research to combine conventional theories with options presented by new data. This paper presumes there will be three types of research based on big data from MOOCs platforms. Such research will mainly be on: the MOOCs platform itself; MOOCs design and delivery; EDM and learning behavior analytics.