Using Clickstream to Understand Learning Paths and the Network Structure of Learning Resources: Using MOOC as an Example

Abstract

Massive open online courses (MOOCs) have attracted great attention from the public and learners. However, their high dropout rates have been criticized over the past few years. There has been a body of work dedicated to using learning analytics to provide feedback for instructors and designers to improve learners’ engagement and retention rates. However, the open and flexible nature of MOOCs has often been overlooked in these analytics studies. In this work, we used clickstream data to construct a flow network model in order to identify MOOC learners’ learning paths and the network structure of available learning resources from an open system perspective. We found that learners tend to adopt linear learning paths within chapters and continue to watch video lectures in next chapters instead of taking quizzes at the end of the chapters, and they rarely review previous chapters during studies. We also found that learning paths of all learners have formed a centralised network structure which implies that certain learning resources in such a network structure dominate the way in which learners learn in MOOCs.