Understanding learner engagement and the effect of course structure in massive open online courses

Abstract

In this thesis, we analyze learner performance in two edX programming courses. We look at many dierent types of learners, such as learners who have taken both intro and advanced courses, learners who opt to pay for certification, learners who take the experimental self-paced course, learners who eventually become community teaching assistants, and learners who take the course after the implementation of gating. In particular, we focus on repeat learners, or learners who have taken the course multiple times. When courses undergo a change from semester to semester, for example changing the pacing of the course or making certain content only available to paid users, it can be very useful to look at learners who were in the course before and after this change. Our goal is to gain a baseline understanding of how dierent factors affect learner behavior and how a few changes that edX has made to courses affect learner performance. With the best understanding of how learners interact with and complete courses, edX instructors will be able to provide the best possible online education experience for their learners.