CRI researchers Jake Wintermute and Ariel Linder, along with CRI alumnus Matthieu Cisel, have published a new study on the dynamics of dropout in online courses. The manuscript, entitled “A survival model for course-course interactions in a Massive Open Online Course platform”, was published last week in the journal PLoS ONE.
Digital learning, once an emerging trend on the bangs of education, has suddenly become the new norm. Yet little is known about the success factors of online courses. The authors set out to create a mathematical description of MOOC certification rates that could tell us more about why students persist or drop out. They were inspired by the “survival models” that have often been used to describe the lifespan and longevity of living organisms.
“It was a collaboration very specific to the CRI environment,” explains Jake Wintermute, lead author of the study. “We had people studying complex systems, aging and education, but it's only at CRI that you see all these different ideas working together.
The study was based on a large dataset comprising over 1,000,000 registrations to France Université Numérique (FUN), a French MOOC platform. On average, only 8.1 % of registrations resulted in a certificate of completion. Wintermute and colleagues identified user- and course-specific factors that correlate with a completed course. But the most important factor seems to be registration waves: when users register for several courses in a short space of time.
“We've found that students who enroll in many courses get about the same number of certificates as students who enroll in few courses,” Wintermute explains. “It's as if students who start out taking many courses eventually choose to focus on just one. It makes sense when you consider how easy it is to sign up for these free, open-access courses. There's really no reason not to sign up for a course that looks interesting, and there are no drop-out fees if you change your mind.
Beyond the ease of access and dropout of multiple enrollments, the study revealed several more complex factors that influence completion rates. For example, students who choose to enroll in difficult courses, defined as courses with low completion rates, tend to do better in all their other courses.
“We hope this study will pave the way for a more integrated understanding of online education,” said Wintermute. “Everything you learn influences everything you learn elsewhere. If we can stop looking at courses in isolation, we may be able to give students the preparation they need to avoid burnout.”
Reference
Wintermute, E. H., Cisel, M. & Lindner, A. B. A survival model for course-course interactions in a Massive Open Online Course platform. PLoS ONE 16, e0245718 (2021).
