Attrition patterns amongst participant groups in Massive Open Online Courses

Abstract

In their current instantiation, Massive Open Online Courses (MOOCs) represent a form of educational delivery with little barrier to entry. In part due to the ease of enrolling, low completion rates have been a notable characteristic of MOOCs with as few as 5 - 10% of enrolled students gaining a certificate of completion. Beyond these observations, little is known about the patterns of attrition, nor how they may vary amongst different groups of MOOC participants. Activity data was analysed from log records collected from 42 MOOCs run by Stanford University between 2011 and 2013 on Coursera. A novel analysis was applied to the data which showed that attrition varies little between those that audited the course and those that actively participated in assessments. The results indicate that attrition is not influenced by levels of participation in a MOOC and attrition is further not greatly influenced by attributes of the course.