By Justin Reich
The signature critique of massive open online courses (MOOCs) is that they have low certification rates. From the first mega-courses offered at Stanford and MIT, commentators have observed that a relatively low percentages of students who ever register for a course go on to earn a certificate. Across hundreds of MOOCs now completed, certification rates typically range from 2 to 10 percent when we divide the number of certificate earners by the total number of students who have ever registered for a course.
Some have responded to these concerns by arguing that this way of calculating MOOC certification rates is misleading because it does not account for student intention. Students register for MOOCs for many reasons, and many students have no intention of completing the courses in which they enroll. For most MOOCs, the only way to “shop” is to sign up, so many students who register do so only to evaluate the course. Others who register for a MOOC intend only to audit the course or complete a section of the course, so they register with an intent to participate and learn but not with an intent to complete.
As some have argued, if residential college course completion rates were calculated in the same manner as MOOC completion rates, we would have to divide the number of people who pass a residential course not by the number enrolled in the course at the add/drop deadline, but by the number of people who ever applied to enter the university. A better approach might be to calculate MOOC completion rates as a percentage of students who enrolled in a course with the intention to complete the course and earn a certificate.
Toward a Measure of Intention
Daphne Koller and her colleagues were among the first to make this rebuttal in their EDUCAUSE Review Online article, “Intention and Retention in Massive Open Online Courses.” In the article, Koller and colleagues investigated some of the underlying mechanisms of certification rates by examining attrition rates — that is, the rate at which students stop participating in a course.
The researchers examined MOOC attrition by analyzing the number of hours of lecture video watched by students in Coursera courses. They argued that the distribution of hours of video viewed in these courses could be modeled with a two-component mixture model, in which students are hypothesized to come from one of two groups: a high-retention or a low-retention group. In this model, each group is estimated to have a consistent attrition rate: a high attrition rate for the low-retention group and a low attrition rate for the high-retention group. These differences in attrition over time explain, by the course’s end, the differences in certification rate.
[ Full article available at EDUCAUSE Online: http://www.educause.edu/ero/article/mooc-completion-and-retention-context-student-intent ]