Shorter Jonathan Haber: If I just make numerical assumptions as favorable as possible, I can get a MOOC class completion rate up to 48%:
Using the number Coursera sent him of “Total Registered Students” (i.e., the number of people who hit the Enroll button) as a denominator does indeed give you a completion percentage of 5 percent. And if you instead use Total Active Students (the number of students who logged onto the site at least once after registering) that completion rate climbs to 10 percent (still within the range MOOC critics use when they complain about attrition).
But if you use the number of students who watched at least one video as your denominator, completion percentages climb to 15 percent. And if you make the assumption that only students who complete at least one assignment (even a short quiz at the end of lesson 1) should be considered serious enrollees, his completion rate skyrockets to 48 percent.
I do wonder if completion rate is the same as passing rate, but Haber never tells us, and there’s no link to check the background data. Basically, by handwaving as hard as he can, Haber has taken the completion rate for a MOOC course from completely appalling to just abysmally bad.