One widely quoted dropout figure for students in massive open online courses is 90 percent. The number would be staggeringly high for a traditional class and has been used to cast doubt on the promise of MOOCs.
The number is simple to come up with: take the number of users who register for a course and compare it to the number still participating at the end. But is it fair?

Some researchers say MOOC dropout figures being bandied about do little to describe why hundreds of thousands of people across the world are signing up for MOOCs in the first place. All but a few of the courses offered by MOOC providers are free and don’t earn students any college credit. There are also no enforced prerequisites as there are for normal college courses.

That’s why it may not make sense to compare the number who register to the number who finish. The widely cited numbers may be “largely missing the point,” said Andrew Ho, a Harvard University assistant professor of education who is involved in some MOOC-related research.

“Measuring the MOOC Dropout Rate,” Ry Rivard

I suppose that MOOC hyperbole has a predictable half-life; after a certain amount of time, the claims decay down to the point where we might consider them realistic. The first MOOC claims were rooted in the huge numbers of students signing up for them, a set of numbers used to sell the idea of the MOOC to an increasing number of schools. It was instant karma at work, a quick fix that must have made these schools feel both generous and modern.

I think there might be real research potential in these courses, simply because their sheer size and the resources behind them allow for some interesting experiments and for large-scale data collection. Or, rather, they could facilitate this sort of research if they were able to keep more than 10% of their students in the classroom for the duration of the class. If they cannot, well, MOOC’s may turn out to more (expensive) fad than intellectual meme.