MOOCs: More Data, More Answers, More Questions

01/23/2014 04:57 pm ET | Updated Mar 25, 2014

To a certain extent, the unstated storyline associated with the two most famous MOOC statistics (100,000+ plus enrollments and 95 percent drop-out rates) never made much sense.

The implication (certainly by MOOC boosters) was that so many people hitting the enroll button on Udacity's, Coursera's or edX's website translated to tens of thousands of students ready to do college-level work in an online course, regardless of the challenges inherent in teaching and learning at a massive scale. And the fact that only one in 20 actually went the distance to earn a certificate meant (at least for MOOC critics) that for every 100 students who signed up for a MOOC, 95 didn't like what they saw and bagged out.

I wrote earlier about how much these stunning numbers don't tell us about the behavior of students enrolled in a massive online class. And with the release of new research by HarvardX and MITx, we now have a better understanding of what students are doing when they participate in a MOOC (other than signing up for it and either completing or not completing it).

There has actually been other research based on data pulled from MOOC databases used to try to figure out what was going on among the student population (one of my favorite being this piece co-authored by, among other people, the founders of Coursera). But the HarvardX/MITx results -- just released this week -- provides the most systematic analysis yet of what is going on among my MOOC classmates.

This overview report summarizes what was discovered by analyzing six Harvard and 11 MIT courses that have been delivered to completion at least once. And if you're as big a data dweeb as I am, you may want to look at the individual reports from Harvard and MIT associated with each course.

It is actually quite valuable to at least think about each course as a separate entity, rather than lump them (and every other massive online class) together under the label "MOOC," since (as the researchers themselves highlight), no two MOOCs are quite the same. Some, for example, try to replicate a full-semester course as closely as possible with technology the only limiting factor in including all the syllabus material one would find in a classroom version of the course. In contrast (as I discovered during my One Year BA project) some MOOCs are created by professors who, liberated from the tyranny of the semester, want to teach a subset of what they might teach in a full credit class.

If you look at some of the data charts I included in this piece (drawn directly from two of the HarvardX research reports), you'll discover that every type of student behavior you can imagine seems to be represented in the makeup of a MOOC class. For example, a category the researchers call an "Optimizer" represents the person who does the absolute minimum to earn a certificate. And while such an individual could be denigrated as a "cheater," I think they more represent the type of homework hacking phenomenon I talked about a few weeks back.

More interesting are those who watch most or all of the course videos who don't pass (or even bother with) graded assessments. While the report refers to them as "Listeners," "Auditors" would probably be a better description since they are the equivalent of those who might sit through the lectures in a traditional classroom-based course where they would be counted as a category separate from "drop-outs."

Personally, I find the "Completionists" -- people who engage with all the course material and earn a grade far higher than is necessary to earn a certificate -- the most interesting. For these are people who, like me, signed up to take one or more MOOCs in order to learn as much as they could. And given the relatively low threshold for passing (most MOOCs have cut-off scores of 50-60 percent), for those devoted to learning passing is just a welcome byproduct.

Having gotten past the notion of MOOCs destroying the educational multiverse (for good or ill), it is now time for a more measured conversation about what this extraordinary educational technology can and cannot do, now that we know it isn't about to replace traditional higher education overnight. And measurement is always a matter of data combined with analysis, with an added dose of humility regarding what we still do not know (and might never now with certainty).