Musing on MOOCs

Mark Guzdial’s blog contains a number of posts where he looks at Massive Open On-line Courses (MOOCs) but a recent one on questionable student behaviour made me think about how students act and, from the link where students sign up multiple times so that they can accumulate a ‘perfect’ score for one of their doppelgängers, why a student would go to so much trouble in a course. As the post that Mark refers to asks, is this a student retaking the course/redoing an assignment until they achieve mastery (which is highly desirable) or are they recording their attempts and finding the right answer through exhausting the search space (which is not productive and starts to look like cheating, if it isn’t actually cheating – it’s certainly against the terms of service of the courses.)

Why is this important? It’s important because MOOCs look great in terms of investment and return. Set up a MOOC and you can have 100,000 students enrol! One instructor, maybe a handful of TAs, some courseware – 100,000 students! (Some of the administrators in my building have just had to break out the smelling salts at the thought of income to expenditure ratio.) Of course, this assumes that we’re charging, which most don’t just for participation although you may get charged a fee for anything that allows you to derive accreditation. It also assumes that 100,000 students turns into some reasonable number of completions, which it also doesn’t and, as has been discussed elsewhere, plagiarism/copying is a pretty big problem.

Hang on. The course is free. It’s voluntary to sign-up to in the vast majority of cases. Why are people carrying out this kind of behaviour in a voluntary, zero-cost course? One influence is possible future accreditation, where students regard their previous efforts as a dry-run to get a high percentage outcome on a course from a prestigious institution. I’ll leave those last two words hanging there while I talk about James Joyce for a moment.

If you know of James Joyce, or you’ve read any James Joyce, you may be able to guess the question that I’m about to ask.

“Have you read and finished Ulysses?”

Joyce’s Ulysses is regarded as one of the best English-language novels of the 20th Century. However, at over 250,000 words long (that’s longer than the longest Harry Potter, by the way, and about half the count of Lord of the Rings) , full of experimental techniques, complexity and a stream-of-consciousness structure, it isn’t exactly accessible to a vast number of readers. But, because it is widely regarded as a very important novel, it is often a book that people are planning to read. Or, having started, that they plan to finish.

However, the number of people that have actually read Ulysses, all the way through and reading every word, is probably quite small. The whole ‘books I claim to have read’ effect is discussed reasonably often. From that link:

Asked if they had ever claimed to read a book when they had not, 65% of respondents said yes and 42% said they had falsely claimed to have read Orwell’s classic [1984] in order to impress. This is followed by Tolstoy’s War and Peace (31%), James Joyce’s Ulysses (25%) and the Bible (24%).

So, having possibly neither started nor finished, they claim that they have read it, because of the prestige of the work. 42% of people claim to, but haven’t read 1984, which, compared to Ulysses, is positively a pamphlet – a bus ticket aphorism in terms of relative length and readability. And we see that the other three books on the list are large, long and somewhat ponderous. (Sorry, Tolstoy, but we don’t all get locked into our dachas for 6 months when it snows.) 1984, of course, is in the public eye because of the ‘Big Brother’ associations and the on-going misinterpretation of the work as predictive, rather than as an insightful and brooding reflection of Eric’s dislike of the BBC and post-war London. (Sorry, that’s a bit glib, but I’m trying to keep it short.)

I have read Ulysses but I think it fair to say that I read it, and forced myself to complete it, for entirely the wrong reasons. Now that I enjoy the work of the Modernists far more, I’m planning to return to Ulysses and see how much I enjoy the journey this time – especially as I shall be reading it for my own reasons. But, the first time, I read it and completed it because of the prestige of the work and because I wanted to see what all the fuss was about. (Hint: it’s about a day in Dublin.)

I think there’s an intersection between the mindset that would make you claim to read a book that you had not, for reasons of prestige rather than purely for tribal membership, and that required to take a MOOC from Stanford, Harvard or Berkeley, and to falsify your progress by copying answers from other people or by solemnly duplicating your identities to accumulate enough answers to be able to ‘graduate’ summa cum laude. In this case, taking a course from one of these august institutions, especially in these days of the necessity of having a college degree for many jobs that have no professional requirement for it, is better than not. Completing assignments to a high standard, however you achieve it, may start to define your worth – this is a conjunction of prestige and tribalism that may one day allow you to become a graduate of University X (even if it is tagged as on-line or there is subsequent charging or marking load for accreditation).

And here we find our strong need for real evidence of the efficacy of the MOOC approach. Let’s assume that we solve the identity problem and can now attach work to a person reliably – how will we measure if someone is seeking mastery or is actually trying to cheat? We can ask that now – is the student who seeks questions to previous examinations testing their understanding and knowledge or conducting a brute force attack against our test bank? If MOOCs can work then the economies of scale make them a valuable tool for education but there are so many confounding factors as we try to assess these new courses: high sign-up rates with very low completion rates, high levels of plagiarism, obvious and detectable levels of gaming and all of this happening before they actually become strong alternatives to the traditional approach.

It would be easy to dismiss my comments as those of a disgruntled traditionalist but that would be wrong. What I need is evidence of what works. I have largely abandoned lectures in favour of collaborative and interactive sessions because the efficacy of the new approach became apparent – through research and evidence. Similarly for my investigation into deadlines and assessment, evidence drove me here.

If MOOCs work, then I would expect to see evidence that they do. If they don’t, then I don’t want students to sign up to something that doesn’t work, potentially at the expense of other educational opportunities that do work, any more than I want someone to stop taking their medication because someone convinces them that unverifiable alternatives are better. If MOOCs don’t quite work yet, by collecting evidence, maybe we can make them work, or part of our other courses, or produce something that benefits all of us.

It’s not about tradition or exclusivity, it’s about finding what works, which is all about collecting evidence, constructing hypotheses and testing them. Then we can find out what actually works.