Some Thoughts on the MOOCopalypse

31 Jan 2014moocs

I was obscurely disappointed to discover that I wasn’t the first person to come up with “MOOCopalypse”. I think I could claim first dibs on “MOOCaclysm” if I wanted, since there are no Google hits for it, but it’s not quite as good.

Unless you’ve been living under a rock for the last couple of years, you’ve probably heard of MOOCs (Massive Open Online Courses). These are courses offered online by universities or other suppliers, intended to bring high-quality higher education resources to a wide and varied audience. I’ve done a few of these courses: Peter Norvig and Sebastian Thrun’s Stanford AI course, one on fantasy and science fiction, and more recently Jeff Ullman’s automata theory course and another on general game playing (about which I’ll write something once I’ve had some time to work on the software I developed as part of the course).

There has been a bit of a hoo-ha about MOOCs recently, mostly because Udacity, the company founded by Sebastian Thrun (High Priest and foremost evangelist of MOOCdom, part-time Googler, ex-Stanford) has stepped back from its original lofty goals to concentrate on vocational and professional training. This is obviously a bit of a comedown: from Opening the Academy to the World to “Introduction to Salesforce App Development” (the first course listed on Udacity’s site as I write this).

There’s been a fair amount of Schadenfreude floating around about this retreat/re-positioning. When he left Stanford to found Udacity, there was a lot of high-minded talk from Thrun about the possibilities of MOOCs that frankly sounded a bit snake-oily. He appeared very well-intentioned, and the mission was undoubtedly worthy, but it was hard to believe that it was really going to work. Recent events, in particular a poorly executed attempt to replace some “normal” courses at San Jose State University with MOOCs, seem to show that this was all another unfortunate case of “all talk, no trousers” from the worthies of Silicon Valley.

So what really went wrong? Tressie McMillan Cottom has done some great work analysing both the SJSU fiasco and the wider question of MOOCs–start here and read everything she has to say. Her analysis is informed by a lot of sociological thinking that I don’t know anything about, and she’s spent a lot more time thinking about these issues than I have, but there are a few obvious things to say about the whole MOOC project that I’d like to cover. My teaching experience is all in physics and mathematics, so is probably of more or less no relevance to any other field.

First, let me say that personally, I really like MOOCs. I’ve learnt quite a bit of interesting material using them. But I’ve had the benefit of ten years of post-secondary education, plus a number of years working as a research scientist. I don’t think I’m really in the target audience for the kind of worldwide dissemination of academic knowledge that MOOCs have been sold as providing. After spending so long “learning to learn”, I can teach myself subjects I’m interested in from books or Wikipedia or even original research literature. But I can only do that because I’ve had the good fortune to be able to spend so much time developing these independent study skills. I also have enough background knowledge that I can usually deal with infelicities in presentation and can usually figure out what it’s safe to ignore, what I need to do further reading about, and so on. These are all skills that it takes time to develop.

The first and most obvious problem with MOOCs involves material barriers to access. These come in at least two forms. If you don’t have easy access to a computer and a fast internet connection, most online courses aren’t going to be usable. If you need to use a computer in a public library, it’s not going to be practical to watch 3+ hours of video lectures per week and to do homework exercises as well. The second form of material barrier is that of time. Even if you have access to a computer and internet connection, if you have a full-time job (or jobs) and/or caring responsibilities, simply making the time in your schedule to follow a MOOC may be impossible. I’m an independent software contractor, which means that I can convince myself that time spent on relevant (or semi-relevant!) MOOCs can be counted as continuing professional education so it’s (sort of) work time, but if I had a “real” full-time job, I would find it hard to do these things. If I had a couple of low-paid jobs plus child-care responsibilities, it would be absolutely impossible. Any approach to education that doesn’t address these accessibility issues is never going to be universal, and is never going to meet the originally stated goals of the MOOC movement.

I’ve almost certainly done a lot less teaching than Thrun, but I have done a bit, and I’ve taught in a range of different settings. I’ve done university lecturing in Canada (a second year differential equations course that was fun for everyone involved, and an instance of Calculus I that was as painful for my students as for me); I’ve done tutorial teaching in Oxford (teaching mathematical physics and quantum mechanics to undergraduates on a one-to-one or one-instructor-to-two-students basis); and I spent a few months teaching maths in a classroom setting in a further education college in Wales (this was a course primarily intended for mature students who wanted to return to university to study biology after a period in work, and was intended to equip the students with the mathematical knowledge they would need to start an undergraduate degree course).

What did I learn from those teaching experiences? The first thing is that lecturing is relatively ineffective as a means of transferring knowledge from instructure to students. As an instructor confronted with a room of 80-90 students, all you can do is present the material in a fairly conventional way. There’s a constant conflict between presenting more examples (students in beginning undergraduate mathematics courses love examples) and going too fast. There’s also the need to be entertaining enough to keep students engaged with the material. You can try as hard as you like to make the material “relevant” (whatever that means), but the sad truth is that, for Calculus I, for example, you do need to talk about ten different kinds of subsitutions for different kinds of integrals, and that is unavoidably tedious. And of what benefit to students are lectures really? A good lecturer can give a different viewpoint on the material compared to the textbook, but that presupposes that the students have the time and inclination to read the textbook to give themselves a baseline for that different viewpoint.

I would have liked to try a “flipped classroom” approach, but that wasn’t really feasible in the setting we had.

So lecturing isn’t particularly effective. This seems to correspond with what other people find (whatever you think of the “Learning Pyramid”, putting lecturing at the bottom in terms of effectiveness seems about right). What about the most effective teaching I’ve done? Without a doubt, this was the small group tutorial teaching I did in Oxford. In this setting, you have one or two students who’ve done some work (in this case, it was undergraduate physics work of various kinds), handed it in for you to mark, and you meet for an hour or so to talk about the topic at hand. I was on the receiving end of this instructional style as an undergraduate myself, and it can be quite intimidating. There’s nowhere to hide if you don’t know something. Given a well-intentioned instructor though, it can be amazingly effective. If a student doesn’t understand a concept the way that it’s been presented in lectures or in their textbooks, the instructor can explain it some other way, using some other example that might be more accessible.

It’s worth saying that this instructional setting can also be stressful for the instructor: there’s absolutely no possibility of bluffing–you need to know the material backwards and forwards and moreover, you need to know seven different ways to look at anything so that you choose another angle to explain things if the first few approaches don’t take. The peak of my understanding of quantum mechanics was definitely the two or three months when I was teaching it to undergraduates at St. Catherine’s in Oxford. It’s a really great feeling to see understanding dawn in someone as you explain really quite difficult concepts to them, adapting your explanations to your model of their understanding. It feels like real TEACHING! That’s an experience that’s sadly not available in most other settings.

When I was teaching in Canada, I really tried to push the benefits of office hours to students. There were a few who turned up religiously at every session, and I really think they got some benefit from it. Office hours aren’t quite as good as tutorial-style teaching, mostly because you have many more students to deal with, so it’s hard to remember what issues individual students have trouble with, but you can use time at the blackboard with students to try to develop some problem-solving skills and “learning to learn” and “learning to think” skills. When a student asks “How do I solve this problem?”, you can hand them a piece of chalk, sit yourself down and say, “Well, how would you start?”. From there, you can guide them as little or as much as they need, you can suggest things, you can ask them why they take a particular approach to a problem (and so diagnose misconceptions that they might have) and, if you do it right (and you’re lucky), you can give the student the feeling of accomplishment of solving a problem that they didn’t know how to solve.

So, what does this all have to do with MOOCs? Most MOOCs have video lectures, on-line homeworks composed of multiple choice or short answer questions (or, in some cases, peer marked written answers), and discussion forums (which may or may not be monitored by the instructors). There may be a recommended textbook (which may be another financial barrier to entry for some students) or lecture notes available as well. There’s no space in this setup for small group teaching or interactive feedback and explanation. You watch the lectures, try the exercises, and look for help on the discussion forums. If you’re lucky, your fellow students or the instructors will answer your questions. I’d contend that in that setting, there’s really relatively little teaching going on.

What’s more, the resources offered by MOOCs are wholely inadequate for the claimed target audience. Video lectures, self study with exercises and notes and some non-realtime online discussion might be enough for those of us fortunate enough to have been primed for it by a lifetime of “in person” education, but for students who haven’t benefited from that, these resources just aren’t enough. There are a whole range of “soft” skills that are needed: you need to know how to listen to lectures, to deal with errors by the lecturer, to know when obscure points can be disregarded until later, to know how to look for and how to understand further information if the lectures aren’t enough or if you have lacunae in your knowledge that you need to fill to understand the lectures. Those independent study skills aren’t things you can teach using the current MOOC model. Once you have those skills, MOOCs can be pretty useful.

Which all makes me wonder just what the hell was Thrun thinking? The promise of MOOCs providing high quality higher education to everyone seems like a pipe dream. The students most in need of access to MOOCs are the students least likely to have the material resources to access them. The students who could benefit most from MOOCs are the students least likely to have the pre-requisite skills needed to benefit from the limited range of teaching possibilities currently offered by MOOCs. It just seems kind of silly to make these great claims for an unproved teaching platform, one that on the face of it simply doesn’t have a hope of meeting the needs of the people who need it most.

I don’t for a moment doubt Thrun’s good intentions, but good intentions don’t get you very far. The SJSU situation, where all of these difficulties of access came to a head because of an attempt to replace some conventional courses with MOOCs (with students paying for them), seems like the result of gross hubris on the part of Thrun’s company, Udacity. At least Thrun now seems to recognise this. He said, “We were on the front pages of newspapers and magazines, and at the same time, I was realizing, we don’t educate people as others wished, or as I wished. We have a lousy product.” It’s a shame he and Udacity didn’t recognise that before they exposed a whole bunch of SJSU students to their lousy product. These things have real-life consequences for real students and you shouldn’t use students as experimental subjects to help improve your product, which is what seems to be happening.

One final question that I’ve been mulling over a little: Can MOOCs really help with broadening access to education? Setting aside the material barriers to access, which are important but not something that MOOCs are any kind of solution for, what about the other barriers? I could parade my ignorance of educational psychology and pull some ideas out of my hat, but it turns out that there are people who do actual real research on these questions! If you’re at all interested in this question, you should head to another of Tressie McMillan Cottom’s articles here and read some of the links. There’s a lot of good stuff there. It seems that there’s some hope. It would be depressing if technology couldn’t help. But it’s going to require something a little more subtle than gung-ho evangelism and “disruption”.