Tuesday, January 21, 2014

MOOC-rize this!

Participant-centered learning is not scalable, so it's MOOC-resistant.

A couple of colleagues (in different fields) have shared MOOC-related worries with me. The logic goes, our research jobs are funded to a large extent by teaching, and if the need for teachers disappears, many schools will stop hiring expensive research faculty. Cathy "Mathbabe" O'Neil suspects MOOCs will have tragic consequences for mathematics research.

I'm not convinced.

As I see it, there are three main MOOC threats to traditional higher education: cost-effectiveness, brand equity of the schools offering the MOOCs, and quality of content. There's also one main visible weakness, certification.

Cost-effectiveness. The cost-effectiveness of MOOCs is the main argument I hear for "the end of universities as we know them," to which I say: if you can replace class X with videos of lectures and computer-graded problems sets, good riddance to class X.

Distance learning is an old proposition, it started with something called "a book." MOOCs add better media, the possibility of computer graded problem sets (for some fields, and requiring a significant investment in problem set design), and tutoring or discussion affordances.

But here's the crux: the scalable parts of MOOCs are the easy part of education. The hard part is motivating students, interacting with them and being responsive to their questions, taking the time to understand the reason for their incomprehension, and reacting in real time to information they bring into the class or developments in the field.

So, while MOOCs will work really well for highly motivated, studious students (nerds like me), the average student will need more personalized attention than is cost-effective to offer in large scale.

Repeat after me: Personalized attention doesn't scale.

True, many classes in many institutions of higher learning don't deliver anything more than the scalable parts of the MOOCs; no personalized attention or significant interaction with the students at all. Those classes are ripe for replacement by MOOCs, and that's good.

This is what gets me steamed about Mathbabe's post: if the professors don't add value to a student reading the textbook and solving the problem sets (that in many cases are straight off teachers' manuals from the textbook publisher and graded by teaching assistants), then what is the purpose of hiring someone with a deep understanding of the field a/k/a a research faculty member?

The answer to that lies in the value of an instructor with a deep understanding of the field to manage participant-centered learning (now called "flipped classroom" but in fact the only way anyone ever really learned any technical material was by practicing it).


Brand equity. Who wouldn't rather say "I took the Caltech Machine Learning course" rather than "I took the Cal State-Moraga Machine Learning course"? This is indeed a problem, but to a large extent it's a matter of brand credibility footprint, not a technological issue.

With prestigious schools creating extension campuses and joint ventures with other universities, MOOCs are only a small part of the problem. And let's remember that brand extensions are not one-way propositions; MOOC-rizing courses may dilute a school's brand equity. (So may having extension campuses, of course. Armani Exchange doesn't help the brand equity of Armani.)

Talking to some Hahvahd B.S. colleagues, I got the distinct impression that they believe the student physical presence in their Cambridge (Allston, really) campus is an essential part of the brand identity, one that they are not willing to compromise on. I'd venture that at Hahvahd B.S.  they know a thing or two about the network and identity dimensions of brand equity.

So, I agree that the brand equity is an issue, but more because of extension campuses and joint ventures than MOOCs, since the brand credibility footprint is much more likely to encompass the former than the latter. (Says the visiting professor at TheLisbonMBA, a joint venture of UCP, UNL, and MIT.)


Quality. Obviously there's a difference between the quality of the classes taught at Caltech and at [the fictional] Cal State-Moraga; and that is part of the brand equity of Caltech. But the real question is whether the students of CS-Moraga are going to benefit from a class that was designed for Caltech students more than from one that was designed specifically for them.

Note that this immediately raises the question of whether CS-Moraga classes are customized to their student population (that is now, before being MOOC-rized). And that's again the issue of what faculty are doing at CS-Moraga: if they rely on the textbook and the teachers' materials provided by the textbook publisher in order to save themselves the trouble of actually preparing a class, then as I said above, good riddance.

On the other hand, in participant-centered learning the instruction follows from the participants' needs and skills, moving at their pace, therefore for good quality the instructor must have a broad training in the general field and a deep understanding of the materials of the class.

It's incumbent upon the faculty to make itself more valuable than a cost-effective MOOC, or a textbook for that matter. Otherwise, it's their own fault if they're MOOC-rized


Certification. Certification of knowledge is the weak point of MOOCs as they currently exist, but it's important to note two issues with this.

First, certification cannot be the only function of universities or research faculty, as certification alone doesn't require the large infrastructure and cost of a university or the need for broad research programs.

Second, and much more critical, if the MOOC certification weakness is part of the advantage of a traditional university, that weakness ends if universities stop taking their certification responsibilities seriously. When some schools graduate computer engineers who never wrote a program that passed a compiler's syntax check, let alone run, let alone run correctly or efficiently — to choose an example I heard from someone I trust — then the credibility of universities as certification mechanisms comes into question, and their advantage vis-a-vis MOOCs in this regard evaporates.

(Yes, there's a third possible issue, that of MOOCs adding some sort of credible certification. I believe that that's a long way off, given how it would require (a) an infrastructure to prevent fraud; (b) some sort of long-term evaluation, since not everything can be certified with a short test; and (c) legal protection in case of unacceptable demographic results in aggregate, which universities seem to have had grandfathered in, but other institutions have found themselves liable to.)



I for one welcome our new MOOC multimedia limited-interaction e-textbooks for the 21st Century. As a complement to real instruction: customized, personal, and responsive. And as a mechanism for making universities take certification seriously.