The Economics of Online Education

The Economist covers the economics of online education:

Alex Tabarrok…reckons the most salient feature of the online course is its rock-bottom marginal cost: teaching additional students is virtually free. prices converge towards marginal cost, there will be little scope for undercutting the competition. Instead MOOCs are likely to compete on quality…Higher production costs are a small price to pay to attract much greater numbers of students. Such markets often evolve into winner-take-all, “superstar” competitions. The best courses attract the most customers and profit handsomely as a result. In this respect online education may more closely resemble information industries such as film-making than service industries such as hair-cutting.

The market for textbooks already fits this description. New textbooks are costly to write and design but can be reproduced fairly cheaply. Not surprisingly, only four introductory economic texts account for half of the American market, according to Mr Tabarrok. Indeed, says Tyler Cowen, a co-founder of Marginal Revolution University, it is possible that textbook publishers are better equipped than universities to develop MOOCs profitably.

I agree also with a point made by Caroline Hoxby:

Less selective institutions are close substitutes for MOOCs. Course content is often standardised and interaction with professors is limited in order to keep costs down.

…Elite institutions face very different circumstances, Ms Hoxby reckons. They operate like venture-capital firms, offering subsidised, labour-intensive education to highly qualified students. They aim to cultivate a sense of belonging and gratitude in students in order to recoup their investment decades later in the form of donations from successful alumni…. For top schools, the best bet may simply be to preserve their exclusivity.


I'm taking my first MOOC now through EdX and it's not bad but I really don't see how this is a replacement for any kind of higher edcuation, if I didn't already know at last 50% of the material I think I#d be pretty lost, it's really more akin to the sort of autodictative learning which has gone on forever - doesn't work for most people and you don't learn as much as you think.

They're still in their infancy. I haven't taken one, but any thoughts on exactly what is missing from your EdX experience?

It's too much just like reading a book, except on a computer - if I were taking a real class at a university I would head to the library to study and make a point of keeping my smartphone off and never ever bringing a laptop. Just distractions. The main problem is that compared to a normal lecture I see fewer examples being done with the idea that you just rewatch the same video multipe times, which I don't think is a substitute for a longer lecture with more examples.
For me the class is easy, I know a good chunk of the material already and it's aimed at a fairly low level. However if I were taking this course without my background I would probably give up.

CBBB I am curious which class you are taking because based on your comments your experience seems very different from mine (also from edx). I have taken a couple introductory physics classes from MITx and an introductory chemistry class, also from MITx. In all 3 classes I've taken the video lectures are from the actual class, and each one is included in its entirety.

You mention below that another problem with the classes is that it's too much like reading a book, which I also have not experienced in the classes I've completed. The homeworks reveal a lot of material I thought I knew but didn't, and were the best part of the courses IMO. I would estimate that 80% of my "learning" occurred through the homework assignments and exams that moved me from "thinking I know something" to "actually knowing something" (or at least in that direction).

I'm not doubting what you say is true, but I think it may be more a function of that particular class than of edx or MOOCs more generally.

CBBB is right and we all know it. Most people lack the motivation and self-discipline to teach themselves entirely new subjects while sitting alone at their computers. "Will I ever need this? Will I ever be asked to prove I know this? Will anyone ever be impressed by my knowing this? Will anyone I care about care if I don't know it?" If the answer to any of these questions ever seems like it might be "no", boom, window closed, game over. Or rather game on. Time for some XBox!

Though the upfront costs do tend to be more than just a $4 app and youtube - more than 1200 inQbation project hours aren't that cheap at NoVa rates -

Of course, TANSTAAFL applies to any online course - it is just that with MRU, it is not easy to discover who is the one paying, as the original creation story was obviously intended for nothing but PR purposes.

I'm sure you have a point..... Well, not really sure.

It faces an uphill battle, but online education will prevail. As anyone who reads a book has figured out, there is nothing magical about knowledge that requires it to be imparted by a human.

OTOH, there's no way ever that I would have had the motivation & discipline to self-study a tenth of the stuff I did with just a book, a video or a website.

Yes exactly very few people succeed at being autodictats. I don't see how MOOCs change the equation much at all. When you are learning a complicated and difficult subject you don't know what you don't know and that is why just picking up a book often doesn't work, unless you already know a good chunk of what's in the book and just need to expand upon your base knowledge.

Yeah, I get why most of the attention now is focused the MOOCs themselves, but I find the future role of teachers (maybe best described as coaches) to be more interesting and indicative of how things will turn out. Better teachers, and better classmates, will still result in a better equipped student.

MOOCs are really just very, very advanced textbooks. Obviously there is a lot to be gained there, but I don't see how they can fundamentally alter the educational landscape in the way that some people seem to think.

Try Duolingo. Online education can be quite addictive, provided that people who design it can free their minds from the "it must look like a textbook" approach.

Well language learning requires a very different approach than other topics.

But doesn't that just mean they need to add incentives - like grades and degrees and such - to motivate you to get your work done? All my college education was basically self teaching with the benefit of a structured format and concern for grades to keep me somewhat on task.

But then again a lot of people don't do their work in normal brick and mortar colleges.

This is exactly what the complaint by CBBB looks like- without the chance to pay for a class and get a failing grade, he would never do the work. Indeed, he isn't failing at the MOOC he is taking, and claims it is just because he already knows 50% of the material.

I guess I have never understood people who claim to learn more in a lecture in a classroom than they ever do from reading the textbook and working through the problems.


Are you, by chance, an engineer?

Anyhow, most people aren't autodidacts. Period. Also, some people genuinely absorb and retain information better getting it live and in-person from another human. That's just how they're wired.

Being an autodidact means you can also really just learn the parts and things you want to learn without bothering with the rest. You can end up with huge blind spots that way.

Well you're clearly discounting labs, and the presence of physical objects. Some people's jobs still involve manipulating objects (or themselves) in the physical world.

One parallel here may be the investment banking business circa 2000. At that point in time all the banks were convinced they were in the middle of a huge transformation of their business because of the internet bubble. Goldman started a large skunk works effort to develop a high end Charles Schwab (then canned it). Morgan Stanley was trying to put all of its capital markets origination services online (then canned it). There were hundreds of strategic re-engineering meetings. Eventually the investment banks realized that they were better off keeping their opaque and exclusive business models, and to use the internet only in ways that benefited them, not their clients. They were right. Most of the client facing investment banking, capital markets and OTC trading businesses that are profitable haven't changed much through the internet era.

I like that Hoxby distinguishes between non-selective institutions -community colleges with non-traditional students - and selective institutions. Bold predictions about MOOCs never seem to recognize that these operate in very, very different markets. I'd be very scared if I were running a community college right now, or a 3rd-tier public univ with huge lecture classes. Top-tier colleges? Not so much.

Right. Which is exactly why you see places like Southern New Hampshire University take the plunge so aggressively.

I very much enjoy on-line classes, those I take being college courses offered on i-tunes university. Of course, I don't really "take" the courses as much as watch the lectures and read recommended texts. Which brings me to my comment: on-line education and the Socratic method are mutually exclusive. I'm a lawyer, so my memories of school are memories of the Socratic method, the predominant method for teaching the law (at least it was when I attended law school many years ago), but my memories of college courses, the ones worth remembering anyway, are of a lively interaction between the lecturer and the students in the classroom, the lecturer challenging (intellectually) the students and the students challenging the lecturer. Watching my first on-line course I was taken aback by how little interaction there was between the lecturer and the students in the classroom. Sure, the lectures are entertaining but what's the point of sitting in the classroom. Of course, that's the point: there is no point of sitting in the classroom if the lecture is the sole method for "teaching" the students. Hence, on-line education.

The marginal cost of conventional education is not that high if merely stuffing more students into giant lecture halls would do the job. It's the grading, labs, projects, office hours, tutorials, libraries, journals, field-trips, workshops etc. that add to the marginal cost.

When we applaud MOOCs for rock-bottom marginal costs are we forgetting this?

Yes, in addition to the other comments about the difficulty of effectively learning online (due to the requirement that the students be able to focus themselves), the cost issue is the other half of the MOOC myth.

Zero marginal cost? Sure, if we like having the 90%-95% dropout rates that MOOCs currently have. True education requires giving the students help. If that were false, libraries would have replaced schools centuries ago. But most students can't teach themselves from books, and they won't learn much from MOOCs either, unless the MOOC provides the stuff that a classroom provides -- the ability to interact with other students, TAs, and if the class is small enough the professor (but in MOOCs the class will not be small enough), etc. And boom, there goes to zero marginal cost. Effective education requires both concentration on the part of the students, and resources to be deployed by the educator.

Is there really more opportunity to interact with your professor at Harvard than at a community college? The Harvard professor may be more interested in research than in answering your questions.

Or even more likely, the Harvard professor isn't even giving the lecture. His TA is.

It's not the professor who matters so much (although they do matter some). It's the students sitting next to you, and who you interact with outside of class. And the fact that you're in a classroom, with almost no choice but to pay attention.

Yes there are some students who fall asleep in class or who play solitaire laptop or whatever.

But that's why bricks-and-mortar classrooms are still advantageous over online classes -- because if a student is that inattentive in the classroom, imagine how inattentive they'll be perched in front of a computer. Cat videos and facebook, rather than learning, will be the result.

Of course the main reason why MOOCs won't go anywhere is that the REAL purpose of formal education is to get you into the guilds. You want to be an accountant, lawyer, doctor, professional engineer? You have to go to a real school and get a real degree. Tyler Cowen ignores all this of course and lives in a fantasy world where education and careers are all about learning actual skills and the only profession out there is computer programmer.

"Less selective institutions are close substitutes for MOOCs. Course content is often standardised and interaction with professors is limited in order to keep costs down."

I hypothesize that the faculty of the directional state university in my town have vastly more direct contact/interaction with undergraduate students than 90% of the faculty at George Mason, and that course content is as varied.

Let's be realisitic here though, is GMU what one would consider selective? The school I graduated from is ranked way higher than GMU and that place wasn't remotely selective in my opinion.

GMU is considered "Selective" by US News & World Report. The institution in my town is much closer to open admission.

My point is that less-selective teaching institutions probably offer considerably more, not less, student-faculty interaction than R1 and R2s. Freshman and sophomore-level classes at GMU can probably be replaced more easily by a MOOC than one at a community college or Southwest Bumblehole State University.

The key to making MOOCs more appealing is for there to be tests (by an accredited notional body) that then ranks the students and awards them certificates based how much they know (much like a CPA license exams). Without that, there won't be much appeal to MOOCs because there is not way for employers to determine how good MOOCs are at teaching students useful skills.

If that sort of accreditation does work, MOOCs might be pretty redundant anyways. A good textbook, already free lecture videos & one of the many excellent forums online will get you most of the way.

Online education, already working wonders!

My main concern with online education is cheating. Distance creates lots of temptations -- just ask many couples. I do not care how much you lock down the computer or get students to take exams in proctored settings. The temptation to cheat on on homeworks, discussion papers (often used as a substitute for in class discussions), and longer papers is strong, especially when many of these can be easily farmed out online to a freelancer. In a 30-person classroom setting, if a student does badly on an exam but aces everything else, a professor can quickly suss what is going on: eye contact and one-on-one interaction are powerful. In an online setting, the professor has no idea who is doing the homeworks, discussion papers, or projects. Acing these and doing badly on an exam will be enough to get a solid B.

Meh, I'm being a little glib here but your argument basically reduces to cheating in universities doesn't work because the instructor can peer into the students' eyes and see their souls. Cheating on homeworks and projects is far easier at a university than online it seems to me: for one, "cheating" in person can be a bit of a gray area - what happens if you form a study group and let everybody else do the work while you just copy down the answer? To seek out an answer online you must more clearly venture into "cheating" territory and it's hard to justify to yourself that you're not cheating. Second, many MOOCs offer randomized numbers on homeworks and exams, which of course does not eliminate cheating but it means you must at least apply somebody else's answer to your own problem.

Also, I think you underestimate the amount of supervision that is provided at universities. 1 professor cannot monitor a class of 100+. Or at the other end, Caltech uses take-home tests. Are you worried about the quality of work submitted there?

This article should have been written around this time last year -- disappointing to see The Economist so behind the higher ed trade press on this.

There won't be a winner-take-all market for the "highest quality courses" because the market for stand-alone courses (already well-developed, BTW) is very low-margin. The people who pay a lot of money on courses with actual production values/attached tutoring aren't even in higher education -- they're homeschoolers. Of course, as Udacity figured out, there's also a market for courses in corporate training. This market is almost meaningless in higher education. The use of the textbook analogy to justify a winner-take-all market for MOOCs is also strange -- people buy textbooks in order to participate in the market for traditional higher education (decidedly NOT winner-take-all). And textbooks themselves clearly "failed" to disrupt the medieval lecture model (terrified faculty of the time actually thought this might happen, BTW).

Courses are not valued. The weak appeal of courses (of any quality) in higher education is why even non-MOOC online ed operations are desperate to get into the PROGRAMS business. Traditional universities that are intentional about getting online (vs. those that just let faculty early-adopters drive it) start by putting their most popular (i.e., profitable) and largest gen ed courses online to cater to traditional F2F students who want flexibility. The next step up the value chain is to offer a degree program, NOT increasingly niche (and decreasingly profitable) courses. This is why MOOC providers like Coursera can only make money off of shilling textbooks -- because they have no credible way to get to program or credential level. Even if you count Georgia Tech's CS master's (which is no longer even a MOOC, just a cheap online degree that borrows some of the staffing principles of a MOOC) as a proof of concept, no one has put forth a credible business model for a credential.

The weirdest thing about this article is that it focuses on the extremely well-trodden supply-side argument for MOOCs (Baumol's cost disease?!!? OMG!), but ignores the total lack of demand-side of enthusiasm for MOOCs. Not only do the traditional-aged undergrads in the system not buy into it (the available data suggest most MOOC-takers are experienced professionals who already have bachelor's degrees), but basically no institutions (except Antioch Univ., which I believe, is still basically bankrupt) have agreed to license or franchise a MOOC. Aren't we a bit late in the game to expect everyone (or anyone) to suddenly turn around and agree that MOOCs are actually really great and totally worth paying real money for?

It's good to see some pushback against MOOCs. I've had one that was good - in part because it had a good quizzes online set-up plus ample participation from the teacher - but most of them have left me really unimpressed as to whether they're effectively teaching people anything. If I had more free-time, I'd take the in-person classes over them every time.

I don't have a problem with MOOCs as "a thing that exists in the world," but I think that their actual capacity is being way, way oversold.

They do a few specialized things very well (like reach students who live in remote places but happen to possess fast internet connections, computers, and good English skills), but they do a lot of things poorly (like discussion), and a lot of things not at all (like any science lab activities that require equipment beyond what is available in the average kitchen/garage). Some people learn well from them (like people who learn best from books), and a lot of people don't (like people who learn best from hands-on lab activities, or people who learn best through face-to-face interaction). They give you access to some useful things (like famous lecturers) and to other things poorly or not at all (like the social-professional networks that successful undergraduates develop and utilize in post-undergrad life).

It's obviously to the benefit of MOOCs that they cost so little, but there are some pretty clear ways in which you get what you pay for.

Out of curiosity, what are the 4 textbooks that have a combined 50 per cent market share?

I've taken a bunch of MOOCs now, passed some "with distinction," and quit others. I learned something from both experiences. Obviously finishing a course I learned the most, but even the "quits" taught me about the domain, and why I might not want to pursue it further.

I hated the "push back" phase, the anti-MOOC hype. I thought that it was either teachers looking to preserve their teaching experience (at a cost to broader student bodies) or people misunderstanding the benefits of a "quit" course.

Say you have 100,000 MOOC "starts," and only "2000" completions. The push-backers would tell us that is a huge failure, but (1) 2000 is still a big class, with more passes than any conventional school ever manages, and (2) you need some subtlety on what a "quit" means. How far in? Every lecture watched and every quiz taken added to the gross national education. And it was free, a secondary effect of producing the class for those core completions.

IOW, I am broadly with The Economist, and other early observers, who thought MOOCs "would change everything."

They will, but the technology acceptance curve applies. There is always a trough of disillusionment.

re: Dropout rates. I wonder if dropout rates are high for MOOCs in part because people sign up them just for curiosity? I've taken 2 classes on coursera, and am in process of taking a third. There's nothing related to a career in any of them. (I've taken a python video game programming class, a songwriting course, and am taking gamification now). I've signed up for a few that after I watched the first week or two of lectures I realized I wasn't interested in, wasn't equipped for, or was motivated to pursue something else - and immediately un-enrolled. There was one class I was interested in but after I finished the songwriting class I knew I wanted to focus on that for a while. To do the classes right I find I have to spend 10-20 hours per week - sometimes more for some assignments - so I'm unable to approach them lightly and if my heart's not in it, well it's just best to come back to it later.

I don't consider it a negative that I drop out of some of them though. To me it was interesting just to wander through the class to see what it was about. Not sure how typical I am, but I do think there may be a large # of window-shoppers who want to see what a class is about and fully expect that they might not complete the class.

Comments for this post are closed