The shift to on-line education can happen gradually and easily

I left the following comment on Bryan Caplan’s blog post:

You don’t need to overturn all convention.  The top schools could shift at the margin, as they have many times in the past, and suddenly the conformist thing to do is to have ?? percent of your classes be on-line, and so on.  In virtually any other context you would see the flexibility of the market here!  No major credentials need to collapse, if it turns out that cannot happen easily.

This is a phantom issue, raised by many people but not thought through deeply enough.  Markets convexify (sometimes).

It is fine to argue “on-line education is not in fact more efficient.”  It is much harder to argue “if it is efficient, conformity pressures will keep it out of the market.”  Don’t confuse the former case with the latter.


Following up on your earlier post on Harvard, doesn't this have the potential to destroy that alumni influence?

Comments for this post are closed

There is a market for it and U of Phoneix and others have pretty much tapped it.
Did that reduce attendance at tradional college's? I don't think much.
So administrators everywhere are needlessly freaking out.
I am not buying. BTW, did you hear that recent population increases will lead to massive starvation?

Comments for this post are closed

Two thoughts:

a) our current system is optimized for upper middle class kids who weren't screwups during their high school years. Open courseware from MIT & Stanford can signal ability that was less discernable during adolescence.

b) as a business owner in tech, I see a lot of resumes. Given the pace of change, the two elements I'm looking for are initiative and capacity for self-learning. Especially for recent grads or career changes, evidence of seeking out high fiber online opportunities sends me a signal that a more gaudy set of traditional credentials doesn't.


Comments for this post are closed

Your argument is in line with the gradual but ultimately disruptive pattern predicted by Clayton Christensen in "Disrupting Class."

One bit of evidence for innovation on the margins is Stanfords' rigorous and accredited Online High School ( Not only does Stanford OHS provide students with an accredited high school diploma with a strong brand, but it also allows students to take many university level courses.

It will be interesting to see how long it takes other elite universities to sponsor similar programs.

Comments for this post are closed

There seems to be a direct analogy with "An Economist Gets Lunch."

You write: "The problem [with restaurants that have lots of attractive women] is that they will attract a lot of men to the restaurant, whether or not the place serves excellent food. That allows the restaurant to cut back on the quality of the food."

The analogy is obvious. If HYP is in the business of peddling prestige, then that's what their resources will be dedicated to; the education is a sideline to be cut back if possible. Yes, if online education gave better education at no cost, it would make sense for Harvard to jump on board. But there IS a cost - moving courses online looks tawdry and breaks with tradition, both of which shatter the facade of elite academic pomp. Kind of like a hip bistro adding sausage and grits to the menu non-ironically.

Comments for this post are closed

But keep in mind that online education can take a variety of forms. The least revolutionary model is one that mimics traditional classroom instruction (and retains most of its inefficiencies). In that case, the current student-teacher ratio is roughly preserved, there is a standard academic calendar, and 'chat sessions' that take the place of classes where attendance and participation count in the grade. This inefficient, hamstrung form of online education is the one that will be favored by accreditation committees and demanded by teacher's unions because it will protect the interests those currently in power in the educational establishment.

Comments for this post are closed

Not sure if this will be of interest, but a fair number of high end professional schools put video versions of lectures online as well as delivering them in person.

Comments for this post are closed

I guess I'm late to the game here, but the problem I see with online education (from an efficiency standpoint) is that it's very efficient from the point of view of issuing diplomas, but maybe not so efficient when it comes to actually educating students. If we view a university as a diploma factory (aka signalling mechanism) then maybe online is the way to go. If it's a tool to educate students...then maybe not.

How do you see online education being inefficient regarding "actually educating students"? Is it that it's more difficult to measure whether the student actually learned the material? That's not really a bigger problem than in physical classrooms if it's implemented well, and those schools which don't implement well will suffer reputational damage. Or is it because fewer students will complete the course? If most of the cost savings from online courses are passed on to the students, then there's not so much loss to a student who starts a course and drops it, especially compared to doing so at a traditional university. (If the online courses have more schedule flexibility than physical ones, then the opportunity time cost of dropping a class is significantly lower, too.)

I suspect there's some educational benefit to things like group activities, back-and-forth dialogue with instructors during class, the availability of office hours. Also I suspect distance learning may lead to a higher prevalence of academic fraud.

Depending how the classes are structured, dialog with the instructor and/or other students is possible. It's also not always available at physical colleges - I had a course from which I learned a lot, where the only back and forth was three people getting to ask questions in the last 10 minutes of class, each one triggering a 3-minute mini-lecture. The lectures were *dense*, too. But anyone who paid attention and could absorb that much material got a lot out of the class.

I would count the difficulty of setting up group activities as a *benefit* of online teaching. Too much group work ends up with the person most worried about her grade doing most of the group's work.

Comments for this post are closed

I concur with Anthony that group work is overrated. Probably true cooperative learning is beneficial, but group assignments rarely work out that way.

I think there are benefits of being there in person. One of the biggest is that being part of a class is that there is more accountability. Most people feel somewhat ashamed to blow off a class when they have to meet with the instructor in person; they also don't want to look ignorant in front of their classmates when called upon.

The fraud point is an interesting one. However, it seems to me that technology could solve this problem pretty well. Anyway, studies have suggested that the rate of fraud is already very high in traditional education.

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

Who would trust a pilot who got his pilot certification in an online school?

Who would trust a builder, welder, machinist, plumber who graduated from an online school?

Who would trust a pastor or priest who was trained online in an online seminary?

Who would hire a chemist who graduated from an online university?

Who would want their kids taught by teachers who were trained and certified online?

Degrees and certification for almost every field requires real experimental learning in the real world.

Look what happened when physicists started writing programs to determine risk in mortgages, derivatives, spreads in hedges.

Physicists spend time in labs learning to believe in the iron law of mother nature, and understand the only way a reading is wrong is from human or equipment failure, but given a large number of reading from different instruments, the law of large numbers ensures the true reading can be determined.

Mother nature never lies.

And told the market never lies because of the large numbers, a $1,000,000 mortgage on real estate that can be knocked out in 6 months for $400,000 issued to a borrower with no verified income is obviously worth at least a $1,000,000 less some fees, because the market never lies and that property will always be worth more than a $1,000,000 and will always go up - the iron laws of the universe is: lies never enter the market, real estate never depreciates or deteriorates, capital value inflation is infinite just like the expanding universe.

Lots of builders, lots of real estate brokers, lots of mortgage originators, bankers processors, lots of property assessors, lots of investors knew the market was nuts from 2003-2007, but went along because the quants said the market was reflecting the true reality of what was on the ground. After all, the quants had science, math, computers, theory behind them and were thus infallible. No 70s NASA scientist engineer laid off who got into real estate, mortgage origination, building developments, could ever have written the code that the quants wrote in the late 90s and early 00s - their experience would prevent them from trusting the data, or the theory, the data seemed to support. After all, after a career in real estate, mortgage banking, and property development, the one iron law would be "people always lie, sometime."

I'm not arguing against online education, but I am arguing that online schools are so limited in what they can deliver, its like saying "computer automated manufacturing eliminates all the manual labor". That's not even close to reality.

+1, with a caveat.

If you currently have a job, and want to advance within it, then under some circumstances, I think online is a teacher getting a masters degree in education, possibly a manager getting an MBA (online with weekend meetings).

I can also envision online education being used for different purposes as well.

For example, instead of an SAT math test, a high school junior takes an online math course from Stanford, MIT, etc. and takes a monitored test at the end of the summer. The student learns what it will be like, the school learns if the student can handle it.

Or, the school uses the online course as a screening device to get the best and the brightest. For example, Stanford is currently doing an artificial intelligence course online. I would be interested in recruiting those students who took the online course, took the monitored test, and applied to my school.

I don't think online will measure up, so to speak, unless and until a complementary product is created: some universal measurement that measures class outputs of both in class and on line classes. So, for example, if ALL MBAs were required to take a test to graduate--those from Harvard, those from X University, etc.--and those from an online program, then I could measure across all schools, much as we measure people today with Bar Exams or Medical Boards or Certification boards. But, until those products (measuring tools) are in place, online has a long way to go, and in neither case (online or in class) you still need to interview the graduate, and hopefully use them as an intern first.

Comments for this post are closed

Who would trust a pilot who got his pilot certification in an online school?
Who would trust a builder, welder, machinist, plumber who graduated from an online school?

Online ground schools to prepare students for exams already exist. The same goes for building trades where there are many online courses available for preparing for various state licensing exams. Would I want a teacher who had an 100% online degree? That's the wrong question. The right question is whether I would want a teacher who took her first two years of general education courses and her 'pedagogy / history of education' courses online and then did her student teaching with an experienced mentor? Why not?

Comments for this post are closed

This is due in part to signalling. We perceive that the only people getting online degrees are folks who couldn't get into "real" schools. So we assume they're somehow defective.

Let me construct a hypothetical. Suppose an elite university (say MIT) changed nothing about its admission requirements, but started dividing each entering class randomly into "distance" and "non-distance" halves. The distance guys never actually make the trip to Boston. They live wherever they want to live, and take online versions of all their courses, taught by the same faculty who handle the regular versions. Care is taken to ensure that the online classes have equivalent grade averages and pass rates to the non-online versions. Ignore for now that many of those placed in the "online" track would probably withdraw and head elsewhere; suppose they all stayed.

You're now presented with two recent graduates. Both had identical credentials coming out of high school and both have graduated with identical GPAs over the exact same coursework. The only difference is that one was in the "online" track and the other the traditional track. Without knowing anything else about them, on average, how likely would you estimate it to be the case that the "online" guy is less well-educated and prepared than the traditional guy?

It's an honest question. I have misgivings about online education; maybe it's the case that 9 times out of 10 the online guy is less well prepared. Hard to say without doing an actual experiment. (Which is hard.)

Signalling is used only when you cannot measure directly the attributes of an agent.

If you can establish measurement devices across both online and inperson classes--they take the same test--signalling is irrelevant, as the test is the vehicle to measure.

Okay. Maybe I misused the term. How about this: I know that it's usually harder to get into traditional degree programs than online programs. Because of this, someone with a traditional degree would likely be more capable than someone with an online degree even if the educational "effectiveness" of each program were identical. Some of the perception that online programs are inferior is probably due to the students who enter these programs being inferior (in an aggregate sense- I'm not intending to demean any individual) before even entering.

Comments for this post are closed

Comments for this post are closed

Labs are hard to do online; they're also hard to monitor if done elsewhere. This is important for some subjects, less so for others.

Comments for this post are closed

Comments for this post are closed

"Lots of builders, lots of real estate brokers, lots of mortgage originators, bankers processors, lots of property assessors, lots of investors knew the market was nuts from 2003-2007, but went along because the quants said the market was reflecting the true reality of what was on the ground."

The builders, real estate brokers, mortgage originators, bank processors and property assessors had no reasons to care why the market was acting the way that it did. More work meant more income, and there was only modest downside to them.

Likewise, the quants did what they were told at the direction of people in heads I win, tails you lose compensation arrangements themselves and within the boundaries afforded to them. It wasn't the ex-physicists who lured other people into putting money into investments that were more risky than they appeared.

The people who screwed up were the investors who were ultimately left holding the bag on mortgage backed securities (many of whom were guilty mostly of misinterpreting deceptively described and evaluated collateral value and counterparty risk who relied on third parties with a great deal of prestige had advising them was safer than it really was), and people who were buying real estate for investment purposes (who were mostly guilty of excess optimism about their ability to time a bubble that it was rational to participate in if you could get out before the bubble burst). Residential homeowners in places like California where the boom was biggest, with non-recouse mortgages, experienced little downside from paying to much with little down payment on credit.

Lenders setting underwriting standards during a rising market that provided an ample equity cushion after just a few months, also weren't terribly irrational in not being unduly concerned about the ability of borrowers to repay. They were basically hard money lenders with no right to a deficiency judgment in practice anyway, and in that economic climate, documenting ability to pay when troubled borrowers can easily resort to their collateral doesn't make sense. And, if real estate prices has stagnated until reality caught up with the prices, instead of busting, neither the lenders nor the real estate investors, would have been burned to badly - even the excess optimists weren't necessarily optimistic about the prices being reality based so much as they were optimistic about the manner in which realistic prices would reassert themselves (i.e. about a hard v. a soft landing).

Comments for this post are closed

I'm not sure how this relates. Your point seems to be "don't rely on computers too much", but we're not talking about computers doing the teaching; it is just a different format for human instructors to deliver information.

As for "would you trust a pilot who went to an online school", you could just as easily say "would you trust a pilot who only learned in a classroom and never flew a plane". This is about replacing one part of education, traditional classroom instruction, with another format. No one thinks it can replace the hands-on experience portion that is integral in many fields.

Comments for this post are closed

Comments for this post are closed

Actually, speaking as someone who took the Stanford Intro to AI class this fall online, I think it's possible to learn MORE from an online class, if the technology is used properly. Some things I noticed:

* You can watch a given part of a lecture over and over as many times as you want, this is very helpful when it comes time to do homework.

* The forums connected to the lectures had a huge participation rate and lots and lots more questions were raised and answered than could possibly have been part of a "normal" class (with links to helpful material, too).

* The profs did a one-hour video answering questions every now and then, with the community deciding what questions were most important.

* Tests were online, so you check your grades and assignments any time.

* Lectures didn't use a white board or a screen capture system. Instead they used a simple piece of paper with a webcam pointed at it. Simple stuff was just written on the paper, but more complex diagrams could be printed out and shoved under the camera, which had to be a lot easier for the lecturer than drawing on a white board.

The course didn't do anything to prevent cheating or verify users, but that wouldn't be particularly hard to add (just issue PKI cards and allow access to everyone's webcam during testing and you would be done).

Kahn Academy is awesome too. I think education is going to be deeply affected by this stuff. I don't know if it will matter to top universities like Harvard, but to lower-tier schools, it's going to be transformative.

Comments for this post are closed

"Markets convexify" ... could someone elucidate!

Firm A offers blue cupcakes. Firm B offers yellow cupcakes. Both do well.

Firm C sees this, enters and offers green cupcakes; also does well.

[Or, in above story, replace 'Firm B offers yellow cupcakes. Both do well.' with 'Tyler identifies latent demand for yellow cupcakes.']

[Or, in above story, replace 'Firm C sees this, enters and offers green cupcakes; also does well.' with 'Firm A expands offering to green cupcakes; cleans up.']

Comments for this post are closed

Comments for this post are closed

Comments for this post are closed

The future as envisaged by Tyler becomes more plausible and more complex once you separate testing from education. Suppose you can learn online all you want (watching stellar MIT profs, Kahn Academy, whatever) but the testing is done independently.

Maybe the MIT profs give you individual feedback on your progress from a distance; maybe they answer your emails; maybe they read your papers themselves rather than put them through some program.

Think UK civil service exams, French Grande Ecole entrance, etc. You (can choose to) prep for those on your own, or enrol in some online class that helps.

Separate brands arise along the testing and the education dimensions. Duke finds itself having/wanting to prep students for the MIT (or Toyota) engineering certificate. A signal of quality (for a program, not an individual student) is that you open the exam to everyone.

You could have three dimensions/brands: education, the setting of the test, the taking of the test (there are testing centres, like for the GMAT, except all the centres do is ensure that nobody cheats).

It's quite exciting, and I speak as a lowly non-tenure-track economist.

I agree, completely: Unbundling, for short.

Comments for this post are closed

Unfortunately, we're not going to see testing unbundled like that any time soon. As I'm sure I've mentioned before, we still have Griggs v. Duke Power. Further, there is still the fact that higher ed is a government-sponsored (via the Department of Education) guild/cartel. Just as any college not approved by the DOE-approved accrediting bodies is dismissed as a "diploma mill", and any accrediting body not DOE-approved is an "accreditation mill", so will any test (provided it survives the inevitable "the test is racist!" lawsuits) be rejected if not appoved by the Guild. I don't see this changing so long as there is still a Department of Education.

Comments for this post are closed

Comments for this post are closed

The argument that traditional education is selling "confirmity" that gives it an edge over non-traditional education may have some merit at the low end of the education market (certainly that is what makes an ordinary high school diploma a much more valuable credential than a GED in the job market, and arguably that is something that makes a brick and mortar higher education at a conventional open admissions higher educational institution more valuable than an online degree), but "conformity" per se, is not the edge that Harvard and Stanford are offering.

Some reasonably traditional format, selective colleges and universities, if anything, encourage non-confirmity and reward eccentricity in ways that no work place would ever tolerate. In a traditional brick and mortar college setting, only sixteen hours a week or so are spent in class anyway, and the number of weeks in the school year is often shorter than in K-12 as well. With planning and determination, a student can set up a schedule so as to avoid having to show up to class before ten in the morning, or with class attendance only two or three days a week. Many college and university classes impose no penalty at all if you skip class so long as you do tolerably well on the exams and homework, and a fair number of respectable brick and mortar colleges and universities make lectures in lecture classes available for viewing online at times other than the regularly scheduled class time. A degree from a selective college or university with medicore academic performance can even have more value that a degree with honors from another institution, and not just due to grade inflation effects.

A big part of what a prestigous traditional school offers than a non-traditional one does not (even if it is prestigous), for traditional aged full time residential students, is networking opportunties with fellow members of the likely to succeed crowd inside and outside of class, with opportunities for friendships, romantic partnerships, and experience in leadership roles in extracurricular settings involving talented peers. It isn't impossible to develop some kind of networking/peer relationship in an online setting (I saw my own students do it when I taught graduate students online, most of whom were online out of logistical necessity). The intense focus on selectivity in college admissions, in part, reflects this reality. One wants to go to a selective undergraduate program as much to be around a particular set of peers as to be around a particular set of professors. If one wants to make the argument lurking behind the "conformity" case, one would argue that prestigous colleges and universities are engaged for their full time residential traditional students an opportunity to participate in the social class socialization process, which is something quite different than "conformity" and influences where prestigous institutions might go with online education (i.e. they will offer it to people who don't need social class socialization, and to people who a receiving that component of the process where they are -- for example, offering math and econ clases online to students engaged in legislative or Big Business internships sponsored by the institution, or offering English literature classes to students working on internships at CERN).

Comments for this post are closed

Maybe we should see the historical and current fragmentation of higher ed into different segments evidence that there isn't a single market and that we are likely to see similar fragmentation in uses of online education. Why would the ability to put material online suddenly erase the current segments?

Comments for this post are closed

If top schools try to maximise "Brand" value for alumni, they will need to manage both scarcity and quality. this means they will probably only allow for on-line individual courses, not degrees. Very much like the buying of shares by companies that try to increase shareholder value.

This means that new entrants, or entrants not limited by their need to keep the returns to investment of alumni maximal, will suddenly be there delivering high value education on-line. The implosion of the " brands" seems possible at that stage.

if on-line education turns out to be a black swan in education, you read it here first.

Comments for this post are closed

Comments for this post are closed