Completion rates for MOOCs don’t tell us much (yet)

Recent analyses released online has shown that completion rates for massively open online courses (MOOCs) are less than 10%. Some commentators take low completion rates as an indication of serious pedagogical and/or economic issues with MOOCs.

While I don’t think that MOOC’s are as revolutionary as perhaps their most ardent backers do, I also don’t think these completion rates tell us much (yet). Right now, my understanding is that almost everyone enrolling in these courses is not doing so for credit and are paying nothing to take the course. The cost to signing up and not finishing it is zero aside from whatever guilt the person may feel for doing so. It’s likely that many of the people who sign up to take the course do not even intend to finish it. MOOCs are still a novelty, so they attract a good deal of curiosity. They sign up, look around and watch a video or two.

That said, I guess much depends on when payments are expected. Say a college allows a certificate verifying MOOC completion as a substitute for taking a particular course. If students pay for their certificates after the course is completed, I imagine course completion rates will remain low (although not as low as they are now). But if students are required to pay upfront, and hence make a financial commitment to course completion, then completion of a MOOC could be comparable to completion of any other online course.

These completion rates for MOOCs aren’t so much a function of anything inherent to the medium, but are instead a function of how the medium is currently implemented.

All of that said, I think Arnold Kling is right that customized multimedia textbooks stand a better chance of revolutionizing education than do MOOCs.


College as ritual

I’ve been thinking a lot lately about how the effect of the proliferation of online education on the current structure of academia.

  • The positive question: Some people think that academia is ripe for creative destruction, and that online education will transform the academy the way the Internet and digital music devices transformed the music industry. Fewer jobs for professors but higher incomes for the star teachers and researchers. How likely is this to happen? And also, why hasn’t it happened already?
  • The normative question: Will this be a good thing? Is creative destruction through automation the antidote to cost inflation / Baumol’s disease in higher education?

I want to focus on the first question. Some prognosticators say that companies like Udacity and Coursera will transform higher education the way the iPod transformed the music industry. But why hasn’t this happened already? In the face of the high cost of higher education, why aren’t more students and the legislators who represent them asking for cheaper online alternatives? Sure, there’s some of this going on now, but why the time lag? After all, correspondence courses have been around for a while.

The music world was transformed by recording devices a long time ago. Why didn’t cheap, easily distributable media transform higher ed as well?  Why haven’t live lectures been replaced by tapes and correspondence courses in the same way that live music declined after the affordable recordings became widely available?

Some have suggested that the answer is about motivation. An analogy is made to a personal trainer. A personal trainer gets people to engage in much more strenuous activity than they otherwise would. Meanwhile, the naturally motivated athletes are amused that people would pay so much money to someone who can tell you to do things you can learn to do yourself. In the same way, college motivates the majority of us who could not motivate themselves to get through online courses sitting in their home by themselves.

Some have argued that a traditional college degree remains important in the face of more efficient alternatives largely because of what it signals to employers. When making hiring decisions, employers seek quick-and-easy indicators of cognitive skill, diligence and conformity. Online degrees haven’t traditionally been a good indicator of any of those three things. Perhaps the connotation of online courses will change since many if not all of the most prestigious universities are now participating in one way or another in MOOCs. But even then, we’re a long, long way from a tipping point in which taking a bunch of online courses in lieu of pursuing a conventional degree could possibly signal conformity.

I can think of another reason that might explain the continuing dominance of the traditional college model. Perhaps we should think about the question more as an anthropologist might.

Going to college is a ritual, a rite of passage. (It might be good here to compare college to rites of passage in smaller societies.)

Going to college (exclusively) online is certainly a viable option for some, but perhaps the bricks-and-mortar version remains dominant in part for the same reasons that many able-bodied religious people go to church rather than watch a church service on TV.

It’s partly about the experience of course. The college experience is I think for many students a haven of hopes and possibilities away from the hum-drum of standard work life. Students may like how college makes them feel about themselves. Even if they aren’t doing too well in school or aren’t particularly interested in school, going to school raises their status in their community and contributes to a feeling that they are improving themselves.

College graduation is vitally important to the identities of many people. It signals an arrival, the meaning of which goes well beyond the additional money one can earn. I think of parents crying at the graduation ceremony. Can online degrees produce this kind of meaningful experience?  Can online degrees have the same impact on identity? Maybe, but it might require a dramatic shift in what college signifies for people.

Take the productivity gains as pay increases, not just as increases in whiz bang

Not enough people take advantage of cheap prices of modern machinery. There are lower end cell phone and tablet models that allow you do essentially everything that the higher end models allow you to do at a fraction of the price. You can get pretty much all the functionality you need at prices ranging from 1/2 to 1/100th the price. Several points :

– The existence of smaller tablets obviate the need for a smart phone. I bought a cell phone for $5 and it allows me the same conveniences as phones that cost 100 times more. What I especially like about the $5 phone is that I don’t care too much if I lose it or break it. I really don’t see the point in buying a phone for more than $15. Also I prefer a phone with just two apps: making phone calls and an address book. I don’t want to read emails or other documents on a surface smaller than the size of my hand. That’s what a tablet is for.

– The only reason to buy an expensive Ipad I can see is for purposes of conspicuous consumption. (Okay partly for tech whiz-bang, etc.) That’s not exactly a minor reason — its the reason people buy most things. Even people who say they aren’t buying something for this reason generally are. But I wanted to have all periodicals and books at my fingertips everywhere there’s wifi, and the $200 Kindle Fire provides all of that at half the price of the cheapest Ipad.

– Cars can be driven much longer than people usually drive them. I drive a 14 year old Honda. It still runs well. I don’t think 10 years is old for a car anymore, functionally speaking. (This is one of those things that makes it difficult to estimate inflation and changes in living standards over time) . Also the Car Talk guys have shown that its generally wrong to think that buytming a new car is the better decision financially when your old car starts needing major repairs. Given the premium that new cars command, fixing your old car is almost always cheaper. (I think often people may use thus reasoning as a rationale because they just want a new car to experience a new car; if they went through the expenses they would realize its foolhardy.)

A lot of people could improve their standard of living by befriending a trusty mechanic and buying a car from him, rather than leasing a new car from a dealership. If you know how to find a reliable used car (not easy but doable) , then buying or leasing a new car bears a close resemblance to spending $4000 for gold rims, detailing and vibrating stereo equipment for an aging Lexus. Different demographics make these decisions but perhaps the reasons aren’t as different as sometimes imagined.

-Another thing: dont replace cosmetic things on the car. Occasionally, various ornamental pieces come off the inside of my car. This isn’t just due to age ; its also because I don’t treat things delicately. I don’t really care too much about this stuff so I wouldn’t think of replacing the pieces.

I wish that a liberal arts college degree certified the possession of certain minimal skills

I teach in a sociology department at a regional university. There are of course thousands of students graduating from similar programs in similar universities around the country. What do such students know when they graduate? What have they learned?

My sense is that there is great heterogeneity in what such graduates know and what they have learned. Some have learned quite a bit about how to think in a more analytical way and about how to write. Others, it seems, not so much. It’s not so much that the threshold for passing is low as that the threshold appears to vary considerably by professor and by program. For this reason, I think that a degree in a field like sociology from a regional university tells an employer little if anything about what a student knows or what skills he or she has to offer. An employer has to look a bit deeper for this. ( note I am NOT saying that a college degree signifies nothing. A college degree does signify to employers that a person will submit to authority and conform to expectations to some degree. What I’m saying is that a liberal arts college degree doesn’t certify that a person has acquire analytical or writing skills above any reasonable, identifiable threshold.)

One could challenge this notion by noting that students must maintain a minimum GPA and must avoid failing required classes. But GPA doesn’t tell us anything specific about what skills a student has mastered. To be sure, GPA probably correlates fairly strongly with such skills one might expect from college graduates in the US such as the capacity to construct an English sentence that is essentially free of grammatical and spelling errors.

But the fact that GPA correlates with the capacity to write a sentence free of major errors hints at the problem itself. Does GPA correlate with the knowledge that 2+2 =4? No, we can safely assume that there is no variation among college graduates in the ability to answer that question correctly. Thus, the correlation between answering that question correctly and GPA is likely to be essentially zero. I don’t know much about the standards set for high school graduation in various States, but I do think there is the expectation for mastery of certain basic rules of writing. (The fact that a considerable number of students are permitted to graduate high school without such skills is no reason to then permit them to go on to graduate from college with these deficits essentially intact.)

In my view, GPA shouldn’t correlate very highly with the capacity to write a sentence free of major grammatical errors because there shouldn’t be much variation among college graduates in this capacity. A degree from an accredited college or university should certify that a person possesses skills that exceed the thresholds set for high schools around the nation.

But the threshold for a college degree does not necessarily exceed the standard for high school graduation. Instead, the minimum threshold is perhaps consistent with Woody Allen’s maxim that 90% of life is showing up.

The reasons are clear enough. What incentive do professors have to assign failing grades to student work that falls below the threshold of many high school curricula? Very few. And when a student or their family is shelling out 10k a year, and the student is going through all the motions, how many professors can assign an F when the work is seriously substandard? In my experience, academics in the field with which I am familiar tend to be softies, myself included. (I think this is partly why I feel like I only started feeling socially at home in grad school– people were just a lot nicer there than what I had experienced previously.)

And university administrators are often primarily concerned about maintaining or boosting enrollment. The viability of the college and the careers of administrators often depend at least partly on the success by which they meet enrollment objectives. This is understandable. And professors salaries and the quality of their work environments often depend to some degree on enrollment. Even if professors weren’t softies, their jobs depend on enrollments.

I don’t think there are many A’s for effort. Most professors will make distinctions between work of varying quality. But I think there a lot of C-‘s and D’s for effort.

This isn’t necessarily a problem in and of itself. But it becomes a problem when universities establish laundry lists of impressive learning outcomes that they apparently expect students to meet. We should at least be honest about it — either acknowledging the situation by lowering officially lowering the bar or raising the standards for graduation, or perhaps a little bit of both.

The lack of honesty about what the minimally passing student knows is very unhealthy from a pedagogical point of view. One of the most important things to learn is to know what you don’t know. My sense is that many students have an inflated sense of their own writing skills (in part due to the self-esteem movement which minimized criticism.)

How could college be structured to ensure mastery of a certain minimum level of skills?

First I think students shouldn’t advance through college by accumulating course credits. Instead, it makes more sense for students to advance through college by mastering certain skills. Students would only be permitted to advance to a higher level if they had demonstrated that certain skills had been acquired. Student evaluation would have to be handled by disinterested parties, perhaps people outside the university. A minimum score on the essay writing test of critical thinking the College Learning Assessment (CLA) could be one sub requirement.

In this model, courses would only be useful insofar as they helped a student improve their skills. So students wouldn’t have an incentive to search around for the easiest courses to take. Instead, students would have an incentive to take the courses that best prepared them for the certification test they had to take.

The idea of courses themselves might have to be scrapped. There is no reason to believe that the structure of college instruction of which the course is the basic unit, is necessarily the structure that maximizes learning potential.

Think about it this way. Say college costs $10,000 for the year, including two semesters as well as some classes over the summer. Now, say you wanted to learn something they didn’t require any particularly special equipment (e.g. this is like a liberal arts degree — no labs or microscopes or corpses needs to dissect) and you had an equivalent amount of money — $10,000 — allocated for this purpose. Put aside for a moment the all important credential that a college confers. Just focus on the learning itself.

To take a concrete example, say I wanted to learn more about statistics just because I’m interested in it and J had 10k to spend in helping me toward this goal. Would I spend this money on:

-renting a special place for me to study outside my home?
-subsidize research about esoteric areas of statistics?

No, what I would do is spend the money on a group of tutors who could help me through basic coursework. You can hire a decent tutor for $40 an hour. Okay, maybe PhDs would command more but I don’t need a PhD to teach me master’s level statistics. I just need someone who understands the material, communicates well and has A LOT of patience. A good undergrad statistics major could teach me a lot.

With 10k, I could afford to pay tutors to meet with me one on one for an hour a day, four days a week for 50 weeks out of the year and have $40 a week left over to pay my tutors to coordinate their lesson plans with each other.

I wouldn’t pay anyone to lecture to me in a room with 20 other students. That would be an obvious waste of money. There was a time in college when I naively thought I would major in economics and have a nicely remunerated position, like the guys from Hong Kong in my dorm. On the first day of the third class in the sequence of courses in the major, only minutes after I first learned the guns and butter model, the lecture became more highly mathematical than my brain could handle without careful, excruciatingly slow tutelage– and as other students raised their hands with questions I didn’t understand, I sat there, lost and thinking about what course I would replace it with. Lectures are useless and even counterproductive for the slow, plodding student. For the quick study, they are often a waste of time. Lectures can be masterful demonstrations of teaching that generate an unknown amount of actual learning. Great lectures are fun to give and to hear, but right now if you want to hear inspiring lectures by gifted speakers on all manner of subjects, you needn’t pay or leave your room.

If I paid a tutor to teach me one on one about the questions those students were asking in that Econ class, I would probably have a good chance of getting it after sn hour or two. Instead, an hour and a half lost on a lecture as a the professor forged ahead. (This is not to say I could have been any good at economics, but rather, that a dollar spent on one on one tutoring by an advanced economics major would have been much more helpful in helping me learn the material than a dollar spent to pay someone to lecture ti me and a group of students with greater math ability.

So I think that the amount of learning in college would increase if:

1. Graduation was contingent on skills mastery as determined by disinterested parties, rather than by accumulating college credit.  (What if a student had already mastered the skills, or did so after a short time?  Good, then they pass the test early and get their degree.  Don’t waste the time and money of these students.  Give them their degree and let them move on.)

2. Tuition money was spent in a way that maximized learning.  Automate what can be automated with relatively minimal loss of quality: lectures, computer-checked exams, etc. Maximize time with one-on-one tutoring, which can be affordable and effective when conducted by talented individuals without terminal degrees.  What better way to keep students engaged and on task than to require them to meet with instructors one-on-one nearly every day of the week?

(Where does this leave the professors with PhDs?  I’m not sure, but thinking about what would maximize student learning.  There’s no reason to believe that the interests of currently employed faculty are aligned with the interests of students.)

Investment bankers become venal pricks when they hang around traders

The revelations in the “resignation-by-op-ed” by a Goldman Sachs VP named Greg Smith are of the dog-bites-man variety.  Although Smith’s missive may very well damage the firm, I’m skeptical.

Here’s an excerpt of the letter by Goldman VP Greg Smith:

“Make the client the focal point of your business again ….. [w]ithout clients you will not make money. In fact, you will not exist. Weed out the morally bankrupt people, no matter how much money they make for the firm. And get the culture right again, so people want to work here for the right reasons.”

According to Smith, what Goldman is doing is both morally wrong and short sighted, although the short sighted part is only an issue for those who stand to lose if the long term viability of the firm is compromised by its present actions.

The culture among traders of client manipulation that, according to Smith, has recently come to dominate Goldman, was memorably described by Michael Lewis in his first book Liars Poker, about his experience at Salomon Brothers in the early 1980s. (From that book: After Lewis managed to successfully unload some troublesome assets onto a unsuspecting client, he received a phone call from a higher up awarding him the coveted title of “big swinging dick.” It brought tears to his eyes.)

Salomon eventually blew up but not because clients pulled their funds after realizing it wasn’t acting in their interests, but instead because they apparently mishandled the junk bond business. Which brings me to a more general point:

The predatory culture of trading firms going back to Salomon is well known, and these firms don’t seem to live or die in accordance with whether they live up or down to this reputation.

However much this bad publicity will cost Goldman, I doubt many of those costs will result from clients suddenly realizing that Goldman is screwing them over. My understanding is that many of Goldman clients are institutional investors (pension funds, who arent likely to be completely naive. I doubt that sophisticated institutional investors are ignorant of the fact that Goldman’s interests are not entirely aligned with their own. The key phrase is “not entirely aligned”. Client-server relationships entered into in a voluntary way by knowing participants will typically involve some overlap of interests. If interests did not align at all, it is unlikely the relationship would have formed in the first place. In areas where interests are not aligned, the client would have to assume from the beginning that Goldman or any other firm would act in its interests and not the client’s. At the very least, investor clients would take adv ice of any firm with a grain of salt.

Since I’m on the market to buy a house, I think of an analogous situation as a realtor and his / her client. Obviously my realtor’s interests are not entirely aligned with my own. I would be a fool to think or act like they were. Its in a realtor’s interest to get deals done because that’s the only time she gets paid. So a buy side realtor will naturally be less concerned with getting the lowest possible price for her client. And a sell side realtor will naturally be less concerned with getting the highest possible price for her client. (Levitt did a study in which he found thqt realtors wait a bit longer and get more money when they are selling their own houses han when they are selling a client’s house.) If a realtor tells me “you’re unlikely to find a home like this for x price in x neighborhood,” I’m going to consider it in light of the conflict of interest that exists.  I know that my realtor will naturally tend to be more eager to close a deal than to get the lowest possible price.

Yet I retain my realtor because I believe her services are valuable to me.  Simply because she doesn’t fight for the cheapest price possible doesn’t mean that I would be better off without her services.  Our interests are in fact in substantial alignment.   My realtor is (presumably) highly motivated to find me the kind of house that I want for a price I can afford because that is the surest way for her to consummate the deal and get paid.

I imagine that Goldman’s clients would be in a somewhat analogous position with respect to the storied investment bank.  If Goldman provides a service that they cannot get elsewhere, then it is worth it to them to retain Goldman.  I don’t know much about the financial industry, but I imagine there is considerable overlap in interests between client and server, as with realtor and her customer.

The Rent is Too Damn High

In a short ebook that came out a few days ago (The Rent is Too Damn High) journalist Matt Yglesias advances the argument that widespread restrictions on housing supply in the United States have much higher costs than people realize.  I was predisposed toward this argument before I started the book, so it is hard for me to judge this book in a dispassionate way.  I’m thinking, “how do these arguments gain traction?” not “are these arguments true?”

Yglesias focused on how restrictions on housing drive prices up.  Probably in the interest of keeping the ebook short, Yglesias didn’t talk about the perverse effects of supply restrictions on the quality of housing.

In NYC, a combination of restrictions on new construction and rent result in very low apartment vacancy rates.  Vacancy rates for cheaper apartments will be particularly low.  This is a problem for apartment quality.  It would be interesting to see scatterplot of “volume of complaints about landlords” by vacancy rate across US cities.  Vacancy rates discipline the landlord; they keep him working for you.  The only way that landlords will consistently maintain their properties is if they are concerned that people might move out.  But when vacancy rates are extremely low, landlords don’t have to worry about that.  So they take longer to fix the heat when it breaks in the winter.  They don’t deal with the roach or rodent problem.  Etc.

I’m willing to bet there’s a lot of complaints about landlords in NYC relative to other parts of the country.  When I lived in NYC, bad landlords were often a subject of discussion.  But no one seemed to think much about why the landlords were bad.  It was just taken for granted that landlords tended to be an especially evil form of capitalist.  But when I lived in central Pennsylvania, I didn’t hear this, despite the fact that the landlord was also non-resident and managed many buildings across several states.  What makes landlords worse in NYC than in central PA or elsewhere is the combination of a low vacancy rate and rent stabilization laws.  Additional supply disciplines landlords.  Without it, the quality of housing for the vast majority of the renting population will be low.