This Is the End


Markets, Risk and Human Interaction

January 15, 2013

Why College: Learn, Signal, Network, Party

An article today in the New York Times discusses the growth of on-line courses, and the plan by the California higher education system to integrate some of these courses into the brick and mortar curriculum. On-line courses allow students around the globe to experience the same courses, taught by the same, world-renowned professors, as those who have jumped the hurdle to get into prestigious universities and paid $40,000 a year to study there.

In some cases it is even better than physically sitting in the real lecture. A friend of mine who teaches a popular MBA class at Harvard gave a set of his lectures for one of these programs. He was set up in a lecture hall-turned production studio, complete with a set of bright-eyed students. (The bright-eyed students – models hired by the company to give the ambience of a classroom). The end result has the production quality of an HBO series. It was watched by over 150,000 students in the first few months after its release, compared to under a thousand a year at Harvard – and at Harvard, the class is oversubscribed.

For the technical subjects like computer science, or, for that matter, accounting, the on-line programs may have a shot at competing with what you get in the brick and mortar world. Beyond the lectures, which you can capture as well on-line as in the lecture hall (where the class size is so large that direct interaction with the professor is all but impossible), most of the learning in computer science comes from interaction with the computer, with other students, and with instructors in small study sessions. The various on-line programs are creating solutions to allow learning through all of these channels.

If on-line programs end up developing to prepare students as well as the college classroom does, will they supplant enough of the function of the university to have a material effect on college-level education? Will they affect the limitations that arise from the tight admissions and hefty price tags for the top-tier schools? Will they shutter lower-tier schools? That is going to be a big area of debate for the next decade as these on-line programs develop and refine themselves, and as we see the quality of the product that emerges.

I want to as the next question: If it does as well in teaching what a student needs to know for these technical fields, or, for that matter if it does it even better -- we have to admit that as a possibility as well-- what else remains of the college experience that it fails to replicate? I think the U.S. college system is founded on four objectives: Learn, signal, network, and party. Once we have learning out of the way, what becomes of the other three?

One thing you do not get if you take Stanford classes on-line that you do get if you graduate from Stanford, even if you master the material just as well, is a Stanford degree. And the value of that degree is as a signal to the job market that you are smart – though what it should signal is something more modest, which is that you did very well in high school – and that you are well trained. This signal is missing for those who do the courses on-line.

But what is the value of the signaling? It is not very valuable if either, a) there is a low-cost, accurate method to gauge ability before hiring, or b) it takes time to gauge ability, but the employee is earning his keep while that ability is assessed.

The first approach works for many technical fields, especially when it comes to computer science. Set someone down to write a few programs and it quickly becomes evident whether they really know what they are doing. Throw in some math problems to verify that they really know their differential equations (though why anyone cares at this point is a bit of a mystery to me). And maybe add in some brain teasers for fun. Granted it is not perfect, but neither is the signal coming from where you graduated and your GPA.

Perhaps this process will become institutionalized, for example by allowing students in the on-line courses to take the same tests as the brick and mortar students in a controlled, verifiable setting. If you studied the same thing and did as well on the tests, then why should an employer care about the technicalities of the degree? What if they can get the on-line student on board for a few thousand dollars less? If it turns out that the market prices the difference at, say, $10,000 a year, and if after a few years the work history dominates the degree in assessing your future, then the $160,000 extra investment does not look too good.

The second approach, used in less technical fields like the entertainment industry, is to let graduates do apprenticeships where the cost of hiring and firing is low. A Harvard degree might give a bit of an edge for this sort of a job, but it is not a guarantee someone will have the difficult-to-define abilities for the job. So the entry hurdle is low, undergraduates from all over can get low-paying or non-paying work doing menial but useful tasks, and over the course of what might be several years they are either moved up the entry-level ladder or are weeded out. The proliferation of internships we have seen over the past few years might be a cyclical phenomenon resulting from the poor job market, but it also might be an emerging structural part of the post-college employment process, to some extent supplanting the college's signaling function. Although this has not been the model for technical fields, it could provide a second approach to deal with the noisier signals that come from the on-line programs.

A century ago, or even less, the old college tie and the old boy network were the main take-aways from the college years. Even today, the network with other students might be the most valuable part of a top-tier MBA. And no doubt it is also valuable in a technical field. But networking now has a life of its own. Supported by the machinery of social networks, on-line programs are already working on building these networks, ones that can be far more extensive than within any university. And when it comes to technical areas – and for that matter, even in non-technical areas – networking and interacting virtually is pretty much the way things go even with the student sitting three seats away in the real classroom.

Once you take the learning, signaling and networking out of the picture, what is left? Well, it might be what, in reality, is the largest component of what goes on in our undergraduate system: taking four years off to have a good time. For many students, the history and English classes are the dues they have to pay.

I think you could pull out half of those who go to college, put some of them into a year or two of vocational training (on-line or not), put others directly into the workplace in an apprenticeship or internship, and see a positive effect for the economy. So although partying is the one aspect of the college experience that cannot be replicated by an on-line higher education, that will be a net gain to society.

Note: One obvious caveat in all of this is that although on-line courses have a good shot with technical subjects, for liberal arts and humanities the task is more difficult. These require small group discussions and intense interaction with the professor. I think for those who seek such an education, the model also is different from that of the large real universities. To my mind, the model is that of any of a few dozen small liberal art colleges with an interactive and focused curriculum.


  1. There is another factor that at least used to be available but probably is not as much anymore that could be called mentoring. Ludwig Wittgenstein believed strongly that the chief purpose of higher education was working through real problems with an expert in a small group setting, which was the format at Oxford and Cambridge at the time, for example. That is seldom available in contemporary setting other than on occasion with one's thesis adviser. By even with online education, some mentoring solution could be designed to replace the old in-residence system. It would increase the cost but it would be available much more widely, even globally, opening a door that would otherwise be closed to many students.

  2. I could see mentoring programs becoming an institution just as the testing programs I mention in the post, with instructors that are the same sort of Ph.D. students or post-grads as teach the small problem sessions in the non-virtual courses. They might be on-line as well, or done in person. The students might pay for the sessions, just as they might for the institutionalized testing, but the cost would be far less than the university, perhaps by a factor of ten.

  3. Rick, perhaps there is another side to this thing, that I did not consider until I looked at your post. A few weeks ago, I went through one of the Stanford classes on-line (Fourier) and was really enthused about the experience (mind you, I'm four decades out of college and had a technical career - patent in data/noise analysis, even). The topic may have been unusual in being technical, but as the prof expanded upon the subject, he did side talks about the mathematical (including historical and foundation'l view) derivations at various points (almost every hour). Many times, this led to (culminated in) comments suggesting (to me) how all of these abstracted models have stacked upon each other, sometimes in a very subtle circularity that no one seems to care about (except some mathematicians).

    Except for, too, the autodidactic mind. I had earlier asked how higher ed (such as the MITS and Harvards) could help those who self learn. First, we would have to look at the autodidact's role (see my link) and why it is necessary. Mind you, the buck as the measuring device (pointing, of course, to FB and other supposed successes of that nature) would have to go out the window. What happened to learning for learning's sake (joy -- what I see are uptight, competitors, everywhere -- talk about denseness!).

    Jamie was bemoaning the career focus of his folks, running around like chickens without heads (his words), when the whale appeared. But, education and business are to blame (for the most part).

    One thing about the true autodidact is that he/she will listen. But, we form our own decisions (hence, mentors have to tread carefully -- actually, who can mentor a self-learner if the person is not such). One might say that these the cats in world of dogs (joke). Some professionals talk neuro/psycho issues (ah, let's medicate away tendencies that just might be of great use to the world) in this regard. That, to me, is mere pushing toward the center as if that is the source of peace in the world (by the way, I have never met a violent autodidact, yet -- so, the behavior does imply some type of talent and ability to grasp - perhaps, it's a matter of comprehension - von Neumann said that one could not get such with mathematics, well, let's change it so that it is more correct).

    Also, enough of the self-learners make it to adult (and effective) hood so that we can look at the situation and its importance. Wikipedia has done a fairly good look at this (here is a list).

    By the way, being autodidact does not imply bullheadedness, inability, or such. Rather, it may be some type of higher-order cognitive processing that ought to get more attention. To me, it's the more important thing in the world: given the idiots on Capitol Hill, the Jamies, and a whole lot more.

  4. I think these online schools could add the party element to their curriculum; hire social directors to host parties in cities with high concentrations of students. (kind of like what Yelp does for its elite level reviewers) I would think a school like University of Phoenix probably has enough students in every major city in the US to rent out a private room at a restaurant or bar at least once a month.

  5. Nice article, thanks for the information.


Note: Only a member of this blog may post a comment.