Friday, January 18, 2013

The Learning Technologist Becomes A Luddite

Norma sent a link to The End of the University as We Know It, by Nathan Harden and in her subject line said, "Interesting Article." I interpreted that as a request, first to read it, then to offer up my opinion on it. It took a few days. I don't process things very quickly now. Let's get to the punchline straight away. I gave it my thumbs down. That gets me to my title. If Harden's piece is indicative of up-to-date thinking about learning technology, then by comparison I'm a Luddite.

I don't think I've really changed that much in my views, certainly a bit more crotchety with age and with that more skeptical on the usability front. Take m-Learning, for example. (Please!) For quite some time I've been a big advocate of large font on the screen, particularly for material that is conceptually challenging. When the brain strains the eyes shouldn't. But if that becomes a stern requirement, it means that smart phones are really not the right environment for textual content. (I have the Kindle application on my iPhone. You can make the font to satisfy my requirement, so it is surprisingly readable. I've read a chapter of Csikszentmihalyi's Flow that way. But doing so tends to drain the battery.) In particular, how does one read the captions in a talking head video when viewing the video on the phone? Or, if the video is a screen capture of something with technical detail, how does one mentally process the content of what's on the screen? So I've got some usability criticisms regarding what is currently being preached, though that in itself doesn't make me a Luddite.

My initial thought in reading Harden's piece was to go back to the net.Learning documentary and ask, aside from the consequences of Moore's Law, what really has changed in regard to eLearning in the 15 years since that was made. In that documentary, which sought balance in its presentation, (it was, after all, a product of PBS which seems to have balance a stern requirement) the Luddite view was embodied in Neil Postman. Postman had been an instructor on the CBS TV program, Sunrise Semester, an early attempt to use television as a formal educational medium. If memory serves, it was a failure and thereafter Postman became a critic of the use of technology in instruction. At the time net.Learning appeared I thought of Postman not just as a Luddite but also as a pompous and pedantic person not to be taken seriously. I certainly have a pedantic streak in me. My pompous aspect is still a work in progress. In any event, my plan was to find the critique offered in net.Learning and see if it still seemed to have validity. (At the essay by Postman I linked to, near the end of the piece he makes some interesting points, in my view, but there is nothing specifically about video.) That would have made my job easier. Why write a long critique of Harden's essay if somebody else has already done so to good effect?

Alas, PBS has retired the net.Learning site. (That's too bad, since the picture of students on the site was from my intermediate microeconomics class.) I did learn that there is a video still in circulation based on the documentary. For folks at Illinois, the Library has a copy that can be accessed online. It is under password protection, however, so not available to the general public. I did watch it. Unfortunately, the segment with Postman was omitted. A good bit of what's there looks like an infomercial for the U of I's online efforts circa 1998. Burks is featured prominently. They show him on the road getting access to the network while at the airport or in the cab to and fro. (If memory serves, Burks was using ATT Worldnet at the time, a dial-up service. They didn't explain how he got online without actually having a phone line at his beck and call.) And they showed Burks doing one of his presentations where the PowerPoint slide had a graph of the growth of Internet usage measure in gazillions, which had the audience chuckling. There was a segment on NetMath mostly featuring the happenings at a rural high school in Illinois that didn't have the resources to teach a Calculus class itself. Near the end of that Jerry Uhl (now deceased) and Godfather of NetMath makes an appearance. And they've got Leigh Estabrook (former Dean of the Graduate School of Library and Information Science) talking about the LEEP program, via a Library student from Alaska, where they practice intensive face to face meeting of students both in the summer as predecessor to and then subsequently one weekend a semester to sustain the online interaction that happens the rest of the time. Estabrook also states unequivocally that going online doesn't enable increasing class size. The LEEP program was an early user of live synchronous sessions online, in addition to substantial asynchronous interaction and the facilitation that the faculty provided in their classes was arduous.

There is also a nice segment with Peggy Lant who became something of an evangelist for the use of technology in instruction and in particular for students writing online and responding to writing online. As a consequence of that documentary, we got Peggy to be one of our featured plenary speakers at FSI (Faculty Summer Institute). If memory serves, the first time she spoke the atmosphere was emotionally charged and some science faculty member in the audience (Peggy teaches English) challenged her in some of the assumptions but by session's end had come at least part way to her way of thinking. It was a truly excellent session.

The last segment of the video is with a student taking a correspondence course via VHS videotape viewed on his television. He is a widower with two teenage children, a farmer by occupation, a Captain in the National Guard who needs to earn his Bachelor's Degree to move up to the rank of major - the quintessential adult learner. He dislikes the videos - they're too slow and there is no way to ask questions. He says that when he on occasion goes for live sessions on a weekend, most of the students don't ask any questions, but he does. This bit is the closest in the program to offering a critique of Harden, though the gamut of issues with online learning are presented through the various segments.

Since the segment with Postman is nowhere to be found and since many people will not have access to the video, below I will offer up my critique of Harden's essay. It has two interrelated parts, one on pedagogy, the other on economics. Normally I begin with economic arguments, since as an economist that is where my opinion should carry the most authority. But in this case I'll make the economic argument at the end because it will be much easier to make after the pedagogical points have been introduced.

A good place to start with the pedagogy is Chickering and Ehrmanns Implementing the Seven Principles: Technology as Lever. I will focus on just two of the seven principles, #1 Good practice encourages contacts between students and faculty, and #4 Good practice gives prompt feedback. In effect, Harden argues that MOOCs have rendered principle #1 obsolete. Via automated conditional feedback or artificial intelligence and through the use of a social network that connects peer students taking the course, principle #4 can be satisfied while abandoning principle number #1.

It is incumbent on us to examine this proposition more closely. The nature of the subject matter is of consequence here as is the intellectual capacity of the student. Other factors may also matter but let's stick with these two as they are sufficient to make the point. Automated feedback may indeed be sufficient to analyze whether a piece of computer code written by a student has any bugs in it or to resolve whether a numerical solution to a system of equations solved by a student is correct. It is these sort of "analytic" domains of study where the case for MOOCs is probably strongest. In contrast, if what a student produces is primarily narrative in nature, including proofs of mathematical propositions in "why" questions posed even in analytic domains, then the efficacy of automated feedback is much more limited and human feedback therefore becomes more important. One then should ask whether there is sufficient expertise among peers in the class to provide that sort of feedback. It may be that there is. But, if not, then the role of the faculty member as the resident expert and the need for that faculty member to facilitate the discussion among the students becomes paramount. In this case, the world looks much more as Leigh Estabrook described it where class size needs to be limited because faculty facilitation is a labor intensive activity.

In my reading of the Harden essay, he simply ignores the subject matter issue. When I was a freshman back in fall 1972, Calculus was offered as a self-paced course. There were optional lectures. There were six exams to pass. If you failed one you could try again a second time. The exams were graded by upper level undergraduate students who knew the subject matter cold. This was MOOC-like efficiency before the Internet was around and I thought it a reasonable approach for the particular class. But no other courses were offered this way. Might it be that we arrive at a core group of classes that can be taught effectively as MOOCs and that's as far as we ever get?

Let's turn to the student capability. I will present two stark types of students, easy to refer to, but each type more extreme than is realistic. The first is the student as a latter day young Abe Lincoln. The myth about Lincoln the learner, you'll recall, is that he walked the many miles to and from school during the day and read by candlelight at night, presumably reading on subject matter outside the school curriculum. No extrinsic motivation was necessary for Lincoln. He was entirely driven by his own desire to learn.  Further, no feedback was needed for his nighttime reading. He was a genius and could process what he read on his own. If all the students are Lincoln-like in their approach and their intellectual capacities, MOOCs will work wonderfully. There'd be no reason to worry even about principle #4 being satisfied. The entire issue would be to provide sufficient breadth of high level content. A more recent articulation of this view is by the lead character in the movie Good Will Hunting, in the scene in the bar where he asserts the key to his education was a library card and late charges, orders of magnitude less expensive than the Harvard tuition his soon to be girlfriend was paying. Now we have videos replacing books in the argument, but otherwise it really is the same thing.

The other type I'll also take from movies, this time its the students in Ben Stein's classroom in Ferris Bueller's Day Off. These kids couldn't care less. Nothing will get their intellectual motor running. They have a Little Rascals' view of school, but now they're in high school so the antics are different. Some of these kids nonetheless end up going to college. They're the ones who provide the subject matter for Academically Adrift. Harden argues it's like throwing money down the drain to give these students a big bucks residential undergraduate experience. In good Marie Antoinette style he argues instead, "Let them eat MOOCs!"

Apart from the elitism in the argument, Harden doesn't fail to mention that he himself went to Yale, it is conceptually flawed. Our own type, presumably somewhere in between the two extremes, is not exogenous as Harden presumes but is strongly influenced by the academic environment the student confronts. While almost none of us are close to being latter day Lincoln's we have the capacity to exercise our intellectual capacities far greater than we actually do, which is one of the main themes of Flow, mentioned above. The key, then, is to convince the student that doing so will provide a good deal of satisfaction and, indeed, that striving for this provides a raison d'ĂȘtre. This, then, gives a rationale for college, quite apart from whatever vocational education the undergraduate experience provides in addition.

I've elaborated on this primary purpose a few years ago in a piece entitled The Purpose of General Education. That piece represents an ideal, one we are falling far short of today. We should ask, would a widespread deployment of MOOCs as an alternative to the residential experience do a reasonably good job in approximating the ideal? Let me note that there already is a different alternative, instructor facilitated online learning with modest class sizes. That segment of higher education has witness rapid growth. The bulk of the learners in online programs are non-traditional students meaning they are older, quite possibly have parental responsibilities, hold down a job, and so must balance their education with their other life obligations. In contrast, the traditional 18 - 22 year old student is single, works at most a part time job, and is free from other obligations so school is the focus.

Harden doesn't really concern himself with the non-traditional learner so he doesn't ask whether MOOCs might replace the distance learning that had become quite popular since the net.Learning documentary and of which the LEEP program represents a particular variant. I don't know the answer to that question. I know that six months to a year ago that question was of some substantial interest on the Sloan-C listserv. But most of these older sort of online programs, particularly at the undergraduate level, have far more modest tuition (and of course no room and board cost) than their residential college counterparts. So Harden's focus is on the latter, because that's where the big cost savings with education appear to be had. In the rest of this piece I'll follow Harden in that focus.

One significant aspect of traditional learners is their immaturity.  It manifests in a variety of ways - a focus on their new found freedom gained by living away from home rather than on their studies, a lack of a sense of responsibility to their parents when it is the parents who are paying the tuition, and a misguided view of school as a passport only, not an end in itself, a way to discover means for self-expression. Couple this with the observation that traditional learners nowadays are digital natives and that their use of technology when they were younger was primarily for recreation - think computer games and Facebook. Now ask the following question. If you want to create a transformative educational experience for these learners, should the learning environment aimed at encouraging the transformation be the same as the recreational environment or not?

In fall of 2008 when I was still working and we had just moved a large Finance class to blended format using videos of the lecture to replace one hour a week (out of three) of in class meeting time, we got an overwhelming response at the time to bring back the face to face lecture and ditch the videos. Some of this was our implementation; we had yet to move far down the learning curve in how to do a blended course effectively so the online pieces of the course didn't fit well at the time with the face-to-face components and the students were letting us know that. (This has since improved and the students no longer appear rebellious about the course.) But another issue was about them. The students wanted to have their class time scheduled and in that sense going to class is a personal commitment mechanism. Immature students need such personal commitment devices. This is an argument that sometimes constraint is better than freedom. Asynchronous learning may be wonderful for those who will act responsibly regardless of the mode of instruction. For other students, however, scheduled class meeting times provides a benefit.

There is a further issue that students are prone to multitask if they are sitting in front of an electronic device. Their attention becomes divided. I'm reminded about the old moron joke where the guy was staring at the can of frozen orange juice because it said, "concentrate." Indeed, the instruction should encourage the students toward thinking deeply and it should encourage students to put in enough time on task to master what they are studying. We know, however, that students seriously underestimate how long it takes to learn something, when to start writing a term paper, allowing for the possibility they might get stuck in working on a project, etc. With mature learners the technology might be neutral on these issues. But with traditional learners isn't it reasonable to expect pernicious consequence from going all MOOC? Harden doesn't take up this question. He presents MOOCs as if they are a free lunch. They aren't. There's good reason to expect that if MOOCs were widely deployed as an alternative to residential education they'd actually make student performance worse. (I should note that at places like the U of I much of general education is taught in large lecture mode and the large lecture may be equally as bad as the MOOC in encouraging the multitasking and thereby dividing the student's attention. So I'm not arguing here as a way to defend current practice, but only as a way to challenge the Brave New World thinking that Harden is advancing.)

This brings me to the next point, which is that you're a Luddite if you are ok in using the technology but think it largely should be a complement to face-to-face instruction, not a substitute for it. The folks who are talking about flipped classrooms are not getting any cost savings from the activity. Presumably they are doing it to further engage the students and because the old hypothesis, students did the assigned readings before coming to class, was measurably proving to be false. So maybe because the videos are easier to get at the meat of the subject and shorter in duration students will access them ahead of time and then be ready for class activity, which now has a social aspect to it mixed with the learning and therefore being prepared is a way of respecting one's classmates.

But worrying about quality of instruction puts you behind the times. With the hyperinflation in college costs we've experienced over the years, and as Harden mentioned the high debt burden that many recent graduates have to carry as a consequence of the hyperinflation, producing a cost effective solution to college is the new game in town and MOOCs look to be the game winner.  We've reached the point in the piece where I can now put forward the economic arguments.

Economists distinguish investment made under a perfect capital markets assumption, which leads to the socially efficient investment level, from investment done under a liquidity constraint, which leads to a smaller investment than is efficient, because that's all that can be afforded.  The traditional model for funding undergraduate education was to have third parties fund the investment, in the case of  public universities that was mainly tax payers, and thereby approximate the efficient investment level in human capital investment.  Presumably the enhanced productivity that the added human capital created would generate greater earnings for both the graduate and that person's employer.  This, in turn, would generate greater tax revenues and the entire investment would therefore pay for itself.

This way of thinking didn't survive the hyperinflation and the Reagan Revolution, so the cost share borne by tuition rose precipitously over the last 30 years and for a middle class family the price of a college education has become near unaffordable, especially in the absence of scholarships.  These middle class families are liquidity constrained and the ones for whom a decent low cost solution is being targeted.   It is important to observe in the economics of this that the investment in human capital temporally precedes the return from that investment, so while the cost of the education is near certain the return is far from it.  It is an investment that is hard for families of modest means to diversify.  So, from a social view, the wrong people are bearing the return risk.  One can make an argument based on moral hazard that it is necessary for the student and their families to bear the return risk, lest the student not try hard to be productive as possible post graduation.  Personally, I find that argument unpersuasive.  Not all immature 18 year olds will turn into industrious and hard working 25 year olds, having first entered the labor market.  But most will and the reasons have much more to do with upbringing than with debt overhang.  I know the default rate on student loans is quite high now, but much of that has to do with graduates being unable to find work, in my view. Solve the employment problem for new college grads and you will solve the loan default issue.

In these circumstances, from an individual student perspective or from the perspective of their immediate families, a low cost solution is very attractive.  Indeed, for the last fifty years or so, there has been at least one alternative low cost solution.  Community colleges and/or less highly ranked four-year public institutions have carried low tuition rates. From an economic efficiency perspective, one needs to consider lifetime earnings appropriately discounted from each possible solution and net tuition out.  The path that gives the higher net benefit is the efficient one to choose.  The liquidity constrained outcome happens when for a particular student an expensive college education is the efficient solution but the kid goes to community college instead or opts to delay college entirely and work full time at a modestly paying job, because that's all that can be afforded at the time. (People may do good works of very important social value but that carry little to no monetary reward for themselves.  If that social value can be monetized, the same sort of calculation can be done to determine the efficient education path for the individual.  If not, then for now it is safer to ignore that case here.  Regardless, what comes next in my post won't speak to that situation.  I simply want to mention here that the case for a high quality education is broader than what I discuss below.)

One wonders in the liquidity constrained case whether some third party other than taxpayers might step in to move the student's education to its economically efficient level.  The obvious candidate for such a third party is the future employer.  Presumably, the future employer is able to internalize the productivity gains from the the enhanced human capital, though it likely means the employee would be attached to the firm for some time after graduation, perhaps with the student and the future employer writing some long term contract concerning work, wages, and buyout from the relationship.

Obviously future employers wouldn't do this to be good Samaritans.  They would do it only if it makes sense from a long term profitability point of view.  So we should ask whether it makes sense or not in that dimension.  This is a different sort of criticism against Harden's piece.  He makes his prediction looking only within Higher Education.  He doesn't look at all at the determinants of demand for its graduates.  Presumably, whatever does come next does a reasonable job of matching supply with demand.

The current labor market seems to be characterized by excess supply in aggregate, which is why the unemployment rate remains high, but there is excess demand for certain types of jobs, vacancies that are hard to fill because people with the right skills appear to be in short supply.  Standard economic theory suggests that the wages and benefits for those jobs that are hard to fill should rise - employers who are desperate to fill such positions will bid people away from other places where they are already working.  It is only after a round or two of this type of musical chairs where it may begin to become apparent to the employers that the skill set is scarce in aggregate.  It is at this point where employers begin to entertain the idea of expanding the supply via their own investments.  This, at least, gives a first impetus for employer pay of tuition.

My friend and former colleague, Al Roth, who recently won the Nobel Prize in Economics (shared with Lloyd Shapley) is an expert on "matching markets" where factors other than price determine how matches get determined.  He has argued quite convincingly that if the matching mechanism is not stable, then matches start earlier and earlier.  (Think about big time College Basketball where it is not uncommon now for a coach to recruit an eighth grader who shows promise, while such behavior would have been uncommon if not nonexistent in the 1970s or earlier.)  If Roth's prediction holds true for the job types that are in scarce supply, this gives a further impetus for employer pay of tuition.

Note that employer pay is not unusual at all for executive education, where the student is already an employee, in their mid thirties to mid forties, with proven worth already and yet many years ahead of further productivity.  Younger potential employees have even more years of future productivity ahead of them.  What they don't yet have is proven worth.  So one needs to imagine mechanisms for establishing that.  Having internships of some sort or early entry modestly paying jobs that can signify the employees worth are candidates for such identification.  There would be more risk in paying for the education of such individuals, as there would be substantially less experience on which to base the judgement.  So the employer would need to be able to diversify that risk and would itself need to not be liquidity constrained.  This adds yet another factor of the current labor market into the story.  One reads periodically that Corporate America is sitting on a pile of cash, on the order of $1 trillion.  These are funds waiting to find a good investment opportunity.  What I'm suggesting here is that one such opportunity is to ensure a talented supply of younger workers who can sustain these companies in the years ahead.

Let me close with one last observation.  There is much being made at present about coupling MOOCs with some sort of certification that comes out of testing.  Certification is being regarded as a kind of magic elixir, because it in itself offers the proven worth of the student and therefore that the approach with MOOCs can resolve the aforementioned excess labor demand for certain job categories.  Unfortunately, certification does no such thing.  Certification provides an academic credential.  The correlation between someone having that academic credential and productivity on the job is something employers need to determine, based on their experience hiring people with the credential.  That productivity is determined in addition, presumably, by what else the employee knows, the employee's capacity to learn further, and the employee's drive in doing the work.  The credential communicates something, but that is far from all that the employer wants to know about the employee.  For years Microsoft has offered training coupled with credentialing .  My understanding of that is that Microsoft credentials were reasonably effective indicators for entry level positions, but not of much worth beyond that. precisely because the non-credential factors matter more in higher level positions.

Harden talks about a bubble in higher education, referring to the high tuition rates and the recent high growth rate in them.  But maybe it is MOOCs that are the bubble.  Many people seem to be bullish on them now.  Presumably the tonic to a bull stampede is to look closely at fundamentals, which is what I've tried to do in this piece. Harden claims to know the future.  I do not.  It is easy to see possibility, much harder to determine likelihood.  The normal approach to uncertainty is diversify and hedge.  The futurists, like Harden, want us to bet the house.   I'm not at all comfortable doing that.  If that makes me a Luddite, so be it. 

No comments: