Gabriel Rossman, a sociologist at UCLA and a friend, kindly agreed to write a guest post on higher education. He blogs regularly at the excellent Code and Culture.
Recently Mark Bauerlein did something rather mischievous in trying to measure how much value society actually gets for its investment in the research function of universities and found that in most cases nobody reads the research. Bauerlein emphasizes the humanities, wherein research largely consists of providing exegesis on an already well-understood canon. Nonetheless, you can also imagine applying his critique in somewhat attenuated form to the social sciences and STEM fields. If Bauerlein is right and the modal academic creates little or no valuable research, this suggests that the light teaching loads — four courses a year is standard at research schools — is probably not a very efficient way to organize the higher education system.
Walter Russell Mead took up this theme and ran with it, suggesting that in an era of austerity it is more or less inevitable that all but a handful of rich private universities will have to become more efficient at undergraduate education by greatly reducing PhD programs, which means both fewer grad students and fewer faculty with the light teaching loads that allow them time to do research. Actually, one could still get this outcome even if the median academic’s research is valuable. We need only suppose that the social value of research is difficult to measure or demonstrate to a skeptical public facing rising tuition and conflicting demands on state spending. We’ve seen this recently in Texas, where UT-Austin is a legitimately world class research institution that produces research of both intrinsic merit and economic spillover value for Texas tech startups. Nonetheless Governor Perry has been pressuring the university to charge extremely low tuition, which in practice would mean very high teaching loads that would turn UT-Austin into a third-tier land grant college.
Hence Mead seems to be thinking along Herb Stein’s logic that if something can’t go along forever, it won’t. This may be true, but inertia can be more powerful than you might expect. We can see why by considering the two possible routes by which we might arrive at the new Meadian equilibrium: reform of incumbents and disruptive innovation by new market entrants.
Incumbent schools have an extremely strong preference for a research mission and are resistant to any attempts to changing it. Indeed, there is a pronounced tendency for schools to expand the research mission over time. For instance, undergraduate departments often seek to develop graduate programs. The basic reason for this is the fact of and resistance to downward mobility by the faculty. Top departments produce more PhDs than they hire which means that almost all academics were educated at schools more prestigious than the ones they work at. (Notably, in Mead’s scenario this structural downward mobility would be exacerbated). Academics develop strong expectations of the faculty role in graduate school and carry this into their teaching careers. Furthermore, the academic labor market is structured around research in part because it is valued more highly, but also because it is intrinsically easier to measure research than teaching productivity. Thus in order to get promoted within one’s institution or to attract outside offers, one will tend to emphasize research — and resist attempts to push one towards emphasizing teaching. This will be especially acute in transitional periods. It may well be that in the long run research expectations will adjust if we move from a four course teaching load to, say, a six course teaching load. However academics intuitively understand that this adjustment will not be instantaneous and in the the medium run there will be a learning process for provosts and hiring committees during which most of the junior faculty with these higher loads will be denied tenure or fail to get external offers because they failed to meet the standards of research that were reasonable in an era of light teaching loads.
Another aspect of Mead’s projection is that with higher teaching loads there will be less demand for new PhDs, which implies fewer graduate programs, which in turn implies more faculty time to devote to undergraduate courses and even less demand for new PhDs. This model assumes that aspiring academics are rational and ignores that there is already a great excess of unemployable PhDs (to say nothing of drop-outs from PhD programs). It is with good reason that Fabio Rojas made “Do Not Go to Graduate School” the first chapter of his handbook for grad students. Nonetheless there seems to be an inexhaustible supply of talented people continuing to apply to graduate school. Moreover, competition between these aspirants is becoming more intense and in the last ten years it has become much more common for people to apply to PhD programs after first getting a terminal MA, which in effect means that people are now paying $20,000 for the privilege of even applying to PhD programs. That people pursue academic careers even under such unfavorable circumstances shows just how much slack there is on the supply side of the academic labor market and how graduate programs could persist even as demand for new PhDs diminishes as undergraduate course loads rise for existing faculty.
If the incumbent institutions are unlikely to reform themselves, this still suggests that we may see disruptive innovation from new entrants. Most versions of this involve distance learning, which takes the lecture/discussion model to its logical limit by having an unlimited number of students watching a professor lecture and then interact with teaching assistants. There are two problems standing in the way of such disruptive innovation.
One is the issue of adverse selection. Disruptive innovations typically start down-market and work their way up. However education is a market where the inputs (that is, the students themselves) matter as much as the service and this will doom any entrant who starts too far down-market. The most convincing part of Tyler Cowen’s The Great Stagnation is his argument that we have already recruited all of our young people who are easily educable and the marginal college matriculant struggles in school. Most of the pupils at for-profit colleges like The University of Phoenix have been marginal students who lack the preparation and/or full-time commitment to pursue coursework on anything like the relatively advantageous terms facing traditional students. Not surprisingly, the vast majority of these students fail and end up remaining uncredentialed and unskilled, but now with the added burden of student loan debt. Even if we put aside the question of how well the for-profits educate these students, they seem to be doing them a disservice merely by recruiting them. This constellation of dropout, debt, and default has promoted a backlash against these schools among policymakers (which the industry has resisted with an impressive lobbying and PR effort).
The other issue with disruptive innovation is that there is a problem of certification. Yale, MIT, and Stanford have been leaders in putting their course materials online where anyone can study for free, but ultimately most people do not pursue education for the sheer joy of it but to leverage in the labor market. Unfortunately one cannot apply for a job and tell the human resources person “I don’t have a degree but I spend a lot of time in the library, I’ve watched dozens of courses on iTunes U, and trust me I have really retained the material.” Hence any radical alternative to our education system must take seriously the “signaling” model of higher education, which holds that education is less about educating students than it is about screening them. However excessive this might be for the purpose, employers are used to this system and would need an alternative way to screen labor. Hence some visions of disruptive innovation include a system of degree exit or occupational certification exams. Transitioning from a credential system to an exam system faces severe legal obstacles. Under Duke v. Griggs, an employer trying to use a test which has disparate impact by race (which most tests do) bears the burden of proof to justify use of the test. Likewise, as any university professor can tell you, the Americans with Disabilities Act makes a mockery of testing by requiring special accommodations for students who claim vague learning disabilities and one can count on such claims becoming more common as the tests become more high stakes. This issue is even worse with Breimhorst v ETS, a case that prohibits flagging exams taken under special circumstances. Thus we might imagine employers considering the use of tests to identify promising autodidacts (who would be attractive as their lack of student loan debt would make them cheap). Such employers will have to worry about lawsuits on the one hand and dubious test scores on the other, all of which makes sticking to the traditional BA system relatively attractive.
Both the primary stakeholders (ie, the faculty) and the downstream consumers (ie, employers) of higher education have reasons for favoring the current system. The former can be expected to fight tenaciously to preserve the current system and the latter to be sluggish to embrace alternatives. As such, notwithstanding the apparent contradictions identified by Professors Bauerlein and Mead, the research university will probably continue as a mass institution for the next few decades.
Editor’s note: Gabriel’s observations concerning the barriers to an exam system strike me as particularly salient to those who aim to “disrupt” higher education. Can we imagine Congress revising the Americans with Disabilities Act, and finding ways to address concerns regarding disparate impact and the flagging of exams? And would these theoretical efforts pass constitutional muster?