January 27, 2012
The cost of knowledge
Did you know that Congress is considering prohibiting the free dissemination of knowledge produced by federal research dollars? That's what the Research Works Act would do. The bill is backed by companies like Elsevier, who profit mainly from the volunteer labor of scientists (who produce the research and who vet the research for quality), and thus have a vested interest in preventing the free exchange of knowledge [1,2], or at least in extracting rents from everyone involved in it.
Other commercial publishers may not be as bad as Elsevier, but there are serious problems with them as well. Computers have arguably reduced the valued added by commercial publishers because they allow scientists to do themselves many of the tasks that publishers used to perform (like typesetting, spell checking, etc.), and they have virtually eliminated the cost of distribution and storage. Prof. Michael Eisen, writing in the New York Times , laid out the case for why open access publishing is not only realistic, but also morally responsible. To be honest, I am deeply sympathetic with these arguments and am reminded of them whenever I try to access journals when I'm off campus.
More recently, the Fields Medalist Tim Gowers  has started a petition to let working scientists declare their opposition to the Research Works Act  by promising to boycott Elsevier. (See also his explanation of why he's doing this.)
You can help by declaring that you (1) won't publish with them, (2) wont' referee for them, and/or (3) won't do editorial work for them. Please consider signing up, and also encouraging your colleagues to do the same:
Finally, John Baez has some thoughtful analysis of the problem, its origin and some potential solutions on his blog.
Tip to Slashdot.
Update 27 February: Elsevier has dropped support for the Research Works Act, and has written a letter to the mathematics community. The claim they will now reduce the overall cost they impose on the mathematics community, but in fact, this is merely a cynical sop because mathematics is a tiny part of Elsevier's portfolio.
 Elsevier used to make some money from the military arms trade, but partly due to a furor raised by scientists, it eventually cut its ties to the international arms fairs in 2008. Given Elsevier's history, it seems unlikely that they would have made this choice without the public pressure the furor generated.
 Elsevier is perhaps the worst offender in the private scientific publishing industry. Their journals (even the crappy ones) typically cost significantly more than other private or non-profit publishers, they've even been caught taking money from the pharmaceutical industry in exchange for creating fake medical journals in which to publish fake research, and a few of their journals have been implicated in more academic types of fraud.
 Michael Eisen is one of the founders of the highly regarded Public Library of Science (PLoS), an open-access scientific publishing group. PLoS's founding story is relevant: it is well respected in the scientific community because many of its original journals were started by the members of journal editorial boards for Elsevier, who resigned en masse in protest over Elsevier's odious practices.
 The Fields Medal is a bit like a Nobel Prize for Mathematics.
 The bill's name really is a lovely example of Orwellian double speak.
January 13, 2012
A crisis in higher education?
Attention conservation notice: 3200 cranky words on the PhD over-supply "crisis."
Higher education is always having crises, it seems. Some of this is probably a kind of marketing strategy, because genuinely serious problems are so systemic and slow-moving that it's easy to ignore them, or because you can't get people to pay attention in today's saturated media environment without a large dose of hyperbole. But, one "crisis" in particular did catch my attention over the past few years: the harsh market faced by relatively new PhDs seeking faculty jobs . Nature did a full spread on the future of the PhD, The Economist weighed in with their own coverage, and I'm sure the Chronicle of Higher Education has done a number of stories on the topic. Now, Online PhD has done its own report, in the form of a slick infographic packed with grim factoids and numbers.
What most of these perspectives miss, and what makes some of their analysis a little shrill, is the historical context of higher education and its growth trajectory over the past 70 years . The overall upward trend in PhD production over this time period can be blamed on huge increases in federal funding for research, on huge growth in the number of students getting undergrad degrees, on a vast broadening of higher education as a whole and on intensified competition between research-oriented universities.
The role of competition, I think, is under appreciated: many more universities now produce PhDs and many more programs are genuinely good than was the case before federal funding for higher education began surging after World War II. The result is a larger and more competitive market for those PhDs, especially the ones produced by the best programs. (The same is true for funding sources: the pie has grown, but the number of people competing for a slice has grown much faster.) In many ways, this is a good thing for higher education overall, since students can receive a good or even a great education at a wider variety of places. That is, higher education is probably genuinely less elitist and genuinely more accessible and inclusive. There's also a cost, however, in brutal treatment that even successful candidates experience on the market and in the colossal waste of human potential from the many talented individuals who fail to find good jobs. (More on "good" jobs in a moment.)
That being said, increased production of PhDs doesn't necessarily mean increased competition. If the number of tenure-track faculty positions increases at the same rate as the production of PhDs, then in principle competition could remain flat. This point gets a lot of attention in the popular discussion and the argument is often that if only we could increase the number of tenure-track lines, everything would be great. But this obscures the complexity of the problem. First, faculty lines are largely paid for by undergraduate (or professional degree) tuition , so increasing the size of the faculty requires increasing the size of the undergraduate population, which has its own problems.  Second, part of the modern National Science Foundation's mandate is actually to overproduce graduate students , and this is largely at the behest of Congress . Alternatively, we could solve the over-supply of PhDs by reducing the overall production, but this would negatively impact the amount of research being produced (since it's largely done by PhD students), the teaching of many departments (again, often done by PhD students) and would reduce the supply of highly educated individuals to non-academic professions.
Third, not all tenure-track positions are equally desirable, not all PhDs are equally competitive, and growth in the most desirable slots and most competitive people has not been uniform. This is a tricky problem to explain but let's put it this way: I would not be surprised to learn that the 10 best programs in the country (in any particular field) collectively produce enough PhDs each year to fill every advertised faculty line at every other university, even the not-so-great ones . This means that the lack-of-tenure-track-jobs / overproduction-of-PhDs "crisis" is not one that everyone feels equally, which complicates the conclusion that it is universally a problem. In fact, a tight job market for faculty positions has some benefits, at least collectively. One is that lower-quality places can pick up relatively better qualified people than they would if the top-ranked departments had enough extra lines to hire all the top applicants. Over time, an over-supply of good PhDs may be necessary to raise the quality of the worst-performing institutions, although this effect may only be observable in the long run. 
Fourth, the history of higher education as an industry is a series of large expansions and contractions, and the effects of these are often felt and distributed unevenly . Life and job prospects for faculty in expanding fields are good, but are hard during contractions. (These effects are surely amplified for young scholars and so one possibility would be better knowing and advertising the true employment prospects for graduates; but maybe not .) It's not entirely clear to me that academia is actually experiencing a contraction, despite the federal budget travails. A more truthful statement may be that higher education is restructuring, which brings us to the issue of "good" jobs versus "bad" jobs.
It's true that universities (at least in the US) are increasingly made up of two types of faculty, those either with or who are eligible for tenure ("tenure track"; a population that is, at best, growing fairly slowly) and those without or who can never receive tenure (teaching faculty, adjuncts, etc.). The latter group is much larger now than it used to be, but it's not very well integrated into the decision-making systems of universities, and this, I think, leads to some level of systemic abuse. In the long run, it seems likely that these groups will become better integrated into the decision-make system, which will reduce the abuse . But a more interesting question, I think, is why has this population grown so much so recently?
The role that universities play in society is changing, and I think the growth of these lower-quality jobs reflects this shift. The US economy overall has shifted significantly toward service-sector jobs and the growth in adjunct and teaching positions at universities should probably be seen as the higher education equivalent. This may be driven in part by the commoditization of a bachelors degree (which is primarily what non-tenure-track faculty help produce), which society has demanded and the universities have embraced (especially the big public universities and the non-endowed private colleges, where increased enrollment means increased tuition revenue). For their part, colleges and universities are figuring out that they can produce an apparently equivalent "product" at significantly lower cost by relying more heavily on non-tenure track faculty [11,12]. It seems telling that losses of tenure-track lines are often at colleges and universities well below the "top tier", where the struggle for product differentiation and the negative consequences of price competition are likely stronger. So, it seems reasonable to expect growth in these "bad" jobs in places where the service rendered (education provided) is less specialized, e.g., entry- or lower-level undergraduate classes where the material is highly standardized and probably does not require the best of the best minds to teach.
Another aspect is that tenure is not just about protecting faculty from being fired for political reasons. Tenure also allows universities to fulfill their mission toward preserving knowledge because tenured faculty will be around for a long time, communicating their vast and detailed knowledge to the next generation. Eliminating tenure lines may basically mean that an institution is giving up some or all of its commitment to the knowledge preservation mission. This is surely a loss for society as a whole, but it does raise the interesting question about which institutions are best positioned to fulfill that mission -- it may be that the institutions who are giving up on it were not doing a very good job at it in the first place. The fact that tenure lines are mainly (but not always) being lost from the lower-ranked institutions suggests that the top places are largely still committed to this mission, even if they are retrenching to some degree (perhaps because of the shifting demands on bachelor degree production described above).
So, let's take stock. Is there a crisis? Not in the usual definition of the word, no. But, there are serious issues that we should consider, and these tap deep into both the mission and purpose of higher education and its relationship to society as a whole.
The good things for society about the current system are that the over-supply of PhDs produces a steady stream of highly educated people for other industries and government to use. The over-supply means that low-quality departments will tend to improve over time because they can hire better people than their peers tend to produce. The over-supply also means that the best or most desirable departments will also tend to improve over time because they can select their new hires from the very best of the very best. For scholarship in general, this is a good thing. The over-supply means that academia has a large supply of low-cost skilled labor (graduate students) for producing research, educating younger students, etc. And, the over-supply means that academia has an adequate supply of potential faculty to facilitate restructuring needs, i.e., responding to the changing demands from society and the changing roles of universities.
The bad things are that the over-supply is a colossal waste of human potential for people who aspire to be faculty but who ultimately fail to find employment. For instance, many very talented individuals will spend substantial time in low-paying, low-benefits temporary employment (graduate students, postdocs, adjuncts, research faculty positions, etc.) only to discover years or decades later that these years are now counted against them on the job market (and not just in the academic market). The over-supply makes the individual experience of finding a job fairly brutal and with a high error rate (many people who should get faculty jobs do not ). Success also comes with a cost in the form of moving a very large distance (the faculty job market is one of the few truly national labor markets). The over-supply has made it easy for susceptible colleges and universities to slowly replace their tenure track faculty with non-tenure faculty with less autonomy, less security, lower pay and lower benefits, which ultimately means these institutions basically abandon one of their missions: preserving human knowledge. It also makes the institution less democratic, which likely has a negative impact on the campus culture and the educational environment.
Just as this situation did not appear suddenly, I don't think it will change significantly in the near future. Although Congress is a powerful voice in higher education, and has had a direct role in creating the over-supply, the large and complex ecology of higher education institutions, society itself and the economy as a whole are also key players. What happens will depend on their interactions, and lobbying Congress alone may lead to unexpected and undesirable results . In the near term, I think the over-supply will persist (and if anything the job market will become even more competitive, but again this is not a completely bad thing), the number of non-tenured positions will continue to increase (mainly at lower-ranked institutions or for teaching intro classes at the higher-ranked places), undergraduate degrees will become even more comoditized, and the mission of knowledge preservation will be increasingly concentrated among the better or more financially stable institutions.
One long-term consequence is a partitioning of the faculty at research universities into "research" faculty (tenure-track faculty who do research and teach mainly graduate and upper-level undergraduate courses, of which I am one) and "teaching" faculty (non-tenure track faculty who teach heavy course loads of lower-level undergraduate classes), but that does seem like the way things are going . I wish that research universities (and tenure-track faculty) would treat the non-tenure groups with more respect and include them more directly into the decision-making processes. And, I hope that we can find better ways of encouraging the very best young scholars to stick with it, even though the system will likely become only more brutal in the future .
To end on a more positive note, one genuinely beneficial thing we as academics could do would be to encourage our PhD students to consider non-academic trajectories. That is, I don't think we should view the PhD as being exclusively an academic degree, and we could strive to teach our PhD students a combination of both academic and practical skills. This would increase their options on the job market, which may reduce the overall brutality that individuals currently experience.
 Partly because I was in that market myself. And, now being in a tenure-track position at a good university, I'm lucky enough to be on the other side of that harrowing process. Had I written this a couple of years ago, I'm sure I would have said slightly different things.
 These are well covered by Roger Geiger's excellent and authoritative books on the evolution of the American university system, in the post-war period and since 1990. These books are highly recommended. Geiger takes what should be a dry and boring subject and makes it a fascinating and insightful story.
 This is true at both public and private universities. The only place it's less accurate is in medical research schools, where faculty lines are often funded out of "soft" money from research grants. (Some are still funded from medical school tuition revenue, so the coupling with student populations is not eliminated.) The result is that these faculty lines are mainly sensitive to changes in federal funding levels.
 Another complicating factor is that tenure lines are traditionally tied to departments, and their number depends on student demand for those courses offered by that department. That is, teaching is still a labor-constrained activity. The division of that labor into departments means that growth in faculty lines is driven by changes in the popularity of different disciplines. The benefits for the faculty job market created by overall growth in student enrollments will thus be distributed unevenly.
There are at least two factors that decouple the number of faculty lines and the popularity of the field: tenure, which means departments tend to shrink very slowly in response to decreasing popularity while the reverse is not true, and the commitment that all universities have to the preservation and production of knowledge, which means even an unpopular department may be maintained as a kind of cultural memory device.
 This is done partly through direct support to students (about 15% of the budget) and partly through grants (50% of the budget); typically, large portions of grants are in fact used to support graduate students by paying them as research assistants.
 Apparently, NSF has always struggled to justify its budget to Congress, who generally has an uncomfortable relationship with the idea of supporting basic research for the sake of humanity. For NSF, historically "supporting education," and more recently "supporting economic growth" (a.k.a. technology transfer), have been a successful arguments, and these are reflected in funding priorities.
 This is almost surely true in Computer Science, where some of the best programs are also some of the largest. For example, MIT and CMU collectively have about 250 faculty; if they each produce a single graduate each year, that would be enough to place one person at each of the other 200-odd PhD-granting Computer Science departments in North America. The per-faculty production rate is probably not so high, the overall volume may be so if we account for other top places like Stanford, Washington, Princeton, etc. If we include the fact that not every department hires every year, it seems entirely reasonable that the top 10 places could fill the entire annual market demand themselves.
 This effect probably happens faster for newer fields, e.g., complex systems. The best universities are all fairly sensitive to their perceived prestige and quality, and for them, it doesn't make strategic sense to risk their current standing with risky investments in untested fields. This means that lower-ranked universities who place smart bets can move up (at least during the time it takes for a field to become established enough that the top places start poaching the best people).
 Physics experienced a long expansion, but that had largely run its course in the United States by the time Congress trashed the Superconducting Super Collider in 1993. In contrast, biomedical research has been expanding fairly steadily for 70 years, which is probably why it dominates federal science budgets. The "golden age" of science was really the post-war and Sputnik eras, when federal spending was expanding faster than universities could satisfy the demand for research. The 1970s were apparently a fairly broad contraction, because Congress moved to limit the growth in science budgets (for instance, NASA's budget peaked in the early 1970s) and because student enrollment growth tempered. Since then, the expansions and contractions have been less even.
 On the other hand, had anyone convincingly explained to my younger self just what life would be like in my first few years as a professor, I may have decided to try a different career path. More generally, ignorance may be a crucial part of what makes the whole system work: it allows us to unknowingly take foolish risks that sometimes yield truly remarkable, or at least highly improbable, results. At the collective level, youthful foolishness may be essential to keeping the quality of the faculty high despite the brutality of the career path.
 Of course, in the meantime it's terrible that some institutions and individuals are taking advantage of the current powerlessness of these groups. They can and should be integrated into the academy and given a voice.
 Some graduate and professional degrees also show evidence of becoming commodities, for instance, MBAs. It's not clear that PhDs are facing similar pressures, although in my darker moments I believe it.
 From this perspective, things like Stanford's free online courses may be a truly disruptive innovation. They offer the possibility of dramatically lowered cost, dramatically increased "production" and they seem to require a currently specialized set of skills. Of course, their success could destroy what remains of the tenure track at smaller or less unique institutions.
 As I've learned, it's a highly stochastic and error-prone process. Departments tend to decide ahead of time to hire in a particular area, and this means top-notch candidates from outside that area are potentially passed over for less-top-notch candidates within the target area. The decision of which area to hire in is often driven by internal politics (which "group" is stronger, which has a louder voice, "who's turn" it is) or existing curricular needs rather than meritocratic or strategic concerns. And, even within a given area, it can be difficult to accurately access true merit and relative quality, particularly for junior positions where each candidate's track record is, by definition, relatively short.
Did I mention that I'm reading PhD applications this week? Ugh.
 It certainly has in the past.
 Ironically, dividing the teaching and research between different groups of faculty is mostly how the Germans used to do things. Now, the Germans are Americanizing their system to some degree, while we Americans seem to be Germanizing ours.
 From my perspective, "early career" funding, fellowship and other young faculty support mechanisms seem to be wholly inadequate (in size and scope) and the easiest way to get them is to propose highly incremental, highly risk-averse research. This does not seem to be serving the right goals or to be teaching young faculty the right lessons about scholarship.