©Deirdre Nansen McCloskey | COPYRIGHTED MATERIAL


A Review of In Pursuit of the Ph.D. by William Bowen and Neil Rudenstine, Princeton, N.J.: Princeton University Press, 1992. 442 pages.

by D.N. McCloskey
Change, Jan-Feb 1994
Filed under articles [academic policy] [reviews]

The former president of Harvard, Derek Bok, called graduate education, so seldom scrutinized, "the soft underbelly of the research university." Now two giants of academic administration, William Bowen and Neil Rudenstine, have scrutinized and scrutinized, in 400 big pages and many scores of charts and graphs.

Bowen is an economist, the president of the Mellon Foundation, and the former provost and president of Princeton. Neil Rudenstine is an English professor, the new president of Harvard and former provost at Princeton, and second in command at Mellon. The Mellon Foundation supported their project, which was started in 1989 and completed, among their many other duties, in record time. The book claims to be the first comprehensive study of graduate education since Berelson's in 1960.

The study is "focused on the top echelon of graduate programs," namely, the authors, "Ten Universities" group--Princeton, Harvard, Yale, Columbia, Cornell (that is, most of the Ivy League), Stanford, Chicago, Berkeley, Michigan, and North Carolina. Remember this list. And it focuses on six fields, namely, English, history, and political science (known euphoniously as "EHP"), with some attention to math and physics ("MP"), plus economics. The EHP fields were the most intensively studied. Remember that, too .

The main question the authors ask is, "Is it possible to find ways to reduce attrition and time-to-degree, while simultaneously sustaining, and perhaps improving, quality?" Their approach to answering it, bracketing quality, is demographic. The book treats graduate school as a population (death is graduation with a PhD; a life getting quickly to death by PhD is optimal). Part 1 of the three parts rolls out data on death rates by academic field, which have been collected by the National Research Council since the 1920s, and data on births, populations, and deaths for the Ten Universities group.

Part 1, called "Trends in Graduate Education," surveys the demographic conditions--in economic terms, it is supply and demand, though the book does not pretend to econometric precision. Their demographic approach leads the authors to downplay the role of demand for professors in creating the graduate-study boom of the '60s, and indeed the book turns away from economic arguments.

For example, in the '60s, "large increases in the sizes of the underlying pools of potential candidates for graduate degrees were more dominant." This "more dominant" trend just happened, it would appear; the authors say nothing of its occurring in response to incentives. Undergraduates of the '50s and '60s, Bowen and Rudenstine say, just happened to major in fields that led easily to birth, death, and transfiguration through a PhD in the traditional liberal arts.

After an allowance for the draft-inspired demographic uptick of the late '60s, the "flight from the arts and sciences" in graduate schools during the '70s therefore is said to explain the contraction on the supply side--although Bowen and Rudenstine would not deny that the depressed job market on the demand side was important, too. (Oddly, they do not emphasize that the flight from the arts and sciences at the undergraduate level directly reduced the demand for teachers in their six fields.)

The chapters in Part 1 are unsurprising, as the authors themselves frequently remark. The best PhD programs by prestige in the six fields tend to be the largest. Yes. Losses of big national fellowship programs were tough on graduate support at elite universities. Uh-huh. Employment in the old fields was uncertain at best. Yup.

Part 2, "Factors Affecting Outcomes," focuses on time-to-degree, the obsession of the book. The authors claim that "the decision to focus much of this study on completion rates and time-to-degree implies no lack of concern for the qualitative dimension of teaching and learning at the graduate level." Here's "quality" again, but whatever concern they claim to have for it, it is plain that at least Bowen's heart lies firmly in the demography. Indeed, the way most graduate schools measure "quality," and the only way Bowen and Rudenstine do, is to compile GRE scores and other rankings at the beginning of graduate study. The economics here is strange. One would think that the right way to measure the quality of a program would be its value added at the end, not the cost of its inputs at the beginning. Goldsmiths are not necessarily more valuable than blacksmiths. We want to educate people, not merely place crowns on the heads of the "brightest" by measures satisfactory at age 21.

The results of focusing on the numbers, especially the numbers on time-to-degree, are not, as the authors claim, "surprising and significant." It is not surprising to anyone familiar with academic life since the founding of the University of Berlin that half of the entering graduate students do not get degrees. Nor is it surprising that the highest level of higher education takes a long time to complete; it has since Plato's Academy. Bowen and Rudenstine themselves conclude--on the basis of some canny demographic reasoning presented in Appendix D--that the official measures of time-to-degree exaggerated its rise during the bust of the '70s. There was in fact only a 15 to 20 percent rise, rather than the 40 percent reported in the newspapers and believed by some deans. The authors concede that "lengthy time-to-degree is a problem with old (and deep) roots; it is not a new phenomenon." If it is not new, then is it a "problem"?

The authors thrash about at the beginning of Part 2 trying to devise a standard whereby a 50 percent attrition rate can be called a problem. They fail because they do not think like economists here. The best they can manage is to cast envious glances at professional schools (three times in four pages) and assert nervously that "the completion rates . . . seem low by almost any standard," without providing a relevant one. Later they regress to mere over statement and alarm: "The direction of change [of attrition late in programs] is unmistakable, and the absolute numbers are high enough to be grounds for serious concern." Why exactly "high enough"? What are the "grounds for serious concern"? No standard is provided.

The correct standard is that of investment and the search for gain. That half of businesses fail in the first few years of their existence does not imply that the investments were foolish in prospect, or easily second-guessed. That half of entering graduate students decide on second thought that a life of teaching and scholarship is not for them is no more alarming than the statistics on business failure, and no more likely to be improved by sagacious second-guessing from the executive suite.

A key "finding" in the book is that smaller graduate programs have shorter time-to-degree than larger ones. One gets the impression that neither Bowen nor Rudenstine has run a graduate program: their surprise at the obvious will strike veterans as odd. Any director of a large graduate program (I was director of graduate studies in economics at the University of Chicago during the '70s, when the program was among the largest) could tell Bowen and Rudenstine why big programs have longer time-to-degree. In order to be large, a large program does not select. (In view of the inherent difficulty of accurately predicting at age 21 what the relevant performance will be at age 30, not to speak of 50, the lack of selection is not socially unwise.)

If a school selects more stringently, as Harvard economics did in the '70s and Chicago did in the '80s, two things happen. First, however poor the selection mechanism is at age 21, the department gets better students who complete degrees. Bowen and Rudenstine claim that they can extract the effect of better selection by comparing recipients of national fellowships with the rest of the students and using rankings for admission gathered from a few elite schools. As one of the rankers, l doubt it.

Second, and Bowen and Rudenstine merely touch on this, smaller programs treat students better. In them, at least until the graduate dean wakes up, the ratio of students to faculty is low. At its peak, Chicago had 200 graduate students in residence and a faculty of 25 in a program that emphasized graduate workshops. When it got smaller, it could offer better support--financial and emotional and intellectual.

There's a more worrisome oversight in the authors, analysis of large and small programs. Many of the larger programs, unsurprisingly, are convenient to large cities (thus Maryland at College Park, Rutgers at New Brunswick, or Chicago again), and therefore attract more part-time, partially committed graduate students. Older people with families who are uncertain about educational goals take longer to get a degree. What is worrisome is that Bowen and Rudenstine do not know the lives of the people they are writing about. They have not looked. In the terminology of social work, their statistics and their questions are "provider driven."

In this second part, too, Bowen and Rudenstine collect confirmations from their Ten Universities and three fields plus three (EHP plus MP and economics) of the usual attrition at the point of thesis. Administrators measure it and talk about it, but like the weather no one does anything about it. A decade ago the University of Chicago's Baker Report (not mentioned in the book) made a convincing case that we need to do something about the intellectual life of advanced graduate students. Bowen and Rudenstine commend the departments that have "instituted dissertation workshops to help combat [the isolation of doctoral research]."

But their main effort goes into gathering unremarkable and irrelevant statistics. Where cliches of administration need criticism, the book does not give it, statistically or verbally. The surface rhetoric is that of the Scientific Study examining The Data. Yet Bowen and Rudenstine do not put their prejudices in jeopardy. They knew from the president's office before consulting the data that large programs are better (any graduate dean, of course, could tell them the prejudice is false). They knew that time-to-degree is a terrible, terrible problem. False. They knew that the conventional fields of study in 1958 represent the whole. False again. Bowen and Rudenstine have not tested their prejudices against the evidence. They have opted to collect statistics that are beside the point, to dazzle the non-statistician with a plating of Science.

The statistical finding that small programs have better records of completion, unsurprising though it is, contradicts the usual administrative rhetoric in other parts of the book about "critical mass" in the scale of graduate programs. Characteristically, the metaphor of "critical mass" is here untested, though defended in footnotes as "reasonable" on grounds unspecified. The great and good know it is true, and that must suffice, data or no. The book exemplifies how much administrators are taken with certain figures of speech manipulated without thought: "critical mass," for example, or "specialization" or "building on strength."

Part 3, "Policies and Program Design," is also narrower than its title. It is not about policies and program design in American graduate education. It examines what the elite universities and the elite private foundations can accomplish with financial aid. "Money plainly matters," we are unsurprisingly told. But in their elite sample, "the form of financial support appears to make less difference than has commonly been supposed." Bowen and Rudenstine argue plausibly that at the elite universities the teaching assistants get a more structured life and have more faculty contact than do students who do not teach. Therefore they complete faster.

But their argument that teaching assistantships are therefore as good as free rides will not stretch beyond the elite, as they themselves admit uncomfortably at the end. Three sections of freshman English a term are not what the professor ordered for scholarly productivity at the University of South Carolina. Are Bowen and Rudenstine seeing into the lives of most American graduate students, or merely processing the statistics of students at Harvard, Berkeley, and Princeton? They appear to have written the book without venturing out into the hallways and seminar rooms or over to graduate housing.

Chapter 13, on departmental culture, while not earthshaking, is the best in the book. It escapes briefly from the scientism of the rest and asks questions worth answering. We learn--or perhaps come to appreciate its prevalence--that students in English and history are worried by the more sophisticated dissertations they are asked to write nowadays. We learn, from Nerad and Cerny, that "students in the humanities and social sciences . . . felt lost in the transition from what they called a class-taking person, to a 'book-writing person'." But the quality of this last substantive chapter does not outweigh the pointless data shuffling of the 12 chapters that come before.

The book is a failure. The prose is adequate, especially when Rudenstine wields the pen. The authors are often intelligent, sometimes about their evidence, as when noting that "it is generally more useful to group observations by . . . the year in which [students] embarked on graduate studies" than by their year of completion: it is the economic conditions at the decision points, not at the end, that should matter. Rudenstine, in Appendix F, gives a useful and uncommonly sensible view of the turmoil over "theory" in the interpretive sciences.

But the book does not meet the standards of its pretensions and does not yield benefits matching its social cost. As Elizabeth Hunt pointed out in a penetrating article in Lingua Franca a few months ago, the book has become the bible of graduate deans but doesn't deserve its canonical status. It's an apocryphal scandal.

Consider the economic accounting that underlies the analysis. Bowen and Rudenstine assert that "students and institutions invest massive amounts of time and other resources, and it seems only reasonable that there be standards of collective accountability." Such accounting, which omits the word "students" when it comes down to policy, is used to justify the main proposal, namely, that institutions like the Mellon Foundation should do the collective accounting, and then, with the federal government in line, should give large sums to the elite portion of graduate education. After all, "universities [by which they evidently mean Stanford and Princeton] are already carrying by far the largest share of the costs of graduate education in the humanities and social sciences."

On elementary grounds the accounting is erroneous. It is erroneous because the students themselves, as every economist understands and as Bowen and Rudenstine themselves emphasize in Chapter 9, pay the opportunity cost of graduate education. Students, time is by far the largest component of the costs of graduate education, especially in equipment-cheap fields like English or mathematics. (The talk throughout of "time and other resources," by the way, sounds strange: what's wrong with money as a yardstick for "time and other resources"?) Though some anti-intellectual journalists and other writers would have us believe otherwise, graduate students are competent people who if not in graduate school would be running things at big salaries. The cost is their salaries forgone. But the choice and the cost fall mainly on the individual student deciding to study. They do not fall mainly on the universities or the general public. The universities and the general public get cheaper teaching assistants and laboratory assistants than they would otherwise. Most of the cost falls on those who decide to incur it, the graduate students themselves.

Bowen and Rudenstine reckon that in their sample the number of graduate-student years to produce a PhD, allowing for attrition, is "stuck at an exceedingly high level": 13.4 years in English, history, and political science and 8.4 years for mathematics and physics. As I've said, they do not justify the rhetoric of "exceedingly high" or "stuck." But in any case it is the choice of the graduate students to become stuck. And, especially outside the elite of full-ride fellowships, the students who are stuck pay for it. There is no case here for "collective accountability." The graduate students, except at the elite universities, are not being supported by the Mellon Foundation. They are not middle-class welfare cases who might justify a stern rhetoric of "accountability." They work for a living and pay the opportunity cost of not working at a higher-paying job. Bowen and Rudenstine assert that "the average societal investment of student time in graduate study is high in relation to the number of PhDs conferred" (italics added). What is it, then, Mr. Educational Economist--society at large or the student? Accountant, balance thyself.

The biggest problem in the book, however, is not how the sample is analyzed. The biggest problem is the character of the sample itself. The book scrutinizes, remember, six conventional fields of study at ten elite universities. Both choices are indefensible in a study of American doctoral education as a whole, which is what the book claims to be. Bowen and Rudenstine admit that "we have looked carefully at only a small number of graduate programs . . . located within a clearly unrepresentative group of universities." Refreshing candor. But then from such a sample they presume for hundreds of pages to offer a "discussion of policy making in graduate education." "Graduate education"--not "graduate education at the Best Schools and in the Best Fields."

Choosing English, history, political science, physics, mathematics, and economics is to choose a point of view. One cannot avoid choosing. In economic terms, the six fields are the market basket, so to speak, in an index number. Someone who viewed graduate education in the '70s from the Department of Classics, for instance, would have a very different view of the world than would someone looking out from the Department of Orthopedics.

But some choices of point of view are better than others. The choice made by Bowen and Rudenstine is poor. The point of view embodied in choosing their Six Fields is that of the East/West Coast educational establishment in 1958. One could be more precise by saying that it is a Princeton index, noting that the authors spent most of their seat time at Princeton: in addition to administering there, Rudenstine got a Princeton BA before his (Harvard) PhD; and Bowen, its long-time president, got his PhD there. Princeton does not have graduate schools of law, medicine, or business. (A joke among university administrators is that the president of Princeton is the only one smiling in a collective photograph of the Ivy presidents. Why's that? Because he's the only one without a medical school.) Therefore, in a Princeton index, the boom in doctoral study outside the EHP, MP, and economics fields is simply not measured.

Business fields such as accounting and management boomed in numbers and in intellectual quality at the time the Six Fields were faltering. No business, please; we're Princetonians. Medical schools produce professors of medicine, and nursing schools produce professors of nursing. You wouldn't know it here. In the story of doctoral education as told by Bowen and Rudenstine, a field like communication studies, mainly Midwestern though almost as old as Eastern political science as a separate field---and these days paying more attention to politics than the political scientists do---has no chance of being represented. It, too, boomed in the '70s, though it was sat on by administrators of the old school.

And the choice of the Six Fields of 1958 means of course that interdisciplinary fields, such as comparative literature and parts of applied mathematics, which often have fine if small programs, are ignored. Film studies is doubtless most vulgar, and best left to places remote from Boston and New York like Iowa and the University of Southern California (though consider NYU). Bowen and Rudenstine's choice of index gives interdisciplinarity, like Rodney Dangerfield, no respect. One would not know from their book that most intellectual advances have come from the crossbreeding of disciplines, from biochemistry to social science history.

The sample of fields, then, is poor. Most fundamentally, the authors, index of fields, marking off the seams of the universe in 1958, has the problem, as we say in economics, of any "Laspeyres" index number: it reflects the tastes and conditions of the past, not the present. From the government's statistics on degrees granted (not given in the book), one finds that the Six Fields represented 26 percent of all PhDs in 1933-34, 20 percent in 1958-59, and 14 percent in 1989. It's not the size of the sample that's the problem, but its drift into bias. Yes, English, history, political science, math, physics, and economics may have represented fairly the world of graduate schools in 1933. No, they do not in 1993.

The most damaging error of sampling, however, is to leave it at the elite Ten Universities (reduced in Chapter 5, astonishingly, to the five Ivies among them). Bowen and Rudenstine defend their decision to study only the "top echelon" in a few sentences scattered here and there. It will not do. As they themselves point out (though I do not recall them giving the statistics), more and more PhDs are coming from non-elite institutions and more and more from places other than the 1958 elite. In 1933-34 the Elite Ten graduated fully 35 percent of the PhDs, in 1958 still nearly a quarter. But in 1989 the share had dropped to 12 percent. In 1994 a history department that wants the best candidate cannot merely make a long distance call to the Harvard Department of History. It must look to graduate students from Santa Barbara and Minnesota and other barbarous places.

Contrary to their confident and final dismissal of the issue in a dependent clause on page 6 of the book, their choice of the Elite Ten certainly does "raise worries about |elitism.'" It has been some time, Bowen and Rudenstine need to learn, since American higher education consisted chiefly of clubbable chaps. Do our authors know about scanners at supermarket checkout counters?

Bowen and Rudenstine want to draw conclusions about doctoral education, but have sampled from an irrelevant universe. They might as well have sampled from Oxford colleges in classics. The sample has no validity, technically speaking, and the conclusions drawn from it do not apply to graduate education as a whole. It is wrong to claim, as does the publicity flyer for the book, that it "is the most comprehensive analysis of the effectiveness of doctoral education in the arts and sciences ever undertaken." On the contrary, it is the narrowest. The study therefore is lacking in point.

One example among many in the book in which findings from the elite sample cannot be generalized, as I've noted, is the effect of full-ride fellowships on time-to-degree. At Princeton and Yale, three-quarters of the students are supported mainly by full-ride fellowships. The authors do not emphasize that in the real world, represented here by the University of North Carolina, 3 percent are so supported. Yet they find space to discuss the bracing effects on time-to-degree of doing a little teaching at Yale. Wait a minute. Does anyone doubt that if Wayne State University could offer full-ride fellowships on the scale that even nearby Michigan can, its time-to-degree would fall precipitously?

Bowen and Rudenstine observe that the various attempts by foundations to reduce time-to-degree at elite universities have not worked. Nonetheless they recommend that the foundations and the federal government give portable fellowships in the humanities, which will of course end up in the coffers of the elite universities (that's economics). I have a suggestion for improving the record of the foundations and the government: give the money instead to the Wayne States of the world.

Occasionally the thought that something is wrong flits into their consciousness. "These nine universities [the nine of the Elite Ten for which attrition rates could be collected] are by no means representative of the full range of institutions rewarding doctorates." By no means, indeed. When they can't get a good sample from the Ten Universities they get a sample from "men [sic] who won Danforth Fellowships." There's the common touch.

Most of the teaching in American colleges is and will be done by graduates of the programs ignored by the Mellon Foundation. Most of them will not come from the Elite Ten. Most of them will not be conventional students going to school at conventional ages from backgrounds that would make admissions officers of Ivy League graduate departments in 1958 comfortable. Most of them are not going to be white male Anglo honors graduates in mathematics from Yale. Many will be 40-year-old nurse supervisors going back for PhDs in nursing at the University of Texas at Austin. Or late bloomers who become professors of history by way of Case Western after the Army. Or PhDs in accounting from Louisiana State with backgrounds in English from Southeast Missouri.

Bowen and Rudenstine pay conventional obeisance, of course, to a wider academy. No one can be an administrator nowadays without retailing the usual cant, copied out from editorials in The New York Times. But their hearts are not in it. The structure of their study and the character of their recommendations show that they have no place in graduate schools for people who got there by any but the routes conventional in 1958 along the Boston-Washington corridor. They recommend that the federal government finance more fellowships in the humanities but argue against "institutionally targeted programs" (read: programs aimed at the places where most of the graduate students are now educated): "the NDEA program of the 1960s . . . [encouraged] the further development of programs that have remained below critical mass." Heaven forbid that we have programs below critical mass, whatever exactly the phrase may mean.

Bowen and Rudenstine repeatedly aver that "programs can be too small to be effective, and the data in Chapter 4 suggest that this is a serious issue." The data show nothing of the kind; they document the number of programs "below critical mass" using an arbitrary definition and without showing that low critical mass so defined has bad effects. A candid title for the book would have been Irrelevant Statistics for a Few Established Fields at a Few Universities We at Princeton and the Mellon Foundation Think are Respectable.

In truth, the self-interested character of the book is embarrassing. It is a tract in favor of continuing the support for elite graduate programs by the government and by elite foundations. That graduate deans---elite and non-elite-across the country are taking it seriously is distressing. They are buying into the unadventuresomeness of their betters. Bowen and Rudenstine congratulate the national foundations, which they run, for being "at least occasionally adventuresome, willing to place bets that might be inappropriate for governmental entities." On the contrary, it is notorious that the private foundations are less adventuresome in educational and scholarly programs than the National Science Foundation and much less so than the public universities. The authors suggest that the Manhattan foundations have a self-image of daring entrepreneurship making a difference for graduate education on the margin. My Lord. Some daring. Some margin.

When all is said and done, the main recommendation is that graduate students pull up their socks and get on with their Ph.D. theses. No doubt Princeton has tea-drinkers on full-ride fellowships who need such bracing advice. It, and the book, are irrelevant to the rest of graduate education, roughly 95% of it.