Doctoral Education:
The Short-Term Crisis Vs. Long-Term Challenge

Roger Geiger

Tables

In 1995, doctoral education in the United States came under the most intense scrutiny in more than two decades. An authoritative report by a Panel of the Committee on Science, Engineering, and Public Policy advocated a "reshaping" of graduate education to better align "the education of scientists and engineers with changes in the conduct of science and in societal needs" (COSEPUP, 1995). Two empirical studies purported to find "overproduction" of PhDs ranging from 12% to 22% (ACS, 1995; Massy & Goldman, 1995). Such evidence has been seized upon by commentators as evidence of a "Malthusian crisis" in doctoral education (Ausubel, 1996; Holden, 1995; Syverson, 1995). They charge that academic science will continue to reproduce itself in the form of new PhDs until that population is restrained by the academic equivalent of Malthus's "preventive checks" (Malthus, 1803). This argument is thus predicated, on the one hand, on poor employment prospects for PhDs in (science and engineering (S&E) fields. On the other hand, it posits a process of overproduction rooted in the academic department: Because of the connection between graduate education and research, departments are allegedly unable to restrict PhD numbers through the exercise of what Malthus would have called "moral restraint." [End Page 239]

Conversely, compared with arrangements for doctoral education in other advanced nations, the American graduate school has been described as a "tower of strength" (Clark, 1995, p. 116). The key to that strength, or efficacy, lies with the capacity of U.S. departments to combine advanced specialized teaching with hands-on research experience. Thus, metaphorically, the challenge to would-be reformers of the doctorate is no less than reshaping a tower of strength. Before addressing this specific issue, however, the allegation that there are too many PhDs demands scrutiny.

PhD Supply and Demand

A simulation of supply and demand by Massy and Goldman (1995) reported an estimated 22% overproduction of S&E doctorates. The authors interpreted this surplus as a long-term structural imbalance that they attributed to an "open-loop" system in which quantities produced were "determined by suppliers according to their own agendas rather than as a function of output-market demand" (p. I-3). The evidence for this conclusion is drawn from 1980-1991 data, and thus does not include the 1970s when top-tier departments deliberately limited their intake of students in response to weak markets for graduates (Bowen & Rudenstine, 1992).

The distinctive contribution of this study is to model the supply of doctorates as a function of the expanding teaching and research needs of departments. Their simulation of demand for doctoral graduates, however, is static rather than dynamic (Syverson, 1995). They simulate the replacement rate for currently employed PhDs by assuming that "today's employment levels reasonably approximate a steady state" (p. I-23). By combining a dynamic supply model with a static demand model, the authors produce the same kind of disequilibrium hypothesized by Thomas Malthus (1803) using a geometrical model for population growth and an arithmetical one for food supply.

A comprehensive survey of the employment of recent chemistry doctorates by the American Chemical Society (ACS, 1995) estimated that 12% more chemistry PhDs (250 out of 2,046 annual degrees) were being graduated than the market could absorb. Contrary to the original expectations of the investigators, however, similar employment results were found for graduates of all levels of programs. The market for chemistry doctorates was thus segmented and presented no obvious place for cutbacks. Instead of recommending a reduction of the number of PhDs, the ACS also endorsed a "reshaping" strategy to enhance the employability of graduates.

The rise in the number of scientists holding postdoctoral appointments has additionally been cited as a symptom of overproduction (Holden, 1995). When measured against the expansion of research, however, this trend seems normal. Postdoctoral appointments in academic institutions grew by 63.7% [End Page 240] between 1982 and 1992, compared with an increase in real expenditures for academic research and development (R&D) of 81.1%. Total postdoctorates grew by 65.3% versus real growth in national Basic Research of 58.7% (1981-1991) (NSF, 1993). If postdoctoral researchers are actual contributors to scientific research, rather than scientists on welfare, their increased numbers do not seem out of line with the need for their services.

More generally, the recent existence of excess numbers of doctoral graduates can be interpreted only as a manifestation of an underlying structural imbalance if one assumes that current conditions are permanent. Malthusians emphasize two sea changes in the environment for science: (a) the end of the Cold War and presumably much of the research it fostered, and (b) the current mania in Congress for budget balancing with consequent bleak prospects for federal support of academic research. From these circumstances, they conclude that growth in American research and the demand for the current number of PhD recipients have disappeared "forever" (Goodstein, 1995, p. 213).

On the other hand, one might question the appropriateness of a static or contracting research system for the nation's future. The United States is a post-industrial society, increasingly dependant on the utilization of sophisticated knowledge; thus, its comparative advantage in the global economy rests largely on cognitive skills, leadership in innovative technologies, and the productivity achieved primarily through the application of expertise. Why, then, is American society not consuming ever greater amounts of graduate education?

Historical Dynamics of Graduate Education

Doctoral education in this country grew rapidly--and unexpectedly--in the 1990s. However, this expansion occurred after two decades of virtual stagnation. In this respect, doctoral education has largely mirrored trends for degrees at all levels of higher education. This fact is a marked departure from the historical relationship between graduate and undergraduate education. [End Page 241]

The number of bachelors' degrees awarded by American colleges and universities increased by 1800% from 1920 to 1975, but both masters' and doctoral degrees increased substantially more. To provide meaningful comparisons, these figures have been broken down into smaller periods. (See Table 1.) The relationships among degrees emerges as a fairly consistent pattern until 1975. ¹

For more than half a century, the demand for graduate degrees rose faster than the "at-risk" population of recent college graduates. Since 1975, however, graduate and undergraduate degrees have risen at about the same overall rate. This in itself represents a departure from the historical pattern; but on closer examination, doctoral education fared even worse.

In 1973 almost 28,000 doctorates were awarded to U.S. citizens. ² That figure has not been equaled since. It fell as low as 23,000 in the mid-1980s and, in 1994, had climbed back to the 27,000 level. Thus, growth in U.S. doctoral education over the last two decades has come entirely from noncitizens, the vast majority on temporary student visas. Table 2 depicts the changes since 1985.

After reaching a low point in the early 1980s, doctoral awards began to rise at an accelerating pace. However, two thirds of this growth came from noncitizens studying science or engineering. This aspect should be factored into any case for overproduction. Two dynamics are clearly at work. If the [End Page 242] size of graduate student cohorts is partially determined by departmental slots created for teaching and research assistants, the supply of qualified students to fill those slots has been sustained by the growing propensity of foreign nationals to seek doctorates in American universities.

This phenomenon has been a cause for concern on a number of counts; but on the whole, it has been an enormous boon to this country. The influx of foreign doctoral students has acted as a double shock-absorber for American graduate schools. First, these students cushioned the downturn in doctoral enrollments by allowing many S&E departments to maintain a critical mass of graduate students. Second, in a difficult job market, international students possess an exit option not available to U.S. citizens. An estimated 50% return to their homes, but anecdotal evidence indicates that the current dearth of employment opportunities has induced even larger numbers to leave the United States. In addition, from a human-capital perspective, those who remain in this country represent an economic windfall.

Thus, foreign doctoral recipients account for much of the current oversupply of PhDs. For example, the ACS (1995) figure of 250 annual surplus chemistry PhDs should be considered in light of 800 annual PhDs awarded to holders of temporary visas. These international PhDs are also far more likely to take postdoctoral appointments, thereby postponing decisions about their permanent place of work. In fact, about half of postdoctoral slots are now filled by noncitizens; and they accounted for 70% of the growth in university postdoctoral fellows from 1982 to 1992 (COSEPUP, 1995).

In addition, the large number of international students reflects conditions elsewhere in the world over which the United States has little control. Close to half of foreign doctoral students are East Asians, mostly Chinese. [End Page 243] Among foreign S&E graduates indicating a desire to remain in this country in 1991, 29% were from China, 13% from Taiwan, and 12% from India (NSF, 1993). This situation may not persist too much longer (Syverson, 1996). There is thus good reason to be concerned about the prolonged stagnation underlying citizen participation in doctoral education.

Changes in Fields of Study

Some indication of the contemporary relation of graduate education to labor markets emerges from comparing changes in the output of graduate degrees in different fields. Tables 3 and 4 present the fields that gained and lost the greatest number of annual degrees from 1976, when stagnation settled in, to 1993.

These data for masters' degrees support the premise that our economy has had an increasing need for highly educated workers. The MBA is the credential that exhibited the greatest growth by far, but the other expanding fields were in areas of employment based on sophisticated knowledge. Moreover, these 80,000+ additional degrees in expanding fields by themselves represent about double the rate of increase of real GDP. What appears to be responsible for the overall stagnation, then, is the collapse of demand for masters' degrees in education and, to a lesser extent, academic fields largely patronized by educators and librarians.

In doctoral education, the pattern appears at first glance to be similar. (See Table 4.) In shrinking fields, the anemic demand for college and university faculty has undoubtedly depressed the number of PhDs in letters and social sciences (although not proportionally to the fall in demand). In expanding fields, the dynamic growth of research in the 1980s encouraged additional degrees in the physical and life sciences and, to some extent, [End Page 244] engineering. This growth, as has been seen, was chiefly supplied by foreign nationals. Unlike masters' degrees, the basic structure of doctoral education makes it far more resistant to economic pressures. These interdependent structural features can be summarized as: (a) the admissions queue, (b) departmental control, (c) the quality imperative, and (d) sponsorship.

Structural Features of Doctoral Education

The Queuing Pattern of Demand

Doctoral education in most disciplines is dominated by a national competition for the best and the brightest students. The most highly regarded programs accept an optimal number of the most qualified applicants; and once those places are taken, demand spills over to the next programs in the queue (Breneman, 1975; Geiger, 1993, p. 222). The most highly rated programs have the largest enrollments and support the most students with fellowships or research assistantships. Departments aim to achieve an optimal level of graduate enrollment, which is determined by the number of faculty and the amount of student support. This feature of doctoral education provides great stability for the top programs.

Table 5 reveals the freedom of action enjoyed by prestigious departments. After voluntarily absorbing the bulk of the cutbacks in the 1970s (Bowen & Rudenstine, 1992), they increased doctoral awards at roughly twice the rate of lower-rated departments in the 1980s. The gains in Category V, however, suggest the existence of regional or local markets for doctoral education, a phenomenon that also emerged from the ACS study (1995). The rules for competing in the national admissions queue are set by the top departments, but regional programs may possess more latitude for innovation. [End Page 245]

Departmental Control of Doctoral Education

The ability of departments to shape graduate programs to conform with faculty needs and disciplinary conditions has been repeatedly shown (Berelson, 1960; Bowen & Rudenstine, 1992; Breneman, 1970; Clark, 1995; Geiger, 1993). This fact is of the utmost importance for contemplating future directions of graduate study, since experience shows that innovations that were not welcomed at the departmental level did not fare well. In particular, departments enforce the next imperative of graduate education--the overriding importance of quality considerations.

The Quality Imperative in Doctoral Education

To be a doctoral student is to run a series of qualitative gauntlets, from the moment of application through the final thesis defense and even beyond, to letters of recommendation. Faculty, too, of course, undergo qualitative scrutiny--of their own work, that of their students, and of the standing of their programs (NRC, 1995). This incessant concern for quality, while surely one of the great strengths of American graduate education, has historically exerted pressure to raise expectations and prolong doctoral programs. Despite the standardizing role of university graduate schools, the quality imperative has been largely interpreted and enforced at the departmental level. These qualitative concerns, in particular, shape and buttress the requirements for the doctoral dissertation--the defining feature of doctoral education, for better or for worse.

Institutional Sponsorship of Doctoral Students

Doctoral education in the United States has traditionally required some degree of sponsorship; however, the situation today is that almost all full-time doctoral students need substantial amounts of support (Bowen & Rudenstine 1992; NSF, 1993, pp. 54-56). Graduate education has become more expensive, even while previous sources of student support have shrunk. As a result, the burden of supporting graduate students now falls heavily on universities. S&E graduate students are predominantly supported with research [End Page 246] funds, others earn their keep as teaching assistants, but large numbers of graduate students are guaranteed support that comes in part from general or earmarked institutional funds. Over the past generation, the increase in the cost of supporting a graduate student has become daunting. (See Table 6.) No matter how these figures are juggled, doctoral education is at least four times more expensive today than in 1960.

These four structural features are mutually reinforcing: the high cost of sponsorship, the quality imperative, and departmental sovereignty, all buttress the hierarchical queuing pattern. They also reinforce the status quo. Thus, in predicting future levels of doctoral degrees the NCES (1993) projected that the recent stagnation would persist into the 21st century. ³

A Strategy for Expanding Graduate Education

On the other hand, if one asks how graduate education might return to its historical pattern of relative growth, then it would seem that structural adjustments would have to be made. To contemplate such changes, three fundamental issues need to be confronted: finding additional clientele whom doctoral education might serve, how to minimize economic disincentives and how to work with, rather than against, the quality imperative and the prestige hierarchy.

Additional Clientele Doctoral Education Might Serve

Doctoral education has been historically oriented toward two overlapping professions--university teaching and scientific research. Although predictions have been made that a boom in faculty hiring is just around the corner, those expectations will most likely be disappointed. The reservoir of potential faculty is large, and the capability of universities to expand payrolls is small. Demand for college faculty may improve in the near future, but such a change cannot be expected to drive significant enrollment growth in graduate schools. The outlook for research employment is probably somewhat better, even though there is a glut today. Still, moderate growth in this area will require only moderate numbers of additional PhDs.

If doctoral education is to grow significantly, it will have to follow the path of masters' degrees--i.e., it will have to become, in some fields, a valued credential in the private sector. This is already the case in chemistry, clinical psychology, and engineering. What would it take for such demand to spread to additional fields? The secular advance of knowledge and particularly the growing salience of science-based technologies favor this development. Opposing it, however, are the horrendous economics of doctoral education. [End Page 247]

Minimizing Economic Disincentives

The high cost of doctoral studies is only part of the economic disincentive to graduate study. A second factor is the excessive time-cost to individuals--five or more productive years. It is these opportunity costs that make the economic returns to doctoral education poor in general but actually poorest in the most lucrative fields, such as engineering or business. Thus, the only way to diminish these disincentives, and to fashion economically viable doctoral degrees, is to provide a more streamlined version of the doctorate--in other words, to shorten the time-to-degree. A large literature exists on this subject (Bowen & Rudenstine, 1992; Tuckman, Coyle, & Bae, 1990), but previous attempts to introduce greater efficiency in graduate education, such as the Doctor of Science, have run afoul of the quality imperative (Geiger, 1993).

The Quality Imperative and the Prestige Hierarchy

The third question that must be dealt with is how to work with, instead of against, the quality imperative and the prestige hierarchy. The currently strong departments may be expected to dominate preparation for careers in basic research and especially for faculty positions in research universities. While greater efficiency might be desirable, preparation for those positions will likely remain a long, costly, and rigorous process. However, to avoid having those programs act as a straitjacket for all doctoral programs, the doctoral process ought to be separated into what are now in fact three distinct phases: the predissertation stage, the dissertation stage, and the postdoctoral stage.

The PhD as it stands today represents too much training for many potential consumers of graduate education; yet it is too little training for its traditional role of preparing future faculty. Instead of a single template, it would be more desirable to recognize and validate each of these three levels of accomplishment. It is beyond the scope of this paper to design or describe a new structure of graduate education--something that would evolve in any case from trial and error rather than a priori invention. Several critical features of a three-phase structure, however, seem logical deductions from the preceding analysis.

1. Completion of predissertation work should represent the end of a definite stage of training and a decision point about future careers (Griffiths, 1995). ⁴ In economically oriented fields in particular, a significant portion of students might enter the workforce at this point with a credential that [End Page 248] certified greater capacities than current masters' degrees. Such a development presupposes that the market will supply desirable jobs, for example, in sales, marketing, or management related to advanced technologies. Departments would have to design such programs in ways that would avoid any stigma of failure for such students. This could probably be achieved by having students pass qualifying examinations to mark the completion of the predissertation stage. The path to employment might also be made more attractive by creating realistic possibilities for later degree enhancement. Some fields might provide for students to present a portfolio of professional accomplishments and petition for a professional doctorate at some later date, as is currently the practice in performing arts. Additional provisions might facilitate later work toward the completion of a dissertation.

Such arrangements, though probably meeting resistance in the pure sciences, would be most effective and most needed in more applied or professional fields where economic disincentives today constitute an obstacle to expanding the incidence of doctoral education.

2. A decision point after the predissertation stage could promote interdisciplinary competence by encouraging students to move to adjacent fields. The COSEPUP report (1995) strongly endorsed this notion; however, it is doubtful if such results would occur with any frequency unless this decision point were made a mandatory structural feature as suggested above. The real aim here is to promote what sociologist of science Joseph Ben-David (1991, pp. 60-70) identified as "role hybridization." He found that striking advancements in 19th-century science sometimes occurred when scientists entered new fields or turned to applied problems because their careers were blocked in traditional subjects. A reasonable goal of redistributing students after the predissertation stage would be likely to reproduce such a process. Blockages (or few openings) would be most likely in traditional fields experiencing slow growth; more openings or opportunities would emerge in new, expanding areas. A forced migration across scientific fields under these circumstances could work to everyone's advantage.

3. The decision point at the transition from PhD to postdoctorate, under current circumstances, probably extends already lengthy times-to-degree. Because of the stiff competition for the most attractive postdoctoral appointments, students prolong dissertation research to strengthen their credentials. Because students at this phase are productive research assistants, contributing to the output of laboratories, such extensions dovetail with the interests of their mentors. The process also harmonizes with the qualitative imperative, thus encouraging ever higher, more time-consuming, attainments.

This cycle cannot be easily broken in the absence of greater demand to lure doctoral candidates from their laboratories. The frank recognition that the definitive competition for faculty positions takes place at the postdoctoral level (Zumeta, 1985) is nevertheless the prerequisite for addressing this problem. [End Page 249] Given this reality, academic departments ought to be more sympathetic toward streamlining doctoral programs, optimizing requirements, and circumscribing the extent of the PhD thesis.

None of these changes would be easy to implement. They are offered, not as quick fixes, but as desirable directions for change to redress two decades of relative stagnation and to do so in a manner consistent with the basic nature and inherent strengths of American graduate education.

Roger Geiger is Professor-in-Charge of the Higher Education Program at Pennsylvania State University, University Park, and the author of Research and Relevant Knowledge: American Research Universities Since World War II (1993).

Notes

1. Bowen and Rudenstine (1992, chap. 3) explore the "BA-PhD nexus" using an 8-year lag. These data shed light on the "PhD proclivity" of given cohorts of BAs. Table 1, on the other hand, presents the surprisingly similar overall trends in coeval numbers of BA, MA, and PhD degrees.

2. Doctoral data is frequently presented for "U.S. residents," a category that includes foreign citizens with permanent visas and, sometimes, unknowns. The permanent visa category has been volatile in recent years (Syverson, 1996) and unknowns fluctuate randomly. Thus, this study examines only "U.S. citizens."

3. NCES (1993) projected annual awards of 41,400 doctorates--about the 1994 level--until 2003.

4. The principal reports on doctoral education all endorse this point (and the one that follows). However, they largely assume that students will make such choices on their own or with improved information. The thrust of this analysis is that the current structure of doctoral education--a single-lane, five+-year trip--makes such choices quite unlikely unless a structural break is inserted between the predissertation and dissertation stages.

References

ACS. American Chemical Society. (1995). Employment patterns of recent doctorates in chemistry: Institutional perspectives and imperatives for change. Washington, D.C.: American Chemical Society.

Ausubel, J. H. (1996, February 5). Malthus and graduate students: Checks on burgeoning ranks of PhD's. The Scientist, p. 11.

Ben-David, J. (1991). Scientific growth. Berkeley: University of California Press, 60-70.

Berelson, B. (1960). Graduate education in the United States. New York: McGraw-Hill.

Bowen, W. G., & Rudenstine, N. L. (1992). In pursuit of the PhD. Princeton, NJ: Princeton University Press.

Breneman, D. W. (1970). The PhD production process: A study of departmental behavior. Unpublished doctoral dissertation, University of California, Berkeley.

Breneman, D. W. (1975). Graduate school adjustments to the 'new depression' in higher education. Washington, D.C.: National Board on Graduate Education.

Clark, B. R. (1995). Places of inquiry: Research and advanced education in modern universities. Los Angeles: University of California Press.

COSEPUP. Committee on Science, Engineering, and Public Policy. (1995). Reshaping the graduate education of scientists and engineers. Washington, D.C.: National Academy Press.

Geiger, R. (1993). Research and relevant knowledge: American research universities since World War II. New York: Oxford University Press.

Geiger, R., & Feller, I. (1993, May). The dispersion of academic research during the 1980s. University Park: Institute for Policy Research and Evaluation, Pennsylvania State University.

Goodstein, D. L. (1995). The era of exponential expansion: Implications of its end for student support. In A. H. Teich, S. D. Nelson, & C. McEnaney (Eds.), AAAS Science and Technology Policy Yearbook (pp. 211-216). Washington, D.C.: American Association for the Advancement of Science.

Griffiths, P. (1995, July 13). "Testimony." Committee on Science, House of Representatives, 104 Sess., Vol. 19:30-45.

Holden, C. (1995). Is it time to begin PhD population control? Science 270, 123-128.

Malthus, T. R. (1803/1992). An essay on the principle of population. (2nd ed.). New York: Cambridge University Press. (Originally published in 1803)

Massy, W. F., & Goldman, C. A. (1995). The production and utilization of science and engineering doctorates in the United States. Menlo Park, CA: Stanford Institute for Higher Education Research.

NCES. National Center for Education Statistics. (1993). Earned degrees conferred: Projections of education statistics to 2004. Washington, D.C.: Department of Education.

NSF. National Science Foundation. (1993). Science and engineering indicators, 1993. Washington, D.C.: National Science Board.

NRC. National Research Council. (1995). Research-doctorate programs in the United States. Washington, D.C.: National Academy Press.

Syverson, P. D. (1995, August). When simulation becomes reality: Press reaction to Massy/Goldman study creates erroneous message. CGS Communicator, 28(8), 8-10.

Syverson, P. D. (1996, February). Profound changes underway in flow of international students, Reports Open Doors 1994-95. CGS Communicator, 29(2), 7-9.

Tuckman, H., Coyle, S. & Bae, Y. (1990). On time to the doctorate. Washington, D.C.: National Academy Press.

Zumeta, W. (1985). Extending the educational ladder. Lexington, MA: Heath.

[an error occurred while processing this directive].