Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 85
Trends in Federal Support of Research and Graduate Education 6 Findings, Conclusions, and Recommendations This report updates and extends the previous analysis of trends in federal research funding to FY 1999, the latest year for which there are data on actual obligations for research by field. It has looked more closely than the 1999 study at trends in basic and applied research, research performed by universities and colleges, and graduate education. In addition, we have examined changes in the structure of agency support of some fields and changes in the research portfolios of some of the agencies with the largest research budgets. The key findings, conclusions, and recommendations resulting from this study are presented below. FINDINGS Agency Research Budgets Are Up What has changed and not changed since the previous STEP Board analysis? First, federal research funding in the aggregate turned a corner in FY 1998. After 5 years of stagnation, total expenditures were up 4.5 percent in FY 1998 over their level in 1993. A year later, in FY 1999, they were up 11.7 percent. By 1999 the research budget of every major R&D funding federal agency was increasing again and, with the exceptions of the Departments of Defense and Interior, was larger than in 1993. FY 2000 and FY 2001 saw continued growth in budget authority for research. Second, increases in appropriations to the National Institutes of Health kept federal research funding from falling lower in the mid-1990s and accounted for 61.8 percent of the net growth in research spending from FY 1997 to FY 1999. Indeed, the rate of NIH budget growth doubled in 1999, the first year of the 5-year campaign to double NIH’s budget. The annual increase in NIH spending on research, which was between 4 and 6 percent in the 1996–1998 fiscal years, jumped to 12.5 percent in FY 1999 and was projected to be 11.9 percent in 2000. NIH’s share of federal research expenditures increased from 32.1 percent in 1993 to 38.4 percent in 1999 and an estimated 40.4 percent in 2000. Substantial increases in NIH budget authority appropriated in FY 2001 and proposed by the current administration for 2002 and 2003 promise to sustain this pace of growth. Research Fields Continue to Diverge The sharp divergence in support of different fields of research that developed after 1993, although moderated, has continued. The life sciences received 46 percent of federal funding for research in 1999, compared with 40 percent in 1993. During the same period, the share of the federal portfolio represented by the physical sciences and engineering went from 37 to 31 percent. More recent actions on federal budgets for research, including doubling of the NIH budget over the 5 years ending in FY 2003, will increase the current divergence between the life sciences and other fields unless other fields receive substantially larger increases than proposed. More specifically, whereas 12 of the 22 fields examined had suffered real loss of support in the mid-1990s (four by 20 percent or more), by FY 1999 the number of fields with reduced support was seven, but of these five were down 20 percent or more—physics, geological sciences, and chemical, electrical, and mechanical engineering. The fields of chemical and mechanical engineering and geological sciences had less funding in 1999 than in 1997. Other fields that failed to increase or had less funding after 1997 included astronomy, chemistry, and atmospheric sciences. One field that had increased funding in the mid-1990s, materials engineering, experienced declining support at the end of the decade. Its funding was 14.0 percent larger in 1997 than in 1993, but that margin fell to 3.0 percent in 1998 and 1.5 percent in 1999. The fields whose support was up in 1997 and has continued to increase include aeronautical, astronautical,
OCR for page 86
Trends in Federal Support of Research and Graduate Education civil engineering, and other engineering, biological and medical sciences, computer science, and oceanography. Of these, the number of fields whose support was up 20 percent or more from 1993 levels increased from one in 1997 to six in 1999. Funding of some fields increased somewhat from 1997 to 1999 but not enough to raise them back up to their 1993 levels. Those include electrical engineering and physics. Fields that, like overall research expenditures, turned a corner were environmental biology, agricultural sciences, mathematics, social sciences, and psychology. Their funding, which was less in 1997 than in 1993, exceeded the 1993 level by 1999. NIH growth accounts for a large part but not all of the increased support of the biological and medical sciences. DOD and VA also increased their funding of those fields. The decline in the support of many of the physical science and engineering fields is partly attributable to the fact that the budgets of their principal sponsoring agencies (e.g., DOD, DOE, and NASA) did not fare as well as the NIH budget and partly to the fact that the agencies with growing budgets, especially NIH and NSF, did not increase their support of those fields and in some cases reduced it. At the same time, some fields—e.g., computer science, oceanography, and aeronautical engineering—experienced substantial growth, even though their largest 1993 funders were agencies with shrinking budgets—e.g., DOD and NASA. They did so by maintaining their level of funding of agencies with declining budgets and by picking up additional support from other agencies. The patterns in federal funding of basic research and research performed at universities are somewhat more favorable than the trend in total research support, suggesting that by the late 1990s agencies were tending to protect basic and university research relative to applied research and other performers. At the aggregate level, funding of basic research was 16.6 percent larger in 1999 than in 1993, compared with 6.8 percent for applied research. University research was 19.9 percent more in 1999 than in 1993, compared with 7.2 percent for all other performers (e.g., industry, federal laboratories, other nonprofit research institutions). Basic and university-performed research are also characterized by sharp divergence among fields, however. In basic research, 14 of the 22 fields had more funding in 1999 than in 1993, compared with 11 in 1997, and the number with 20 percent or more funding increased from five in 1997 to 8 in 1999. But basic research funding was less in eight fields, three by 20 percent or more (chemical and mechanical engineering and geological sciences). In university research, 15 of 22 fields had more funding in 1999 than in 1993, nine by 20 percent or more, compared with 10 and 4 fields, respectively, in 1997. The amount of university funding remained less in seven fields, two of them by 20 percent or more (mechanical engineering and geological sciences). In most fields, trends in basic research funding were similar to those for total research. Where total funding was up, basic research funding was also up, and vice versa. There were some interesting discrepancies between overall and university research trends, however. For example, although total funding of chemical engineering research was down substantially in 1999 compared with 1993 (by 25.9 percent), chemical engineering research at universities was up slightly (by 2.2 percent). And while mathematics research was up by 6.4 percent overall, mathematics research at universities was down by 13.5 percent. Production of Doctoral Scientists and Engineers Is Down The number of Ph.D.’s awarded in science and engineering by U.S. colleges and universities declined 5 percent from 1998 to 1999. The number of Ph.D.’s awarded in the sciences peaked in 1998 at 21,379 and declined 3.6 percent to 20,616 in 1999. The number of Ph.D.’s in engineering peaked earlier in 1996 at 6,305 and has since declined by 15.4 percent to 5,337 in 1999. Because in most fields it takes 7 or more years to complete Ph.D. requirements, these declines must be attributable largely to factors other than changes in federal research support. Nevertheless, in the years ahead the ongoing decline in enrollment in most fields will reinforce the drop in graduate school output of Ph.D.’s. Sharp Differences in Graduate Enrollment Trends Among Fields From 1993 to 1999, trends in federal funding for university research, full-time graduate enrollment, and numbers of doctorate recipients reveal two divergent patterns among science and engineering fields. Fields in which federal funding for university research was down from 1993 to 1997 have nearly all had declines in both graduate enrollments and doctorate recipients from 1993 to 1999. Fields with increasing federal funding for university research, however, exhibit a range of divergent trends in graduate enrollment and doctorate production. These trends depend on a variety of factors, including the state of both the industrial and academic research labor markets and the supply of undergraduates. As funding for most of the physical, environmental, mathematical, and social sciences declined in the 1990s, so did the number of graduate students in these fields, the number of students federally supported, and the number of federally funded research assistants (RAs). In physics, geology, atmospheric sciences, and mathematics, the decline in the number of federally funded RAs was approximately 20 percent between 1993 and 1999. Nevertheless, two fields with increasing research support—astronomy and ocean sciences—also experienced reductions in federally funded graduate students, although less drastic.
OCR for page 87
Trends in Federal Support of Research and Graduate Education In engineering the pattern was similar with the exception of electrical engineering, where the number of federally funded RAs (especially through NSF) increased as research support declined. Federally supported graduate students in computer science increased along with research funding, as expected. In the life sciences, biology and the medical sciences exhibit different trends in graduate enrollment although both fields are benefiting from increasing federal research support. The number of graduate students in the biological sciences grew only marginally in the 1990s and the number of federally supported RAs actually declined. On the other hand, the number of postdoctoral fellows in the biological sciences has increased. In the medical sciences overall graduate enrollment and the number of RAs in particular grew nearly 40 percent. Trends in Nonfederal Research Support Together, states, philanthropies, foundations, other nonprofit institutions, and industry are sources of 63 percent of the nation’s basic and applied research spending, and their share increased in the 1990s as federal expenditures reached a plateau. Although the data are much more limited, it appears that states and philanthropies have shared the research priorities of the federal government in the last decade. For both states and foundations, biomedical research consumes a majority of research funding and has grown at a faster rate than support of other scientific and engineering fields. If anything, this orientation is reinforced by patterns in the growing number and size of individual donations to research and research facilities and in the disposition of funds received by the states in settling their suits against the tobacco manufacturers. Data on the composition of industry-funded research are not comparable to the data on federal expenditures because they are classified by the industry sector of the reporting parent firms, not by product line or constituent business, let alone by research discipline. Nevertheless some observations on the 1990s are appropriate and relevant. First, only a few industrial sectors are research intensive. Pharmaceutical industry research spending was the highest as a percent of sales of any industrial sector and has been growing rapidly. On the other hand, the information technology sector is spending more on research absolutely and has had a higher rate of growth. For example, real spending on R&D by the electronic components industry increased 17 percent from 1996 to 1998, in contrast to the sharp drop in federal support of electrical engineering research. Nevertheless, except for a few industries such as pharmaceuticals, only a small fraction (less than 5 percent in computers and semiconductors, for example) of all corporate research and development is basic research. Moreover, private research investment is quite volatile, sometimes subject to wide fluctuation from year to year with or independent of the business cycle. CONCLUSIONS The recent shift in composition of the federal research portfolio is significant. Although nonfederal entities increased their share of national funding for R&D from 60 to 74 percent between 1990 and 2000, the government still provides almost one-half of all basic research support and nearly one-third of total research support. Reductions in federal funding of fields of the magnitude that occurred in several fields in the 1990s have national impact, unless there are corresponding increases in funding from nonfederal sources. There is little evidence of compensating actions by states, foundations, or the private sector. Industry has been investing more in R&D but little of it supports long-term research except in a few cases such as pharmaceuticals. The funding trends leading to shifts in the federal research portfolio will continue under the admini stration’s budget plan, especially the build up in funding of the biomedical sciences relative to other most other fields. They will continue for several more years, at least until the fulfillment of the campaign to double the NIH budget from 1998 to 2003. The administration’s request for NIH for FY 2002 would increase its budget authority for research by 12.9 percent over the FY 2001 level in constant dollars, and reduce all other non-defense research by 1.5 percent. As a result of the strategic policy review, DOD’s research budget is likely to increase again; but based on trends in the department’s portfolio from 1993 to 1999 there is little indication that funding for fields previously cut would be rebuilt. There are compelling reasons for the federal govern ment to invest across the range of scientific and engi neering disciplines.1 The most important problems in science are increasingly interdisciplinary. Examples include genomics and bioinformatics, which rely on mathematics and computer science as much as biology for progress; nanotechnology, which depends on chemistry and chemical engineering, physics, materials science and technology, and electrical engineering; and understanding of climate change, which relies on collaboration among oceanographers, atmospheric chemists, geologists and geophysicists, paleontologists, and computer scientists.2 Historically, of course, progress in physics and chemistry 1 The rationale for a diverse portfolio is articulated in NAS, NAE, IOM. 1993. Science, Technology, and the Federal Government: National Goals for a New Era. Washington, D.C.: National Academy Press; and National Research Council. Allocating Federal Funds for Science and Technology, 1995. Washington, D.C.: National Academy Press. 2 Donald Kennedy, “A Budget Out of Balance,” Science, 291 (23 March 2001):2337.
OCR for page 88
Trends in Federal Support of Research and Graduate Education made critical contributions to the development and biotechnology and genetic engineering. The development of magnetic resonance imaging (MRI) used extensively in medical diagnoses was based on developments in physics, mathematics, and computer science. Another reason for investing across a wide range of science and engineering disciplines derives from the high level of uncertainty associated with science. It is not possible to know where breakthroughs will occur or what practical applications they may have when they do occur. Important advances in one field sometimes come from apparently unrelated work in another field. For example, who knew in 1945 that the discovery of nuclear magnetic resonance in condensed matter by basic research physicists would lead to the development of MRI technology 30 years later?3 Because of increasing interdisciplinarity and uncertainty about when and where advances will take place or pay off it is prudent to invest in a broad portfolio of research activities. Successive reports by committees of the National Research Council/National Academy of Sciences have recommended as an explicit goal of research policy maintaining U.S. parity with or superiority over other countries’ capabilities in all major fields of science and engineering.4 Private sector groups such as the Committee for Economic Development and the Council on Competitiveness have also called for sustaining federal support of the full range of research fields.5 There is cause for concern about the current and prospective allocation of funding among fields in the federal research portfolio, in particular, with respect to most of the physical sciences and engineering, whose funding, in contrast with the biomedical sciences, has with few exceptions stagnated or declined. We are not suggesting that every field of research merits constantly increasing or even stable support. Portfolio management should not be viewed in static terms, i.e., a single year’s budget, nor in isolation from all other sources of research support—states, institutions, philanthropies, and industry. Nevertheless, it is not clear that the current allocation is optimal from a national viewpoint. It is also not necessarily optimal from the standpoint of advances in biomedical research or of computer science research, another field in which federal funding has increased substantially relative to other fields. Improved health and a strong information technology industry will rely on progress in a range of fields of fundamental research, including physics, chemistry, electrical engineering, and chemical engineering, all fields with less funding at the end of the 1990s.6 Similarly, it may not make sense to cut geology research at a time of renewed concern about how to increase production of fossil fuels while minimizing environmental damage. Although it may be wise policy to reduce the linkage between research funding and training support,7 re search allocation decisions should take into account the need for trained people in a field. Although federal funding is one factor among many in determining graduate enrollments and production of Ph.D.’s in a field, enrollments and the number of Ph.D.’s awarded were generally down in fields that had less federal funding in 1999 than in 1993, reducing the supply of new talent for positions in industry, academia, and other employment sectors. Curtailing research in a field may constrict the supply of trained people who are capable of exploiting emerging research opportunities. This effect is both direct, in that federal funding of university research supports the education of a significant number of graduate students in most fields, and indirect, in signaling to prospective graduate students that some fields offer poor career opportunities. Many graduates with master’s or doctoral degrees in science or engineering work in industry, including the majority with doctorates in engineering, chemistry, and computer science and 40 percent of those with doctorates in physics and astronomy. Most of the rest work in universities, where they conduct research and train the next generation of scientists and engineers.8 The current system for allocating research funding does not necessarily ensure that national priorities are taken into account. In the highly decentralized U.S. system of support for science and engineering, most research funding is tied to the missions of federal agencies rather than national needs more broadly conceived, such as technological innovation and economic growth. If a mission changes—for example, defense strategy in the post 3 National Academy of Sciences. March 2001. A Life-Saving Window on the Mind and Body: The Development of Magnetic Resonance Imaging. Washington, D.C.: National Academy of Sciences. At: www/beyonddiscovery.org/beyond/BeyondDiscovery.nsf/files/PDFMRI.pdf/$file/MRIPDF.pdf. 4 National Academy of Sciences, National Academy of Engineering, and Institute of Medicine. 1993. Science, Technology, and the Federal Government: National Goals for a New Era. Washington, D.C.: National Academy Press; and National Research Council. 1995. Allocating Federal Funds. Washington, D.C.: National Academy Press. 5 Committee for Economic Development. 1998. America’s Basic Research: Prosperity Through Discovery, pp. 34–35. New York: Committee for Economic Development; Council on Competitiveness. 2001. U.S. Competitiveness 2001: Strengths, Vulnerabilities and Long-Term Priorities, pp. 38–41. Washington, D.C.: Council on Competitiveness. 6 Harold Varmus. March 22, 1999. “The Impact of Physics on Biology and Medicine.” Plenary Talk, Centennial Meeting of the American Physical Society, Atlanta, At: www.mskcc.org/medical_professionals/president_s_pages/speeches/the_impact_of_physics_on_biology_and_medicine.html. 7 A position taken by the Committee on Science, Engineering, and Public Policy in its report, Reshaping the Graduate Education of Scientists and Engineers, Washington, D.C.: National Academy Press, 1995. 8 National Science Foundation. 2001. Characteristics of Doctoral Scientists and Engineers: 1999 (Early Release Tables), Table 7. At: www.nsf.gov/sbe/srs/srs01406/tables/tab7.xls.
OCR for page 89
Trends in Federal Support of Research and Graduate Education Cold War world—the mix of sponsored research may change and the support of certain fields of research may decline for reasons that are entirely defensible in terms of the affected agency’s priorities but not in terms of the research opportunities in and productivity of those fields and their potential contributions to other national goals.9 In the mid-1990s, when research budgets were flat or shrinking, agency leaders and congressional overseers had to make choices about which research to fund and which to sacrifice. The evidence of priority setting within agencies is encouraging. As the shift in the Defense Department’s focus illustrates, agencies did not simply spread their research budget losses or gains evenly across research fields. But it appears that the decline in support of certain fields was unplanned and unevaluated from the perspective of their research productivity, production of knowledge and scientific and engineering talent relevant to progress in other fields, and contributions to other national needs. Some fields, such as computer science, that might have been more adversely affected by dependence on agencies with declining research budgets or changing priorities were able to increase funding by shifting or diversifying their sources of support among federal agencies. Others, such as electrical engineering, were not able to find other support. Improvements in data and analysis would support a better informed process of allocating federal funding for research. Current surveys are valuable and underutilized tools for assessing the nation’s allocation of resources to the conduct of science and development of technology, but their utility could be improved by modest changes in the surveys and in the presentation of their results. Moreover, there are significant gaps in information, especially on non-university performers of federal research and on non-federal research sponsors—states, philanthropic institutions, and businesses at a fine level of detail. There needs to be a good deal more qualitative evaluation of the output of research fields and the effects on outputs of changes in funding levels as well as more rigorous analysis of the influences on the supply of and demand for scientists and engineers with advanced training. RECOMMENDATIONS Evaluations and Adjustment of the Research Portfolio This report documents large shifts in federal research funding that occurred in the mid-1990s, when federal funding was flat for several years and that for the most part have persisted, although federal funding began to increase again after 1997. The decade ended with the support of five fields in the physical sciences and engineering below their funding levels in 1993 and several other fields at about the same levels of funding, whereas support of a few fields increased substantially. The evidence suggests that the increases for a few fields were the product of deliberately chosen priorities of Congress and the administration, but the decline in support of other fields was more the product of isolated decisions of agency officials and congressional committees focused primarily, albeit appropriately, on particular agencies’ missions rather than on the productivity or quality of work being done in those disciplines or their potential contributions to broader national goals. More work needs to be done to determine how the fields with declining support were affected and what budget adjustments need to be made. This requires some sort of centralized review. Given the imperfect correspondence between how agency research budgets have fared and how research fields’ support and graduate training have fared, simply increasing the research funding of certain agencies (e.g., DOD, DOE, or NSF), irrespective of how they have been allocating research funds, may not by itself shift funding to fields with declining support. There is, however, an accepted mechanism for establishing research priorities across agencies. It involves the President’s selection of an area of research emphasis—for example, high performance computing or global climate change—and mobilization of the resources of the Executive Office of the President, especially the Office of Science and Technology Policy and the Office of Management and Budget, to evaluate needs and opportunities, determine current spending patterns, and assign new resources. For the FY 2001 budget the directors of OSTP and OMB included balance in the government-wide research portfolio as a criterion for making R&D budget decisions. As a result, the President’s budget proposal that year did provide increased funding for some agencies, in part to bolster support of certain fields.10 In the early 1970s, in circumstances similar to current ones, when funding for physical sciences and engineering research was reduced by cuts in the DOD, NASA, and Atomic Energy Commission budgets, OMB and Congress encouraged NSF to seek additional funding equal to about 10 percent of its budget to support scientifically valuable programs that were being dropped by other agencies. The appropriators obliged.11 Other reports have urged OSTP or its director, the President’s Science and Technology Adviser, and OMB to 9 National Science Board. March 28, 2001. “The Scientific Allocation of Scientific Resources” [Discussion Draft for Comment], pp. 3. 10 Neal Lane and Jacob J.Lew. June 3, 1999. “FY 2001 Interagency Research and Development Priorities” [Memorandum for the Heads of Executive Departments and Agencies]. 11 Milton Lomask. 1976. A Minor Miracle: An Informal History of the National Science Foundation. NSF 76–18. Washington, D.C.: U.S. Government Printing Office.
OCR for page 90
Trends in Federal Support of Research and Graduate Education take the lead in reviewing the federal research portfolio with respect to national goals rather than departmental or agency priorities alone. The NRC committee chaired by Frank Press called on OMB and OSTP to determine if the aggregate budget for science and technology would provide the resources to enable the United States to perform at a world class level in all major research fields and to be preeminent in selected fields. It urged Congress to examine the total resources budgeted for science and technology before parcelling out the budget to the appropriations subcommittees for consideration.12 Most recently, the United States Commission for National Security/21st Century, chaired by former Senators Gary Hart and Warren Rudman, called for better coordination of R&D efforts within the executive branch and Congress.13 Recommendation 1. The White House Office of Science and Technology Policy (OSTP) and the Office of Management and Budget (OMB), with assistance from federal agencies and appropriate advisory bodies, should evaluate the federal research portfolio, with an initial focus on fields related to industrial performance and other national priorities and a recent history of declining funding. Examples are physics, electrical engineering, chemistry, chemical engineering, mechanical engineering, and geological sciences. Fields with flat funding or only small real increases through the 1990s also merit attention. These include materials engineering, atmospheric sciences, mathematics, psychology, and astronomy. The conclusions of the evaluation should be reflected in budget allocations. Recommendation 2. Congress should conduct its own evaluation of the federal research portfolio through the budget, appropriations, or authorization committees. Recommendation 3. For the longer term, the executive branch and Congress should sponsor the following types of studies: (1) in-depth qualitative case studies of selected fields, taking into account not only funding trends across federal agencies and nonfederal supporters and international comparisons but also subtler differences in the foci, time horizons, and other research characteristics that are obscured by quantitative data; (2) studies of agency research portfolios and decision making to understand the reasons for shifts in funding by field and the extent to which the health of individual fields and interrelationships among fields are taken into account; and (3) studies of methodologies for allocating federal research funding according to national rather than merely departmental criteria and priorities. Recommendation 4. The executive branch and Congress should institutionalize processes for conducting and, if necessary, acting on an integrated analysis of the federal budget for research, by field as well as by agency, national purpose, and other perspectives. Data Improvements This report uses a valuable federal research funding data set initiated by NSF in 1970 and annually updated through a survey of agencies that support R&D. Data on support by broad and detailed fields of research at both the basic and applied levels are available by department and agency, including major subunits. For the six largest R&D agencies, these data are available for one category of performer—universities and colleges. A number of other NSF surveys on research and development spending and on the training and employment of scientists and engineers are also valuable tools for assessing the nation’s allocation of resources to the science and technology enterprise. In addition to the perennial issue of how rapidly data can be collected, verified, and published,14 several factors stand in the way of these data being readily accessible by and highly useful to policy makers. The following observations for the most part have been made by other reports and the committee’s recommendations anticipated by other groups, including the Academies’ Science, Technology and Economic Policy Board.15 Data need to be presented in a manageable and meaningful form. Among other steps, expenditure data should be reported in constant dollars to show real trends unaffected by inflation. More information should be available on performers of federally funded research and development other than universities and colleges. In particular it would be useful to 12 National Research Council. 1995. Allocating Federal Funds for Science and Technology, pp. 8–14. Washington, D.C.: National Academy Press. 13 The Hart-Rudman Commission calls for doubling the U.S. R&D budget and strengthening the capacity of OSTP to coordinate agency R&D activities, but notes that currently the Science and Technology’s Adviser’s Office is inadequately funded, staffed, and used to fulfill its functions. 14 At the time of completion of the review of this report (June 2001), the most recent data on actual federal R&D obligations are for FY 1999, ending September 30 of that year. 15 For example, National Research Council. 2000. Measuring the Science and Engineering Enterprise. Washington, D.C.: National Academy Press; and National Research Council. 1997. Industrial Research and Innovation Indicators. Washington, D.C.: National Academy Press.
OCR for page 91
Trends in Federal Support of Research and Graduate Education know the field allocation of funds spent by government laboratories, in industry, and by nonprofit institutions. More information should be available on nonfederal sponsors of research and development. State governments have been surveyed only once in recent years. Philanthropic contributions are reported only for major foundations and not in a form consistent with federal statistics on research funding. Although it may not be possible to ascertain the field allocation of industrial research funding, it should be possible to derive a more accurate picture of the composition of industrial R&D than classification of corporate-level reporting by major industrial sector permits. It should be easier to link related data sets—for example, research funding and graduate student enrollment by field. This requires use of the same classification of research fields and definitions of research activity across surveys. While continuity of data series is important for evaluating long-term trends, data also need to reflect contemporary reality including the emergence of new fields of research and the reorientation of others. As our understanding has grown of the contribution of science and engineering to economic performance and other national goals, so has the importance of good data. Our national data sources need to be expanded and improved to support better policy making. Recommendation 5. NSF should annually report and interpret data from its survey of federal R&D obligations in a form (e.g., adjusted for inflation) and on a schedule useful to policy makers. Improvements in the data that should be given careful consideration include reporting of data on university research support by all agencies that support a major share of research in certain fields (e.g., Department of Interior in geological science and DOC in oceanography), obtaining data by field on performers other than universities (e.g., in industry and government laboratories), evaluating and revising the field classification, and making the field classification and research typology uniform across surveys (e.g., the surveys of academic R&D expenditures and earned doctorates as well as the survey of federal R&D obligations). Agencies should make sure that the data they provide NSF are accurate and timely. Recommendation 6. Although it may be impractical to obtain data on industrial R&D spending by research field, NSF should administer the Industrial R&D survey at the business unit level to make data on the composition of private R&D more meaningful. Recommendation 7. NSF should consider ways of obtaining data on the allocation of state expenditures on a regular basis. Recommendation 8. The philanthropic community should cooperate in collecting and publishing data on a basis comparable to federal research statistics. Analytical Improvements The analysis reported here is simply a more thorough collection and integration of existing data. It raises as many questions as it answers. One direction for improved analysis helpful to policy makers is to focus on innovation results and to develop better measures of effort than funding inputs and formal patent outputs. The funding trends observed in this report are only one aspect of innovation. They have important implications, but determining what difference the funding trends are making is a much more ambitious but important task. This is true even if the objective is to understand the impact of funding trends on research performance. Recent Academy experiments in international benchmarking of scientific performance in diverse disciplines has nevertheless shown that this can be done relatively quickly at modest expense.16 It is also important to explore more carefully the interrelationships between federal research funding and the development and use of human resources. The correlations between trends in funding and trends in graduate education documented in this report are intriguing, but many more factors are involved. Those factors include population flows (supply of baccalaureates in science and engineering), employment demands for trained personnel by field, and nonfederal sources of graduate support. One important question to address is the extent to which federal research funding determines the number of advanced science and engineering degrees produced, compared with the need for such personnel in the workforce. Recommendation 9. NSF and other federal agencies funding research should support benchmarking studies that compare inputs and outputs across countries and sponsor other efforts to develop techniques for assessing the productivity of various fields of research. Recommendation 10. NSF should continue and expand its efforts to develop innovation indicators other than 16 National Academy of Sciences, National Academy of Engineering, and Institute of Medicine. 2000. Experiments in International Benchmarking of U.S. Research Fields, Washington, D.C.: National Academy Press.
OCR for page 92
Trends in Federal Support of Research and Graduate Education R&D expenditure inputs, collect data on them, and fund researchers to analyze them. Other agencies (e.g., NASA, DOD, DOE, and the National Institute of Standards and Technology) interested in the role of federal research in technological innovation, could fund or jointly fund such analyses. Recommendation 11. Researchers, professional societies, industry associations, and federal research agencies should explore the relationships between federal research funding and other factors (e.g., population flows through the educational system, domestic and foreign student demand, labor market conditions, etc.) in the development and use of scientific and engineering talent. Only then can we evaluate the trends in student enrollment and in graduate study programs’ output and determine how to influence those trends if that is the conclusion of the analysis.
Representative terms from entire chapter: