APPENDIX A
RECENT TRENDS IN FEDERAL SPENDING ON SCIENTIFIC AND ENGINEERING RESEARCH: IMPACTS ON RESEARCH FIELDS AND GRADUATE TRAINING
Michael McGeary and Stephen A. Merrill
Aggregate and Agency Trends
The federal government has funded a large share of national research and development (R&D) since World War II. It was the largest funder until 1980 when it was surpassed by private industry. In 1998 (the most recent year for which expenditure data are available) it still provided 40.9 percent of all funding for research, basic and applied, carried out in university laboratories and medical centers, industrial and federal laboratories, and other research facilities in the United States.1
Real growth in the federal R&D budget—that is, growth in excess of inflation—began to level off in the late 1980s, and after 1992 it fell as part of the effort to reduce the federal budget deficit.2 According to data collected by the National Science Foundation (NSF) on actual research obligations, federal spending on the research part of R&D peaked in 1993 and by 1997 was 2.2 percent less in real terms.3
1 |
That is 56.7 percent of basic research and 30.0 percent of applied research. Industry spending on research did not exceed the federal government's support until 1995. Calculated from National Science Foundation (1998a, Tables B-2B (basic research) and B-3B (applied research)). See Box A-1 for definitions of basic research, applied research, and development. |
2 |
Budget authority for R&D fell 8.8 percent in real terms between 1992, its historical high point, and 1997 (AAAS, 1998). Budget authority is legal authority to incur financial obligations that will result in outlays. |
3 |
Obligations are commitments to spend money, although actual payment may be made later, for example, under multiyear contracts. |
As Table A-1 indicates, the trend in research funding has not been uniform across agencies.
Much of the decline has been in defense research, because of the end of the Cold War and changed national security requirements. The Department of Defense's (DOD) support of research was down substantially in both relative and absolute terms between 1993 and 1997. Other agencies that spent less in real terms in 1997 than in 1993 included the Department of Energy (DOE), U.S. Department of Agriculture (USDA), and the Department of the Interior (DOI). But what would have been a $2 billion decline was largely offset by increases in other agencies, especially the National Institutes of Health (NIH), NSF, and the National Aeronautics and Space Administration (NASA), so that the net decrease was just $173.6 million in 1998 dollars.
Federal support of basic research did not decrease between 1993 and 1997. Although the aggregate level of basic research funding dipped for several years after 1993, it was up again slightly—by 1.5 percent—in 1997 over 1993 in real terms. This occurred in part because of real growth in research spending by agencies that favor basic research, NIH and NSF, which offset decreases at DOD, DOE, and other agencies with shrinking research budgets. In 1993, 66 percent of NIH's research budget and 93 percent of NSF's were classified as basic research;
TABLE A-1 Trends in Federal Research Funding Obligations, FY 1993-1997 (millions of 1998 dollars)
|
|
|
Change, 1993-1997 |
|
|
FY 1993 |
FY 1997 |
Amount |
Percent |
DOD |
5,353.7 |
3,882.0 |
-1,471.6 |
-27.5 |
NASA |
3,971.7 |
4,264.2 |
292.5 |
7.4 |
DOE |
3,850.3 |
3,635.3 |
-214.9 |
-5.6 |
DHHSa |
10,288.3 |
11,440.9 |
1,152.5 |
11.2 |
NIH |
9,668.2 |
10,719.0 |
1,050.9 |
10.9 |
NSF |
2,106.2 |
2,291.1 |
184.9 |
8.8 |
USDA |
1,400.9 |
1,314.6 |
-86.3 |
-6.2 |
DOI |
649.2 |
563.1 |
-86.1 |
-13.3 |
EPAb |
404.6 |
416.9 |
12.3 |
3.0 |
DOCc |
651.5 |
823.4 |
171.9 |
26.4 |
Others |
1,419.4 |
1,290.7 |
-128.7 |
-9.1 |
Total research |
30,095.7 |
29,922.1 |
-173.6 |
-0.6 |
aDepartment of Health and Human Services. bEnvironmental Protection Agency. cDepartment of Commerce. Note: Constant-dollar conversions were made using the Gross Domestic Product (GDP) deflators in OMB (1998, Table 10.1). Source: National Science Foundation (1998b). |
BOX A-1 Definitions of R&D Common definitions of basic research, applied research, and development are used by the Office of Management and Budget (OMB), the American Association for the Advancement of Science (AAAS), and NSF, the sources of data in this paper. NSF uses the same definitions in its survey of industry. They are also generally consistent with international definitions. The objective of basic research is to gain more comprehensive knowledge or understanding of the subject under study, without specific applications in mind. The objective of applied research is to gain knowledge or understanding to meet a specific recognized need. Development is the systematic use of the knowledge or understanding gained from research directed toward the production of useful materials, devices, systems, or methods.4 |
the percentages in 1997 were largely unchanged. As a group, agencies other than NIH and NSF spent 7.7 percent less on basic research in 1997 than in 1993 in real terms.
In short, although federally funded research does not appear to have suffered greatly from the decline in R&D that took place after 1992, the overall average downturn and the modest recovery since 1996 obscure the fact that research spending by some agencies has declined much more than others. Moreover, the agencies with declining of stagnant research budgets also turn out to be the primary funders of certain fields of research (Table A-2). In 1993, for example, DOD provided the majority of federal support of research in electrical engineering (82 percent), mechanical engineering (75 percent), materials engineering (73 percent), and computer science (57 percent). DOE provided the majority of funding for physics research (62 percent) and was the single largest supporter of chemical engineering (42 percent) and chemistry (29 percent). NASA provided the majority of funding for four other fields: aeronautical engineering (81 percent), astronautical engineering (79 percent), astronomy (76 percent), and atmospheric sciences (52 percent).
Issues
It was inevitable that R&D expenditures would be affected by the bipartisan consensus to reduce the budget deficit, and it is not very surprising that agency
TABLE A-2 Top Federal Funders of Research by Field, FY 1993 (percent of total federal funding)
R&D budgets would change with circumstances as historic as the dissolution of the Soviet Union. Agency missions change and with them the resources for functions that support the missions. Nevertheless, the trends in agency budgets have raised three concerns:
1. |
First, that the nation's capacity and productivity in fundamental longer-range research may be harmed by shorter-term trends, including declining support of research by certain federal mission agencies, especially if there is a trend in private industry to focus on projects with nearer-term payoffs.5 |
2. |
Second, that because of the dependence of certain research fields such as physics, engineering, computer science, and mathematics on agencies with declining budgets, their health could be endangered for reasons unrelated to |
5 |
See Appendix B for an assessment of the evidence on this point. |
| |
3. |
Third, that because research grants to university investigators are the principal source of federal funds for the production of highly trained people in those science and engineering disciplines, graduate training could also be curtailed inadvertently. |
Funding Trends by Research Field
We have seen that the research budgets of some of the largest R&D agencies have fallen or have not grown significantly. Some of those agencies, notably DOD, DOE, and NASA, provide the majority of funding in most fields of engineering, physical sciences, mathematics, and computer sciences. Some of the fields, for example, physics, computer science, electrical engineering, and materials engineering, are important to innovation in information technologies and to national economic performance generally. This situation raises a set of important questions:
- How and to what extent have changes in federal agencies' research budgets affected the funding of fields dependent on them?
- Are the changes invariably in the same direction?
- Is there evidence of an effort to protect certain performers (e.g. universities) even in fields experiencing declining federal support overall?
- Are there cases in which one agency has compensated for reduced support by another agency?
- Is there evidence that a mechanism exists for making such adjustments to maintain a balanced research portfolio?
- Do changes in graduate student support parallel changes in levels of research funding?
This paper relies on a series of annual surveys conducted by NSF's Division of Science Resources Studies (SRS) to attempt a preliminary test of these questions. The SRS survey of federal funds for R&D includes retrospective reports of agencies' actual obligations by fiscal year. The information is collected in a number of relevant categories that can be cross-tabulated in useful ways for an analysis of trends in federal research funding. For example, federal obligations for research are classified as basic research or applied research in 19 natural science and engineering fields.6 Research performed by universities and colleges
6 |
The NSF survey also includes obligations for research in the social and behavioral sciences. In addition, the agencies report funding of research ''not elsewhere classified'' for each broad field, such as life sciences and engineering, and for research that cannot be attributed to any broad field (see Table A-3 and footnote 9). |
but not federally funded research performed in industry or government laboratories can be separately tabulated by field and/or whether it is basic or applied research.7 The ability to consider funding by agency and field of research makes it possible to assess whether the trend in an agency's level of research funding is correlated with trends in the funding of fields of which that agency has been the primary funder.
SRS also conducts an annual survey of university departments that includes reports of the principal source of support for each graduate student in science and engineering, including federally funded mechanisms such as fellowships, traineeships, and research assistantships. This information is reported for most of the research fields included in the federal funds report. Thus, trends in federal support of graduate students in those fields can be related to trends in the overall number of graduate students.
Both surveys—of federal R&D obligations and graduate student support—lag events because they are retrospective. For example, NSF recently released data from the survey of actual obligations during FY 1997 approximately 15 months after the end of the fiscal year. This permits analysis of trends in funding by field and agency during the period of budget cuts—fiscal years 1993 to 1997. A data series of just five years is too short a period to be very sure lasting trends exhibited by individual fields will be, but it is sufficient to derive tentative conclusions with regard to the broad questions posed above.8
Another obvious caveat is that quantitative changes in funding reveal little about the character of the research being cut or increased and therefore little about the qualitative effects of the changes in funding. It may be that a small change in one agency's support would have a far more profound effect on research and training in a field than a more substantial change in funding of research with a different character. The evolving orientation of some research fields is also difficult or impossible to discern from the data, although there is some effort to capture inter- and multidisciplinary research.9
7 |
Although the nomenclature corresponds to academic disciplines and departments, in NSF's Federal Funds Survey it is applied to R&D performed in industry and government laboratories and by other nonprofit institutions. Obligations for "development" are not reported by field. |
8 |
The surveys ask agencies to estimate future research allocations for seven broad fields for research—e.g., physical sciences, life sciences, math/computer sciences, engineering, etc.—but these estimates are too general to be very useful to this analysis of impacts on specific fields. |
9 |
In the Federal Funds Survey, multidisciplinary or interdisciplinary projects that do not fall within one of the broad fields of science (e.g., engineering, physical sciences, life sciences) are to be reported as "Other science, n.e.c.," where "n.e.c." means ''not elsewhere classified." Multidisciplinary projects that fall within a broad field and single-discipline projects that cannot be classified within one of the listed subfield categories are to be reported as ''Engineering, n.e.c." or "Mathematics & computer sciences, n.e.c.," etc. As shown in Table A-3, the n.e.c. categories are quite large in engineering and environmental science and much smaller in other major fields. They have been growing in most major fields but not very rapidly. All "n.e.c." research combined was 12.8 percent of total federal research in 1993 and 14.1 percent in 1997. |
There is also a complication with the data reported by NSF on the survey. Beginning with FY 1996, NSF changed its procedures for classifying research obligations by field. The change most affected engineering. The amount classified as "engineering, n.e.c." went from about 20 percent of the total in 1995 to about 40 percent after that year. Mechanical engineering went from 13 percent to 2 percent of engineering research funding. The physical sciences were also affected. The amount classified as "physical sciences, n.e.c" increased from about 9 percent in 1995 to 26 percent in 1997. It appears that this involved moving research previously classified astronomy, chemistry, and physics into the n.e.c. category. Environmental sciences were also affected, except that atmospheric, geology, and ''environmental sciences, n.e.c." research were each apparently reclassified as oceanography. As a result, as we analyze affected fields such as mechanical engineering, we must take into account the extent to which NSF's new classification scheme might affect the results. The change will also affect our ability to ascertain how much NSF might have compensated for cutbacks in the support of fields by other agencies.
The research obligation trends by field are summarized in Table A-3. The findings reported in the following section are for a number of fields related to industrial activity.10
Fields with Declining Support
Fields whose federal support declined from 1993 to 1997 included electrical engineering, mechanical engineering, physics, chemical engineering, chemistry, and geology. Although several of these fields had DOD or DOE as a dominant11 funder in 1993, others (chemistry, chemical engineering, and geology) had quite diversified support.
Electrical Engineering
In 1993 DOD provided 82 percent of the federal funding for electrical engineering. DOD support dropped 40.3 percent between 1993 and 1997 in real terms, which accounted for most of the net drop of 35.7 percent in federal support of the field (see Figure A-1 and Table A-4). In this case there were simultaneous decreases at DOE and NSF. The only increase of significance—90.4 percent—was at the Department of Commerce (DOC) presumably because of the growth of the Advanced Technology Program at the National Institute for Standards and
10 |
Data for fields not included here (i.e., astronomy, astronautical engineering, agricultural sciences, environmental biology, psychology, and social sciences) are available at www4.nas.edu/pd/step/23ba.nsf, as are tables in current dollars for all of the fields in the survey. |
11 |
"Dominant" funder is defined here as providing 50 percent or more of the federal support for a research field. |
TABLE A-3 Constant Dollar Changes in Federal Obligations for Research, Selected Fields, All Versus University Performers, FY 1993–1997
(millions of constant 1998 dollars)
|
All Performers |
|
Universities |
|
||
|
1993 |
1997 |
% Change |
1993 |
1997 |
% Change |
All fields |
30,095.7 |
29,922.1 |
-0.6 |
10,681.9 |
10,984.3 |
2.8 |
Engineering, total |
6,154.9 |
5,798.1 |
-5.8 |
993.8 |
1,006.7 |
1.3 |
Aeronautical |
1,335.2 |
1,378.3 |
3.2 |
58.3 |
50.2 |
-13.8 |
Astronautical |
552.9 |
607.6 |
9.9 |
22.8 |
17.9 |
-21.5 |
Chemical |
274.8 |
239.4 |
-12.9 |
73.2 |
63.9 |
-12.7 |
Civil |
282.0 |
280.8 |
-0.4 |
42.2 |
46.0 |
8.9 |
Electrical |
986.6 |
634.0 |
-35.7 |
219.0 |
149.1 |
-31.9 |
Mechanical |
522.6 |
259.4 |
-50.4 |
131.3 |
77.6 |
-40.9 |
Metallurgy/materials |
778.6 |
877.0 |
12.6 |
224.5 |
266.2 |
18.5 |
Engineering, n.e.c. |
1,422.4 |
1,521.5 |
7.0 |
222.4 |
335.8 |
51.0 |
Physical Sciences, total |
4,954.7 |
4,227.3 |
-14.7 |
1,312.4 |
1,186.0 |
-9.6 |
Astronomy |
768.1 |
789.4 |
2.8 |
134.4 |
171.2 |
27.4 |
Chemistry |
943.7 |
861.5 |
-8.7 |
389.4 |
347.4 |
-10.8 |
Physics |
2,952.9 |
2,106.8 |
-28.7 |
680.5 |
532.0 |
-21.8 |
Physical sciences, n.e.c. |
290.1 |
469.6 |
61.9 |
108.1 |
135.4 |
25.3 |
Life Sciences, total |
12,056.1 |
12,901.3 |
7.0 |
6,156.1 |
6,690.5 |
8.7 |
Biological sciences |
5,360.8 |
5,421.0 |
1.1 |
3,072.7 |
3,566.9 |
16.1 |
Environmental biology |
622.8 |
593.6 |
-4.7 |
179.6 |
149.4 |
-16.8 |
Agricultural sciences |
797.9 |
653.5 |
-18.1 |
167.0 |
176.4 |
5.6 |
Medical sciences |
4,929.4 |
5,637.6 |
14.4 |
2,614.9 |
2,585.4 |
-1.1 |
Life sciences, n.e.c. |
345.0 |
595.7 |
72.6 |
121.9 |
212.5 |
74.3 |
Mathematics & Computer Sciences, total |
1,371.5 |
1,703.5 |
24.2 |
548.5 |
582.4 |
6.2 |
Mathematics |
325.5 |
307.3 |
-5.6 |
152.1 |
130.7 |
-14.0 |
Computer sciences |
924.6 |
1,288.7 |
39.4 |
378.1 |
427.2 |
13.0 |
Mathematics & computer sciences, n.e.c. |
121.4 |
107.5 |
-11.5 |
18.4 |
24.6 |
33.8 |
Environmental Sciences, total |
2,919.4 |
3,103.4 |
6.3 |
660.8 |
684.8 |
3.6 |
Atmospheric |
1,101.4 |
1,186.4 |
7.7 |
178.6 |
209.7 |
17.4 |
Geological |
893.0 |
704.9 |
-21.1 |
211.0 |
126.6 |
-40.0 |
Oceanography |
523.1 |
609.0 |
16.4 |
164.2 |
219.5 |
33.7 |
Environmental sciences, n.e.c. |
402.0 |
603.0 |
50.0 |
107.0 |
129.1 |
20.7 |
Technology (NIST), but DOC's program is so small that nearly doubling it only reduced the drop that would have occurred in overall federal support by about 6.4 percentage points.
The results were similar for electrical engineering research conducted at universities and colleges (see Table A-5). The falloff in DOD support accounted for most of the net drop of 31.9 percent in federal funding of university research in electrical engineering in 1997. NSF also reduced its support by 25.4 percent.
TABLE A-4 Federal Obligations for Electrical Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-5 Federal Obligations for University Research in Electrical Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Mechanical Engineering
Federal support of mechanical engineering research followed a pattern very similar to electrical engineering. DOD, which provided 75 percent of all federal funding for mechanical engineering research in 1993, reduced its real level of funding of the field by 52.4 percent in 1997 (Figure A-2 and Table A-6). The next two largest federal funders, DOE and NSF, also reported that they had reduced their levels of support. As a result, net federal support of mechanical engineering research was 50.4 percent less in 1997 than in 1993 in real terms, as reported in the NSF survey. Mechanical engineering, however, is one of the fields affected by NSF's change in classification procedures. If we assume that
the amount of research NSF classified as mechanical engineering in 1993 has stayed at the same funding level in real terms, then overall federal support fell by 44.3 percent in 1997.
The same agencies, DOD, DOE, and NSF, reduced their support of university research (see Table A-7). Net federal support for mechanical engineering research was 40.9 percent less in 1997 than in 1993. If NSF funding is held constant in real terms, the drop is 20.9 percent.
TABLE A-6 Federal Obligations for Mechanical Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-7 Federal Obligations for University Research in Mechanical Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Physics
DOE is the largest federal funder of physics research, accounting for 62 percent in 1993, followed by DOD, which supported 18 percent. Physics research at DOE and DOD sustained real cuts of 28.6 and 63.2 percent, respectively (see Figure A-3 and Table A-8). NSF also reduced its support of physics research by 28.0 percent. Although several other agencies (e.g., Commerce, NASA) maintained or increased their funding of physics research, federal funding fell 28.7 percent between 1993 and 1997. University physics research did a little better, losing 12.5 percent of its federal support in real terms in the same period (see
Table A-9). Holding NSF constant, because some research classified as physics in 1993 was reclassified as "physical sciences, n.e.c." in 1997, the drop in federal funding would be 26.9 percent in total support and 6.3 percent in university support.
TABLE A-8 Federal Obligations for Physics Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-9 Federal Obligations for University Research in Physics, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Mathematics
DOD support for mathematics research by all performers suffered a steep decline (nearly 50 percent) between 1993 and 1997, but this was partially offset by increased support by DOE and DHHS (not NIH) and to a lesser extent NASA, so that the overall drop was only 5.6 percent. NSF support remained flat (see Figure A-4 and Table A-10).
Support of mathematics research at universities fell by 14.1 percent between 1993 and 1997, and in this case DOE and NSF support also declined while NASA and NIH support increased (see Table A-11).
TABLE A-10 Federal Obligations for Mathematics Research (basic and applied), by All Performers, by Field and Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-11 Federal Obligations for University Mathematics Research (basic and applied), by Field and Agency, FY 1990–1997 (millions of constant 1998 dollars)
Chemistry
The base of support for chemistry is broad; six agencies provided between 5 and 30 percent of the funding each in 1993. The level of federal support for chemistry research was reduced by 8.7 percent between 1993 and 1997 in real terms (see Figure A-5 and Table A-12). Spending was down at DOD (by 26.6 percent), DOE (by 12.5 percent), DHHS/NIH (by 5.7 percent), and NSF (by 3.5 percent). There were increases at Commerce (27.1 percent), Interior (5.8 percent), and NASA (3.0 percent), but none were large in absolute terms. The pattern in federal support of university chemistry research was very similar (10.8
percent) (Table A-13). Holding NSF constant, because some research classified as chemistry in 1993 was apparently reclassified as ''physical sciences, n.e.c'' after 1995, reduces the federal cuts slightly, to 8.2 percent overall and 9.2 percent in university research.
TABLE A-12 Federal Obligations for Chemistry Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-13 Federal Obligations for University Research in Chemistry, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Chemical Engineering
Support of chemical engineering research is spread among a number of agencies: DOE (42 percent), DOD (28 percent), and NSF (16 percent). Between 1993 and 1997, DOD reduced its support substantially, and NSF also provided less support, in real terms (see Figure A-6 and Tables A-14 and A-15). Increases at a number of other agencies did not offset the declines. Federal spending on chemical engineering research was 12.9 percent less in 1997 than in 1993 in real terms. NSF's new classification procedures did not noticeably affect the level of NSF support of chemical engineering.
TABLE A-14 Federal Obligations for Chemical Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-15 Federal Obligations for University Research in Chemical Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Civil Engineering
DOD support for civil engineering research by all performers declined 44.2 percent between 1993 and 1997, but this was almost entirely offset by increased support by NSF, the Department of Commerce, and other agencies, so that total funding was down only 0.4 percent. Nevertheless, 1993 funding of civil engi-
neering research was low compared to previous and subsequent years, and the general trend in federal funding of civil engineering research has been down in the 1990s (Figure A-7 and Table A-16). As with chemical engineering, (and unlike mechanical engineering), NSF support of civil engineering was not noticeably affected by the advent of new classification procedures in 1996.
For university research in civil engineering, a nearly 30 percent increase by NSF more than offset reduced DOD support, so that funding was up 8.8 percent overall between 1993 and 1997 (see Table A-17).
TABLE A-16 Federal Obligations for Civil Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-17 Federal Obligations for University Research in Civil Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Geology
Another field with support spread among a number of agencies (five accounted for between 10 and 33 percent each in 1993), geology experienced a 21.1 percent real drop in federal research spending between 1993 and 1997 (see Figure A-8 and Table A-18). The Department of the Interior, the largest funder, and NASA increased their levels of support during the period, by 3.8 and 6.2 percent,
respectively. NSF joined DOD, DOE, and EPA in providing less support—33.6, 86.9, 55.2, and 26.0 percent, respectively. The NSF drop was probably due in part to the 1996 change in classification procedures. If NSF funding were held constant, the overall federal drop in support would be closer to 15 percent.
Federal spending on university geology research was 40.0 percent less in 1997 than in 1993 in real terms, as every agency involved in geology research that was surveyed reduced its level of support (see Table A-19). It should be noted, however, that the Department of the Interior is not surveyed for spending on university research and that the Interior Department accounts for more than two-fifths of the overall federal investment in geology research (see Table A-18). Presumably if the Interior Department numbers were included, the drop in federal
support of university geology research would be reduced, even though the U.S. Geological Survey has a relatively small extramural research program. At least some of the decrease in NSF support probably stems from the 1996 change in classification. If NSF were held constant, the overall decrease in federal funding of university geology would be closer to 27 percent.
TABLE A-18 Federal Obligations for Geological Sciences Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-19 Federal Obligations for University Research in Geological Sciences, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Fields with Increasing Support
As might be expected, the fields of biological and medical sciences benefited from the 10.9 percent real growth in the National Institutes of Health between 1993 and 1997, although to very different degrees. Aeronautical engineering benefited from the modest growth in NASA's research budget, especially from 1996 to 1997, but university research was reduced in contrast to the pattern in most other fields. Contrary to trends in the research budgets of the dominant funding agencies, however, the fields of computer sciences and materials engineering showed significant growth in federal support. Oceanography, the only other field showing substantial growth, is supported by several agencies.
Biological Sciences
DHHS, led by NIH, funded 82 percent of the biological sciences in 1993. NIH funding of the biological sciences was $4.5 billion in 1997, 2.1 percent more than in 1993 in real terms (see Figure A-9 and Table A-20). There were increases
TABLE A-20 Federal Obligations for Biological Sciences Research, All Performers, by Agency, FY 1990-1997 (millions of constant 1998 dollars)
from much smaller bases at the Department of Commerce, NSF, NASA, and the Department of the Interior, which were more than offset by declines at DOD and USDA. The net increase in federal support for biological sciences was 1.1 percent between 1993 and 1997 in real terms.
Federal funding of university biological sciences research fared better. It was 16.1 percent larger in 1997 than in 1993 in real terms (see Table A-21). The increase in NIH funding accounted for nearly all of the increase. There were small increases at NASA, NSF, and DOD, which were more than offset by a drop in DOE support.
Medical Sciences
DHHS, again led by NIH, provided 84 percent of the support for research in the medical sciences in 1993. NIH increased its level of support for medical research by 19.4 percent between 1993 and 1997 (see Figure A-10 and Table A-22). Although several other agencies decreased their levels of support, especially DOD and the Department of Veterans Affairs (included in "All others"), there was a net increase in federal funding of medical sciences research of 14.4 percent between 1993 and 1997 in real terms.
TABLE A-21 Federal Obligations for University Research in Biological Sciences, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-22 Federal Obligations for Medical Sciences Research, All Performers, by Agency, FY 1990–1997 (millions of dollars)
The annual level of federal funding of university medical research fell by 1.1 percent in real terms between 1993 and 1996 (see Table A-23). Increased support by DOD (9.0 percent) and DOE (99.5 percent) (the latter two from small bases) was more than offset by decreases at NIH and NASA.
Aeronautical Engineering
NASA accounted for 81 percent of the federal support for aeronautical engineering research in 1993 (DOD supported most of the rest). NASA's total budget for research was 7.4 percent more in 1997 than in 1993, and it increased aeronautical engineering research by 6.1 percent (see Figure A-11 and Table A-24). Nevertheless, the increase occurred in a single year, 1997. Previously, NASA's funding of aeronautical engineering was falling. Several other agencies reduced support, especially DOD, and the field grew by only 3.2 percent overall. If the elimination of NSF support was only apparent, because the research was reclassified as ''engineering, n.e.c.,'' aeronautical research would have increased slightly more (by 3.5 percent if NSF support is held constant).
In contrast with many other fields, however, university research was not protected. University aeronautics was cut 13.9 percent in real terms between 1993
TABLE A-23 Federal Obligations for University Research in Medical Sciences Research, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
and 1997 (see Table A-25). However, the amount of funding going to universities was small to begin with, $52 million in 1993, just 4.4 percent of all federally funded aeronautical engineering research. Also, if NSF support were held constant, the cut would be less (8.7 percent).
TABLE A-24 Federal Obligations for Aeronautical Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-25 Federal Obligations for University Research in Aeronautical Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Computer Science
Not all research fields in which DOD is the main funder experienced decreased federal funding. One of these exceptions was computer sciences, for which DOD provided 57 percent of the federal support in 1993. DOD support fell between 1993 and 1997 but by less than 1 percent (Figure A-12 and Table A-26). But due to increases at several other agencies, total federal support of computer science research increased more than 39 percent in real terms. In absolute terms there were substantial increases in computer science funding at NSF and DOE in 1996, as well as smaller increases at the Department of Commerce and DHHS. This growth may have been driven in part by the interagency high-performance computing and communications initiative of the National Science and Technology Council. The increase in computer science research funding by DOE in 1996 was due to a large increase in computing by DOE's atomic weapons program. The large increase in NSF support appears to be real and not an artifact of changed classification procedures. The level of NSF support of mathematics and of "mathematics & computer sciences, n.e.c." did not change substantially from 1995 to 1996.
DOD also reduced slightly its support of university research in computer sciences between 1993 and 1997, from $210 million to $204 million or by 2.9 percent (see Table A-27). In fact, university computer science funding increased by 13.0 percent in real terms, with NSF accounting for nearly all of the increase. But if the NSF increase is all or mostly due to changes in the classification and reporting of ongoing activities, the increase might actually have been much smaller.
TABLE A-26 Federal Obligations for Computer Sciences Research All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-27 Federal Obligations for University Research in Computer Sciences, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Materials Engineering
Materials research, broadly defined, is an economically important area of research. NSF does not track the field, a relatively new one, but it does track funding for a major component, materials engineering. DOD is the dominant sponsor, providing 73 percent of federal funding in 1993. By 1997, DOD had reduced its annual commitment by 28 percent in real terms (see Figure A-13 and Table A-28). The Department of the Interior also reduced its support from about $25 million a year to zero when the Bureau of Mines was abolished. In this case, however, other agencies stepped up their support substantially, DOE by 275 percent, NSF by 356 percent, and the Department of Commerce by 85 percent (the latter two from small bases). As a result, real spending on materials engineering
TABLE A-28 Federal Obligations for Metallurgy & Materials Engineering Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
research was 12.6 percent greater in 1997 than in 1993 and the field had a broader base of support among the agencies.
The DOE increase began suddenly in 1996, when its spending on materials engineering tripled to $264 million, and may reflect a recategorization of existing activities, not a real increase. If the DOE increase were ignored, the field may not have experienced a real increase at all. On the other hand, the general level of NSF funding was lower after 1995 by about $40 million, indicating that materials engineering was affected by the 1996 change in classification procedures at NSF. If so, real federal funding of materials engineering research was greater than the survey indicates, offsetting the DOE changes.
Federal funding of university research in materials engineering increased 18.6 percent in real terms between 1993 and 1997 (see Table A-29). DOD's contribution was down by 28.6 percent compared with 1993 in real terms. Despite the huge increase in overall support for the field beginning in 1996, DOE's support of university research was much less—5.6 percent. The net increase in support came from a substantial increase in funding by NSF—372 percent from 1993 to 1997 in real terms. The increase may have been larger, because the 1996 change in classification procedures at NSF shifted at least some work from the materials engineering category to "engineering, n.e.c."
TABLE A-29 Federal Obligations for University Research in Metallurgy & Materials Engineering, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
Oceanography
Oceanography research receives most of its federal support from five agencies: Department of Commerce (32 percent), NASA (23 percent), DOD (19 percent), NSF (17 percent), and Department of the Interior (8 percent) (in 1993). Total federal funding was 16.4 percent greater in 1997 than in 1993 in real terms (see Figure A-14 and Table A-30). Federal spending levels fell at Commerce and Interior, by 17.8 and 68.8 percent, respectively. DOD increased its level of support substantially, by 23.8 percent. NSF more than doubled its level of spending on oceanography beginning in 1996, which accounted for more than the net increase. As noted above, however, NSF changed its procedures for classifying research by field in 1996. Although NSF support for environmental sciences overall was about the same in 1995 and 1996, there were sudden decreases in the levels of support for atmospheric and geological sciences and in "environmental sciences, n.e.c." that equaled the sudden jump in the level of oceanography research that year. If NSF funding of oceanography had stayed about even in real terms, federal support for oceanography research would have increased only by a few percent.
A similar pattern held in federal support of university research in oceanography. It was 36.6 percent larger in 1997 than in 1993 in real terms (see Table A-31). Again, a substantial jump in NSF spending beginning in 1996 accounted for all of the net increase, and most or all of that increase may be due to changes in how NSF categorizes research by field. Support from the other agencies was
TABLE A-30 Federal Obligations for Oceanography Research, All Performers, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
TABLE A-31 Federal Obligations for University Research in Oceanography, by Agency, FY 1990–1997 (millions of constant 1998 dollars)
either about the same or reduced. Also, the size of the increase might be less if spending on university oceanography by the Department of the Interior, which reduced its overall support of oceanography research during the time period, were included.
Trends in Graduate Enrollment
The NSF survey of the numbers of graduate students by source of support publishes most of the data by broad area of science and engineering, such as physical sciences or earth, atmospheric, and ocean sciences, rather than by academic field, such as physics and chemistry or geology and oceanography. The former categories are too general to associate changes in federal research funding with changes in graduate training by field. Table A-32 presents data on sources of federal support of full-time graduate students in science and engineering in all universities and colleges in five fields that correspond to the categories used above: computer science, biological sciences, chemical engineering, electrical engineering, and mechanical engineering.
The changes are mostly in the expected direction (see Table A-33). Where federal support of university research in a field was down sharply between 1993 and 1997 in real terms, graduate enrollments were generally down also. In chemical engineering, for example, federal funding of university research declined by nearly 13 percent between 1993 and 1997 in real terms (from Table A-15). The number of graduate students in chemical engineering also declined during the same period, by 5 percent (the number of graduate students whose main source of support was federal declined more) (Table A-32). In mechanical engineering,
TABLE A-32 Full-Time Graduate Students in Science and Engineering, Doctoral Institutions, by Field and Agency of Support: 1990–1997 (number)
|
|
|
|
|
|
|
|
|
Change, 1993–1997 |
|
Field and Source of Support |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
Number |
Percent |
All S&E Fields |
|
|
|
|
|
|
|
|
|
|
All federally supported |
51,706 |
55,166 |
57,326 |
59,469 |
59,788 |
58,539 |
57,094 |
55,902 |
-3,567 |
-6.0 |
DOD |
8,425 |
8,640 |
8,753 |
9,220 |
8,951 |
8,834 |
8,317 |
8,568 |
-652 |
-7.1 |
DHHS, total |
14,522 |
15,381 |
16,154 |
16,701 |
17,035 |
17,024 |
16,817 |
16,610 |
-91 |
-0.5 |
NIH |
13,298 |
14,081 |
14,930 |
15,749 |
15,843 |
15,600 |
15,411 |
15,181 |
-568 |
-3.6 |
NSF |
11,932 |
12,542 |
13,223 |
13,325 |
13,656 |
13,447 |
13,258 |
13,169 |
-156 |
-1.2 |
USDA |
2,629 |
2,982 |
3,099 |
3,202 |
3,297 |
3,152 |
2,902 |
2,560 |
-642 |
-20.0 |
Other federal |
14,198 |
15,621 |
16,097 |
17,021 |
16,849 |
16,082 |
15,800 |
14,995 |
-2,026 |
-11.9 |
All nonfederally supported |
198,824 |
205,988 |
215,816 |
216,827 |
215,536 |
210,874 |
209,138 |
206,259 |
-10,568 |
-4.9 |
Total, all sources of support |
250,530 |
261,154 |
273,142 |
276,296 |
275,324 |
269,413 |
266,232 |
262,161 |
-14,135 |
-5.1 |
Computer Sciences |
|
|
|
|
|
|
|
|
|
|
All federally supported |
2,376 |
2,501 |
2,595 |
2,883 |
3,007 |
3,109 |
3,074 |
3,114 |
231 |
8.0 |
DOD |
1,127 |
1,125 |
1,171 |
1,314 |
1,311 |
1,390 |
1,335 |
1,411 |
97 |
7.4 |
DHHS, total |
68 |
71 |
100 |
103 |
97 |
110 |
113 |
106 |
3 |
2.9 |
NIH |
61 |
65 |
97 |
95 |
89 |
92 |
80 |
78 |
-17 |
-17.9 |
NSF |
818 |
896 |
958 |
1,003 |
1,040 |
1,050 |
1,044 |
1,076 |
73 |
7.3 |
USDA |
14 |
8 |
6 |
7 |
9 |
13 |
13 |
12 |
5 |
71.4 |
Other federal |
349 |
401 |
360 |
456 |
550 |
546 |
569 |
509 |
53 |
11.6 |
All nonfederally supported |
12,522 |
12,369 |
13,174 |
12,944 |
12,382 |
12,107 |
12,559 |
13,374 |
430 |
3.3 |
Total, all sources of support |
14,898 |
14,870 |
15,769 |
15,827 |
15,389 |
15,216 |
15,633 |
16,488 |
661 |
4.2 |
|
|
|
|
|
|
|
|
|
Change, 1993–1997 |
|
Field and Source of Support |
1990 |
1991 |
1992 |
1993 |
1994 |
1995 |
1996 |
1997 |
Number |
Percent |
Biological Sciences |
|
|
|
|
|
|
|
|
|
|
All federally supported |
13,519 |
14,372 |
15,002 |
16,048 |
16,583 |
16,492 |
16,435 |
16,111 |
63 |
0.4 |
DOD |
222 |
220 |
229 |
269 |
346 |
354 |
358 |
358 |
89 |
33.1 |
DHHS, total |
9,908 |
10,616 |
10,989 |
11,554 |
11,927 |
12,004 |
11,819 |
11,723 |
169 |
1.5 |
NIH |
9,542 |
10,164 |
10,575 |
11,236 |
11,412 |
11,332 |
11,230 |
11,141 |
-95 |
-0.8 |
NSF |
1,201 |
1,269 |
1,333 |
1,362 |
1,369 |
1,399 |
1,484 |
1,518 |
156 |
11.5 |
USDA |
960 |
1,060 |
1,078 |
1,215 |
1,280 |
1,078 |
1,065 |
921 |
-294 |
-24.2 |
Other federal |
1,228 |
1,207 |
1,373 |
1,648 |
1,661 |
1,657 |
1,709 |
1,591 |
-57 |
-3.5 |
All nonfederally supported |
26,076 |
26,910 |
27,830 |
28,611 |
29,627 |
29,943 |
29,402 |
28,953 |
342 |
1.2 |
Total, all sources of support |
39,595 |
41,282 |
42,832 |
44,659 |
46,210 |
46,435 |
45,837 |
45,064 |
405 |
0.9 |
Chemical Engineering |
|
|
|
|
|
|
|
|
|
|
All federally supported |
1,528 |
1,592 |
1,643 |
1,735 |
1,706 |
1,740 |
1,769 |
1,622 |
-113 |
-6.5 |
DOD |
98 |
134 |
123 |
138 |
128 |
164 |
166 |
158 |
20 |
14.5 |
DHHS, total |
121 |
132 |
166 |
141 |
173 |
163 |
162 |
137 |
-4 |
-2.8 |
NIH |
119 |
128 |
146 |
119 |
143 |
136 |
148 |
121 |
2 |
1.7 |
NSF |
678 |
663 |
696 |
726 |
747 |
759 |
776 |
705 |
-21 |
-2.9 |
USDA |
27 |
16 |
38 |
52 |
54 |
66 |
34 |
25 |
-27 |
-51.9 |
Other federal |
604 |
647 |
620 |
678 |
604 |
588 |
631 |
597 |
-81 |
-11.9 |
All nonfederally supported |
3,892 |
4,161 |
4,257 |
4,267 |
4,348 |
4,165 |
4,087 |
4,091 |
-176 |
-4.1 |
Total, all sources of support |
5,420 |
5,753 |
5,900 |
6,002 |
6,054 |
5,905 |
5,856 |
5,713 |
-289 |
-4.8 |
Electrical Engineering |
|
|
|
|
|
|
|
|
|
|
All federally supported |
3,529 |
3,791 |
3,986 |
4,033 |
3,974 |
4,004 |
3,927 |
4,338 |
305 |
7.6 |
DOD |
1,636 |
1,623 |
1,680 |
1,734 |
1,649 |
1,699 |
1,647 |
1,961 |
227 |
13.1 |
DHHS, total |
77 |
85 |
90 |
103 |
100 |
113 |
93 |
113 |
10 |
9.7 |
NIH |
55 |
76 |
81 |
86 |
83 |
84 |
71 |
68 |
-18 |
-20.9 |
NSF |
1,131 |
1,253 |
1,360 |
1,260 |
1,266 |
1,298 |
1,298 |
1,324 |
64 |
5.1 |
USDA |
16 |
18 |
24 |
9 |
24 |
33 |
12 |
14 |
5 |
55.6 |
Other federal |
669 |
812 |
832 |
927 |
935 |
861 |
877 |
926 |
-1 |
-0.1 |
All nonfederally supported |
14,428 |
15,408 |
16,246 |
15,636 |
14,745 |
13,565 |
13,462 |
13,963 |
-1,673 |
-10.7 |
Total, all sources of support |
17,957 |
19,199 |
20,232 |
19,669 |
18,719 |
17,569 |
17,389 |
18,301 |
-1,368 |
-7.0 |
TABLE A-33 Changes in Graduate Enrollment and Federal Funding of University Research, Selected Fields, FY 1993–1997 (full-time graduate students in doctorate-granting institutions and federal research in constant dollars)
|
Number of Graduate Students |
|||
Field |
All Sources of Funding (%) |
Federally Funded (%) |
Nonfederally Funded (%) |
Federal funding of research in Universities (%) |
All S&E fields |
-5.1 |
-6.0 |
-4.9 |
2.8 |
Computer sciences |
4.2 |
8.0 |
3.3 |
13.0 |
Biological sciences |
0.9 |
0.4 |
1.2 |
a14.3 |
Chemical engineering |
-4.8 |
-6.5 |
-4.1 |
-12.7 |
Electrical engineering |
-7.0 |
7.6 |
-10.7 |
-31.9 |
Mechanical engineering |
-16.1 |
-13.3 |
-17.0 |
-40.9 |
a Includes environmental biology because graduate students in environmental biology are included in biological sciences in the survey of graduate students and postdoctorates. Notes: (1) Ninety-three percent of all full-time graduate students, and 99 percent of federally supported full-time graduate students, were enrolled at doctorate-granting institutions in 1997; (2) Graduate students are reported according to the source of "the largest amount" of support received in fall 1997; (3) Federal support of university research is reported only for six agencies: USDA, DOD, DOE, DHHS, NASA, and NSF. SOURCES: (1) Graduate students: Table A-29; (2) federal funding: Table A-3 (constant dollars). |
federal funding of university research declined by more than 40 percent (from Table A-7). Not surprisingly, then, the number of graduate students funded by federal agencies and by nonfederal sources was down, by 13 and 17 percent, respectively (Table A-32).
Electrical engineering is harder to interpret. Real federal support of academic research in electrical engineering declined by 32 percent between 1993 and 1997 (from Table A-5). Although the overall number of graduate students declined, as might be expected, the number of federally supported graduate students increased (Table A-32).
Federal support of computer science increased between 1993 and 1997 in real terms (the NSF survey reported an increase of nearly 13 percent, but it was probably less for reasons explained earlier) (Table A-27), and the number of federally funded graduate students in computer science also increased, by 8 percent (see Table A-33). The number of nonfederally funded graduate students also increased but by a smaller percentage (3 percent). Since they constituted the vast majority—more than 80 percent—of the graduate students in computer science, overall enrollment went up about 4 percent (Table A-32).
In the biological sciences, even though the level of federal support of university research was up substantially in 1997 compared with 1993 in real terms—more than 14 percent (including environmental biology)—the number of gradu-
ate students in biology hardly increased at all—by less than 1 percent. Federally supported graduate students increased even more slowly—by less than 0.5 percent.
The lack of consistent correlations between changes in federal research funding and changes in graduate enrollments is probably attributable to several factors. First, in most fields the federal government supports a fairly small proportion of graduate students. Second, there is bound to be a lag in the effect of federal research funding changes on graduate enrollments as current students work their way through graduate school and before the decisions of prospective students not to enter graduate school in a shrinking field can be felt. The data series is too short to ascertain whether there is a lagged effect. Third, although much federal support of graduate students comes thorough research grant funding in the form of research assistantships, some of it comes through specific fellowship and traineeship programs that are not counted as R&D. DHHS has the largest such programs, and NSF also has substantial fellowship and traineeship programs. If agencies protected education and training programs relative to research grant programs in budget downsizing, the number of federally funded graduate students could increase as research funding remained the same or declined.
Observations
In many fields of science and engineering a single mission agency is the principal source (50 percent or more) of federal support for research. It is natural to assume that, if the research budgets of such agencies are shrinking, the fields they support will be similarly affected. This analysis shows that it is not necessarily true that an agency's level of funding for a particular field goes down or up with the fortunes of the agency's overall research budget. Agencies do not cut or increase programs across the board. They may, and often do, protect some fields from cuts, cut them less, or even let them grow despite a shrinking research budget. Of course, that implies even deeper funding cuts in other fields supported by the shrinking research budget. Conversely, an agency with an increasing budget may hold down or even cut the level of support for one or more fields and give larger increases to other fields. Thus, it is important to assess funding trends for agencies and fields together.
In the period 1993 to 1997 the fields disadvantaged by changes in agency budgets included most fields of engineering (apart from aeronautical and materials engineering), physics, mathematics, and chemistry. The field that benefited most by the changes during this period was medical science. Fields whose funding fortunes did not comport with those of their major sponsor were computer science and materials engineering, which grew despite DOD cuts.
Although university research increased slightly from 1993 to 1997, support by field was also variable. In the 15 research fields examined in this paper, federal spending increased in only six and was lower in nine fields. In many
fields university research support fared better than overall research funding—that is, it decreased by a smaller percentage (chemical engineering, electrical engineering, mechanical engineering, physics, and geology) or increased by a larger percentage (metallurgy & materials engineering, biological sciences, atmospheric science and oceanography), but there were exceptions (chemistry, mathematics, and computer science). In two fields (medical sciences and aeronautical engineering), university research support declined while overall support increased.
Although there are instances in which a second agency has "picked up the slack" in a field receiving less support elsewhere, no single agency has explicitly or implicitly played that role, in contrast to the early 1970s when in response to cutbacks at NASA and DOD, NSF was charged by Congress to serve as the balance wheel of the system and the NSF budget was increased to pick up some of the researchers previously funded by NASA and DOD. In contrast, during the 1990s, NSF reduced or held flat its levels of support for several fields being cut by mission agencies, including chemical engineering, electrical engineering, mechanical engineering, physics, mathematics, and geology. NSF did increase its level of support for some fields being cut by DOD—materials engineering and computer sciences. The 1996 changes in how NSF classifies research by field make it difficult to interpret some of the changes from 1993 to 1997, especially in mechanical and materials engineering. Overall, however, NSF funding appears to have followed (or led) aggregate trends, generally boosting funding for fields that prospered and reducing funding for fields that suffered reductions.
It should be acknowledged that, although NSF's research budget has increased faster than inflation, the percentage increases would not be nearly enough to compensate for the magnitude of cutbacks in some fields, given the relatively modest size of NSF's budget. NSF provides support for most fields of science and engineering but typically accounts for 10 to 15 percent or less of total federal research funding in an individual field.
The reverse of stepping up funding to compensate for a drop in another agency's support of an important field would be for an agency under budget pressure to cut back because of increases elsewhere in the federal government. This may have been a factor in DOD's reductions in the biological sciences and medical research.
It is probably too early to see the effects of changes in levels of federal funding of research, even large ones, on graduate enrollments, let alone degree attainment rates. Such effects, if they do appear, will probably be attenuated because federal funding is not the principal source of support for graduate students in most fields.
Finally, what about the argument that having diversified agency research support is preferable to being dependent on a single major source of support because it is less vulnerable to sudden budget changes? This analysis provides little support for that hypothesis. One field with a broad base of federal funding support came out of the downturn in research funding between 1993 and 1997
with a larger budget in real terms (oceanography, +16.4 percent), but as explained above, most or all of that increase may be more apparent than real, resulting from changes in NSF classification procedures. The rest had substantially smaller budgets (chemistry, -8.7 percent; chemical engineering, -12.9 percent; geology, -21.1 percent; mathematics, -5.6 percent; and civil engineering, -12.9 percent) (see Figures A-15 and A-16).
References
American Association for the Advancement of Science. 1998. Historical Data on Federal R&D FY 1986-1999. Accessible from the AAAS website: HtmlResAnchor http://www.aaas.org/spp/dspp/rd/rdwwwpg.htm.
National Science Board. 1998. Science & Engineering Indicators--1998. Arlington, VA: National Science Foundation. Accessible from the NSF/SRS website: HtmlResAnchor http://www.nsf.gov/sbe/srs/stats.htm.
National Science Foundation. 1998a. Patterns of R&D Resources: 1998, Early Release Tables. Accessible from NSF/SRS website: HtmlResAnchor http://www.nsf.gov/sbe/srs/srs99402/start.htm.
National Science Foundation. 1998b. Survey of Federal Funds for Research and Development: Fiscal Years 1997, 1998, and 1999. Unpublished tables available from Division of Science Resource Studies, NSF.
National Science Foundation. 1999. Graduate Students and Postdoctorates in Science and Engineering: Fall 1997, Early Release Tables, NSF 99-405. Available at HtmlResAnchor http://www.nsf.gov/sbe/srs/srs99405/start.htm.
Office of Management and Budget. 1998. Historical Tables, Budget of the United States Government: Fiscal Year 1999. Washington, D.C.: U.S. Government Printing Office.