Chapter 5: Academic Research and Development

Financial Resources for Academic R&D

Academic R&D is a significant part of the national R&D enterprise.[1] To carry out world-class research and advance the scientific knowledge base, U.S. academic researchers require financial resources, stability of research support, and research facilities and instrumentation that facilitate high-quality work. Several funding indicators bear on the state of academic R&D, including:

  • The level and stability of overall funding
  • The sources of funding and changes in their relative shares
  • The distribution of funding among the different R&D activities (basic research, applied research, and development)
  • The distribution of funding among S&E broad and detailed fields
  • The distribution of funding across institutions that perform academic R&D and the extent of their participation
  • The role of the federal government as a supporter of academic R&D and the particular roles of the major federal agencies funding this sector
  • The state of the physical infrastructure (research equipment and facilities)

Individually and in combination, these factors influence the evolution of the academic R&D enterprise and, therefore, are the focus of this section. The main findings are as follows:

  • Continued growth in both federal and nonfederal funding of academic R&D
  • A recent increase in the role of the federal government following a steady relative decline, and a corresponding relative decline in the roles of industry and state and local government
  • A substantial increase in National Institutes of Health (NIH) funding relative to the other main federal funding agencies
  • Continued but differential increases in funding for all fields, resulting in a relative shift in the distribution of funds, with increasing shares for the life sciences, engineering, and the computer sciences
  • R&D activity occurring in a wider set of institutions, with the concentration of funds among the top research universities diminishing slightly
  • The share of all annual R&D expenditures spent on research equipment reaching a historic low
  • Continuous growth in academic S&E research space, particularly in the medical and biological sciences
  • The increasingly important role of "cyberinfrastructure" in the conduct of S&E research.

For a discussion of the nature of the data used in this section, see sidebar, "Data Sources for Financial Resources for Academic R&D."

Academic R&D Within the National R&D Enterprise

Academia is widely viewed as important to the nation's overall R&D effort, especially for its contribution to generating new knowledge through basic research. Since 1998, academia has accounted for more than half of the basic research performed in the United States.

In 2004, U.S. academic institutions spent an estimated $42 billion, or $39 billion in constant 2000 dollars, on R&D.[2] Academia's role as an R&D performer has increased during the past three decades, rising from about 10% of all R&D performed in the United States in the early 1970s to an estimated 14% in 2004 (figure 5-1 figure.). For a comparison with other countries, see chapter 4, "International R&D Comparisons."

Character of Work

Academic R&D activities are concentrated at the research (basic and applied) end of the R&D spectrum and do not include much development activity.[3] For the definitions used in National Science Foundation (NSF) surveys and a fuller discussion of these concepts, see chapter 4 sidebar, "Definitions of R&D." Recently, there has been some discussion about whether a shift away from basic research and toward the pursuit of more utilitarian, problem-oriented questions is occurring in academia. (For a brief analysis of this issue, see sidebar "Has Academic R&D Shifted Toward More Applied Work?" later in this chapter.) For academic R&D expenditures in 2004, an estimated 97% went for research (75% for basic and 22% for applied) and 3% for development (figure 5-2 figure.). From the perspective of national research (basic and applied), as opposed to national R&D, academic institutions accounted for an estimated 33% of the U.S. total in 2004. In terms of basic research alone, the academic sector is the country's largest performer, currently accounting for an estimated 54% of the national total. Between the early 1970s and early 1980s, the academic sector's basic research share declined from slightly more to slightly less than one-half of the national total. In the early 1990s, its share of the national total began to increase once again.


Between 1970 and 2004, the average annual R&D growth rate (in constant 2000 dollars) of the academic sector (4.5%) was higher than that of any other R&D-performing sector except the nonprofit sector (4.7%). (See figure 5-3 figure. and appendix table 4-4 Excel table. for time-series data by R&D-performing sector.) As a proportion of gross domestic product (GDP), academic R&D rose from 0.24% to 0.37% during this period, a 50% increase. (See appendix table 4-1 Excel table. for GDP time series.) Between 2000 and 2004, average annual R&D growth was higher in the academic sector (6.3%) than in any other sector except federally funded research and development centers (6.9%).

Top of page.Top of page

Major Funding Sources

The academic sector relies on a variety of funding sources for support of its R&D activities. The federal government continues to provide the majority of funds (figure 5-4 figure.). After declining for almost three decades, with most of the decline occurring during the 1980s, its share has recently begun to increase. In 2003, the federal government accounted for about 62% of the funding for R&D performed in academic institutions, compared with its peak of 69% in 1973 and 58% in 2000 (figure 5-5 figure.; appendix table 5-2 Excel table.).

Federal support of academic R&D is discussed in detail later in this section. The following list summarizes the contributions of other sectors to academic R&D:[4]

  • Institutional funds. In 2003, institutional funds from universities and colleges constituted the second largest source of funding for academic R&D, accounting for 19%, slightly below its peak of 20% in 2001 (appendix table 5-2 Excel table.). Institutional funds encompass two categories: (1) institutionally financed organized research expenditures and (2) unreimbursed indirect costs and related sponsored research. They do not include departmental research and thus exclude funds, notably for faculty salaries, in cases where research activities are not separately budgeted.

    The share of support represented by institutional funds had been increasing during the past three decades, except for a brief downturn in the early 1990s, but recently began to decline in 2001. Institutional R&D funds may be derived from (1) general-purpose state or local government appropriations (particularly for public institutions) or federal appropriations; (2) general-purpose grants from industry, foundations, or other outside sources; (3) tuition and fees; (4) endowment income; and (5) unrestricted gifts. Other potential sources of institutional funds are income from patents or licenses and income from patient care revenues. (See "Patents Awarded to U.S. Universities" later in this chapter for a discussion of patent and licensing income.)

  • State and local government funds. State and local governments provided 7% of academic R&D funding in 2003. Between 1980 and 2001, the state and local share of academic R&D funding fluctuated between 7% and 8%. However, the share has declined every year since 1996. This share, however, only reflects funds directly targeted to academic R&D activities by state and local governments. It does not include general-purpose state or local government appropriations that academic institutions designate and use to fund separately budgeted research or cover unreimbursed indirect costs.[5] Consequently, the actual contribution of state and local governments to academic R&D is not fully captured here, particularly for public institutions. (See chapter 8 for some indicators of academic R&D by state.)

  • Industry funds. The funds provided for academic R&D by the industrial sector grew at a faster rate than funding from any other source during the 1973–2003 period. However, actual industry funding in inflation-adjusted dollars declined in both 2002 and 2003, the first time such a decline occurred in the past three decades. As a result, industry provided only 5% of academic R&D funding in 2003, a substantial decline from its peak of 7% in 1999. Industrial support accounts for the smallest share of academic R&D funding, and support of academia has never been a major component of industry-funded R&D. In 1994, industry's contribution to academic R&D represented 1.5% of its total support of R&D compared with 1.4% in 1990, 0.9% in 1980, and 0.7% in 1973. Between 1994 and 2004, this share declined from 1.5% to 1.1%. (See appendix table 4-4 Excel table. for time-series data on industry-funded R&D.)

  • Other sources of funds. In 2003, other sources of support accounted for 7% of academic R&D funding, a level that has stayed about the same since 1972. This category of funds includes grants for R&D from nonprofit organizations and voluntary health agencies and gifts from private individuals that are restricted by the donor to the conduct of research, as well as all other sources restricted to research purposes not included in the other categories.

Top of page.Top of page

Expenditures by Field and Funding Source

The distribution of academic R&D funds across S&E disciplines often is the result of numerous, sometimes unrelated, funding decisions rather than an overarching plan. Examining and documenting academic R&D investment patterns across disciplines enables interested parties to assess the balance in the academic R&D portfolio. The majority of academic R&D expenditures in 2003 went to the life sciences, which accounted for 59% of total, federal, and nonfederal academic R&D expenditures (appendix table 5-3 Excel table.). Within the life sciences , the medical sciences accounted for 32% of total academic R&D expenditures and the biological sciences for 18%.[6] The next largest block of total academic R&D expenditures was for engineering: 15% in 2003.

The distribution of federal and nonfederal funding of academic R&D in 2003 varied by field (appendix table 5-4 Excel table.). For example, the federal government funded about three-quarters of academic R&D expenditures in physics, the atmospheric sciences, and aeronautical/astronautical engineering but only about one-third in economics, political science, and the agricultural sciences.

The federally financed fraction of support for each of the broad S&E fields, except for computer sciences, was lower in 2003 than in 1980 (appendix table 5-4 Excel table.).[7] The most dramatic decline occurred in the social sciences, down from 54% in 1980 to 40% in 2003. The overall decline in federal share also holds for all the reported S&E subfields. However, most of the declines occurred in the 1980s, and many fields did not experience declining federal shares after that. In some fields, the federal share of support has increased since 1990.

Although total expenditures for academic R&D in constant 2000 dollars increased in every field between 1973 and 2003 (figure 5-6 figure.; appendix table 5-5 Excel table.), the R&D emphasis of the academic sector, as measured by its S&E field shares, changed during this period (figure 5-7 figure.). Relative shares of academic R&D:

  • Increased for the life sciences, engineering, and computer sciences
  • Remained roughly constant for mathematics
  • Declined for psychology; the earth, atmospheric, and ocean sciences; the physical sciences; and the social sciences

Although the proportion of the total academic R&D funds going to the life sciences increased by about 6 percentage points between 1973 and 2003, from 53% to 59% of academic R&D, the medical sciences' share increased by 10 percentage points during this period, from 22% to 32%, and the shares for the agricultural sciences and biological sciences both declined (appendix table 5-5 Excel table.). The largest declines in the proportion of total academic R&D funds were in the social sciences and physical sciences, which declined by about 4 and 3 percentage points, respectively. When this analysis was limited to the period 1993–2003, similar trends in share changes were seen, with two exceptions: the engineering share declined by almost 1 percentage point, and the psychology share increased slightly.

Top of page.Top of page

Federal Support of Academic R&D

The federal government continues to provide the majority of the funding for academic R&D. Its overall contribution is the combined result of discrete funding decisions for several key R&D-supporting agencies with differing missions. Most of the funding provided by the federal government to academia reflects decisions arrived at through a competitive peer review process. Some of the funds are from long-established programs, such as those of the U.S. Department of Agriculture (USDA), that support academic research through formula funding rather than peer review, and other funds are the result of appropriations that Congress directs federal agencies to award to projects that involve specific institutions. These latter funds are known as congressional earmarks. (See sidebar, "A Brief Look at Congressional Earmarking.") Examining and documenting the funding patterns of the key funding agencies is key to understanding both their roles and that of the federal government overall.

Top Agency Supporters

Six agencies are responsible for most of the federal obligations for academic R&D, providing an estimated 96% of such obligations in FY 2005 (appendix table 5-6 Excel table.). NIH provided an estimated 66% of total federal financing of academic R&D in 2005. An additional 13% was provided by NSF; 7% by the Department of Defense (DOD); 5% by the National Aeronautics and Space Administration (NASA); 3% by the Department of Energy (DOE); and 2% by the USDA.[8] Federal obligations for academic research (i.e., without the development component) are concentrated similarly to those for R&D (appendix table 5-7 Excel table.). Some differences exist, however, because some agencies place greater emphasis on development (e.g., DOD), whereas others place greater emphasis on research (e.g., NIH).

Between 1990 and 2005, NIH's funding of academic R&D increased most rapidly, with an estimated average annual growth rate of 6.1% per year in constant 2000 dollars, increasing its share of federal funding from just above 50% to an estimated 66%. NSF and NASA experienced the next highest annual rates of growth: 3.9% and 3.6%, respectively.

Agency Support by Field

Federal agencies emphasize different S&E fields in their funding of academic research. Several agencies concentrate their funding in one field (e.g., the Department of Health and Human Services [HHS] and USDA in the life sciences and DOE in the physical sciences), whereas NSF, NASA, and DOD have more diversified funding patterns (figure 5-8 figure.; appendix table 5-8 Excel table.). Even though an agency may place a large share of its funds in one field, it may not be a leading contributor to that field, particularly if it does not spend much on academic research (figure 5-9 figure.).

In FY 2003, NSF was the lead federal funding agency for academic research in the physical sciences (31% of total funding); mathematics (69%); the computer sciences (66%); and the earth, atmospheric, and ocean sciences (43%) (appendix table 5-9 Excel table.). DOD and NSF were the lead funding agencies in engineering (37% and 35%, respectively). HHS was the lead funding agency in the life sciences (90%), psychology (96%), and the social sciences (46%). Within the S&E subfields, other agencies took the leading role: DOE in physics (46%), the USDA in the agricultural sciences (99%), and NASA in astronomy (78%) and astronautical engineering (73%). If the analysis is confined to basic academic research, which constituted 62% of federal obligations for academic research in 2003, the lead funding agencies by field differ slightly (table 5-1 table.).

Top of page.Top of page

An Institutional Look at Academic R&D

The previous sections examined R&D for the entire academic sector. This section looks at some of the differences across institution types.

Funding for Public and Private Universities and Colleges

Although public and private universities rely on the same funding sources for their academic R&D, the relative importance of those sources differs substantially for these two types of institutions (figure 5-10 figure.; appendix table 5-10 Excel table.). For all public academic institutions combined, about 9% of R&D funding in 2003, the most recent year for which data are available, came from state and local funds; about 23% from institutional funds; and about 56% from the federal government. Private academic institutions received a much smaller portion of their funds from state and local governments (2%) and institutional sources (10%), and a much larger share from the federal government (74%). The difference in the role of institutional funds at public institutions may largely reflect the substantial amounts of general-purpose state and local government funds that public institutions receive and can decide to use for R&D (although data on such breakdowns are not collected).[9] (For a more detailed discussion of the composition of institutional funds for public and private academic institutions, see sidebar "The Composition of Institutional Academic R&D Funds.")

Both public and private institutions received approximately 5% of their R&D support from industry in 2003. The industry share of support for both public and private institutions decreased between 1993 and 2003, whereas both the federal and institutional shares of support increased.

Distribution of R&D Funds Across Academic Institutions

The distribution of R&D funds across academic institutions has been and continues to be a matter of interest both to those concerned with the academic R&D enterprise and those concerned with local and regional economic development. Most academic R&D is now, and has been historically, concentrated in relatively few of the 3,600 U.S. institutions of higher education.[10] If institutions are ranked by their 2003 R&D expenditures, the top 200 institutions account for about 95% of R&D expenditures that year. (See appendix table 5-11 Excel table. for a more detailed breakdown of the distribution among the top 100 institutions.)

The historic concentration of academic R&D funds diminished slightly between the mid-1980s and mid-1990s but has remained relatively steady since then (figure 5-12 figure.). In 1983, the top 10 institutions received about 20% of the nation's total academic R&D expenditures, compared with 17% in 2003. There was almost no change in the shares of the group of institutions ranked 11–20 and 21–100 during this period. Consequently, the decline in the top 20 institutions' share was offset by an increase in the share of those institutions outside the top 100. This group's share increased from 17% to 20% of total academic R&D funds, signifying a broadening of the base of institutional performers.

It should be noted that the composition of the universities in any particular group is not necessarily the same over time, because mobility occurs within groups. For example, only 5 of the top 10 institutions in 1983 were still in the top 10 in 2003. The discussion later in this chapter in "Spreading Institutional Base of Federally Funded Academic R&D" points to an increasing number of academic institutions receiving federal support for their R&D activities between 1972 and 2002. Many of the newer institutions receiving support are not the traditional Carnegie research and doctorate-granting institutions.

One program with the objective of improving the geographical distribution of federal academic R&D funds is the Experimental Program to Stimulate Competitive Research (EPSCoR). Several federal agencies have established EPSCoR or EPSCoR-like programs. EPSCoR attempts to increase the R&D competitiveness of eligible states through developing and using the science and technology resources resident in a state's major research universities. Eligibility for EPSCoR participation is limited to jurisdictions that have historically received lesser amounts of federal R&D funding and have demonstrated a commitment to develop their research bases and improve the quality of the S&E research conducted at their universities and colleges.

Changes in R&D Expenditures Across Academic Institutions

As academic R&D expenditures grew between 1997 and 2003, more institutions expanded their R&D activities. In FY 2003, as in the 6 preceding years, a greater number of institutions reported increased R&D expenditures than reported decreased R&D expenditures (figure 5-13 figure.). In fact, an examination of the ratio of the number of institutions increasing their expenditures from one year to the next to the number that did not increase their expenditures shows a fairly steady rise in this ratio during this period (figure 5-13 figure.). In FY 1997, approximately 1.4 institutions reported increased expenditures over FY 1996 for each institution that reported either unchanged or decreased R&D expenditures. In FY 2003, 2.9 institutions, more than twice as many as in 1997, increased their R&D expenditures for each institution that did not.[11]

Spreading Institutional Base of Federally Funded Academic R&D

The number of academic institutions receiving federal support for their R&D activities increased fairly steadily between 1972 and 1994, when it reached a peak of 907 institutions. Between 1995 and 2002, the number of institutions receiving federal support fluctuated between 791 and 901 (figure 5-14 figure.).[12] These fluctuations almost exclusively affected institutions of higher education with Carnegie classifications of comprehensive; liberal arts; 2-year community, junior, and technical; and professional and other specialized schools. The number of such institutions receiving federal support more than doubled between 1973 and 1994, rising from 315 to 680. It then dropped to 563 by 1999 before beginning to rise again in the past several years (appendix table 5-12 Excel table.). These institutions' share of federal support also increased between 1972 and 2002, from 9% to just above 14%. The number of Carnegie research and doctorate-granting institutions receiving federal support remained relatively constant during this period.

Top of page.Top of page

Academic R&D Equipment

Research equipment is an integral component of the academic R&D enterprise. This section examines expenditures on research equipment, the federal role in funding these expenditures, and the relation of equipment expenditures to overall R&D expenditures.


In 2003, about $1.8 billion in current funds was spent for academic research equipment. About 81% of these expenditures were concentrated in three fields: the life sciences (45%), engineering (20%), and the physical sciences (16%) (figure 5-15 figure.; appendix table 5-13 Excel table.).

Current fund expenditures for academic research equipment grew at an average annual rate of 4.6% (in constant 2000 dollars) between 1983 and 2003. However, recent annual growth (since 2000) was almost 6%, compared with less than 1% during the 1990s. The growth patterns in S&E fields varied during this period. For example, equipment expenditures for engineering (5.5%) and the biological sciences (5.2%) grew more rapidly during the 1983–2003 period than did those for the social sciences (0.7%) and agricultural sciences (0.5%).

Federal Funding

Federal funds for research equipment are generally received either as part of research grants or as separate equipment grants, depending on the funding policies of the particular federal agencies involved. The importance of federal funding for research equipment varies by field. In 2003, the social sciences received about 45% of their research equipment funds from the federal government; in contrast, federal support accounted for more than 70% of equipment funding in the physical sciences, mathematics, the computer sciences, psychology, and the earth, atmospheric, and ocean sciences (appendix table 5-14 Excel table.). The share of research equipment expenditures funded by the federal government declined from 62% to 55% between 1983 and 2001, but thereafter rose to 63% in 2003. This overall pattern masks different trends in individual S&E fields.

R&D Equipment Intensity

R&D equipment intensity is the percentage of total annual R&D expenditures from current funds devoted to research equipment. This proportion has been declining fairly steadily since reaching its peak in 1986 (7%). By 2003, it had declined to 5% (appendix table 5-15 Excel table.). R&D equipment intensity varies across S&E fields, tending to be higher in the physical sciences (about 9% in 2003) and lower in the social sciences (1%) and psychology (3%).

Several years ago, Congress requested that an NSF National Survey of Academic Research Instrumentation, last conducted in 1994, be reinstated to determine the extent to which a lack of equipment and instrumentation prevents the academic research community from undertaking cutting-edge, world-class science. NSF is investigating the feasibility of obtaining such information.

Top of page.Top of page

Academic R&D Infrastructure

The physical infrastructure of academic institutions is critical to supporting R&D activities. Traditional indicators of the status of the research infrastructure are the amount of research space currently available and the amount of investment in future facilities. Furthermore, the quality of research space is a key factor in the types of research that can be undertaken.

In addition to the traditional "bricks and mortar" research infrastructure, "cyberinfrastructure" is playing an increasingly important role in the conduct of S&E research. Technological advances are significantly changing S&E research methods. In some cases, advanced technology is already changing the role of traditional bricks and mortar facilities. According to the NSF Advisory Panel on Cyberinfrastructure, these advances are not simply changing the conduct of science but are revolutionizing it (NSF 2003). The panel defined cyberinfrastructure as the "infrastructure based upon distributed computer, information and communication technology" (NSF 2003, p 1.2). The report discusses the current and potential future importance of cyberinfrastructure, stating that "digital computation, data, information and networks are now being used to replace and extend traditional efforts in science and engineering research" (NSF 2003, p 1.1).

At this time, how the relationship between cyberinfrastructure and traditional bricks and mortar infrastructure will play out is unknown. Access to high-quality research facilities may become available to researchers located at institutions where traditional research space has not been available. Some institutions now indicate they need less physical space as they begin to conduct research not in their own laboratories or research facilities but through networking and/or high-performance computing, communicating with research facilities thousands of miles away or accessing very large databases generated by advanced data collection technologies.

Bricks and Mortar

Research Space. Research-performing colleges and universities[13] continued to expand their stock of research space in FY 2003 with the largest increase in total research space since 1988. By the end of FY 2003, total research space increased 11% from FY 2001 to approximately 173 million net assignable square feet (NASF)[14] (table 5-2 table.). This increase was substantially greater than any previous 2-year increase since FY 1988 and continued a trend of increases in the amount of academic research space. During this 15-year period, the amount of research space increased biennially at a rate of at least 4%.

Except for the agricultural sciences, all S&E fields experienced increases in research space between FY 2001 and FY 2003. Two fields, the computer sciences and mathematics, experienced the largest increases (but their total space was the smallest among all S&E fields). Social science space increased by 27%. Growth in medical sciences research space, 26%, was the fourth highest, reaching 35 million NASF. Only the biological sciences had more research space (36 million NASF). These two fields, combined with engineering, accounted for 57% of all research space at the end of FY 2003.

Little change occurred in the distribution of research space across S&E fields during this 15-year period. The largest increase in the share of total research space occurred in the medical sciences. However, this share only changed 3 percentage points between 1988 and 2003. The engineering share of total research space increased 2 percentage points. The largest decrease, only 2 percentage points, occurred in the physical sciences.

Construction of Research Space. Universities invested $7.6 billion in FY 2002–03 in the construction of 16 million NASF of research space (appendix tables 5-16 Excel table. and 5-17 Excel table.).[15] Almost half of all universities began construction projects (NSF/SRS forthcoming).

Although universities began construction of research space in all S&E fields in FY 2002–03, the largest share of space under construction (56%) was for research in the medical sciences and biological sciences (appendix table 5-17 Excel table.). Fifty-six percent of research space construction started in FY 2002 or FY 2003 is to be used for research in these two fields. If engineering research space is included, these three fields account for about 70% of the new construction started. Even if some newly constructed space replaces existing space, the share of newly constructed space in the medical sciences (31%) was substantially greater than that of any other field, and therefore would not likely change the overall field distribution. The biological sciences, which had the second largest share, accounted for 25% of newly constructed research space.

If the universities were able to follow through on planned construction for FY 2004 and FY 2005, the medical sciences and biological sciences will likely continue to dominate the share of total research space (appendix table 5-17 Excel table.). Universities plan to construct 19 million NASF of research space during this period at an estimated cost of $9.1 billion. The biological sciences and medical sciences will account for 53% of the planned space and 61% of estimated construction costs (appendix table 5-16 Excel table. and 5-17 Excel table.).

Funds for Construction. Institutions use one or more sources to fund their capital projects, including the federal government, state or local governments, and the institutions' own funds.[16] The federal government's share of total construction funding has been declining and reached its smallest proportion (5%) since 1986–87 in FY 2002–03 (figure 5-16 figure.; appendix table 5-18 Excel table.). During the same period, the institutional share of construction funds increased overall and reached its highest share, 63%, in FY 2002–03.

Over time, the share of institutional funds universities and colleges have allocated for repair/renovation of research space has been consistently greater than the share they have allocated for construction. However, even for repair/renovation, the institutional share of total funds reached its highest level since 1988, 71%, in FY 2002–03 (NSF/SRS forthcoming).

Unmet Needs. Determining the capital infrastructure needs of universities has at least several dimensions. Two indicators of need are the dollar value of deferred projects and the quality of existing space.

Deferred projects are projects in a university's institutional plans that are needed for current program commitments but that have not yet been funded and therefore are not scheduled to begin. Institutions reported approximately $8.4 billion in deferred construction projects in FY 2003 (appendix table 5-16 Excel table.). More than half of this deferred construction was in the biological sciences and medical sciences.

There are no objective criteria to determine how much of a field's research actually requires state-of-the-art space. However, space rated as needing replacement can be seen as an indicator of need. In FY 2003, institutions rated 30% of their existing space as state of the art and 79% as either state of the art or suitable for most levels of research and reported that 5% should be discontinued as research space within the next 2 years (appendix table 5-19 Excel table.).[17] The amount of space needing replacement varied little by field, ranging from 7% in the social sciences and earth, atmospheric, and ocean sciences to 2% in mathematics.

Perhaps not surprisingly, the computer sciences, the field that had the greatest amount of relative growth in research space between FY 2001 and FY 2003, rated the largest percentage of its space as state of the art. The medical sciences, another field that experienced a large increase in the amount of new space during this period, had the second highest amount of space rated as state of the art.

Cyberinfrastructure: Networking

Networking resources are a key component of cyberinfrastructure. Networks allow users and researchers to communicate and transfer data both within a specific institution's boundaries and with others around the world. At many institutions, the same networks are used for multiple academic functions such as instruction, research, and administration.[18]

All academic institutions today have network connections to the commodity Internet, or Internet1, the network commonly known as the Internet. Although Internet connections are used for many purposes (e.g., e-mail, buying books from the campus bookstore), conducting research can require higher capabilities of network connections than other activities.

There are numerous indicators of network capability. One common indicator is bandwidth, or speed. A network's bandwidth can affect the amount and type of research activity accomplished through the network. The faster the network speed, the more capable the network is in handling both large amounts of data and communication traffic and more demanding or sophisticated communications. Whereas a slow network connection might well be able to transmit scientific articles, transmitting scientific instruments located thousands of miles away or accessing large databases demands (among other requirements) high bandwidth or fast speed.

Desktop Connection Speed. The speed of the desktop computer's connection to the campus network will likely differ from that of the campus network's connection to the Internet. Generally, researchers access the Internet from their desktop computers. Therefore, the speed of the desktop connection to the institution's campus network is one useful indicator of an institution's network capability. Desktop connection speeds will vary across an institution. Almost 75% of academic institutions reported the highest operating speed of the majority of their desktop connections (ports) as 100 megabits/second in FY 2003, and 1% reported it as 1 gigabit/second (NSF/SRS forthcoming).

In FY 2003, 76% of non-doctorate-granting institutions had the majority of their desktop connections at 100 megabits/second or faster, compared with 71% of doctorate-granting institutions (appendix table 5-20 Excel table.). However, only 19% of non-doctorate-granting institutions estimated their highest speed as 1 gigabit/second, compared with 46% of doctorate-granting institutions. Most institutions planned to obtain faster connection speeds in FY 2004, and 52% of all institutions estimated that their highest connection speed would be 1 gigabit/second at the end of FY 2004.

Internet Connection Speed. Another critical point is the connection between the institution's campus network and the Internet. At the end of FY 2003, most universities had multiple connections to the Internet at a variety of speeds. The majority (49%) were at the lowest speed, 1.5 megabits/second (i.e., T1 or DS1 lines). The second largest share of connections (22%) was at the next lowest speed, 45 megabits/second (i.e., T3 or DS3 lines). Together, these two speeds accounted for 71% of connections (figure 5-17 figure.; appendix table 5-21 Excel table.). However, at least 6% of connections were at 1 gigabit/second or faster. Doctorate-granting institutions had the largest number of high-speed connections. Although the greatest number of connections was at 1.5 or 45 megabits/second, the highest connection speed was 155 megabits/second or faster at 45% of all institutions and 1 gigabit/second or faster at 12% (table 5-3 table.).

Overall, institutions did not anticipate a large increase in the total number of Internet connections between FY 2003 and FY 2004. However, institutional plans overall called for fewer connections at slower speeds and a larger number at faster speeds, estimating a 4% increase in the number of connections at speeds of 1 gigabit or higher by the end of FY 2004. Both doctorate-granting and non-doctorate-granting institutions anticipated increases in connection speeds. In fact, non-doctorate-granting institutions estimated fewer total connections overall but more at higher speeds. Furthermore, both doctorate- and non-doctorate-granting institutions expected to increase the speed of their highest speed connections by the end of FY 2004.

Wireless and High-Performance Network Connections. In addition to their hardwire network connections, many universities have wireless Internet connections as well as connections to advanced or high-performance networks. High-performance networks are not only faster than the Internet but also have other characteristics important to conducting research. At the end of FY 2003, 65% of academic institutions had connections to Abilene (often called Internet2 ) (NSF/SRS forthcoming), a high-performance network dedicated to research led by a consortium of universities, governments, and private industry. A substantially larger proportion (79%) of doctorate-granting institutions had Abilene connections as compared with non-doctorate-granting institutions (28%).

Although wireless networking is used less frequently for research, universities are moving toward greater institutional coverage by wireless networking. At the end of FY 2003, 67% of institutions had 20% or less of their building areas covered by wireless network connections (NSF/SRS forth-coming). However, less than 30% estimated that their coverage would be 20% or less by the end of FY 2004.

Top of page.Top of page

Top of page.Top of page

Top of page.Top of page

Top of page.Top of page


[1] Federally funded research and development centers (FFRDCs) associated with universities are tallied separately and are examined in greater detail in chapter 4. FFRDCs and other national laboratories (including federal intramural laboratories) also play an important role in academic research and education, providing research opportunities for both students and faculty at academic institutions.

[2] For this discussion, an academic institution is generally defined as an institution that has a doctoral program in science or engineering, is a historically black college or university that expends any amount of separately budgeted R&D in S&E, or is some other institution that spends at least $150,000 for separately budgeted R&D in S&E.

[3] Despite this delineation, the term "R&D" (rather than just "research") is primarily used throughout this discussion because data collected on academic R&D do not always differentiate between research and development. Moreover, it is often difficult to make clear distinctions among basic research, applied research, and development.

[4] The academic R&D funding reported here includes only separately budgeted R&D and institutions' estimates of unreimbursed indirect costs associated with externally funded R&D projects, including mandatory and voluntary cost sharing.

[5] This follows a standard of reporting that assigns funds to the entity that determines how they are to be used rather than to the one that necessarily disburses the funds.

[6] The medical sciences include fields such as pharmacy, veterinary medicine, anesthesiology, and pediatrics. The biological sciences include fields such as microbiology, genetics, biometrics, and ecology. These distinctions may be blurred at times because boundaries between fields often are not well defined.

[7] In this chapter, the broad S&E fields refer to the physical sciences; mathematics; computer sciences; the earth, atmospheric, and ocean sciences; the life sciences; psychology; the social sciences; other sciences (those not elsewhere classified); and engineering. The more disaggregated S&E fields are referred to as "subfields."

[8] The recent creation of the Department of Homeland Security (DHS) should have major implications for the future distribution of federal R&D funds, including federal academic R&D support, among the major R&D funding agencies. DHS's Directorate of Science and Technology is tasked with researching and organizing the scientific, engineering, and technological resources of the United States and leveraging these existing resources into technological tools to help protect the homeland. Universities, the private sector, and the federal laboratories are expected to be important DHS partners in this endeavor.

[9] Another hypothesis is that some of the difference may be due to many public universities not having the incentive to negotiate full recovery of indirect costs of research because the funds are frequently captured by state governments.

[10] The Carnegie Foundation for the Advancement of Teaching classified about 3,600 degree-granting institutions as higher education institutions in 1994. See chapter 2 sidebar, "Carnegie Classification of Academic Institutions," for a brief description of the Carnegie categories. These higher education institutions include 4-year colleges and universities, 2-year community and junior colleges, and specialized schools such as medical and law schools. Not included in this classification scheme are more than 7,000 other postsecondary institutions such as secretarial schools and auto repair schools.

[11] Inflation averaged less than 2% over the period discussed. For an analysis of this trend among the top 200 institutions with the largest R&D expenditures and for a comparison of institutions that increased their R&D expenditures by more than 3% over the preceding year with those that did not, see NSF/SRS 2004.

[12] Although the number of institutions receiving federal R&D support between 1973 and 1994 increased overall, a rather large decline occurred in the early 1980s, most likely due to the fall in federal R&D funding for the social sciences during that period.

[13] Research-performing academic institutions are defined as colleges and universities that grant degrees in science or engineering and expend at least $1 million in R&D funds. Each institution's R&D expenditure is determined through the NSF Survey of Research and Development Expenditures at Universities and Colleges.

[14] Research space here is defined as the space used for sponsored research and development activities at academic institutions that is separately budgeted and accounted for. Research space is measured in NASF, the sum of all areas on all floors of a building assigned to, or available to be assigned to, an occupant for a specific use, such as research or instruction. NASF is measured from the inside faces of walls. Multipurpose space that is at least partially used for research is prorated to reflect the proportion of time and use devoted to research.

[15] Some of this space will likely replace existing space and therefore will not be a net addition to existing stock.

[16] Institutional funds may include operating funds, endowments, tax-exempt bonds and other debt financing, indirect costs recovered from federal grants/contracts, and private donations.

[17] Institutions rated space using four categories: (1) space in superior condition that is suitable for the most scientifically competitive research in the field over the next 2 years; (2) space in satisfactory condition that is suitable for continued use over the next 2 years for most levels of research in the field but that may require minor repairs or renovation; (3) space that requires renovation and that will no longer be suitable for current research without undergoing major renovation with the next 2 years; and (4) space that requires replacement and that should stop being used for current research within the next 2 years.

[18] The "bricks and mortar" section of the Survey of Science and Engineering Research Facilities asked institutions to report on their research space only. The reported figures therefore do not include space used for other purposes such as instruction or administration. In the networking and computing section of the survey, however, respondents were asked to identify all of their computing and networking resources, regardless of whether these resources were used for research.

National Science Board.