Fateful Choices: The Future of the U.S. Academic Research Enterprise (1992)

Achieving such an optimistic and challenging vision for the future will require choosing among a range of options, each of which will have profound implications for the research enterprise in the United States. All those with a stake in the research enterprise—investigators, university administrators, funding agencies and the larger public—need to understand the implications of these difficult choices. Whatever decisions are made will determine the capacity and character of the U.S. research enterprise of the next century.

In the view of the working group, two processes need to begin simultaneously. First, universities and research sponsors need to take immediate, concrete steps to "put their houses in order." The working group believes that decisionmakers at the highest levels need to set overall national research priorities with input from the university and research communities. Universities and funding agencies need to clarify their respective responsibilities for funding university-based research and their need to update their organizational and management strategies. Universities, funding agencies, and professional societies need to adapt to shifting demographics and the changing value systems of many young investigators. Universities need to improve the quality of science and engineering education, especially at the undergraduate level.

Second, all those with a stake in academic research—including the political, corporate, and public interest sectors—should begin to think strategically about options for the future of the research enterprise. To start this process, the working group describes a heuristic framework for considering future options. Central to this framework is a better understanding of the large-scale forces that affect the enterprise: the pace and nature of research, the economy, politics and international events. Based on a consideration of possible interactions of these factors, the working group sets forth several "scenarios" depicting the future size and structure of the enterprise. The working group then identifies key policies or programs, specific to each scenario, that would be required to maintain the quality and productivity of the enterprise. The working group believes that this heuristic framework brings new ideas to the attention of the research community. This is in keeping with the working group's charge to concentrate on issues affecting the enterprise over the long term.

In an earlier discussion paper,¹ the working group identified several trends that it believes to be symptomatic of underlying long-term changes within the enterprise. While the enterprise is not now in a crisis, the working group believes that these changes require urgent attention by all participants in the enterprise—investigators, university administrators, research sponsors.

The near-term decisions outlined below will not come easily. Without them, however, harmful tensions in the enterprise will persist and, equally important, the public support needed to achieve the optimistic vision set out in this paper will not be forthcoming.

It will be necessary in the coming decades to set national priorities for the support and conduct of science and engineering research. If priorities are not established, there will be increasing confusion about and less than optimal investments in frontier research and in research infrastructure of vital importance to the nation.

Growth In The Number Of High-Quality Research Opportunities Is Outpacing Increases In Research Funding. The 1950s and 1960s saw an enormous expansion of both the number of university research personnel and the financial resources for the support of university-based research. Following a general steady-state in funding and personnel during the 1970s, expansion and diversification of university-based research resumed during the past decade. All evidence points to a continuation of the trends of the last 10 years: increased numbers of institutions and people involved in research; greater participation of industry, states, and universities in the support of science and engineering; and enhanced university research capacity designed to boost local, regional, and national economic development.

The growth of the past decade has brought many benefits. The enlarged base of support for research has resulted in major scientific and technological accomplishments, and it has enhanced the nation's research and educational capacity. Serious questions are being raised, however, about the nation's continued willingness to support a growing research enterprise. The enterprise itself is experiencing a number of detrimental tensions that threaten its quality, integrity, and ability to respond to new opportunities and challenges.

At the federal level, demand for financial support for research activities is outpacing the recent increases in research funding by the federal agencies. As a consequence, although the absolute number of federally supported

researchers is higher now than at any point in history, competition for research funding is increasing—for both established and younger investigators. This situation is exacerbated by the rapidly increasing costs of doing research—paying the salaries of scientists and engineers, building and maintaining new research facilities, and purchasing new research instruments— and by the simultaneous emergence of several large-scale research projects within federal agency budgets. As a result, tension is growing among investigators, their parent institutions and the agencies that fund them. The high level of frustration expressed by many scientists is but one highly visible sign of the stress these tensions are creating. ²

The trends of the last 10 years have also put the federally supported peer-review system, used to allocate funds among competing research proposals, under growing strain. Many investigators are having to spend increasing amounts of time as reviewers, leaving less time for their own research. There is also growing concern that greater numbers of conventional proposals are being submitted because investigators fear that innovative, but unorthodox, research projects will not be funded.

Another source of tension is the balance between funding greater numbers of science projects and rebuilding the research infrastructure. Many research facilities are in need of repair, renovation, or replacement. Many laboratories lack state-of-the-art scientific equipment. Fiscal belt-tightening, if continued at the state and federal levels, will undoubtedly compound this problem.

Government Leaders Must Set Broad National Priorities For Research In Consultation With The Individual Scientific And Engineering Disciplines, The Larger Scientific Community, Academic Institutions, And Industry. ³ Such a broad-based process for setting priorities needs to address not only the relative importance of various research projects and programs, but also the funding needs for facilities, instrumentation, education, and training.

Priority-setting at the level of the individual research proposal has worked quite successfully. The research community relies primarily on two criteria to allocate funding: research excellence and impact on the knowledge base, or the potential of a research proposal to expand the horizons of human understanding. The working group strongly believes these two criteria must continue to be the primary basis for making funding decisions in research. Originality, or unique, non-traditional approaches to addressing research questions, will also have to be explicitly considered. This approach will be necessary to address concerns that more "conventional" research proposals are being submitted as a response to increased competition for funding.

Other funding criteria will also play a role in funding decisions. These are: relevance, or having eventual application to human needs; economic promise, or the potential for accelerating growth in the gross national product (GNP); and equity, or the degree to which funding agencies should consider the

geographic location, race, and sex of grant applicants. The challenge for research institutions and agencies that fund research will be to decide on the appropriate mix of these criteria, and to make them explicit. Different criteria will be called for, depending on the goals of those supporting the research.

Within research fields, with a few notable exceptions, individual scientific and engineering disciplines are only now attempting to agree on research priorities and to set forth a strategic vision. To be of use to decision-makers who allocate federal research dollars, such priority-setting will need to be adopted more broadly and updated on a regular basis.

Processes for setting national priorities across research fields or within the enterprise as a whole work less effectively. Within the federal government, there is no coherent, over-arching strategy for research. This lack of coordination at the highest levels of decisionmaking is a serious problem. ^4,5 Although research priority-setting occurs de facto during the federal budget process, with input from the White House, the Office of Management and Budget, the Office of Science and Technology Policy, Congress, and the federal agencies that fund research, priority-setting across agencies occurs only rarely.

This lack of coordination goes beyond the federal level. The nation's research community has conducted little debate about priorities, and there has been resistance to priority-setting efforts. State government and federal agency officials have only recently begun an informal dialogue about those issues. Few academic institutions have engaged in any kind of long-term strategic planning necessary to set priorities for conducting and supporting research.

Because priority-setting entails important trade-offs, the long-term implications of such decisions must be taken into consideration. The need for investing in the research infrastructure, for example, will have to be weighed against funding a larger number of individual research grants. The importance of investing in large, expensive "mega-projects" will need to be balanced against the desirability of funding a number of smaller research projects.

The financial resource base for academic research is becoming increasingly complex. Through their support for research infrastructure and the indirect costs of research, non-federal sectors now play a much more significant role in setting the research agenda in the United States than they have in the past three decades. Ensuring that the essential needs of the enterprise are met within this changing environment requires more explicit focus on the division of federal, state, and university responsibilities in funding academic research.

Over The Past Two Decades, While Federal Investments In Academic Research Have Increased, Non-Federal Funding Has Grown Even Faster. As a consequence, the percentage of total non-federal academic research funds rose from 31 percent in 1970 to 41 percent in 1990, its highest point since the 1950s. ⁶ (See Figure 1, page 14.) The rise of non-federal investments in research is most pronounced in public universities. The non-federal share of research funds for public universities rose to 48 percent in 1990, compared with a 29 percent federal share for private universities. (See Figure 2, page 15, and Figure 3, page 16.)

During the same 20-year period, the estimated share of university-generated funds devoted to research grew from 11 percent to 19 percent. The most significant trend in university funding during this time was the willingness of public universities—especially institutions aspiring to develop a stronger research base—to utilize their own resources to cover part of the indirect costs of externally sponsored research.

Although the overall share of academic research funds contributed by state governments held steady at 8 percent, several states greatly increased their spending for academic research. Industry also took a larger role, more than doubling its share of investment in academic research, from 2.8 percent of the total in 1970 to an estimated 7 percent in 1990.

Non-federal sectors now pay about 41 percent of the total for academic research equipment. (See Figure 4, page 17.) For academic science and engineering facilities, the non-federal share of direct funding has risen to 90 percent. (See Figure 5, page 18.)

Policymakers Need To Rethink The Current Division Of Federal-State And Federal-University Responsibilities For Higher Education In General And Research In Particular. This will require more effective ongoing communication, interaction, and coordination among research sponsors and academic institutions. ⁷

All sponsors of academic research are currently facing financial constraints. Federal appropriations are constrained by large budget deficits and public resistance to raising taxes. State governments—many of which are confronting their own budget problems—are closely evaluating their priorities, including the support of academic research. Foreign governments and foreign-based industries are potentially important sources of funding for U.S. academic research; however, as global economic competition intensifies, foreign investments could be constrained by government policies.

During the next decade, the ability of most universities to increase their own resources in support of research will be limited. It is unlikely that more funds could come from undergraduate student tuition or state monies appropriated for the educational mission. This is because most universities are facing steady student enrollments and flat or reduced state appropriations, and at both private and public universities, public pressure has slowed the pace of tuition increases. More universities are competing with each other

FIGURE 1

Academic R&D Expenditures by Source

Note: Data series within the figures are not overlapped; top line represents total. Financial data are expressed in 1988 constant dollars to reflect real long-term growth trends

Definition of Terms: Academic R&D expenditures include current-fund expenditures within higher education institutions for all research and development activities that are separately budgeted and accounted for. This includes both sponsored research activities (sponsored by federal and non-federal agencies and organizations) and university research separately budgeted under an Internal application of institutional funds; but excludes training. public service, demonstration projects, departmental research not separately budgeted, and FFRDCs. Federal funds include grants and contracts for academic R&D (including direct and reimbursed indirect costs) by agencies of the federal government. State/ Local funds include funds for academic R&D from state, county, municipal, or other local governments and their agencies, including funds for R&D at agricultural and other experiment stations. Industry funds includes all grants and contracts for academic R&D from profit-making organizations, whether engaged in production, distribution, research, service, or other activities Own Funds include institutional funds for separately budgeted research and development, cost-sharing, and under-recovery of indirect costs; they are derived from (1) general purpose state or local government appropriations. (2) general purpose grants from industry, foundations, and other outside sources, (3) tuition and fees, and (4) endowment income. Other sources include grants for academic R&D from non-profit foundations and voluntary health agencies, as well as individual gifts that are restricted by the donor to research.

Source: National Science Foundation, Division of Policy Research and Analysis. Database: CASPAR. Some of the data within this database are estimates, incorporated where there are discontinuities within data series or gaps m data collection. Primary data source: National Science Foundation, Division of Science Resources Studies, Survey of Scientific and Engineering Expenditures at Universities and Colleges.

Public Doctoral Institution R &D Expenditures by Source of Funds

 

Note: Data series within the figures are not overlapped; top line represents total. Financial data are expressed m 1988 constant dollars to reflect real long-term growth trends.

Definition of Terms: Public doctoral institutions are institutions thathave granted an average of 10 or more Ph.D. degrees per year m the natural sciences or engineering over the past two decades, and are under the control of—or affiliated with—federal, state, local, state and local, or state-related agencies; they include 116 institutions. R&D Expenditures include current-fund expenditures within doctoral institutions for all research and development activities that are separately budgeted and accounted for; excluding departmental research not separately budgeted and FFRDCs. Federal funds include grants and contracts for R&D (including direct and reimbursed indirect costs) by agencies of the federal government, excluding funds for FFRDCs. State/Local funds include funds for R&D from state, county, municipal, or other local governments and their agencies, including funds for R&D at agricultural and other experiment stations. Industry funds include all grants and contracts for R&D from profit-making organizations, whether engaged m production, distribution, research, service, or other activities. Own Funds include institutional funds for separately budgeted research and development, cost-sharing, and under-recovery of indirect costs. They are derived from (1) general purpose state or local government appropriations, (2) general purpose grants from industry, foundations, or other outside sources, (3) tuitionand fees, and (4) endowment income Other sources include grants for R&D from non-profit foundations and voluntary health agencies, as well as individual gifts that are restricted by the donor to research.

Source. National Science Foundation, Division of Policy Research and Analysis. Database: CASPAR. Some of the data within this database are estimates, incorporated where there are discontinuities within data series or gaps in data collection. Primary data source: NationalScience Foundation, Division of Science Resources Studies, Survey of Scientific and EngineeringExpenditures at Universities and Colleges.

FIGURE 3

Private Doctoral Institution R & D Expenditures by, Source of Funds

Note: Data series within the figures are not overlapped; top line represents total. Financial data are expressed in 1988 constant dollars to reflect real long-term growth trends.

Definition of Terms: Private doctoral institutions are institutions that have granted an average of 10 or more Ph D degrees per year in the natural sciences or engineering over the past two decades, and are under the control of—or affiliated with—non-profit, independent organizations with or without religious affiliation; they include 69 institutions. R&D expenditures include current-fund expenditures within doctoral institutions for all research and development activities that are separately budgeted and accounted for; excluding departmental research not separately budgeted and FFRDCs. Federal funds include grants and contracts for R&D (including direct and reimbursed indirect costs) by agencies of the federal government, excluding funds for FFRDCs. State/Local funds include funds for R&D from state, county, municipal, or other local governments and their agencies, including funds for R&D at agricultural and other experiment stations Industry funds include all grants and contracts for R&D from profit-making organizations, whether engaged in production, distribution, research, service, or other activities. Own Funds include institutional funds for separately budgeted research and development, cost-sharing, and under-recovery of indirect costs. They are derived from (1) general purpose state or local government appropriations, (2) general purpose grants from industry, foundations, or other outside sources, (3) tuition and fees, and (4) endowment income. Other sources include grants forR&D from non-profit foundations and voluntary health agencies, as well as individual gifts that are restricted by the donor to research.

Source: National Science Foundation, Division of Policy Research and Analysis. Database: CASPAR. Some of the data within this database are estimates, incorporated where there are discontinuities within data series or gaps in data collection Primary data source National Science Foundation, Division of Science Resources Studies, Survey of Scientific and EngineeringExpenditures at Universities and Colleges.

FIGURE 4

Expenditures for Academic Research Equipment by Source of Funds

Note: Data series within the figures are not overlapped, top line represents total. Financial data are expressed m 1988 constant dollars to reflect real long-term growth trends Definition of Terms: Research equipment expenditures include (1) reported expenditures of separately budgeted current-funds for the purchase of research equipment, and (2) estimated capital expenditures for fixed or built-in research equipment and furniture. Federal funds include expenditures for academic research equipment with monies from grants and contracts for academic R&D (including direct and reimbursed redirect costs) by agencies of the federal government; excludes expenditures for FFRDC facilities Other sources include state and local governments, the institution themselves, industry, and other non-profit organizations.

Source. National Science Foundation, Division of Policy Research and Analysis Database CASPAR. Some of the data within this database are estimates, incorporated where there are discontinuities within data series or gaps in data collection. Primary data source: National Science Foundation, Division of Science Resources Studies, Survey of Scientific and Engineering Expenditures at Universities and Colleges

and with other local and national organizations for the same sources of private philanthropy.

In response to these funding pressures, ''leveraging" arrangements and "cost-sharing" requirements have become common components of the research-support system. To the extent that leveraging increases the overall level of funds for research, it is beneficial for the entire academic research enterprise. If academic institutions are pressured to cost-share—both for direct project costs and through contributions to indirect costs—and by doing so must reallocate resources from instructional programs to research, the research enterprise is imperiled in the long-run.

FIGURE 5

Expenditures for Academic Science and Engineering Facilities by Source of Funds

Note: Data series within the figures are not overlapped; top line represents total. Financial data are expressed in 1988 constant dollars to reflect real long-term growth trends. Definition of Terms Academic science and engineering facilities expenditures include capital expenditures for research and instructional facilities, including fixed or built-in equipment, some movable equipment and movable furnishings such as desks, and facilities constructed to house scientific apparatus. Federal funds include expenditures for academic science and engineering facilities with moneys from federal agency contracts in grants. Other sources include state and local governments, the institutions themselves, industry, and other non-profit organizations.

Source: National Science Foundation, Division of Policy Research and Analysis. Database: CASPAR. Some of the data within this database are estimates. incorporated where there are discontinuities within data series or gaps m data collection. Primary data source: NationalScience Foundation, Division of Science Resources Studies, Survey of Scientific and Engineering Expenditures at Universities and Colleges.

Aspiring research universities may be much more able to put together funding packages for cost-sharing, targeted to selected research programs, than more well-established institutions whose resources already are strained by significant, ongoing commitments to a wider range of research fields. Over the long term, cost-sharing requirements could have profound implications for the allocation of national resources among academic research institutions. The potential effects of cost-sharing requirements on the structure of the research enterprise should be explicitly addressed by research institutions and the government agencies and private organizations that support them.

States traditionally have assumed responsibility for funding a major share of the instructional mission of public universities and colleges. In a few re-

search areas, such as agriculture, they also have provided much of the support for research infrastructure and research projects. The federal government, following World War II, assumed major responsibility for supporting research projects and equipment.

In recent years, there has been increasing overlap in the research funding roles of the federal government and the individual states. State governments increasingly are supporting research projects and programs. Some senior federal officials have suggested that the federal government should be involved more directly in the support of higher education in general. At the very least, the states must be full partners in any national dialogue intended to clarify future funding responsibilities for academic research.

The growing complexity and size of the academic research enterprise require new and innovative organizational strategies. They will be necessary if academic research institutions are to balance their commitments to teaching, research, and public service.

The organization and management of universities have become increasingly complex. The changing nature of many areas of research requires that the departmental structure of universities adapt. For example, multi-disciplinary research teams have become more common as more complex topics emerge in science and technology. Larger-size research teams are increasingly evident, as well. The impact of advances in computers and telecommunications on the conduct and organization for research is yet to be fully felt, but it is sure to require new ways of organizing and managing university-based research.

The average age of university faculty is rising because of the rapid surge in student enrollments and faculty hiring 20 years ago, which was followed by a general steady-state in enrollments and a decline in hiring of younger faculty. Conclusive data are lacking to predict the effects of increased faculty retirements over the next decade. It appears, however, that the pace of faculty retirements may very well accelerate. If retirements do increase, the demand for research faculty could outstrip the supply of available research personnel. This could cause a shift of some presently employed, non-tenure track personnel into faculty positions, or it may result in increased recruitment of investigators from industrial laboratories or from foreign universities.

Universities and funding agencies need to consider new approaches to the organization and management of individual research institutions, and also of the larger U.S. research system. During several workshops held by the working group, the following approaches were raised and discussed.

Good Stewardship. The public is increasing its scrutiny of the process of scientific inquiry and of the stewardship of the taxpayer's mon-

ey by academic institutions.⁸ While this public attention is valuable, it is a further source of tension within the system. Better management of university resources will serve the interests of both universities and the public. This should include a systematic process for self-evaluation, constant feedback, and an emphasis on improving quality. Better and more visible university-based oversight practices, particularly those designed to reduce instances of research fraud and the waste of resources in academic research, will substantially strengthen the enterprise.

Decisionmaking Practices. A key challenge for universities is deciding which decisions should be made by the central administration and which should be made in a more decentralized fashion by the faculty. University administrators and faculty both bring valuable assets to the table. Administrators are skilled in balancing a variety of institutional goals. Faculty offer the necessary quality of creativity in their technical disciplines.

While decentralized, faculty-based decisionmaking is preferable as a general strategy, other approaches may be necessary under special circumstances. For example, if the range of funding mechanisms and sponsors becomes overly complex, or if funding itself levels off, a shift to more centralized decisionmaking across the institution may be desirable. In such cases, a major challenge will be to provide faculty with information about, and encourage their participation in, the institution's affairs, particularly in matters of research priority-setting.

In the years ahead, multi-disciplinary research teams will become more common as more complex topics emerge in science and technology. Universities will need to hire new research staff or establish new multi-departmental centers to provide support for these teams. New approaches to organization also will be required as collaborative research arrangements among industry, government, laboratories, and academia become more frequent.

Large-scale, multi-disciplinary research programs may require a more centralized, hierarchical type of management. This will allow strategic decisions and inter-departmental planning to be accomplished more effectively. Further, the increasing regulatory environment affecting several aspects of research (e.g., laboratory animals, carcinogenic substances, and radioactive materials) will often require institutional officials to participate in the selection of research topics and projects. Strong faculty participation in decisionmaking still will be essential, however. Effective communication and cooperation between institutional administrators and faculty is crucial to the success of all changes in academic management.

Non-departmental and Independent Centers. Special concerns may arise regarding the management of non-departmental and independent research centers. Research institutions may need to clarify their

expectations for faculty who work at such centers. For example, faculty obligations to their home departments should be distinguished from those to the non-departmental units that house their research laboratories.

When new centers are being established, universities may want to consider the effects of adding new administrative layers. New centers may be most effectively managed if they are administered and receive a similar degree of oversight as other comparable university functions.

Finally, research institutions may want to eliminate or redirect independent research centers or university-based federal laboratories that have lost their vitality or purpose. The more research institutions depend on these kinds of non-departmental arrangements for conducting research, the more the quality of their output will need to be monitored. Such centers may need to be periodically reevaluated to make sure they are placed within the institution where they will be most effective, to maintain their ties to traditional teaching functions, and to monitor the faculty reward systems.

Consortial Arrangements. The number and type of research consortia formed between or among colleges, universities and federal and industry research laboratories will increase in the next several decades. Just as for new centers, however, rigorous management standards and careful organization and planning will be required. Consortia must not only be productive, but they also must support the other missions of universities.

The current organizational structure of research institutions also may need to change as industries increase their support for university-based research. Academic scientists and engineers will continue to be drawn to join industry research operations, either as consultants or as full-time employees. To adapt successfully to these new linkages and to the changing composition of their research staffs, research institutions will need to put in place flexible management and salary policies.

Flexibility will also be needed as research universities negotiate new arrangements with federal research laboratories. There are mounting pressures to make federal research facilities, especially those housing one-of-a-kind scientific instruments, available to the research community at large. In instances where this has already taken place, university consortia have been formed to manage these "shared resources." Such considerations will also apply to other consortial arrangements with non-federal laboratories.

Tenure Process. The practice of providing unlimited tenure for senior teaching faculty was originally instituted to protect freedom of speech and scholarly inquiry. Academic institutions in the future will need to develop ways to protect such freedoms for their teaching faculty, while at the same time being responsive to rapid

change in the sciences and engineering. Universities and colleges with major research programs may consider instituting modified tenure arrangements for certain research faculty appointments in the sciences and engineering.

While any change in the tenure system will be difficult, one such option would be to appoint research faculty to annually renewable, multi-year terms, an approach already taken by several institutions for non-faculty research personnel. Such "rolling tenure" contracts might give institutions the flexibility to refocus their research programs, while at the same time providing some protection to faculty and staff whose terms are not renewed.

Universities, government agencies, and professional scientific and engineering organizations will need to consider the implications of societal and demographic change for the nation's research enterprise.

Societal and demographic changes occurring in the United States are increasingly reflected within the research enterprise. Academic institutions are faced with declining numbers of students who are interested in science and engineering and the additional but related problem of inadequate pre-college preparation in mathematics and the sciences. Women and minorities constitute a growing share of student enrollments in higher education, yet both groups are currently under-represented as educators, researchers, academic officers, and policymakers in the sciences and engineering.

Younger scientists and engineers bring different sets of experiences and often different expectations to their jobs than do more established researchers. Women and minority, faculty may have cultural values that differ from the majority of U.S. researchers, who are predominantly white males of European descent. The increasing phenomenon of two-income households has changed the personal-support network for many research faculty. With both parents working, child-rearing responsibilities are being shared more equally. In addition, longer commutes for faculty at campuses in large urban areas, combined with increased family responsibilities, exert pressures to adopt work hours similar to other occupations.

Universities and funding agencies must adopt appropriate policies and programs in response to societal and demographic change. To address the declining number of students who are interested in science and engineering, immediate and concerted action by all educational institutions will be needed to encourage qualified students, especially women and minorities, to pursue coursework in the sciences and engineering.

In response to changing family and personal demands of younger investigators, universities need to institute "flexible workplace" policies and programs. These might include hiring, promotion, and tenure policies that

take into account interruptions or slowdowns in the progress of an academic researcher's professional career, and policies that allow for temporary, part-time employment or extended leaves of absence.

The academic research community will need to engage in extensive dialogue about the changing expectations and cultural values of younger researchers. Such a dialogue should address the implications of societal and cultural change for the research agenda in the various disciplines and for the conduct of research itself. Some argue that the research community will have to adapt to the changing culture, organization and personal styles of newer generations of academic investigators. Others, in contrast, suggest that new participants in the enterprise must become better "acculturated" within the traditional "culture of American science" that has evolved over the past century. Whichever approach is taken, the working group believes that several general principles guiding the conduct of research should not change. These include the maintenance of quality through peer review, the unrestricted flow of scientific information, replication of research re-suits, and publication in refereed journals.

In the future, research institutions will need to expend considerable effort to maintain or enhance the quality of science and engineering education at the undergraduate, graduate, and increasingly, at the pre-college level.

Universities' dual missions of research and education are under increasing strain. The United States is unique in its primary reliance on universities for conducting basic research.⁹ The distinguishing feature of U.S. academic research, the linkage of research and education, originated in the earliest American universities during the years following the Civil War. Over the past century, this coupling of functions has led to extraordinary success in the sciences and engineering.

Tensions within the university faculty community, however, have been growing because of recent changes in the research environment. More than ever, university-based research activities are dependent on external funding. Academic researchers are devoting increasing time and effort to obtaining research support. They also are coping with new state and federal regulations that affect the conduct of research. As a result of these and other factors, faculty involved in research generally have less time not only for research, but also for teaching and public service.

Universities need to improve the quality of science and engineering education, especially at the undergraduate level. Colleges and universities will have to take their teaching responsibilities very seriously and put as much creative effort into curricula planning and undergraduate teaching as is put into their research programs. Undergraduate students within research universities often benefit from being in contact with investigators working at the frontiers of science and, in the best of circumstanc-

es, learn how research is really done. Furthermore, public support of institutions of higher education is strongly linked to public perceptions of the quality of the undergraduate teaching mission.

Academic research institutions face a dilemma as they try to improve their undergraduate programs. They must help their faculty, sustain active research programs that are an essential component of graduate education. At the same time, they must make sure that faculty, involved in research have enough time and incentives for high-quality undergraduate teaching.

University resources will have to be used effectively for undergraduate education, for developing new courses and for providing proper educational facilities and equipment. Attention will have to be given to preparing teachers and to providing opportunities for faculty to improve their teaching skills and course work.

The evaluation of teaching skills and of educational programs will become increasingly important. An accurate and fair method of assessing teaching skills will be necessary. Local faculty, peer groups, and current and former students, should take part in this assessment process.

With respect to graduate education, although the average time-to-degree for PhDs has not increased in the natural sciences and engineering as much as it has in the social sciences and the humanities, the increase is costly and should be taken as an additional warning sign of serious problems in the science and engineering pipeline.¹⁰

The working group believes that a significant factor in the lengthening time-to-degree is inadequate funding for graduate fellowships and the direct costs of research. Students who are unsupported or only partially supported take longer to complete their degree. Similarly, students who are dependent for their research activities on under-funded research programs spend considerable time improvising what they could purchase, and waiting to buy from the following year's budget what they need today.

The availability of new academic employment opportunities and better information about industrial research careers would also encourage students to enter science and engineering graduate programs and, once enrolled, to work quickly to complete their degree.

With respect to pre-college education, the public increasingly is looking to the nation's institutions of higher learning for help in defining and organizing pre-college curricula. It should be emphasized that, except for schools of education, this is a new assignment for universities, and one for which they are poorly prepared at present. Over the last decade, numerous attempts by universities to improve pre-college science and mathematics education have been made. The result in some cases is that the best students have arrived at college better-prepared than in the past. This is particularly true of programs designed to attract and retain science and engineering students from previously under-represented groups. The situation is not as hopeful for the average student, however, who arrives at college less well-prepared than in the past.

Ensuring the long-term health of university-based research—its overall quality, originality, productivity, diversity, and social usefulness—will require national policies, programs and resources appropriate to the changing characteristics of the enterprise. Desirable funding levels, required numbers of research personnel, and appropriate decision making mechanisms, for example, will in large part be dependent upon the evolving size of the enterprise and its changing structure—the degree of concentration or diversification of scientific and engineering academic research programs.

The overall size and structure of the enterprise, in turn, will be influenced by a number of large-scale forces which, for the most part, are outside the direct control of those who conduct, fund and, oversee research in institutions of higher learning. All participants in the U.S. research system need to understand these forces and how they may affect the academic enterprise in the future.

The working group developed its own framework for organizing these complex and often subtle issues. (See box on page 26.) First, it identified the large-scale forces it believes will have the most profound impact on the future of academic research.

Second, the working group considered the effects these forces might have on the size and structure of the enterprise. A set of hypothetical "scenarios" for the future of U.S. academic research was developed to illustrate these concepts. Third, the working group considered the long-term consequences of moving toward each scenario. Finally, the working group identified key policy or programmatic requirements, specific to each scenario, that would have to be met to maintain the health of the enterprise in the decades ahead.

The working group identified four large-scale forces—the pace and nature of research, economic conditions, political interests, and the international context—that it believes will have important and powerful effects on the enterprise.

The rapidly growing array of new scientific and technological opportunities and challenges is inexorably pushing the enterprise toward expansion. These opportunities and challenges arise both from within the research community and from society at large.

Genetic engineering, for example, now over 10 years old, has its roots in earlier genetic research conducted during the 1930s and 1940s, and the determination of the structure of DNA and the elucidation of the genetic code in the 1950s and 1960s, respectively. The vigorous pursuit of these discoveries has led to dramatic advancements in many fields of biology and medicine. Similarly, the study of the solid state nature of materials led to the development of transistors, which are now fundamental components of most electronic devices. More recently, molecular modelling is allowing biochemists and pharmacologists to construct new, potentially useful drugs.

Societal demands for solutions to many urgent problems also tend to increase the size of the enterprise. For example, the biomedical research

community was mobilized in the early 1980s when the threat of AIDS became apparent. Congress appropriated additional monies and agencies re-directed existing funds for basic research on the AIDS retrovirus, for drug development, and for prevention. A large cadre of researchers began to specialize in AIDS-related research. These developments were the result of the need to confront a severe epidemic, rather than a logical and orderly extension of ongoing research. Yet, as a consequence, new research avenues are rapidly emerging in the fields of virology, immunology, and genetics.

Recently identified global environmental problems have prompted significant additional investments in research by a number of countries. The need to learn about the causes and consequences of such phenomena as acid rain and global warming will most likely increase the numbers of investigators in relevant scientific and engineering fields.

The increasingly complex nature of research will affect the size and organization of research institutions. Larger and more multi-disciplinary research teams will be necessary for addressing many topics in science and technology. It is not clear whether these teams will be located most appropriately at large institutions with broad-based research portfolios, or among groups of smaller, more narrowly focused research institutions.

The desirable work environment of researchers in the next century is also unclear. The preeminent researchers of the future may need to be more specialized than today's investigators, relying more on cooperative relationships with their peers. On the other hand, to be highly successful, tomorrow's scientists may have to be fiercely independent, isolating themselves at times from the enormous flood of information that flows over modern communication systems.

The strength of the U.S. economy will be an important factor in setting the overall level of resources available for meeting national needs. In this respect, a large and growing national economy would more readily accommodate an expanding research enterprise. Conversely, over the long term, a weak or declining economy would most likely make it difficult to sustain even the current level of research activity in the United States.

Academic research institutions will thus have an increasing interest in the economic vitality of the nation. Their public support will be closely tied to the country's ability to generate wealth through increased industrial competitiveness and work-force productivity.

Universities and colleges involved in research contribute to the nation's economic growth through their role in educating a productive workforce and by creating new knowledge of potential commercial value. Additional economic benefits may result from the impact new or expanded research institutions have on local and regional economic growth.

For all of its economic benefit to the nation, however, academic research will be just one of many factors affecting the size and strength of the U.S.

economy. Industrial productivity, international trade, inflation, and employment also will come into play. Similarly, policies for U.S. research and development will be but one part of overall national economic policy. Fiscal, monetary, and trade policies will be other vital ingredients.

Furthermore, academic research institutions will benefit from a healthy economy only to the extent that the public believes in the social value of the work they perform. Taxpayers will provide substantial financial support for academic research only if there is convincing evidence that research is helping to maintain or improve the quality of life and the standard of living.

In the decades ahead, the unpredictable cross-currents of the U.S. political process will exert a major influence on both the size of the overall research enterprise and its structure. The level and allocation of resources devoted to university-based research will be determined by political decisions made at the local, state, and national levels.

How well academic research fares in the political process will depend largely on public perception of the enterprise, and how these collective feelings are communicated to lawmakers. Supporting research remains a generally popular political position. Society's ability to accept and make use of advances in technology, and public understanding of the value of basic research, will be critical factors in future political support for university-based research.

Those who fund, conduct, and oversee academic research will be able to influence certain aspects of the political process. Opinions expressed by the scientific and engineering communities, for example, can affect decisions by legislators to appropriate funds for specific research projects.

Public spending, however, is a reflection of the priorities that the country places on addressing important national goals. As such, academic research will compete with other needs, such as rebuilding roads and bridges, addressing crime and poverty, assuring high-quality education and health care, and providing for the well-being of the elderly. While steady investments in academic research may very well contribute to the long-term solution of these problems, economic problems, such as deficits and recessions, tend to shorten national political perspectives and to encourage spending on short-term remedies. Academic research requires long-term financial commitment.

The allocation of public funds to research institutions is affected by two powerful and sometimes conflicting national political objectives. The first is to support research of the highest quality. To that end, Congress has authorized several federal agencies to disburse research dollars, which they do with the advice of outside experts. Most federal support for academic research today is allocated, at the project level, through such ''peer review" processes.

A second objective is to enhance the research capacity of individual states and regions. At the national level, the growing federal contribution to ac-

ademic research, approaching $10 billion in 1991, makes the research enterprise more visible politically and thus more subject to political apportionment. In the U.S. Congress, this is evident in the increasing number of non-peer-reviewed line-item appropriations for research facilities and programs at individual universities.

The notion that each state or region should possess a world-class institution of higher education and research has enormous political appeal. There is a generally strong belief that the presence of research universities contributes to regional economic well-being and competitiveness. As a consequence, numerous research programs and research facilities have been made possible by political initiatives launched or supported by individual legislators at the state or federal level. Several state governments have increased resources for building up their flagship campuses. The growth in leveraged local funding for research programs has also created research capacity at scores of universities previously devoted largely to teaching.

In the first three decades following World War II, the United States provided most of the research infrastructure—equipment and facilities—for its own research requirements in academia, industry and government laboratories. The United States also granted much of the world's research community access to its research infrastructure.

As the research capability of other nations increases, however, such research instrumentation is now becoming more equally distributed among countries. Furthermore, as the cost of research infrastructure has increased rapidly, many expensive pieces of equipment are becoming scarce.

This occurs at a time when fierce economic competition is beginning to dominate the international arena. The political response of individual nations and multinational trading blocks could shift the world economy either toward greater openness and interdependence or toward more barriers to the free flow of information and goods. The outcome will have a profound impact on the future size and breadth of the U.S. research effort.

At stake are the ability and willingness of the growing international research community to exchange information and to collaborate in vital research areas. The extent to which U.S. investigators have access to frontier research conducted in other nations has major strategic implications for the academic research enterprise.

A shift toward a more open international research system would give the United States greater flexibility in setting its overall research agenda. The nation would have the option of targeting selected fields of vital strategic interest or those in which it has a comparative advantage. New communication technologies would facilitate collaboration between U.S. and foreign researchers in specific fields. Shared investments in "big-ticket" research instrumentation would become more attractive.

With greater access to foreign research at the frontiers of science and technology, the United States could explicitly decide not to invest in fields in which other countries have a commanding lead, or for which other nations have constructed expensive, one-of-a-kind research facilities not available here. The United States, of course, would still need to maintain a critical mass of expertise in "non-targeted" fields. Without this relatively modest investment, it would be difficult to take advantage of discoveries made abroad and to educate future generations of researchers in these fields.

At the other extreme, barriers to the international flow of research information would likely force the United States to seek research self-sufficiency in almost every field. This approach would require a significant expansion of the research workforce, including the recruitment of science and engineering talent from foreign countries. At the same time, it would necessitate greatly increased resources for research.

Such a massive mobilization of human and financial resources might well be economically or politically unrealistic and, consequently, the ability of the United States to remain a leader in science and technology might be jeopardized. In the long run, international policies that restrict the free flow of fundamental scientific information would be as destructive to this nation's strength as are trade wars.

Whatever the international economic and political climate, achieving an open flow of research information may not be easy. This is primarily because of the structural differences among the research systems of different nations. In the United States, for example, most basic research is conducted by university-based scientists and engineers. The results of that work are disseminated in publicly available journals and, increasingly, through electronic media. In many other countries, by contrast, frontier basic research in targeted fields is performed largely in government or industry laboratories, which also conduct a substantial amount of applied, proprietary research. There is a danger that, in response to growing international economic competition, the results of much foreign-based research may not be promptly or freely released to the public.

It is likely that other nations would resist a demand by the United States to make fully available the research output of their government or industry laboratories. If they were to comply, they probably would do so only if the results of research conducted in U.S. government and industry laboratories were made available. In the short term, this uncomfortable and improbable quid pro quo might be the only way of ensuring a fair exchange of basic research information among countries.

In the long term, an international convergence of national research systems, toward more equitably funded, publicly available research institutions, may help resolve the problem. In the meantime, tensions among national governments, brought about by the differences among the research systems, are likely to continue.

At the level of individual investigators and research projects, however, new communication technologies will facilitate greater worldwide exchange of information, independent of the international economic or political climate. Such interactions will become a pervasive feature of the international research community in the next century.

In the past, major international crises such as wars, political turmoil, or foreign technological breakthroughs served to strengthen U.S. university-based research. For example, the rise of National Socialism in Germany during the 1930s stimulated a stream of emigration by eminent European scientists, many of whom assumed professorships at U.S. universities. World War II and the Soviet launch of Sputnik led to dramatic increases in public funding of U.S. science education and university-based research. Future international crises, especially those involving national technological advantages, could also have a profound influence on public support for U.S. university-based research. While unpredictable, the possibility of such events must be factored into scenarios for the future.

The pace and nature of research, economic conditions, political interests, and the international context will shape the character of the U.S. aca-

demic research enterprise into the next century. These forces will be felt primarily in terms of the size and structure of the enterprise.

During the next several decades, the size of the U.S. academic research enterprise—the number of academic departments and research personnel—may grow, remain in steady-state, or become smaller. (See box on page 31.) It would be possible to have a larger enterprise without creating more research institutions, if the number of departments or personnel within the existing set of institutions were to increase. Similarly, a smaller enterprise could be envisioned while retaining the current number of institutions, if the number of departments or personnel in existing universities were reduced. Independent of the overall size of the enterprise, the distribution of talent and resources among research fields may change.

The structure of the academic research enterprise, expressed as the degree of concentration or diversification in science and engineering talent

among research institutions, also could change over time. (See box on page 32.) The enterprise could become more concentrated within universities recognized for their scientific strength in a broad array of fields. Alternatively, the enterprise could become more diversified, with a dispersion of the nation's academic research activity among institutions with a focus on fewer research fields or sub-fields. This could be accomplished if current large-scale research universities reduced their research portfolios, or if aspiring research universities developed one or more nationally-recognized research programs, or by both factors.

Within this framework, there are nine possible combinations of size and structure. (See box on page 33.) These combinations should be viewed as points on a plane, not as rigid endpoints where U.S. academic research will one day come to rest. These hypothetical shifts in the direction of the enterprise are further characterized as qualitative "scenarios" for the future of U.S. academic research. (See Figure 6 on page 34.) The scenarios are a way of thinking strategically about the different directions in which the four large-scale forces could push the enterprise. (See Tables I-III, pages 35-37.)

FIGURE 6

Scenarios for the 21st Century

The consequences of the enterprise moving in a given direction are profound, in terms of the overall capacity of U.S. academic research, the resources required to support the enterprise, and the organization of decisionmaking.

Expanding the research enterprise would offer the United States greater capacity for addressing new scientific and technological opportunities. With an increasing number of talented research personnel, this country could be more responsive to future challenges to our economy, health, environment, and national defense. To sustain such growth, however, would require a bold national commitment to enhancing our scientific and technological base. Given the rising costs of conducting research, the nation would have to dramatically increase its investments in research facilities, equipment, graduate education, and individual research projects. Given the problems of the U.S. educational system, science and mathematics programs at all educational levels would have to be improved substantially. Without such a long-term national commitment, expanding the enterprise would further exacerbate current tensions and ultimately damage the quality and productivity of U.S. academic research.

Maintaining a steady-state enterprise would barely sustain the current national research capacity. Reallocation of resources and re-deployment of personnel would be the major way new research opportunities and social needs would be addressed. Given the certain expansion of the worldwide research agenda, national decisionmaking would focus on the choice between new opportunities and ongoing research. However, maintaining the current number of research personnel throughout the coming decades would not imply a similar steady-state in financial resources. Given rapidly rising research costs, the enterprise would decline if increased real investments were not forthcoming. Likewise, given students' declining interest in science and mathematics, a significant improvement in the quality of science education would be absolutely essential.

Down-sizing the enterprise would mean surrendering many ongoing research areas in favor of opportunities of paramount national importance. Several serious consequences of such a transition need to be considered. First, the United States would increasingly become dependent on the science and technology of other nations. As described earlier, access to foreign science and technology may be thwarted by international political and economic rivalries and by the structural asymmetries of national research systems. Second, given the rising costs of research, even a down-sized enterprise would require at least sustained current real levels of research funding. Third, for a smooth transition to a smaller enterprise, explicit funding criteria for choosing among research fields would have to be established.

A shift toward a more concentrated research enterprise would, in general, allow for greater national decentralization of decisionmaking to universities and their academic departments. Coordinated planning of research activ-

ities within each university would more likely result in a comprehensive national research portfolio. However, a down-sized enterprise condensed into a smaller set of ''elite" research universities would, in time, most likely become a system of "national universities," resulting in increased decision-making authority at the federal level.

A shift toward a more diversified academic research enterprise would encourage research institutions to focus their resources within areas of research strength. It would achieve the political goal of a broader institutional and geographic distribution of the nation's research capacity. With fewer research programs, however, individual universities would become more vulnerable to shifts in federal funding priorities than if they had broader research portfolios. Institutional decisions to reduce the number of research fields thus entail new risks to universities and to important parts of the academic research enterprise. In addition, significant investments in communications technologies would be essential to allow information exchange and research collaboration among investigators in geographically distant regions.

A down-sized enterprise, simultaneously becoming more diversified, would require new mechanisms for setting national research priorities. This would be needed both to coordinate widely dispersed activities and to ensure there were no important gaps in the nation's research portfolio. This scenario would be the most problem-ridden for individual U.S. investigators. Not only would it reduce the number of research fields, but the universities' central administrations would have to play the precarious role of deciding which research fields to abandon and which to reinvigorate.

A critical challenge for decisionmakers will be to preserve the health of the academic enterprise—its quality, originality, productivity, diversity, and social usefulness—and the United States' standing in the international scientific and technological community. Regardless of the overall direction of the enterprise, U.S. policies and programs must encourage new generations to pursue careers in the sciences and engineering, ensure adequate funding, improve research-related decisionmaking, implement new communication technologies, and facilitate participation in the global research system.

The relative importance of any one of these policies, however, differs according to the size and structure of the enterprise. For each scenario, a unique set of complementary policies and programs is required. (See Figure 7 on page 40.)

If the U.S. academic research enterprise grows significantly larger, the need for more science and engineering personnel will become critical. This requirement is the same for all growth scenarios, regardless of whether the

FIGURE 7

Scenario Policy Requirements

structure of the enterprise is more concentrated or more diversified. (See Figure 8, page 41.)

The United States has two primary sources for enlarging the pool of available research expertise. The first and most desirable is the nation's own citizenry. Successful efforts would have to be made to attract larger numbers of students, including women and minorities, into post-secondary science and engineering education, and to promote their subsequent participation in the research enterprise.

At the graduate level, significantly increased financial support would be needed in the form of stipends, fellowships, and assistantships. At the pre-college level, substantial improvements in science and mathematics programs would be required. This would necessitate dramatic improvements in student attitudes toward science and mathematics. For many, it also would

FIGURE 8

Human Resource Requirements

demand a fundamental recognition of the value of long-term investment in science and education.

The second source of talent would be the potentially large number of foreign-trained scientists and engineers desiring to work in American research institutions. However, reliance on this source is risky. As described earlier, other countries, including many developing nations, are creating their own research systems that will compete directly with the United States for this pool of talent. As other nations increase their dependence on technology and science and face shortages of skilled workers, the competition for human skills is likely to increase.

Should research costs per investigator continue to rise at a rate higher than general inflation in the economy, maintaining a steady-state in the size of the enterprise will require annual real increases in financial support for research projects, facilities, and equipment, as well as for graduate education. If such investments are not forthcoming, the enterprise would likely decline in both quality and capacity.

With rapidly increasing costs, expanding the number of researchers would require substantial real annual increases in funding for university-based research. In an expanding enterprise composed of more large-scale research universities, even greater funding growth would be required to support added duplication of programs across the enterprise. (See Figure 9, page 42.)

FIGURE 9

Funding Requirements

The plausibility of any funding level must be considered within the larger context of national economic growth and political willingness to increase public investments in research. In addition, political decisions to augment university-based research, relative to other research sectors, will also be a factor. If federal laboratories were to assume increased responsibility for costly research facilities and expand access to these facilities by university researchers, a growing academic research enterprise would be more possible. On the other hand, should federal laboratories more actively compete with academic institutions for basic research funding, growth in the academic sector may be more difficult.

Decisions to add or terminate academic research programs, as well as to allocate funds for research infrastructure, currently are made at different levels. The locus of decisionmaking will depend on the changing size and structure of the enterprise. (See Figure 10, page 43.)

If the enterprise were to become more concentrated within large-scale research universities, decisionmaking might best take place at the institutional and departmental levels rather than the national level, because of the breadth of coverage within each university.

If the enterprise were to become more diversified, there would be greater need for national-level decisionmaking to coordinate widely dispersed activities. If the enterprise were simultaneously shrinking in size, new na-

FIGURE 10

Shift Toward National Decisionmaking

tional decisionmaking mechanisms would be absolutely required to ensure that no important gaps occur in the nation's research portfolio.

In an enterprise comprised of more research institutions with limited research portfolios, there also would be greater need for decisionmaking at the level of the central administration of each institution. If these universities were to reduce the number of their research fields, more centralized decisionmaking would become a necessity during the transition.

The conduct of research at all levels is becoming increasingly dependent on telecommunications technologies. Contemporary research problems are being addressed by larger and more complex teams of investigators around the world. In many instances, sophisticated and more expensive research equipment is needed. Telecommunications networks allow widely dispersed research personnel to share data, to collaborate effectively, and to use one-of-a-kind research instrumentation.

If the enterprise were to become more diversified across a broader array of institutions, an effective national telecommunications infrastructure would be absolutely essential. A key component of the infrastructure would be personal computers used as "information ports" to receive and send electronic and voice mail, complex documents, and real-time video images; to access specialized databases and digital libraries; and to perform experiments

FIGURE 11

Requirements for Communication Technologies

using advanced supercomputers, automated research instruments and other modern facilities in remote locations. (See Figure 11, page 44.)

Whatever the size of the U.S. academic enterprise, access to foreign sources of new knowledge and technology will be essential. This will be especially critical as the research systems of other nations grow and become more effective.

If the U.S. enterprise were to shrink over the next several decades, access to foreign research would become absolutely essential for maintaining the overall quality and usefulness of academic research in this country. (See Figure 12, page 45.) In this situation, the United States would have to have full access to the results of basic research conducted by other countries. Unlike the requirements for funding, human resources, decisionmaking, and the use of telecommunications technology, however, the openness of the global research system depends upon decisions made within many nations.

Policymakers must carefully consider the consistency and efficacy of policy proposals affecting human and financial resources, locus of decision-making, communication infrastructure, and openness of the global research system. The proposals must be evaluated in terms of their explicit and im-

FIGURE 12

Required Access to Foreign Research

plicit objectives, assumptions about the size and structure of the overall enterprise, plausibility of complementary policy requirements, and implications for near-term decisionmaking. (See box on page 46.) Inconsistent policies, programs, and resource commitments will lead to chaotic conditions and will have a potentially disastrous effect on the quality of academic research in this country.

The research enterprise of the future will be unlike the one of today. Two visions of the future stand in stark contrast: a more hopeful vision in which the research enterprise sustains leadership within the emerging global research system, provides opportunities for young talent within a diverse population, takes best advantage of frontier technology, and contributes to the vitality and well-being of this country; and a less-hopeful one, in which there is unproductive competition within the research community, an inability to pursue research opportunities of critical importance to the United States, and a gradual decline in international preeminence.

Achieving the more positive vision will require making difficult choices with potentially fateful consequences for the research enterprise.

Unless the U.S. research community adequately responds to changes now occurring within the enterprise, harmful tensions will persist and public support will erode.

Strategic decisions that will determine the future size and structure of the enterprise will be made in this country, either deliberately or inadvertent-

ly. These choices will influence the capacity and structure of research in the United States for decades to come.

These choices are interdependent and will be affected by the complex interplay of large-scale societal forces and specific policy requirements. Inconsistent or contradictory policies will have a decidedly harmful effect on the quality and productivity of U.S. academic research. Thus it is crucial that these choices be made comprehensively and explicitly. Inaction would be the worst possible choice.

1. Government-University-Industry Research Roundtable, Science and Technology in the Academic Enterprise: Status, Trends and Issues, Washington, D.C.: National Academy Press, October 1989.

2. For a review of issues and data regarding the costs of research and proposal success rates, see U.S. Congress, Office of Technology Assessment, "Understanding Research Expenditures," in Federally, Funded Research: Decisions for a Decade, (OTA-SET-490), Washington, D.C.: U.S. Government Printing Office, May 1991, pp. 171-201.

3. "Science: The End of the Frontier?" American Association for the Advancement of Science, January 1991.

4. Decisionmaking processes regarding priority-setting will need to be developed by the parties involved. For further discussion, see Part III, "Charting a New Course," pgs. 53-58.

5. For an example of such priority-setting, see National Research Council, The Decade of Discovery in Astronomy and Radiophysics, Washington, D.C.: National Academy Press, 1991.

6. For recommendations for national priorities within the biomedical research enterprise, see Institute of Medicine, Funding Health Sciences Research: A Strategy to Restore Balance, Washington: National Academy Press, 1990.

7. See National Academy of Sciences, National Academy of Engineering, Institute of Medicine, Federal Science and Technology Budget Priorities —New Perspectives and Procedures, Washington, D.C.: National Academy Press, December 1988.

8. See U.S. Congress, Offence of Technology Assessment, ''Priority Setting in Science," in Federally Funded Research: Decisions for a Decade, (OTA-SET-490), Washington, D.C.: U.S. Government Printing Office, May 1991, pp. 137-167.

9. See Government-University-Industry Research Roundtable, Federal-State Cooperation in Science and Technology Programs, February 1992.

10. Source: National Science Foundation, Division of Policy Research and Analysis.

11. These financial data do not include reimbursement for facilities costs, included within indirect cost payments. For further discussion, see Government-University-Industry Research Roundtable Perspectives on Financing Academic Research Facilities: A Resource for Policy Formulation, Washington, D.C.: National Academy Press, October 1989.

12. Decisionmaking processes regarding funding responsibilities will need to be developed by the relevant parties involved. For further discussion, see Part III, "Charting a New Course," pgs. 53-58.

13. "Leveraging" as used here refers to the bringing together of resources for the support of research from multiple parties—federal agencies, state government, industry, universities, or philanthropies—to the mutual benefit of all parties. "Cost-sharing" refers to a requirement by research funding agencies that grant awardees or research contractors fund a share of project costs as a condition of receiving the award or contract—in essence, an entry fee.

14. For example, remarks by Walter E. Massey, Director, National Science Foundation, at Director's Seminar, Office of Technology Assessment, September 23, 1991, (unpublished).

15. For a review and analysis of recent literature, see U.S. Congress, Office of Technology Assessment, "Human Resources for the Research Workforce," in Federally Funded Research: Decisions for a Decade, (OTA-SET-490), Washington, D.C.: U.S. Government Printing Office, May 1991, pp. 205-230.

16. See, for example: "Indirect Costs: The Gathering Storm," Science , 252:636-638, 1991; "Allegations of University Abuses of Overhead System Continue as House Panel Releases a New List of Embarrassing Items," The Chronicle of Higher Education, May 15, 1991; "OMB to Cap Overhead Payments for Research," The Washington Post, May 16, 1991; and "Overhead Cost Research Dear," Nature, 351:255, 1991.

17. Government-University-Industry Research Roundtable, The Academic Research Enterprise Within the Industrialized Nations: Comparative Perspectives, Washington, D.C.: National Academy Press, March 1990.

18. See H. Tuckman, S. Coyle, and Y. Bae, On Time to the Doctorate , Washington, D.C.: National Academy Press, 1990; and W. G. Bowen, G. Lord, and J.A. Sosa, "Measuring Time to the Doctorate: Reinterpretation of the Evidence," Proceedings of the National Academy of Sciences, USA, Vol. 88, pp. 713-171, February 1991.

19. See U.S. General Accounting Office, Budget Issues: Earmarking in the Federal Government, Washington, D.C., 1990.