CRA Testimony on Scientific Allocation of Scientific Resources


given by Andy van Dam, Thomas J. Watson, Jr., University Professor of Technology and Education and Professor of Computer Science at Brown University

before the National Science Board

May 22, 2001

Good afternoon. My name is Andy van Dam. I am the Thomas J. Watson, Jr., University Professor of Technology and Education and Professor of Computer Science at Brown University. My research interests are in computer graphics, electronic books, and the use of technology to support education. Today, I am representing the Computing Research Association, an association of more than 180 North American academic departments of computer science and computer engineering (CS&CE); laboratories and centers in industry, government, and academia engaging in computing research. I would like to thank the National Science Board for giving me as a representative of the computing research community the opportunity to share my views on the allocation of scientific resources.

I think that the recommendations that are contained in the draft report are sensible, and if implemented, could improve the allocation of research funding. However, we may be promising too much if we claim to offer "scientific allocation of scientific resources" - particularly when we try to establish priorities across disciplines. That's because the United States supports research for a wide variety of reasons, including:

It's not clear to me how one would make a "scientific" tradeoff between developing an AIDS vaccine, building a new telescope, supporting the research needed to make software more dependable and reliable, and improving our energy security. We are in a different world from the management committee of a corporation choosing between a variety of investment opportunities. In the corporate world, decision-makers can look at a quantitative yardstick - estimates of the discounted cash flow of the proposed investments. In contrast, setting priorities between areas of research ultimately requires making tradeoffs between different national goals, such as national security, economic competitiveness, our quality of life, and expanding the frontiers of human knowledge and understanding. As the report acknowledges, the "allocation of funds to research is primarily a political process."

Having said that, there are clearly a number of steps that could be taken to improve the allocation process, and raise the level of debate about competing research priorities. Let me suggest six concrete steps that policy-makers and the research community could take:

1. Identify areas of science and engineering that we are under-investing in because there is no mission agency that is responsible for supporting it, or the relevant agency has little or no capacity to manage a research program. The science and technology needed to support education and life-long learning is a good example. The PCAST concluded that we are investing less than 0.1 percent of total expenditures on education in R&D. Clearly, more research is needed in areas such as basic research in cognitive science and neuro-imaging, to improving the state-of-the-art of learning technologies, to large-scale clinical trials for proposed K-12 educational reform strategies. It's difficult for me to imagine anyone concluding that our current modest investments in learning science and technology at the Department of Education, NSF and NICHD (National Institute of Child Health and Human Development) are adequate, in terms of vision, scale, and rigor. A high-level panel of the kind that the NSB is calling for could point out these kinds of gaps.

2. Expand efforts to set priorities within a given field, or to support a particular set of national policy objectives. The science and engineering community has demonstrated the ability to do this. In my field of information technology, for example, the President's Information Technology Advisory Committee looked at the totality of the government's investment in IT research, and concluded that:

Several years ago, the PCAST completed a similar exercise with respect to the nation's portfolio of energy R&D, and made concrete recommendations concerning the level and composition of government investment needed to reduce greenhouse gases, improve efficiency, and enhance America's energy security. These kinds of efforts should be encouraged. It would also be helpful to compare different efforts to set priorities to determine whether there are any emerging "best practices."

3. Make recommendations that would improve the allocation of resources between extramural and intramural research. An expert panel could summarize whatever literature exists on the relative impact and social return of extramural vs. intramural research. It could identify some general guidelines for determining the most appropriate mix between the two. For example, intramural research may be needed to support very large, long-duration, complex projects. Extramural research, particularly university-based research, clearly has advantages in terms of preparing the next generation workforce and making research results more widely available. Finally, the panel could recommend structural reforms that would improve the allocation of funding between extramural and intramural research. For example, the 1995 National Academy of Science report entitled Allocating Federal Funds for Science and Technology concluded that science and technology budget decisions "should give preference to funding projects and people rather than institutions." The report also concluded that for most science and engineering projects, there should be a "general preference for academic over nonacademic institutions."

4. Establish some guidelines or goals with respect to the size and duration of grants, and funding rate. For example, the average NSF grant is $110,000 and lasts for less than three years. Researchers would be more productive if they did not have to support their labs and research teams by cobbling together lots of small grants and constanty writing proposals. Furthermore, many important challenges in science and engineering will require support for small teams of researchers and "centers of excellence." Limited budgets create enormous pressure to spread the money around. It's hard to change the world with a $110,000/year grant. Success rates are also an important indicator to track. The last NSF Information Technology Research solicitation generated proposals requesting $3.2 billion, of which NSF was able to fund only $90 million. Even assuming that only 30 percent of the proposals were worth funding, this is still an indication that there are more good ideas than can be supported in current budgets.

5. The high-level expert panel should provide some general assessment of the balance of the federal government's research portfolio. For example, I doubt that many leaders in the research community would support the decision to increase the NIH budget by $2.6 billion while cutting the NSF research budget, given that NSF accounts for almost half of the federal funding of non-medical university research. Advocates of biomedical research have been increasingly vocal on the need to support the physical sciences and engineering - given the importance of these disciplines for advances in areas such as imaging, DNA chips, and computer simulation for rational drug design.

6. The high-level panel should be charged with identifying a handful of research initiatives that are strong candidates for increased funding. Again, I believe that making these decisions is a matter of judgment and informed opinion as opposed to a science, but it would be useful to discuss the criteria that might be used to evaluate proposed initiatives. Examples of potential criteria include:

If a set of criteria could be agreed upon, and if a certain amount of money was set aside to support new research initiatives or expansions of existing initiatives, I think that the research community would be motivated to suggest ideas for consideration by federal policy-makers.

In conclusion, I want to commend the members of the NSB for your work in this area. Although we may not be able to create a science for the allocation of scientific resources, there is certainly plenty of room to improve upon the status quo. The research community has an obligation to continue to wrestle with these questions, and to provide the best possible analytical frameworks, solid data, informed opinions, and inspiring visions of the future benefits of investments in science and technology. The Computing Research Association would be happy to provide additional input to the NSB's deliberations on this important topic.

Slides

Document last modified on .