The panel offers the following recommendations based on its review of the STC program:
1. NSF should continue the STC program.
NSF support for centers is based on the premise that some compelling scientific questions can best be addressed through the long-term, coordinated efforts of teams of researchers. STCs are unique among NSF centers in that they create opportunities for research that are unconstrained by the NSF directorate structure. The program also provides a means for the focusing of substantial resources on frontier problems within the directorates; these problems might require facilities beyond the scope of individual grants and require coordination through center funding. The driving forces for creating centers are the scientific importance of such research and the resulting societal benefits.
2. Research and the undergraduate and graduate education linked to it should be the paramount goals of the STC program.
The primary functions of universities are education and research. The STC program's primary goal must be long-term research. Associated with that are undergraduate and graduate education and knowledge exchange with other institutions and industry. Education and the exchange of knowledge make research valuable to society by making sure the knowledge is used.
Recently, the NSF has placed considerable emphasis on K-12 educational programs in the STCs. In our view, these activities should be undertaken in ways
that are natural extensions of particular centers' activities. In general, they should be of secondary importance. If K-12 activities are to be continued, they should not be prescribed. Centers should be allowed to judge the appropriateness of topics and opportunities for K-12 outreach. In addition, a process for evaluating their educational effectiveness needs to be put in place.
3. In future solicitations, NSF should encourage but not require that proposed STCs be multidisciplinary.
The STC program has undergone shifts of vision and goals. Most recently, the changes included an implied requirement for multidisciplinary research. That is too restrictive. The STC program provides a mechanism to address a wide variety of important problems that are not otherwise possible within the NSF structure. They may or may not be multidisciplinary.
The panel endorses the view of the Zare report (NAS 1987):
[Science and technology centers'] primary goal is to exploit opportunities in science where the complexity of the research problems or the resources needed to solve these problems require the advantages of scale, duration, or facilities that can be provided only by the center mode of research. . . . Interdisciplinary research, although essential for the solution of many problems, should be pursued only when there is a demonstrated need or opportunity, not because of current fashions or the enhanced likelihood of funding.
Multidisciplinary research should not be a requirement in future STC solicitations. As indicated in the original solicitation shown in Box 2-1, the STC program should “help maintain U.S. preeminence in science and technology and ensure the requisite pool of scientists with the quality and breadth of experience required to meet the changing needs of science and society. ” The benefit of this view is shown in the current STCs—most of which have provided unique problem-solving abilities but that are both disciplinary and multidisciplinary.
4. The level of funding for the STC program should be maintained to ensure that it retains its strength and vigor.
When the STC program began, there was some concern that substantial funds would be diverted from individual-investigator activities. The panel found that not to be the case. The fraction of NSF's research budget devoted to the STC program is very small—about 2%. Yet, this is the only NSF program that provides open competition across all directorates. In our view, the investment is modest and has paid large dividends. Sufficient funding is available to support a variety of exciting endeavors without taking a great deal of funding from other NSF programs.
The NSF director and the National Science Board need to ensure that, at the
margin, investments in each of the various NSF modes of support—projects, centers, facilities, education, and training—are matched appropriately to the nature of the problems faced in the fields of science and engineering supported by NSF. In other words, an additional dollar of support for individual investigators or facilities should be as well matched to the problems that are best addressed by that mode as an additional dollar of support for centers is matched to the problems best addressed through centers.
In addition, concerns were expressed at the time of program initiation that the program would result in projects that funded weak researchers and would not receive adequate peer review; on the basis of the site-visit reports and the publication records of the STC researchers, this appears not to be the case. However, in the first two rounds, there were far more appropriate and excellent proposals than could be funded. The site visit report, in particular, repeatedly noted that the STC they visited supports some very strong and excellent investigators. Moreover, individual centers are monitored closely, perhaps overreviewed rather than underreviewed. Although not all STCs have been successful and not all fields of research will retain their timeliness over the full 11 years, the STC program is important because it meets an important need in NSF's modes of support.
5. The budget for the STC program should retain a separate identity. Moreover, the tradeoff between this program and other NSF activities should be made at the level of the NSF director.
To fulfill its roles, the STC program must keep a separate and distinct budgetary identity, and the tradeoff of this program with other NSF activities needs to be at the foundation level, with the NSF director responsible for that budgetary decision. The program also needs an advocate outside the individual directorates. We believe that Office of Science and Technology Infrastructure or its equivalent must continue to play a strong role.
6. STC solicitations should be conducted openly across all fields by NSF as a whole (rather than within specific directorates), and existing STCs should be allowed to compete in this open process.
The STC program allows scientists to propose research ideas that are multidisciplinary or that do not fit well within the modes of support provided by the current NSF directorates. The panel believes that it is important to support promising initiatives in science and engineering that cross disciplinary lines or that are otherwise frustrated by the NSF structure. Solicitations for the STC program should be managed by NSF as a whole so that competition can take place across all program directorates and funds can be provided for the fields of greatest opportunity.
Some have suggested that existing STCs not be allowed to recompete, inas-
much as they have already had their chance to explore a given research subject for a long period and others should be provided an opportunity. But some STC subjects are still important to explore and of great relevance, and existing STCs have advantages over new competitors in having had the experience of making an STC work, having the infrastructure in place to support an STC, having collaborative groups already functional, and having management and advisory teams trained and functional. Although that might give them an advantage during the review process over those who are just sending in proposals, the panel believes that existing STCs should be allowed to compete. They should have neither advantages nor penalties in this process. It is the purpose of this open competition of allowing funds to go to the best ideas and not to allow the kind of permanence that other centers seem to achieve.
7. The duration of STC awards should be 10 years. Two periodic solicitations should occur within that period.
The dynamic nature of science and technology requires that NSF accept funding requests on a regular cycle. The panel recommends that solicitation be conducted every 4 years with two solicitations in the 10-year span of each grant. This would allow two years of ramping up of the new programs and two years simultaneous phaseout of the old programs. In addition, it would spread the opportunities for funding and the necessary intense solicitation and review process more evenly over time.
The current duration of 11 years (9 years plus 2 years for phasedown) is working well but could be shortened by 1 year. If future competitions occur in year 8 now or year 7 in the future and decisions are announced in year 9 now (or year 8 in the future), it would allow for the current 2-year phasedown period (and a 2-year startup).
8. The differing roles of Office of Science and Technology Infrastructure and program directorates in the management of STCs are complementary and should be continued.
No program office can cover the enormous scope of scientific activities in the STC program. But there needs to be consistency in the goals of the program and in the procedures used for proposals, recordkeeping, reporting, oversight, and so forth. The best way to achieve that is to continue managing the STC program as it is now, taking advantage of the complementary roles of OSTI and the program directorates.
9. NSF should place greater weight on scientific and administrative leadership in evaluating proposals for STCs and in the periodic reviews of centers.
Scientific leadership is the key to a successful center, especially if it is managing highly complex, multi-investigator, multi-institutional research.
A number of the STCs have had problems in scientific leadership or administrative management. NSF has responded to the concerns well; the centers that have had difficulty were able to make corrections within a short period. The National Academy of Public Administration report indicated that NSF does not have a clear method for removing or replacing an STC director and that the National Science Board should be involved in developing such a process. But the panel believes that the existing process works. We have several examples of cases in which the review process did its job. However, the panel also believes that the evaluation of the scientific leadership and administrative management of a center should be made an explicit part of the site-visit review process, which is not now the case. NSF should immediately respond to adverse site-visit reports and provide the university an opportunity to make corrections; this has generally been done, but it should be more explicit in NSF procedures.
There has been a tendency on NSF's part to conduct a full review of a center even for an administrative problem that does not involve the center's scientific work. That is unnecessary and unduly burdensome. The follow-up reviews should focus on only the aspects of the STC that are of concern.
10. NSF should establish policies allowing center directors to allocate funds and other resources (e.g., staff) both within and among participating institutions, so as to optimize progress toward the center's goals. The limits of this unilateral authority should be clearly defined and procedures to make major reallocations beyond these limits should also be defined.
In some cases, agreements have been made among the institutions setting up a center for a specific allocation of funds to the host institutions. That makes it difficult for centers to shift resources over time to the subjects of greatest need. NSF must require that these agreements —which are often set at the proposal stage—include a process to change allocations. Center directors should have the freedom to move resources to where they are most needed. Given the length of the award for centers, one cannot expect an original allocation always to make sense 5-10 years later.
11. NSF should make every effort needed to coordinate reviews of the centers to avoid redundant data collection and to make previously collected data available to all who have good reason to be interested.
There has been considerable concern that the STCs are overreviewed. The panel believes that the 3-year cycle of the NSF is appropriate. However, the addition of the other reviews (by NAPA, NSF, and the Inspector General, and as a