Science is getting bigger, faster, and more expensive, and simultaneously heightening the expectations of the public. One pivotal factor contributing to these changes IS the expansive growth of the research infrastructure. Borrowing from economics, infrastructure is often viewed as a “common good,” but it is by no means a “free good,” and indeed it represents a huge investment of capital by universities as well as public and private funding agencies. In essence, infrastructure competes for a piece of the same pie as does investigator-initiated research. It is for this reason that researchers are often divided on the merits of investing in research infrastructure.
The psychological sciences are at important juncture in how they choose to prepare for the future. Most researchers rely on various forms of infrastructure that they take for granted, such as access to the Internet, development of state-of-the-art statistical software, and even special training opportunities to acquire new technical skills. During the past decade, the physical and biological sciences have made enormous investments in infrastructure, such as data bases, collaboratories, shared research facilities, e.g., Large Hedron Collider (LHC), that have led to enormous returns in research productivity. A question for the psychological sciences is whether similar investments should be made, and if so, how should projects be prioritized and funded.
I would like to stimulate this debate by briefly reviewing some of the issues that are associated with different forms of infrastructure.
- Large data bases that are publicly accessible on the web provide an opportunity for researchers to pool enormous amounts of data to begin addressing questions that were heretofore unanswerable. A very exciting example of this form of infrastructure is the CHILDES language data base that has been developed at Carnegie Mellon University. Data from hundreds of language acquisition studies are now stored in a common format and provide a much larger corpus of data for answering a host of questions about language acquisition. Similar data bases could be assembled to provide access to data from brain imaging studies, crosscultural studies of cognition, etc .
The technology for developing these data bases and making them accessible on the web is advancing rapidly, and some the most recent advances in digital libraries are already contributing to the organization and availability of large nationally-representative data bases, such as the General Social Survey and the Panel Study of Income Dynamics, supported by the Social, Behavioral, and Economic Sciences (SBE) Directorate at the National Science Foundation (NSF). Of course, the development of these data bases comes with a considerable cost in terms of time and money. In addition, it is necessary for researchers to be willing to share data that they were responsible for collecting. For these reasons, the decision to invest more resources in data bases goes well beyond a question of funding, because it ultimately demands a different set of cultural values for collecting and sharing data.
- Multidisciplinary Centers are becoming increasingly common in science as the questions we address cut across traditional scientific boundaries. The clear advantage of these centers is that they provide an opportunity for bringing together a critical mass of experts interested in common problems, such as violence, environmental decision making, or cognitive science. Prior to coming to NSF, I was skeptical about the added value of centers; however, after observing first-hand the impressive accomplishments of some of the centers supported by SBE, I have been converted. A novel variation on the traditional center is the “virtual center” which is possible via the Internet. Indeed, the availability of the high-speed Internet adds further to the flexibility of centers, because it is now possible to create collaboratories where experiments can be conducted simultaneously at different locations with the same stimulus materials.
One of the greatest advantages of interdisciplinary centers is that they provide new forms of multidisciplinary training to the next generation of scientists in a way that could not be accomplished easily by anyone or small group of researchers. Perhaps, the greatest impediment toward the establishment of these multidisciplinary centers is the coordination necessary between universities and funding agencies to commit sufficient funds and faculty positions to provide researchers with the assurances they need to justify the time and effort involved in applying for such centers. In order to help provide this assurance, NSF funds many different types of centers, including Science and Technology Centers that are eligible for up to $40 million over a period of 10 years.
- Scientific instrumentation and training are complementary investments for assuring that the psychological sciences are supported by the most advanced research equipment as well as the expertise necessary to use this equipment to maximum advantage. At NSF, research instrumentation falls into at least 5 categories: (I) platforms and observational systems (e.g., neural imaging equipment, observational coding systems); (2) computational systems (e.g., supercomputers. mass storage devices, visualization systems); (3) laboratory and analysis systems (electron microscopes, statistical software, image processing); (4) information systems and data bases (digital libraries, large surveys); (5) communications and network systems (vBNS, Internet). Traditionally, the psychological sciences have been avid consumers, but not initiators of research instrumentation. In order to reverse this trend, it is necessary to establish long-range planning and priorities for the discipline. One model for this type of planning is provided by the physical sciences where consortiums of researchers are formed to support the funding and management of new facilities, such as telescopes or particle accelerators. Clearly, a precondition for proposing such facilities are rational management plans that will ensure continued funding and accessibility of the equipment beyond the construction phase or initial purchase of the equipment.
For the past few months, the SBE Directorate has been engaged in a review of its infrastructure support to the scientific community. We anticipate that this review will continue through the summer, and we invite researchers and professional societies to contribute to this process by identifying for themselves how much and what kinds are infrastructure will be needed to continue to push the envelope of their science in the next decade. In order to facilitate this process, the current Social, Behavioral, and Economic Research Division is providing funds for workshops or working groups to canvass researchers and to develop a clearly articulated assessment of needs. It would also be useful to receive input on how such large proposals should be reviewed vis-a-vis research proposals. I believe that it is critical for the psychological sciences to contribute to this review, because the final decision by the SBE Directorate should be informed by the wisdom, experience, and knowledge of all researchers who have a stake in the future of their science.