93 



SUPPLEMENT 3 / 69 



Box n.8 

 Methods for Selecting Federal R&D Performers and Projects 



A number of approaches are used to decide which R&D projects receive federal funds, 

 how much should be spent, and who should conduct the work. The approach used depends 

 on the nature of the work, its relationship to specific government missions, and the history and 

 culture of different research commimities, programs, and agencies. 



Traditionally, agencies such as the National Science Foundation and National Institutes of 

 Health that make grants to universities to support fundamental scientific and engineering re- 

 search have used some form of prospective />eer review to judge the quality of competitively 

 submitted project proposals. Peers are established working scientists or engineers from di- 

 verse research institutions who are deeply knowledgeable about the field of study and who 

 provide disinterested technical judgments as to the competence of the researchers, the scien- 

 tific significance of the proposed work, the soundness of the research plan, and the likelihood 

 of success. Since the early 1980s, NSF has asked peers also to take into account the utility of 

 the proposed research to the nation and its potential for contributing to graduate education 

 and to the infrastructure of science itself. Since the middle 1980s, NSF has used the term merit 

 review to indicate both that proposals are judged on their merits and that NSF program officers 

 also have the authorit>' to take into account various general policies of the Foimdation when 

 making awards. NIH makes limited use of a second level of review by institute councils that 

 take into account national relevance and direction. Some programs in other departments and 

 agencies, including the Department of Energ>-, National Aeronautics and Space Administration, 

 U.S. Department of Agriculture, and Depanment of Defense, employ variants of peer or merit 

 review. TTie various departments and agencies differ in the degree to which their program 

 managers are bound to follow the recommendations of peer and merit reviewers in making 

 awards Practices van,- even within the NSF and NIH across research fields and areas. 



Other agencies, including the Office of Naval Research and Advanced Research Projects 

 Agency, use a strong program manager approach to prospective assessment of the scientific 

 or technical merit of research proposals, particularly those that are of a more fundamental 

 nature. Strong technical staff members have responsibility for being well informed about the 

 state of the an of their specialties and for idcntif>'ing and recruiting investigators to conduct 

 research that they deem to be of greatest importance to the agenc> 's mission. Program manag- 

 ers often devote considerable energy- to soliciting the views of peers about these matters but 

 usually are not bound to heed their advice. 



Agencies seeking to contract for performance of R&D projects of direct interest to the 

 government in mdustrial or other nongovernment organizations typically conduct competitive 

 procurements for R&D services, using government technical employees and, occasionall>', 

 consultants to judge the prospective merit of contract proposals. This approach has much in 

 common with standard procedures used by the federal government to procure other goods 

 and services 



Federal laboratories use several approaches to project seleaion. In most cases, however, 

 on-site technical and unit managers share responsibility with agency program managers 

 for selecting project topics and performers. In some cases, the external p>eer commimity is 

 asked for advice on specific projects, and in other cases, on an entire program of aaivity. 

 Sometimes such advice is obtained on a prospective basis; sometimes it is obtained via formal 

 reviews of ongoing or completed research activities. In some agencies and some programs, 

 proposals to begin new projects at the federal laboratories compete across several laboratories 

 or even with proposals submitted from academia or industry. 



Formula funding is used by a few programs, principalh' in the USDA, to allocate R&D 

 funds among performing institutions such as the land-grant colleges and imiversities. 



Executive agency decisions about R&D allocations to institutions and projects hive in- 

 creasingly been specified in detail by congressional appropriations committees. These alloca- 

 tions often do not reflect the considered judgments of scientific experts or the fimding agen- 

 cies, and they often are determined instead by individual members of Congress acting on be- 

 half of constituents 



