About 2,000 institutions compete for National Science Foundation (NSF) research funds which are awarded mainly through grants to colleges and universities. In fiscal year 1977, academic and other nonprofit institutions received about 10,200 grants totalling over $570 million. In deciding whether to fund proposals, NSF officials get comments from experts (peer reviewers) in the fields of the proposals.

Accountability in the proposal evaluation process is still a problem in spite of actions taken by the NSF over the past 2 years to improve the process. Action review boards, made up of officials in each of the six directorates awarding grants, were established to ensure that proposal evaluations complied with policies and procedures. However, some grants reviewed by the boards lacked proper justification; some grants were excluded from review; reasons for selection of specific peer reviewers and the disposition of reviewer's comments were not adequately documented; there was no periodic reporting of boards' findings; and board members were not independent of the evaluation process. At GAO suggestion, NSF established an Office of Audit and Oversight which samples proposal actions after they are finalized to evaluate documentation and compliance with procedures. However, preaward controls are also needed. Other weaknesses in the process involve lack of consistency among NSF directorates in documenting the bases for selecting peer reviewers and how their comments are handled and officials responsible for approving decisions who are not expert in proposals' subject areas.