National Science Foundation/Office of Science and Technology Policy

Recommendations and Actions

NSF03: Continue Automation of NSF Research Support Functions


In the last several years, the National Science Foundation (NSF) has reengineered many of its business practices in order to handle an increase in workload without a concurrent increase in staff. Five years ago, NSF's interactions with the research community were paper-intensive. NSF would mail thousands of copies of program announcements to researchers throughout the country. Research institutions would respond with 10 to 15 copies of over 30,000 proposals that averaged over 50 pages. Appropriate peer reviewers were selected by NSF scientific staff, and 8 to 12 copies of a proposal would be sent to them. The peer reviewers would mail back their written review of the proposal. Copies of the reviews of proposals would be made for panel reviewers. After a funding decision was made, an award letter would be sent to the research institution and to the principal investigator. In short, NSF was beginning to drown in its paper processes.

The agency committed itself to implementing new technological strategies to attack the paperwork problem and reduce workload burdens for NSF staff and the research community. NSF initiated a research program to address the technological and behavioral issues involved in the design and implementation of a totally electronic proposal, review, and award process. In 1987, this program, Experimental Research in Electronic Submission (EXPRES), began as a cooperative agreement with university and industry partners to explore the research issues involved in implementing a full-scale interoperable system. The goal was to build on the diversity of technology in the university environment rather than to mandate limited standards for interacting with NSF.

EXPRES researchers discovered that major breakthroughs in technology were necessary before the basic concepts of EXPRES could be achieved. In 1987, the technology was not advanced enough to meet the objectives of the research program. Additionally, the university information technology infrastructure was uneven; quality varied significantly even on the same campus. The EXPRES program, however, did validate the NSF goals of electronic document exchange. Although full-scale implementation of NSF's concept was not possible, NSF committed to the pursuit of achievable components of the overall program. The following examples illustrate some of the progress the Foundation has made in these areas:


Several years ago, NSF implemented the Scientific and Technical Information System (STIS) to provide information about NSF and its research programs to its customers.[Endnote 1] This system, which is accessible in a variety of ways, including Internet access and dial-in access, has grown over the years. Today, there are over 2,000 requests for information each day from across the United States. People in over 60 countries have obtained information from the STIS system. The system now provides immediate access to all NSF program announcements, NSF guidelines for the submission of proposals, information about the research being performed by NSF grantees, personnel vacancy announcements, and numerous informational publications.


In coordination with other research agencies, NSF has identified data elements commonly required by each agency and has developed proposal submission standards in a package that may be used by all research institutions to submit proposals.[Endnote 2] This package is available electronically and is intended to facilitate electronic submission of proposals.[Endnote 3]


After careful consideration and coordination with the research community, NSF allows reviewers to send their reviews to NSF by electronic mail. Several technical and legal issues needed to be addressed and resolved, such as the question of electronic signatures. With the cooperation of the research community, many proposal reviews are now received electronically.


Beginning in 1992, NSF implemented a program to electronically mail Federal Cash Transaction Reports to many of the major research institutions receiving NSF funding. This process greatly facilitates the institution's ability to account for expenditures made under NSF grants and expedites posting to the NSF financial accounting system.


While the agency still receives much of its external correspondence by paper, it has sharply reduced the majority of paper-intensive processes that existed five years ago. This was accomplished with state-of-the-art information systems to support mission requirements, the use of bulletin boards and electronic mail to support internal communications requirements, and the modification of proposal processing to make it suitable for automation.

A sign of the success of these initiatives is the increased efficiency in the work processes of the agency. In the last 10 years, the number of proposals received by the agency has more than doubled and its budget has nearly tripled. While this represents a significant increase in workload and accountability, the staffing at NSF has remained virtually constant. The reengineering of business practices using information systems is credited with NSF's ability to do more with less.

NSF continues to reengineer. However, its primary "production'' process, the programmatic reviews of unsolicited proposals, is still too paper-intensive and inflexible. Pilot studies are needed to test advanced information technology and new processes for exchanging information among proposers, reviewers, NSF staff, and the systems which support their work. For full value, a major study of system reengineering to explore ways to tie together processes, program design, proposal acquisition and review, selection decisionmaking, and award administration will be required. If the workload continues to increase at the current rate, increasing efficiency through automation will be the only way to handle it.


1. NSF should continue its efforts to implement advanced information technology in the proposal submission, review, award, and information dissemination process.

NSF should continue, as it has done with STIS, to make use of commonly available low-cost technologies. Advances in commercial word processing, forms applications, and mail-enabled applications will provide many of the tools needed for researchers to prepare and submit grant applications electronically and for NSF to distribute them for review and comment. Public domain search-and-retrieval software, like that used in STIS, is continuing to improve and to be adopted by university researchers. NSF already has responsibility for much of the work in development of the high-speed network that will be needed. Because of its responsibility in the area of computer science, NSF will continue to stay abreast of industry advances in encryption, network security, and digital signature standards. These are necessary for widespread acceptance of electronic grant processing. Researchers must be assured that the technology will protect their proposals against unauthorized access and electronic theft, and universities must be confident that proposals cannot be submitted without proper authorization by designated officials. Demonstrations and pilot projects will be necessary to evaluate the cost-effectiveness of new systems prior to implementation.

2. A formal means for NSF to share developments with other research agencies should be defined.

NSF's achievements in automation should be shared more widely with other agencies. Because of its expertise and experience, in consultation with the Office of Science and Technology Policy, NSF should be designated as the lead agency for automation of grants processing. NSF should use interagency coordinating groups such as the Working Group on Government Information Technology Services proposed in the NPR accompanying report on information technology to interact and form alliances with agencies who have similar needs. In addition, it should use the informal contacts forged over the years on other cooperative efforts to identify agencies that could both contribute and benefit from participation in collaborative efforts.

Through the lead agency mechanism, NSF could be identified as a test site for projects that will demonstrate the efficiencies realized through full- scale use of information technology in grant-making agencies. These pilot projects could focus on all aspects of the process from proposal receipt, review, and award to financial management and dissemination of research results.

3. NSF should be able to compete for funds to accelerate automation of grants processing.

The NPR accompanying report on information technology has proposed that a governmentwide "venture capital'' fund be available to federal agencies for innovative research and that agencies be allowed to keep some of the funds saved as a result of reengineering their business processes. If these suggestions are enacted, NSF should seek such "venture capital'' to accelerate its efforts. NSF should also explore the possibility of entering into interagency agreements with other agencies that might benefit from its work. Cost of development could be shared and more funding would be available to NSF for its development efforts.


Considerable savings in materials, mailing, and storage can be realized by moving to paperless grant processing and information dissemination. However, the government cannot set up a system that would have the effect of excluding potential users. NSF must provide for those universities that do not have the latest technology available. They must also be sensitive to the various groups within the university community who participate in grant preparation. Whereas scientific researchers may have the necessary hardware and software, the administrative side of the university, which must approve budgets and prepare the institutional sections of the proposal, may not.

NSF must balance the need to serve all parties with the need to take advantage of the best technology. With rapidly changing technology, using the least common denominator means building in obsolescence. NSF's approach in STIS was to provide several alternative methods. Like STIS, other new systems must attempt to achieve the advantages of the most advanced approach while making allowance for those who cannot afford to acquire them.

The concern for security cannot be ignored. Stories of computer fraud and break-ins are becoming routine. Advances in data encryption and network security will alleviate some concern. Widespread acceptance of electronic processing may not occur until security improvements are available.

Universities may have to change their internal processing so that the electronic grant proposals can flow through their offices and be properly signed and approved before submission. The need for change within the university may be the most difficult barrier to automating grant processing.

Whereas there are economies for the government in developing a single method for processing all grant proposals and for the university in processing all grants the same way, not all agencies may be satisfied with a generic grant proposal kit. Several research offices joined with NSF in developing their current kit, but at least one preferred to develop a kit to meet its own requirements.


While cost savings are difficult to quantify, savings in postage for the STIS system cited above amount to as much as $1 million per year in addition to the cost of preparing thousands of mailings.

After full implementation, the electronic receipt and review of proposals could result in savings of $1 million or more annually in postage alone. Currently, approximately $1.2 million is spent to mail 10 copies of the 30,000 proposals that are sent out to reviewers. Productivity gains could result in even larger savings. For full implementation, a multiyear investment to purchase equipment, revise procedures, and retrain employees will be required. Therefore, anticipated savings cannot be estimated until pilot projects have been completed. Savings may also accrue to the universities which collectively spend a minimum of $1.2 million to mail proposals to NSF.

Economies in grant processing, combined with the introduction of information technologies in the process for dissemination and use of research results, would yield substantial benefit to the government and the research community. If proposal volumes increase over the next decade as they have in the past, only intensive use of information technology will permit NSF and other agencies to maintain control over rapidly expanding program administrative costs.


1. National Science Foundation, STIS User's Guide, NSF 91-19 (rev.), version 2.1 (Washington, D.C., September 1992).

2. National Science Foundation, Grants for Research and Education in Science and Engineering: A Proposal Guide and Forms Kit, Draft (June 5, 1993), Appendix H.

3. National Science Foundation,The NSF Electronic Proposal Submission Project, Draft (May 5, 1993).

Return To Report Index

Who We Are |||Latest Additions |||Initiatives |||Customer Service |||News Room |||Accomplishments |||Awards |||"How To" Tools |||Library |||Web Links

Reinvention Comments
Technical Comments