recognize its potential applicability to practical problems, which typically requires input from potential users. Expert review should be used to assess both basic-research and applied-research programs. A balance must be achieved between having the most knowledgeable and the most independent individuals serve as reviewers.
Pluralism is a major strength of the US research enterprise. But better communication among agencies would enhance opportunities for collaboration, keep important questions from being overlooked, and reduce inefficient duplication of effort. Identifying a single agency to serve as the focal point for particular fields of research could bring needed cohesion to the federal research effort. In some cases, it may make sense to adopt the model used at the Defense Advanced Research Projects Agency (DARPA), in which the desired end product or technology is defined before research begins, so that research teams can coordinate their efforts to solve the problem.
To improve the effectiveness of federal research and development programs, the federal government could:
Ensure that federal agencies include research programs in their strategic plans and that they evaluate the success of those programs in performance reports.3
Evaluate research in terms of quality, relevance, and leadership. For basic research, include assessments of the historical value of basic research in contributing to national goals.
Evaluate how well research programs develop human resources and the quality, relevance, and leadership of the programs. If federal research activities do not continue to produce a flow of well-educated scientists and engineers, the capability of an agency to fulfill its mission will be compromised and the knowledge learned and technology developed will be lost.
Establish a formal process to identify and coordinate areas of research that are supported by multiple agencies. A lead agency should be identified for each such field, and that agency should be responsible for ensuring that coordination occurs among the agencies.
Investigate and experiment with innovative ways of managing research, such as establishment of long-term research goals, very flat management structures, multidisciplinary teams, and a focus on technology transfer (these are some of the approaches that have met with considerable success at DARPA).4