vation, an effort sponsored by the National Aeronautics and Space Administration (NASA) and the National Science Foundation's (NSF's) Division of Science Resources Studies (SRS). At the NSF's request, the workshop focused on its survey of R&D spending, its recent pilot innovation survey, and the various industrial indicators (patents, patent citations, trade in goods and intangibles, etc.) that are reported every two years in Science and Engineering Indicators.1
The workshop nevertheless took as comprehensive a view as possible of innovation data needs and issues. The principal exception was that data on science and engineering personnel characteristics, as indicators of innovative activity and capacity, were reserved for later consideration. Because of the workshop's broad scope, a number of issues and questions were raised, but discussion was deferred to possible follow-on meetings and working groups. In this way the workshop served as an organizing tool for further efforts to improve the national collection and interpretation of innovation information.
There were several reasons for holding this workshop. First, increasing importance is being attached to technology and innovation in national policy debates about economic performance and international competition. Second, major changes are under way in the structure of advanced industrial economies (for example, the growth of service industries relative to manufacturing industries) and in the composition, location, and organization of industrial science and technology activities. Third, changes in public-sector R&D investment and other microeconomics policies ranging from intellectual property protection to economic regulation and corporate taxation are also occurring. Finally, issues of measurement, data collection, and valuation are receiving more attention internationally, for example, in the Organization for Economic Cooperation and Development (OECD), which has sponsored work to improve understanding of the ''knowledge-based economy."
However timely, this review poses major challenges. First, progress in developing a theoretical understanding of innovation has been hampered by difficulties in measuring and evaluating its outputs as determinants of industrial performance and economic growth. As a consequence, it is uncertain how to capture changes in a volatile industrial environment that gives off mixed signals. For example, on the one hand, U.S. industrial firms appear to be cutting back on long-term research and the infrastructure to support it. On the other hand, the competitive performance of much of U.S. industry appears to have improved significantly in recent years. Second, because of the distance between the language and decisions of policy makers and the language and practical constraints of analysts