ISSUES IN MEASURING INDUSTRIAL RESEARCH-AND-DEVELOPMENT

Effective government science and technology policy requires a reliable database on industrial innovation. This paper examines a number of issues concerning the measurement of industrial R&D by analyzing the relationship between data reported in the NSF/Census RD-1 survey - the basis for the official government R&D statistics - and data reported in two other sources (firms' 10-K reports and a federal procurement database). We advance and test a hypothesis which accounts for systematic discrepancies between RD-1- and 10-K-based estimates of company R&D. We also develop estimates of the extent of measurement error reflected in the various measures of R&D, and suggest some modifications to the RD-1 survey. © 1990.