Us companies file R&D data with the Census Bureau on Form RD-1, and the data are used to create industry aggregates published annually by NSF. Many of those companies also file R&D data with the SEC, on form 10-K. The purpose of this project is to investigate the differences between these two sets of data. We expect the results of this project to lead to improvements in the quality of reported R&D data. R&D-1 is the only source of R&D data for 2600-2800 companies, those that do not report to the SEC. It is also the only source for the breakdown between basic and applied research, and between company funded and government funded industry performed R&D. The definition of R&D on the RD-1 survey is restricted to the hard sciences and engineering, so the survey is unique in that sense. It is the main source of industrial sector R&D data for US government policy makers, and is also the source for the multi-country R&D presentations of the OECD. Whether the unique features of RD-1 are worth preserving, and in what forms, and whether there should be changes in the survey, are data quality questions that are clearly important for both public and private science and technology policy determination. The SEC 10K R&D data results are reported by name of firm and are very up to date, these are advantages over the current RD-1. The project has two parts: a detailed investigation into the differences at the company level, and a set of field interviews with firms in order to discover the sources of unexplained differences in the two R&D reports. The final technical report, executive summary, and abstract should provides a stronger basis for deciding on future industry R&D data collection strategies and more effective government and private use of the resulting data.
