U.S. Surveys

In 1994 the NSF conducted a pilot innovation survey of manufacturing firms to explore the value of such an effort. The Manufacturer's Innovation Survey, administered by the Census Bureau for NSF, asked questions on firm innovations, sources of information, objectives of innovation, acquisition and transfer of technology, and R&D activity. The NSF is considering what steps, if any, to take beyond this experimental effort.

A privately conducted survey, administered in 1994 by a Carnegie Mellon University research team headed by Wesley Cohen, collected information from manufacturing firm R&D laboratory or unit directors on the determinants of R&D activity and performance. Data were collected on such issues as firms' use and perceptions of the effectiveness of different mechanisms for protecting intellectual property in order to appropriate the returns to their innovations, sources of new technical information, interactions with competitors, and the speed at which new products or processes are imitated. Because this survey builds on a 1987 survey conducted at Yale by Richard Levin, Alvin Klevorick, Richard Nelson, and Sidney Winter, which also examined appropriability conditions in the U.S. manufacturing sector, it has been called the "Yale II" project. A comparison of results yields interesting differences. For example, Cohen reported that the preferred means of protecting intellectual capital changed dramatically from 1987 to 1994, with the relative effectiveness of patents declining and that of trade secrecy increasing. Unlike the first Yale survey, the Carnegie Mellon survey instrument was designed to enable comparisons across countries as well as across and within industries. The survey has been duplicated and administered in Japan, and many of the questions have been incorporated in a European questionnaire administered by the research organization MERIT.

On the basis of his experience with the Yale surveys, Cohen offered a number of suggestions for designing and administering a future U.S. innovation survey or surveys:

  • In light of the prevalence of multiproduct firms, most data elements should be gathered at the business unit rather than at the enterprise level; a business unit reflects the firm's activities within a particular industry, which is typically defined by the four- or at least three-digit Standard Industrial Classification.
  • Data on some variables such as R&D and sales should be collected annually. Other variables, such as those representing appropriability conditions or information flows, could be assessed much less frequently, for example, at five-year intervals.
  • Following the Canadian model, innovation surveys should employ several questionnaires addressed to firm officials who have the most relevant knowledge.


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement