- periodic census-type activities. The federal portfolio is not static, however, and should be reconsidered from time to time. To derive maximum value from this national aggregate statistical program, a premium should be placed on the ability to link these data with other innovation and economic data and on consistency over time to produce useable time series and longitudinal data.
Draw on Outside Expertise and Capacities to Improve Data Programs
- Explore public-private partnerships. The use of public-private partnerships and cooperative agreements for data collection and reporting should be explored, particularly in areas where program-specific data or exploratory, developmental efforts are needed. The advantages and disadvantages of such agreements and obstacles to them need to be studied.
- Form a consultative body. There is no standing mechanism for the NSF, Department of Commerce, and other agencies to acquire expert advice on issues related to innovation data collection, integration, and reporting. A specific suggestion was made that a technical advisory committee to the Science Resources Studies Division of the NSF should be reinstituted.
Improve Existing Data
- Broaden industry coverage. Research and innovation databases, such as the NSF's R&D survey, should incorporate emerging, high-growth, technology-intensive industries such as telecommunications and biotechnology and industries across the service sector—financial services, transportation, and retailing, among others. Although the NSF has taken constructive steps to adjust its data collection to changes in the structure of the economy, additional analysis of innovation processes in these industries is needed to inform the development of appropriate survey instruments.
- Collect business segment detail. Existing survey programs, such as the NSF R&D survey, should collect information at the business unit level of corporate activity rather than for the firm as a whole. In an economy in which large, complex, multiproduct firms perform most R&D activities, the business unit represents a more homogeneous set of activities, whereas combining responses on a range of variables for a variety of products and processes will obscure significant industry-specific conditions that affect technological innovation. Moreover, managers at the business unit level are likely to be better informed about innovation-related investments and performance measures than are corporate headquarters officials.