NSF sponsors a variety of surveys designed to collect data on the human resources devoted to science and technology in the United States, including information on worker inputs for R&D. Surveys directed at employers or institutions focus on the amount of time devoted to the performance and management of R&D. These data are reported in terms of person-years, or full-time equivalent (FTE) R&D jobs. Surveys directed at individuals collect data on self-reported primary work activity; that is, the activity on which a scientist/engineer spends the largest proportion of time but that is not necessarily full-time. The 1994 National Patterns was the first to include revised estimates of the total number of scientists and engineers (S&Es) engaged primarily in R&D activities. The national totals include an FTE count of S&Es employed by industry, the total number of Federal employees whose primary work activity is research or development, an FTE estimate of graduate students' research activity, and the number of doctorate-holding S&Es working in educational or nonprofit organizations who self-report their primary work activity as research, development, or (up to 1993) the management of R&D work. These concepts are further described in appendix A.
Approximately 987,700 scientists and engineers were employed in 1995 on R&D activities in the United States (appendix table B-27). This figure reflects an annual growth rate of 1.3 percent from the 1993 level of 962,700. It reflects a 2.1-percent annual growth rate over the 1985 figure of 801,900, the first year for which revised national tabulations have been derived.
In 1995, industry employed 79.9 percent of these R&D personnel. Transportation equipment accounted for 17.6 percent of the industry total (789,500), and nonmanufacturing for 27.0 percent. This stands in sharp contrast to only eight years earlier, 1987, when the transportation equipment industry had nearly twice as many R&D S&Es as nonmanufacturing (187,800 versus 99,200, respectively). The Federal Government employed 5.5 percent (53,900) of the Nation's R&D S&Es in 1995, while the academic and nonprofit sectors accounted for the rest.
In 1981, the number of scientists and engineers engaged in R&D per 10,000 labor force was just under 62. This ratio climbed continually through the 1980s, reached a peak of 76 per 10,000 in 1991, and has not changed significantly since then.
These personnel estimates make it possible to gain a rough perspective on the changing cost of doing research. In 1985, the Nation spent an average of approximately $143,000 on R&D per R&D scientist and engineer, which includes salaries, fringe benefits, materials, supplies, and overhead for R&D activities. By 1995, this cost rose at roughly the same rate as inflation to $185,000. (See appendix table B-23 for industry-specific ratios.)
In 1995, the latest year for available data, there were approximately 484,780 doctoral scientists and engineers employed in the United States (appendix table B-28). This total represents a 2.5-percent annual growth over the 344,000 reported for 1981. Holders of doctorates in sciences in 1995 greatly outnumbered holders of doctorates in engineering, 406,130 versus 78,650 respectively, with the number for sciences including 143,390 under "social and related sciences."
Forty-one percent of all science and engineering doctorate-holders reported R&D as their primary work activity in 1995. Basic research, as a primary activity, accounted for 13.7 percent of all scientists and engineers holding doctorates; applied research 20.2 percent; development 4.9 percent; and design 2.3 percent. Teaching as a primary activity accounted for 22.1 percent of doctoral scientists and engineers, with the remaining 35.7 percent being distributed among management/sales/administration (16.4 percent), computer applications (4.4 percent), and professional services/other (14.9 percent).
Scientists holding doctorates in 1995 were more likely to have basic research as their primary activity (15.3 percent of all scientists hold doctorates) than engineers holding doctorates (4.9 percent). Consequently, in comparison to engineers, scientists holding doctorates were less likely to have applied research, development, or design as a their primary activity. The respective percentages for doctoral scientists and engineers with regard to these primary activities were 19.3 percent versus 24.8 percent for applied research; 3.5 percent versus 11.9 percent for development; and 1.3 percent versus 7.9 percent for design.
Doctoral engineers reported more involvement in management, sales, and administration as a primary work activity (20.1 percent) than doctoral scientists (15.7 per-cent). In contrast, scientists reported more involvement in teaching than engineers, i.e., 23.2 percent versus 16.1 percent, respectively (figure 17).
 See appendix A for details on the FTE R&D scientists and engineers series.
 The category of R&D called "design" refers here to design in the context of engineering, e.g., the design of equipment, processes, structures, and prototype models, as opposed to "design" in other contexts, e.g., the design of entire research programs, expenditures, etc.
 This last category includes "production, operations, maintenance (e.g., truck driving, machine tooling, auto/machine repairing)" and "professional services (health care, counseling, financial services, legal services, etc.)"-see, "Survey of Doctorate Recipients, 1995", page 5, in National Science Foundation, Division of Science Resources Studies, Characteristics of Doctoral Scientists and Engineers in the United States: 1995, NSF 97-319, R. Keith Wilkinson (Arlington, VA 1997).