Physical Sciences: Astronomy, physics, chemistry, oceanography, and geosciences.
Engineering: Biomedical engineering, chemical engineering, electrical engineering, industrial engineering, material sciences, and other fields of engineering.
Life Sciences: Agriculture, biological sciences, and medical sciences.
Social and Behavioral Sciences: Anthropology, economics, geography, political science, psychology, sociology, and other social and behavioral sciences.
Further details on these fields are given in Chapter 3 on Ph.D. production.
While our study examines both science and engineering, for simplicity we sometimes use the shorter term “science” rather than “science and engineering” to refer to the fields combined. Similarly, the term “scientists” is sometimes used as shorthand for “scientists and engineers,” and “scientist” for “scientist or engineer.”
Analyses are based on scientists and engineers who participated in the 1973, 1979, 1989, or 1995 Survey of Doctorate Recipients (NSF1973–1995). In this section we describe this survey, referred to as the SDR, along with other data sources that were used to supplement this key data source.
The scientists and engineers studied in our report were respondents to the Survey of Doctorate Recipients (NSF 1973–1995). Since 1973, with support from the National Science Foundation and other federal sponsors, the National Research Council (NRC) has conducted a biennial survey of doctoral scientists, engineers, and humanists who completed the Survey of Earned Doctorates (discussed on page 17). The sample for the SDR is stratified by year and broad field of Ph.D., gender, and other demographic variables. Responses to the survey are weighted to represent the science and engineering doctoral population. Sample weights are computed as the inverse of the probability of a case being selected from the population with adjustments based on the response rate from that stratum. Currently the SDR samples about 10 percent of the doctorates from U.S. universities who remain in the United States after they receive their degree. In earlier years of the survey, the sample was larger and included some individuals with doctorates from foreign institutions. In 1995 computer assisted telephone interviews (CATI) were used to increase the response rate and two weights were computed to account for differences