Appendix A.
Technical Notes

Survey Description top

All statistical data presented in this paper are from the Survey of Earned Doctorates (SED). This survey, which is conducted annually under the sponsorship of the National Science Foundation (NSF) and four other Federal agencies, is a census of recipients of research doctorates at all accredited universities and colleges in the United States. Research doctorates include doctoral degrees such as the Ph.D. and D.Sc., but exclude first-professional degrees such as the J.D. and the M.D.

The survey data are collected directly from the individual research doctorate recipients. Questionnaires are distributed, with the cooperation of the various graduate schools, to those people completing their research doctorates. The data for a given year include responses from all persons whose doctorates were awarded in the 12-month period ending on June 30 of that year. A copy of the questionnaire used for the 1994-95 survey is attached as Exhibit A.

Approximately 94 percent of the 1994-95 cohort of doctorate recipients responded to the questionnaire. Since partial data from public sources are obtained for survey nonrespondents, the counts for conferred doctorates by field are considered relatively complete. Data for this report were drawn from the responses to items 5, 7, 9, 10, 13, and 17 of the 1995 questionnaire.

Missing Data top

Missing data items are coded as missing and are not imputed. In item 17, respondents were asked to indicate their primary and secondary sources of support and to check all other sources from which support was received. The overall response rate to the sources of support was 94 percent, but only 76 percent reported a primary source of support and 63 percent a secondary source. That is, 63 percent indicated both a primary and secondary source of support, 13 percent indicated a primary source of support, but not a secondary source of support, and an additional 18 percent checked multiple boxes on the source of support question, but did not indicate which were primary or secondary sources of support. Thus, a total of 94 percent either checked a box and/or indicated one or more modes of support as primary or secondary. The 13 percent who indicated a primary source of support but not a secondary source of support includes both people who had no other support and also those who checked other sources of support, but did not designate a secondary source. Respondents not reporting any source are excluded from the tables presenting any source of support, those not reporting a secondary source are excluded from tables reporting secondary source of support, and those not reporting a primary source are excluded from tables reporting primary source of support. See appendix tables A2 and A3 for differences between those missing and not missing primary source of support and any source of support on other variables used in this report.

Item Response rate (percent)
5 (Sex) 100
7 (Citizenship) 97.9
9 (Race/ethnicity) 98.9
13 (Field of study) 100.0
17 (Any source of support) 93.6
17 (Primary source of support) 76.2
17 (Secondary source of support) 63.2

Validity of Data on Sources of Support top

The National Research Council (NRC), at the request of the Federal sponsors of the SED conducted a study in 1994 to assess the validity of item 17, sources of support. In the study, responses to the SED were matched with records of grantors of support money to graduate students. The study found that doctorate recipients can reasonably accurately identify the type of financial support they had in graduate school (e.g., RA, TA) but not necessarily the source of that support (e.g., NSF, National Institutes of Health, Ford Foundation, university funds). (NRC 1994.) For this reason, the 32 possible responses to item 17 were recoded into 7 "modes" of support that reflect the type of funding but not the source of funding. (The question on sources of support was changed in later versions of the SED.)

Data Recodes top

Data from the file were recoded into the categories used in this report as follows.

Support mode was determined from item 17 as follows:

Mode Code on questionnaire:
Fellowship 33[24] , 53, 70, 71, 73, and 78
Traineeship 12, 21, 40, and 44
Research assistantship 11, 22, 32, 52, and 62
Teaching assistantship 10
Own funds 01, 02, and 03
Loans 80, 81, and 89
Other 14, 19, 29, 49, 60, 69, 90, 91, 92, and 99
Missing None specified

Primary mode of support was determined from item 17, source indicated as primary; if no primary source was specified, it was considered missing.

Secondary mode of support was determined from item 17, source indicated as secondary; if no secondary source was specified, it was considered missing.

Discipline was determined from item 13, field of doctorate study. The National Research Council's Office of Scientific and Engineering Personnel field codes used to indicate study field were assigned to the discipline codes reported herein using the NSF Computer-Aided Science Policy Analysis and Research (CASPAR) database crosswalk shown in Exhibit B.

Citizenship was determined from item 7 using the following crosswalk:

Citizenship Code on questionnaire:
U.S. citizen or permanent resident visa 0, 1, and 2
Foreign student (on temporary visa) 3
Missing None specified

Race/ethnicity was determined from items 9 and 10, using the following crosswalk:

Race/ethnicity Code on questionnaire:
American Indian or Alaskan Native (I) Item 9, code 0; item 10, response "no"
Asian or Pacific Islander (A) Item 9, code 1; item 10, response "no"
Black, non-Hispanic (B) Item 9, code 2; item 10, response "no"
White, non-Hispanic (W) Item 9, code 3; item 10, response "no"
Hispanic (H) Item 9, any; Item 10, response "yes," codes 0, 1, or 2
Other (O) None specified or multiple responses

Sex was determined from item 5:

Sex Code on questionnaire:
Male 1
Female 2
Missing None specified

Carnegie codes were assigned to the doctorate-granting institutions reported in item 13 based on the Carnegie classification system. (The Carnegie Foundation 1994)

The doctorate-granting institutions reported in item 13 were categorized as public or private institutions based on their reporting on the institutional control item in the National Center for Education Statistics IPEDS surveys.

Table A1. Number of 1995 S&E Ph.D. recipients by primary source of support and selected characteristics

Table A2. Number and percentage of 1995 S&E Ph.D. recipients by primary mode of support missing, and selected characteristics

Table A3. Number and percentage of 1995 S&E Ph.D. recipients by any mode of support missing, and selected characteristics

Table A4. Percentage distribution by field, race/ethnicity and citizenship of 1995 S&E Ph.D. recipients


[24] A number of these may be "false positives." The NRC Validation Study (NRC 1994) showed that 39 percent of doctorate recipients listing NSF fellowship were not listed in the NSF files as having received one.

Previous Section Top of page Next Section Table of Contents Help SRS Homepage