The estimates reported for the Survey of Earned Doctorates (SED) are simple tabulations of all available information, with no adjustment for nonresponse. Therefore, differences in response rates from year to year can produce numerical fluctuations that are unrelated to real trends.
Although response to the SED has been as high as 95 to 98 percent over time, it declined to 92 percent during the 1980s. In an effort to improve the response rate, the survey methodology was modified in the years after 1989. Response rose, stabilizing around 95 percent from 1991 to 1995. However, the response rates for 1996 and 1997 were 92.8 percent and 91.5 percent, respectively. This year, the response rate again was 91.5 percent. (Note: These percentages represent self-report rates, that is, the proportion of questionnaires completed by doctorate recipients. While survey forms containing partial information filled in by either the doctoral institution or the survey contractor are not included in these rates, tables in this report incorporate the available data from these sources.) The self-report rate for 1998 may increase slightly in the next year if additional questionnaires are received from doctorate recipients. See appendix C for a table giving survey response rates from 1967 to 1998.
Item response rates have shown a pattern of improvement since 1990a natural consequence of the increase in the overall self-report rate, as well as a result of format revisions to the questionnaire and follow-ups for missing information. In 1990, new follow-up procedures were implemented to increase coverage of several variables: birth year, sex, race/ethnicity, citizenship status, country of citizenship, baccalaureate year and institution, and postgraduation plans. Response rates for these variables have since improvedespecially for citizenship and race/ethnicity, resulting in an increase in the reported numbers of minority Ph.D.s. (However, for 1997 and 1998 the citizenship response rate of 92 percent is lower than it was in 1990-1996.) Whether or not individuals completed the survey questionnaire, the following four data items are available for most all recipients: sex and Ph.D. institution, field, and year.
The data for a given year are updated the following year with any responses received after survey closure. Postsurvey adjustment was most significant for 1990 and 1991 Ph.D.s, with the largest impact on the number of blacks. For both of these years, the total number of black Ph.D.s increased by about 7.5 percent in the year after survey closure. The survey cycle was then extended to allow receipt of more follow-up information before closure, resulting in much smaller postsurvey adjustments for the 1992 through 1995 data (a 1.4 percent increase in black Ph.D.s for 1992, a 0.2 percent increase for 1993, a 0.5 percent increase for 1994, and a 1.5 percent increase for 1995).
Adjustments to data are presented in reports subsequent to the initial report for a survey. For example, updates for 1994 appeared in Summary Report 1995. Updates to 1998 data will be presented in next year's report.