APPENDIX A
1991 SURVEY CHANGES
In 1991, a number of methodological changes were made to the Survey of Humanities Doctorates. These changes distinguish 1991 data from those collected in earlier years. The changes are described below.
-
Sampling Frame. The 1991 sampling frame was redefined to include only doctorates who were 75 years of age or younger in September 1991. In prior surveys, the frame had excluded individuals on the basis of years since the award of their degree, not on the basis of age. Under that definition, individuals who had earned their degrees more than 42 years prior to the survey year were excluded (or at about age 72). This change was made to accommodate growing interest in retirement issues.
-
Sample Design and Sample Size. Because of budgetary constraints, the initial 1991 sample was cut in half--from 17,716 to 8,894 sample cases. At the same time, it was restratified into fewer sampling cells and greater homogeneity in sampling rates across strata was introduced. These changes were made to reflect current analytic interests.
-
Response Rates. The resources saved as a result of the sample size reduction were redirected toward increasing the response rate, which had fallen to about 55 percent in 1989. The approach was two-pronged. First, the mail survey was made more productive through (a) extensive efforts to locate and update addresses for individuals in the sample, (b) reformatting of the survey questionnaire (the content did not change), and (c) the use of personalized mailing techniques. Second, a sample of about 60 percent of the nonrespondents was followed up by telephone. As a result of these efforts, the overall response rate to the 1991 survey increased to 78 percent (unweighted). Most likely, this reduced the effects of nonresponse bias in the 1991 survey estimates.7
-
Reference Period. Due to the Change in the schedule for fielding the survey (traditionally, mailing had commenced in March or April of the survey year), the reference date for survey items was moved from February to September. Thus, 31 months elapsed between the 1989 and 1991 surveys, compared with 24 months between previous surveys.
7
A study conducted on a sample of science and engineering doctorates who were nonrespondents to the 1989 SDR showed bias due to the low response rate in several variables, including location, type of employer, primary work activity, and tenure status. The findings were that the size of the U.S. population of doctorates was being overestimated, as were the numbers of those employed in the academic sector and the numbers of those teaching. To the extent that these biases existed in estimates of the humanities population, they should be minimized by the higher response rate in 1991. For additional information, see S. Mitchell and D. Pasquini, Nonresponse Bias in the 1989 Survey of Doctorate Recipients: An Exploratory Study, Office of Scientific and Engineering Personnel, National Research Council, Washington, D.C., 1991.
-
Definition of Degree Field. One additional difference between the 1991 survey and earlier surveys needs to be noted. The humanities sample has always been drawn from a population defined by the degree field chosen by the respondent at the time of degree award (and therefore limited to the DRF taxonomy list that had been used to collect data for the population at that time). However, the 1977, 1979, and 1981 surveys gave respondents in the humanities sample the opportunity to reselect their degree field from a revised list. In subsequent Profile reports, these individuals were counted by field on the basis of their revised responses. However, because revised responses had not been collected from every sample member, an inconsistency was introduced between the “field” used in sampling and reporting. Therefore, the decision was made in 1991 to classify humanities doctorates according to the field they selected at the time they earned their degree.
As a result, doctorates who had revised their degree fields between 1977 and 1981 reverted to the field in existence when they completed the DRF form. Three fields were particularly affected: American history, “other history,” and speech/theater. In American history, doctorates from the earlier cohorts who had revised their field to American history reverted to the “other history” category, because American history was not on the list when they earned their degrees. This caused a significant drop in the number of doctorates reported in American history in 1991, and a related increase in the number reported in “other history.”
In addition, the number of doctorates in speech/theater showed a large increase in 1991 because older doctorates who had earned degrees in audiology and communications, and who had been counted as nonhumanities doctorates in the 1977, 1979, or 1981 reports, reverted to the category “speech.” (This is because audiology and communications were not on the list at the time they earned their degrees.)
Table A-1 illustrates how trend lines were affected by the changes mentioned above. The columns labeled 1977 through 1989 show published numbers by field in those years. In general, each field experienced small but steady growth in the number of doctorates reported in each year. In 1991, however, this trend was reversed. Most fields (with the exception of speech/theater and “other history”) show only slight growth, or even a decline in numbers.
Thus, as indicated in the Notice of Methodological Changes at the beginning of this report, 1991 estimates are not comparable with estimates shown in earlier reports. Readers are cautioned not to display 1991 data beside published data from earlier years in order to examine trends. Instead, readers are referred to the indexed time-series tables in Appendix D for information about how the humanities population has changed over time.
TABLE A-1 Changes in the Humanities Population, by Field of Doctorate, Caused by Methodological Differences (1977-1991)
Field of Doctorate |
1977 |
1981 |
1985 |
1989 |
1991* |
All Fields |
66,400 |
78,600 |
90,600 |
100,700 |
100,300 |
American History |
5,400 |
8,500 |
8,800 |
10,000 |
6,300 |
“Other History” |
11,400 |
11,000 |
12,500 |
12,700 |
15,500 |
Art History |
1,500 |
2,100 |
2,700 |
3,100 |
3,100 |
Music |
3,700 |
5,200 |
6,700 |
8,300 |
8,700 |
Speech/Theater |
3,200 |
3,200 |
3,800 |
4,200 |
5,400 |
Philosophy |
5,400 |
6,200 |
7,000 |
7,500 |
7,500 |
English and Amer Lang/Lit |
18,500 |
21,700 |
23,800 |
26,000 |
25,900 |
Classical Lang/Lit |
1,700 |
1,800 |
1,900 |
2,000 |
2,100 |
Modern Lang/Lit |
11,800 |
14,300 |
16,000 |
17,400 |
16,400 |
“Other Humanities” |
3,800 |
4,600 |
7,500 |
9,600 |
9,500 |
NOTE: These numbers are for the purpose of illustration only and are not valid indicators of trends in the humanities population. For a listing of the fields in “other history” and “other humanities,” see footnote 1 in the Introduction to this report. Numbers are rounded to the nearest hundred; therefore, subcategories may not add to totals. *Estimates for 1991 incorporate the methodological changes enumerated in this appendix; they are based on mail and telephone data. |