Chapter 5:

Academic Research and Development: Financial and Personnel Resources, Integration With Graduate Education, and Outputs

Academic Doctoral Scientists and Engineers

This section examines major trends over the 1973-95 time period regarding the composition of the academic S&E workforce, its primary activities (teaching vis-a-vis research), and the extent of its support by the Federal Government. For a discussion of the nature of the data used here, see "Data Sources: Nature, Problems, and Comparability." [Skip Text Box]

Data Sources: Nature, Problems, and Comparability top

The data used in this section to describe the employment, characteristics, and activities of academic doctoral scientists and engineers derive mainly from the Survey of Doctorate Recipients (SDR) and in part from the National Study of Postsecondary Faculty (NSOPF).

SDR is a sample survey conducted biennially since 1973 under the sponsorship of the National Science Foundation and several other federal agencies. The survey underwent several changes in 1991, and again from 1993 forward; these affect the comparability of data from these years with those of earlier periods. Through 1989, the sample included three major respondent segments: (1) recipients of S&E doctorates from U.S. institutions; (2) a small number of doctorate-holders in other fields who were working in S&E in the survey year; and (3) a small number of people with S&E doctorates from non-U.S. institutions.

Starting with the 1991 sample, only recipients of S&E doctorates from U.S. universities were retained, and persons over 75 years old were ruled out of scope. Further, sampling strata and sample size were reduced in an effort to improve response rates within budget constraints. Other changes in data collection also were introduced, most notably the use of computer-assisted telephone interviewing, which resulted in much higher response rates than had been attained previously. A 31-month interval elapsed between the 1989 and 1991 surveys instead of the usual 24 months; the interval between the 1991 and 1993 surveys was 20 months.

Methodological studies to assess the full impact of these changes on overall estimates and individual data items remain to be conducted. Preliminary investigations suggest that SDR data permit analysis of rough trends, provided comparisons are limited to recipients of S&E doctorates from U.S. institutions. This has been done herein, with data structured in accordance with suggestions offered by the National Research Council's Office of Scientific and Engineering Personnel, which has conducted all of these surveys through 1995. Nevertheless, in the text and tables presented here, apparently interesting but small statistical differences should be treated with caution.

NSOPF is a sample survey that was conducted by the U.S. Department of Education in 1988 and 1993. The two NSOPF survey frames are not comparable. Those with no teaching duties in the fall semester of 1988 were considered out of scope, while the comparable group was included in the 1993 cycle. Internally consistent subsets can be constructed and compared across the two survey years, however. Because the NSOPF estimates of doctoral scientists and engineers agree quite well with those derived from SDR, and since NSOPF contains information on teaching activities that is unavailable from SDR, data from this survey have been used to supplement SDR information.

The Academic Doctoral S&E Workforce[29] top

The total number of scientists and engineers in the U.S. labor force with doctoral degrees from U.S. universities has more than doubled over the past two decades, rising from about 215,000 in 1973 to 475,200 in 1995; the academic component increased from an estimated 118,000 to 217,500.[30] (See text table 5-5.) The rate of academic employment growth, though robust over much of the period, was lower than growth in other sectors. The growth rate for academic employment dropped from nearly 7 percent annually in the early 1970s to just under 1 percent from 1989 onward; consequently, the academic employment share declined from an estimated 55 percent in 1973 to 46 percent in the 1990s.

While the total number of academic doctoral scientists and engineers continued to rise from 206,700 in 1989 to 217,500 in 1995, the number of incumbents holding full-time faculty positions-full, associate, and assistant professors plus instructors-remained roughly stable at between 169,800 and 173,100. (See figure 5-10.) Consequently, the share of full-time faculty among all academic doctoral scientists and engineers declined to an all-time low of 79 percent. This drop continued a downtrend evident since the early 1970s, when this share had stood at 88 percent. Psychology and the physical, environmental, and life sciences experienced particularly substantial shifts toward nonfaculty employment, with full-time faculty percentages dropping by 10 or more points. Developments in the social sciences, mathematics, and engineering were somewhat less pronounced. (See appendix table 5-23.)

The number of incumbents in other types of academic positions-full- and part-time adjunct faculty, lecturer, research and teaching associate, administrator, postdoctorate-grew at a more rapid rate than the number of full-time faculty, increasing from 14,700 in 1973 to 46,200 in 1995. The 1989-95 increase was 25 percent, in contrast to the essentially flat full-time faculty count. Most of the growth in this segment was due to postdoctorate[31] and other full-time appointments; part-time employees accounted for between 2 and 3 percent of the total throughout. (See appendix table 5-23.)

Employment growth was not uniform across different segments of higher education. Universities categorized as research universities in the Carnegie classification system experienced slower growth than other institutions;[32] their doctoral S&E staff increased by 56 percent, from 57,600 in 1973 to 90,100 in 1995. Other universities and colleges combined had twice that rate of increase, as their numbers went from 60,400 in 1973 to 127,400 in 1995. Consequently, the proportion of academic doctoral scientists and engineers employed by research universities dropped from 49 to 41 percent during the period. (See text table 5-6.)

Women in the Academic Doctoral S&E Workforce[33] top

The number of academically employed women with S&E doctorates rose more than fourfold between 1973 and 1995, increasing from about 10,700 to an estimated 52,400. In comparison, the number of men increased by roughly 54 percent over the period, from 107,300 to 165,100. Consequently, men's employment share dropped by 15 percentage points, from 91 percent in 1973 to 76 percent in 1995. Women's gains were especially pronounced in psychology and the life and social sciences, fields where their participation in 1973 had already been the highest. (See appendix table 5-24.)

The recent decline in the full-time faculty component, discussed above, was driven by an estimated 10 percent drop in the number of male full professors since 1989-from 76,300 to 68,800-combined with roughly stable numbers of male associate professors and junior faculty (assistant professors and instructors). (See figure 5-11.) But the number of women serving as full professors, associate professors, and junior faculty-assistant professors and instructors-increased by 30 percent or more during this time. By 1995, women constituted 21 percent of full-time S&E faculty. The number of women also increased faster than the number of men-41 versus 17 percent since 1989-in the other types of academic positions: full- and part-time adjunct faculty, lecturer, research and teaching associate, administrative, and postdoctorate. (See appendix table 5-24.)

Throughout the period, the field distribution of women remained more concentrated than that of men. Fully 84 percent of women scientists and engineers in 1995 were found in three broad fields: life sciences (41 percent), social sciences (23 percent), and psychology (22 percent). In contrast, only 58 percent of men were in these fields in 1995. Conversely, only 8 percent of women, but 19 percent of men, were in the physical and environmental sciences; and just 3 percent of women were in engineering versus 14 percent of men. (See appendix table 5-24.)

These field distributions in academic employment reflect the different field patterns of doctorate degrees earned by men and women. Over the past two decades, increased hiring of women into academia has been commensurate with women's rising proportion of new S&E doctorates. Among recent Ph.D.s in academic employment,[34] women have been represented in rough proportion to their share of newly awarded doctorates in every major field over the entire 1973-95 period. However, their proportion of the doctoral academic S&E workforce-24 percent in 1995-continues to lag their percentage of new S&E doctorates-38 percent. (See text table 5-7.)

Underrepresented Minorities in the Academic Doctoral S&E Workforce[35] top

Academic employment of underrepresented minorities-blacks, Hispanics, Native Americans, and Alaskan Natives[36]-rose to 12,800 in 1995 from 2,400 in 1973. Their employment share rose from 2 percent in 1973 to 6 percent in 1995, the same as their share of full-time faculty positions. Relative gains for underrepresented minorities were greatest in psychology and the social sciences-where their employment share rose from 2 to 8 percent-and in engineering-from 2 to 5 percent. (See appendix table 5-25.)

These low but rising numbers reflect the growing number of S&E Ph.D.s earned by members of underrepresented minorities.[37] For the past two decades, underrepresented minorities have been hired into academic positions at somewhat higher rates than would be expected based on their share of new S&E Ph.D.s awarded. As a consequence, their representation in the total academic workforce has been close to their share of new doctorates. (See text table 5-7.)

The distribution of underrepresented minorities by field is similar to that of whites, with two exceptions. Underrepresented minorities are less likely than whites to be in the life sciences-28 versus 34 percent, and they are more likely to be in psychology and the social sciences-41 versus 33 percent.

Asians in the Academic Doctoral S&E Workforce[38] top

Asians as a group have been quite successful in entering the academic workforce. The number of Asian academic doctoral scientists and engineers rose rapidly between 1973 and 1995, increasing from 5,100 to 22,500 in 1995. This growth more than doubled their employment share: 10 percent in 1995 versus 4 percent in 1973. Asians made especially strong gains in the physical sciences (from 5 percent in 1973 to 14 percent in 1995), computer sciences (from 13 percent in 1985 to 29 percent in 1995),[39] and engineering (from 9 percent in 1973 to 21 percent in 1995). (See appendix table 5-25.)

Asians are increasingly prominent among new Ph.D.s in academia, well in excess of their share of S&E Ph.D.s awarded to U.S. citizens and permanent visa-holders. That is, Asians, more than any other group, tend toward academic employment. By 1995, Asians accounted for nearly one-quarter of all new academic S&E doctorates. (See text table 5-7.)

Fifty-four percent of Asian academic S&E doctorates are in the physical, environmental, and computer sciences; mathematics; or engineering-a much higher proportion than for whites (33 percent) or underrepresented minorities (32 percent). Few Asians enter psychology, and a relatively small proportion is in the social sciences. (See appendix table 5-25.)

Employment Growth by Field top

Academic employment in the physical sciences grew more slowly than in other fields over the 1973-95 period, rising from 22,100 to 29,300-33 percent growth compared to 84 percent for all of S&E combined. As a result, the share of academic doctoral scientists and engineers employed in the physical sciences fell from 19 to 13 percent; this drop was experienced in both physics and chemistry. In contrast, employment in the life sciences increased by more than 100 percent over the period, rising from 34,900 to 71,600; this field's employment share rose from 30 to 33 percent. Other fields experiencing relative gains were engineering and psychology. (See appendix tables 5-24 and 5-25.)

The Shifting Faculty Age Structure top

The rapid pace of hiring of young Ph.D.s into academic faculty positions during the 1960s to accommodate soaring enrollments, combined with slower hiring in later years, resulted in an aging professoriate. (See figure 5-12.) Through the 1980s, a growing proportion of academic faculty was found in the older age brackets. A noteworthy feature of the data involves the upper end of these age distributions. The fraction of total faculty older than 65 has been about 3 percent for the past decade, with 1 percent older than 68 years. By and large, academics tend to retire before that age. (See appendix table 5-26.)

Concerns had been voiced early in the decade about the possible deleterious effects of delayed faculty retirements resulting from the full applicability of provisions of the Age Discrimination in Employment Act to universities and colleges starting in 1994.[40] The concerns focused on the potential ramifications for universities' organizational vitality, institutional flexibility, and financial health. A study by the National Research Council (NRC) concludes that "overall, only a small number of the nation's tenured faculty will continue working in their current positions past age 70" (NRC 1991, p. 29), but adds: "At some research universities a high proportion of faculty would choose to remain employed past age 70 if allowed to do so" (NRC 1991, p. 38).

Recent data indicate, however, that, for the system as a whole, little has changed in the last decade in terms of retirement behavior. (See appendix table 5-27.) Across all of higher education, about 3 percent of full-time faculty stay on beyond age 65. As anticipated by the NRC study, faculty at research universities tend to keep working longer than those elsewhere. But it is also worth noting that research universities have managed to maintain a relatively more balanced age structure than other types of universities and colleges. (See appendix table 5-26.) The faculty age distribution in research universities tended to be older, on average, than that of other academic institutions through the early 1980s, but that tendency has since reversed. By 1995, research universities had a greater share of their full-time faculty in the under-46 age brackets than other institutions, and a slightly greater share in the above-60 ones as well. (See figure 5-13.)

The mean and median ages of full-time doctoral faculty show a clear upward trend from 1973 through 1989, with a flattening thereafter. (See figure 5-14.) This result can now be interpreted in light of the overall number of faculty, which grew through 1989 and has since essentially held steady in the range of 169,800 and 173,100. During the years of growth, the average faculty age climbed from 42.5 to 47.1 years before leveling off. This suggests that for academia as a whole-not necessarily for individual institutions or departments-a rough balance has been maintained between attrition from all causes and new hires. However, the number of replacements from 1989 onward has to be seen in the context of Ph.D. awards which rose by more than one-fifth overall from 1989 to 1995 (up from 22,705 to 27,846) and by 30 percent for U.S. citizens and permanent residents. (This latter growth reflects in part Chinese students' conversion to permanent visa status following the Tiannanmen Square events.) In short, the modest increases in hiring from the late 1980s onward took place against a backdrop of steeply rising numbers of new Ph.D.s.

New Ph.D.s in Academic Employment[41] top

The presence in academic employment of people with newly earned S&E doctorates provides a leading-edge indicator of the future composition of the academic teaching and research workforce. Because of the small number of new Ph.D. recipients entering academic employment relative to the size of the existing workforce, changes in the overall composition of the academic workforce will occur slowly-but are already visible, as noted above.

The number of recent Ph.D.s entering into academia-defined as those who had earned their doctorate in the three years preceding the survey-declined gradually from 25,000 in 1973 through the early 1980s, reaching a low of 20,500. It then rose again through the mid-1990s, reaching 26,900 in 1995. These represent just over half of all recent doctorate-holders. (See appendix table 5-28.) But the meaning of academic "employment" has changed for these young Ph.D.s. Fewer than 45 percent had regular faculty appointments in 1995, compared with over 75 percent in the early 1970s and 57 percent in the mid-1980s. (See figure 5-15.) Since 1973, the proportion of new doctorate recipients holding postdoctorate positions has increased steeply, rising from 13 to 28 percent in 1985 and 40 percent in 1995.[42] The proportion of doctorates in other nonfaculty appointments has also doubled, rising from 8 to 17 percent. (See appendix table 5-29.)

The demographic composition of these recent academic doctorate-holders has shifted noticeably over two decades. The proportion of women has risen from 12 to 38 percent. The proportion of underrepresented minorities has grown from 2 to 7 percent, and of Asians from 5 to 23 percent. (See text table 5-7 and appendix table 5-30.) The field composition of young Ph.D.s reflects the larger employment changes: 38 percent are in the life sciences (up from 28 percent in 1973), 14 percent are in the physical sciences (after dropping from 16 in 1973 to 11 percent in 1985), and 4 percent are in mathematics (down from 9 percent in 1973). But the field distribution of young Ph.D. recipients in full-time faculty positions differs from this total employment picture, with smaller faculty shares in the physical and life sciences and higher fractions in psychology and the social sciences. (See appendix table 5-29.)

Research and Teaching Activities[43] top

In the academic workplace, particularly in universities with a strong research orientation, teaching, research, and research training are often inextricably intertwined. In this way, academic research produces both new knowledge and highly trained personnel. Most academic scientists and engineers do not do either teaching or research, but pursue both activities in a mix that may change with the time of year and the course of their careers. Nevertheless, for the past two decades, a reasonably consistent indicator of the relative balance between teaching and research may be obtained from responses of academic doctoral scientists and engineers to a question about their major work responsibilities. The discussion here commences with an examination of a snapshot of the distribution of research and teaching activities, including undergraduate and graduate teaching, in academia; proceeds to trends in respondents' primary work responsibility; and closes by focusing on trends in primary and secondary responsibilities.

While not directly addressing the synergy between teaching and research, a survey (NSOPF) conducted by the U.S. Department of Education allows examination of the patterns of undergraduate and graduate teaching activities of doctoral academic scientists and engineers, and the extent of their research activities in relation to these teaching duties.

Of the estimated 213,800 doctoral scientists and engineers employed in academic institutions in 1993, 81 percent had some teaching duties in the fall semester of that year: 58 percent taught courses primarily for undergraduates, 25 percent taught courses primarily for graduate students, and 17 percent taught both graduate and undergraduate courses. (See text table 5-8.)

Those who taught undergraduate courses exclusively on average spent an estimated 65 percent of their weekly work time on teaching activities and 22 percent on research. For those with only graduate teaching responsibilities, the corresponding time estimates were 34 and 38 percent, respectively; and for those teaching both undergraduate and graduate students, the percentages were 50 and 27. These time estimates have not changed greatly since 1988.[44]

Primary Work Responsibility: Emphasis on Research top

SDR respondents (see "Data Sources: Nature, Problems, and Comparability") were asked to select their primary work responsibility from a list that includes teaching, various R&D functions, administrative work, consulting, and other activities. A crude but consistent indicator of the relative emphasis on research can be constructed from the responses. The choices in research activities as primary work responsibility reveal two major shifts. First, the relative balance between teaching and research has shifted toward the latter. Second, by this measure, growth of the research function has been especially pronounced outside the ranks of the traditional research universities.

The number of those reporting teaching as their primary work responsibility rose from 73,300 in 1973 to 101,100 in 1985 and has fluctuated around the 100,000 mark since then. In contrast, the number of those identifying research as their primary work responsibility has increased steadily, rising from 27,800 in 1973 to 56,000 in 1985 and 83,000 by 1995. These divergent trends have lowered the proportion of those reporting teaching as their primary work from 62 percent in 1973 to 46 percent in 1995, while the proportion of those reporting research as their primary work has risen from 24 to 38 percent. Those with other types of primary work responsibility-for administrative or managerial functions, service activities, and the like-constituted between 14 and 19 percent of the total over the period. (See appendix table 5-31.)

Employment growth in research universities since the late 1970s has been largely confined to those identifying research as their primary activity. Their number stood at 17,500 in 1973 and 45,900 in 1995, as their share rose from 30 to 51 percent of research universities' doctoral S&E workforce. In contrast, the number of research university faculty for whom teaching was the primary activity rose from 32,300 in 1973 to a high of 39,600 in 1981 before declining to 30,500 in 1995. The number identifying other functions as their primary work responsibility has remained at around 12,000 to 15,000 since the early 1980s. In other types of universities and colleges, a growing number of faculty identified teaching as their primary work activity for much of the two decades; since 1989, this number has fluctuated between roughly 67,000 and 70,000. But those for whom research was the primary work responsibility increased more rapidly and continuously. As their numbers grew from 10,300 in 1973 to 37,100 in 1995, their share rose from 17 to 29 percent. (See appendix table 5-31.)

Besides these institutional differences, there have been field differences as well. (See text table 5-9.) Employment growth from 1973 to 1995 has exceeded 50 percent in most fields,[45] except mathematics (41 percent) and-notably-the physical sciences (17 percent). Growth in teaching (as characterized here) was slower than overall employment growth in every field but the computer sciences; the physical sciences, by this measure, actually experienced negative growth. On the other hand, the number of respondents who designated research as their primary work responsibility quadrupled in engineering and more than tripled in several science fields. In mathematics and the physical sciences, it roughly doubled.

Participation in Research top

Academic work generally entails a more complex mix of functions-teaching, research, administrative work, consulting, public service, among others-than the above-discussed indicator (research as primary work activity) takes into account. A more encompassing measure can be constructed from respondents' choice of research as either a primary or secondary work function; this yields a better lower bound estimate of the broadly defined academic doctoral research workforce.[46] By this measure, an estimated 153,500 academic doctoral scientists and engineers were engaged in R&D in 1995, up from a range of 80,000 to 90,000 during the 1970s.[47] (See figure 5-16.) The number of academic researchers has essentially been stable since the late 1980s, after strong growth in the preceding decade and a half. (See appendix table 5-32.)

Roughly 71 percent of all academic doctoral scientists and engineers reporting primary and secondary work responsibilities in 1995 were engaged in research activities, but this varied by field. At the high end-80 percent-were the environmental sciences; the life and computer sciences and engineering ranged from 75 to 78 percent. Those in the physical sciences, mathematics, psychology, and the social sciences reported the lowest levels of research activity, ranging from 62 to 70 percent.

These field differences in the levels of research intensity have been fairly consistent over time, and the field composition of academic researchers has generally not shifted dramatically. But the relative employment shift noted earlier away from the physical sciences and toward the life sciences is evident in the research workforce as well. The physical science share has declined by 6 percentage points since 1973, and that of the life sciences has increased by 3 percentage points, with marginal gains or losses for the other fields. (See appendix table 5-32.)

Federal Support of Academic Researchers top

In 1995, 39 percent of the academic doctoral scientists and engineers responding to SDR reported receiving funding from the Federal Government during the week of April 15. (See text table 5-10.) This number cannot be easily compared with those from earlier years, which were based on a year-long reference period-49 percent in 1989, 50 percent in 1991-but is in line with SDR estimates for other reference periods shorter than a full year: 37 percent each in 1985 and 1993. If the volume of academic research activity is not uniform over the entire academic year, but varies to accommodate teaching and other activities, a one-week or one-month reference period may well understate the extent of support over an entire academic year. Several pieces of evidence suggest this to be the case.[48]

Just over half (51 percent) of the doctoral scientists and engineers surveyed in the 1993 NSOPF reported having Federal Government funding in the fall semester of that year. This is in line with earlier SDR estimates based on year-long reference periods. The NSOPF estimate, when taken together with information regarding growth in federal funding, suggests that no major changes have occurred since the late 1980s in the number or proportion of researchers supported with federal funds. This tentative conclusion is further bolstered by the steady growth in the number of federally funded research assistants through the 1980s and 1990s.

Notable and persistent field differences exist in the proportion of researchers supported by federal funds.[49] Above the overall S&E average are the life, environmental, and physical sciences and engineering. Clearly below the mean are mathematics, psychology, and the social sciences. The relative position of these fields has not changed substantially over the past two decades. (See text table 5-10.)

Since the late 1980s, a larger fraction of academic researchers has reported federal support from more than one agency. This trend can be observed across most S&E fields. (See appendix table 5-33.) Fields with the highest levels of researchers receiving multi-agency support are the environmental sciences-more than 40 percent-and engineering and the computer and physical sciences-well above 30 percent for each. Single-agency support is most prominent in the life and social sciences, psychology, and mathematics. However, no clear upward trend in multi-agency support is evident since the late 1980s.

The interpretation of these data is ambiguous. They could, for example, indicate greater difficulty in obtaining funding, the growing availability of multiple funding sources, or increasing entrepreneurism by investigators in seeking out funding.

green bar


[29] The academic doctoral S&E workforce includes full, associate, and assistant professors and instructors-collectively defined throughout this section as faculty-lecturers, adjunct faculty, research and teaching associates, administrators, and postdoctorates.

[30] The trend data in this section refer to scientists and engineers with doctorates from U.S. institutions, regardless of their citizenship status. Comparable trend data for Ph.D.-level scientists and engineers with degrees from non-U.S. institutions are not available. A 1993 U.S. Department of Education survey of academic faculty suggests that this component of the academic workforce numbers around 13,000.

[31] For more information on this subject, see "Postdoctoral Appointments" in chapters 2 and 3.

[32] This periodically revised classification describes research universities as institutions with a full range of baccalaureate programs, commitment to graduate education through the doctorate, annual award of at least 50 doctoral degrees, and receipt of federal support of at least $15.5 million (average of 1989 to 1991). See Carnegie Foundation for the Advancement of Teaching (1994).

[33] Also see chapter 3, "Women in the S&E Workforce."

[34] Recent Ph.D.s are those who have earned their doctorates within the past three years.

[35] Also see chapter 3, "Racial/Ethnic Minorities in the S&E Workforce."

[36] There is variation among and within these groups, which are treated here in the aggregate. Appendix table 5-25 provides somewhat more detailed data; the survey sample does not permit further disaggregation. Asians as a group have been quite successful in entering the academic workforce and are treated separately.

[37] This in turn, of course, reflects their increasing participation in higher education and graduate school training. See chapter 2, "Master's Degrees by Race/Ethnicity" and "Doctoral Degrees by Race/Ethnicity."

[38] Again, also see chapter 3, "Racial/Ethnic Minorities in the S&E Workforce."

[39] Pre-1985 estimates are unreliable because of the low number of computer science degree-holders in the sample.

[40] A 1986 amendment to the Age Discrimination in Employment Act of 1967 prohibited mandatory retirement on the basis of age for almost all workers. Higher education institutions were granted an exemption through 1993, allowing termination of employees with unlimited tenure who had reached age 70.

[41] No trend data exist on detailed in- and outflows. The data reported here are "snapshots" of the number and demographic characteristics of doctorate-holders in academic employment who had earned their degree in the three years preceding the survey.

[42] An accurate count of postdoctorates is elusive, and the reported increase may be understated. A postdoctoral appointment is defined here as a temporary position awarded primarily for gaining additional training in research. The actual use of the term, however, varies among disciplines and sectors of employment. In academia, some universities appoint postdoctorates to junior faculty positions which carry fringe benefits; in others, the appointment may be as a research associate. Some postdoctorates thus may not regard themselves as genuinely "employed." Also see "Postdoctoral Appointments" in chapters 2 and 3.

[43] This material is based on individual respondents' reports of their primary and secondary work responsibilities. The data series-which is drawn from SDR-is reasonably consistent for the 1973-89 period: respondents were asked to designate primary and secondary work responsibilities from a list of items, the majority of which remained unchanged. Since 1991, however, primary and secondary work responsibilities have had to be inferred from reports of the activities on which respondents spent the most and second-most amount of their average weekly work time. These two methods yield close-but not identical-results, so the SDR series must be considered a rough indicator only. In addition, some nonrespondents in 1981-87 were sent a shortened version of the questionnaire that did not ask about secondary work responsibility. For these respondents and these years, secondary work responsibility was estimated using full-form responses, based on field and type of position held. This analysis also draws on data from the 1988 and 1993 NSOPF. As noted in "Data Sources: Nature, Problems, and Comparability," the sample estimates of numbers of faculty from this survey differ slightly from those derived from SDR.

[44] Those without fall 1988 teaching responsibility were ruled out of scope in that survey year, but not in 1993. The comparison with 1988 is based only on those 1993 respondents with teaching responsibilities.

[45] Computer science data were not broken out before 1979. The series starts from a very low base and involves a relatively small number of respondents. Thus, the percentage increases in computer science teaching versus research growth must be viewed in this context and are best interpreted only within the field.

[46]  The estimate fails to account for respondents who ranked research third or lower in their ordering of work responsibilities. Additionally, for 1981 through 1985, some respondents who received short forms of the survey questionnaire could not record a secondary work responsibility, thus resulting in a definite undercount for these years. All estimates are calculated based on individuals who provided valid responses to this item.

[47] A rough estimate of the nondoctoral researcher component, excluding graduate research assistants, was derived for 1993 from NSOPF. This study suggests that this component is approximately 10 percent the size of the doctoral research workforce and is concentrated in the medical and health sciences (60 percent), biological sciences (15 percent), and engineering (10 percent).

[48] Indirect evidence that the extent of support is understated can be gleaned from the number of senior scientists and postdoctorates supported on NSF grants. This number is published annually as part of NSF's budget submission. It bears a relatively stable relationship to numbers derived from SDR in 1987, 1989, and 1991, but diverges sharply in 1993. (The figures from the two data sources are never identical, however, since NSF's numbers reflect those funded in a given fiscal year, while SDR numbers reflect those who have support from NSF regardless of when awarded.)

[49] The relative field shares of federally supported researchers appear to be stable across recent survey years, i.e., they are relatively unaffected by changes in the survey reference period. The distribution (but not the estimated number) based on NSF estimates is quite similar.

Previous | Top | Next

Indicators Home | Contents | Help | Comments