FASEB Consensus Conference on Graduate Education

Table of Contents

Assessment of Employment Trends for Biomedical Ph.D.s

Current Data on Careers for Biomedical Ph.D.s

Future Directions for the Employment of Biomedical Scientists

Trends in the Production of Biomedical Ph.D.s

Should There Be National Regulation of Biomedical Ph.D. Production?

Should the Workforce Needs of Schools Determine the Number of Graduate Students?

How Many Students from Other Nations Should We Train?

What Is the Status of Women and Minorities in Biomedical Science?

What Is the Mission for Biomedical Graduate Education?

Should Predoctoral Training Be Broader and More Interdisciplinary?

What Is the Appropriate Time Needed to Complete a Ph.D. Program in the Biomedical Sciences?

Importance of Mentoring in Biomedical Science Education

What Are the Career Opportunities, Other than Academia, Industry, and Government, and Shall We Educate for These Diverse Careers?

What Defines the Success of a Biomedical Training Program?

How Should Graduate Programs Be Reviewed?

What is the Role of Peer-Reviewed, Competitively Awarded NIH and NSF Training Grants in Setting and Maintaining Standards of Excellence in Biomedical Graduate Education?

Executive Summary

During most of the 1970s and 1980s, production of biomedical Ph.D.s in the U.S.A. was fairly constant. From 1985 to 1995, however, there was an increase of more than 50% in the number of biomedical Ph.D.s awarded by U.S. institutions; nearly 70% of this increase can be accounted for by the increase in the number of non-citizens getting their Ph.D. in the U.S.A. What are the driving forces that have resulted in this increase? Can the increased production of biomedical Ph.D.s continue without altering the job market for new graduates? Should this growth be curtailed in order to achieve a new steady state and, if so, at what point?

At the present time, unemployment among U.S. citizens with biomedical Ph.D.s is extremely low, less than 2.0%. However, there have been some important changes in the job market for biomedical Ph.D.s. The total number of biomedical scientists has grown, while the number of faculty positions has remained stable. As a result, faculty positions have declined as a percentage of total employment for biomedical scientists. Jobs in industry have increased and, in the future, might surpass academic jobs as the most prevalent form of employment for U.S. biomedical scientists.

Although nearly all biomedical science Ph.D.s are fully employed, the jobs they hold may not match the faculty positions in prestigious universities that they may have aspired to during their training. We recommend that current data on employment be readily available to the public, and that the employment information be updated regularly. Students and faculty should be made aware of the broad range of career options for a biomedical Ph.D., including opportunities outside the academic sector.

Research in the biomedical sciences is exciting, but we cannot predict precisely its impact on the job market or how much it might spur the creation of new job opportunities. Although the future job market demand cannot be predicted, the future supply can be estimated. Predoctoral applicants making career plans and institutions deciding on the size of their graduate programs should carefully consider current trends, while recognizing that variables dictating job market demands may vary. We oppose external regulation of the size of biomedical Ph.D. programs at the national level, especially since it is impossible to predict job market needs 10 years into the future (the time at which beginning graduate students will be ready to enter the job market). Instead, institutions should self-regulate the size of their graduate programs after consideration of many parameters.

Applicant ability should be the essential criterion for admission to graduate school. Under-qualified predoctoral applicants should not be admitted simply to meet the workforce needs (e.g., teaching or research) of the institution, and alternate ways to fill workforce needs should be considered. The number of non-U.S. predoctoral students should not be capped arbitrarily, and there should be no discrimination with regard to race or gender for admission into graduate programs.

Quality of graduate programs should be promoted by frequent self-study, but redundant reviews by national, regional, and state accreditation bodies should be eliminated wherever possible. The success of a program should be assessed by testing congruence of the program's mission with the career outcomes of its students, and this information should be provided to applicants to the program.

To equip biomedical students for future jobs in research or in transmission of scientific knowledge in a variety of environments, they should be trained in depth in one specific area of biology and also be educated broadly in many other areas in the biological and physical sciences. The predoctoral training experience should continue to focus on independent research, and entire graduate programs should not be totally restructured to train students for the small number of diverse career options currently outside academia, industry, and government. Graduate programs should be sufficiently flexible, however, to allow individual students to broaden their education consonant with their career goals. Moreover, skills such as effective communication, ability to work in a team, and use of modern information technology should be developed. These skills will be valuable to the researcher in a variety of employment contexts. To insure completion of the Ph.D. in no more than 5-6 years, faculty committees should review the progress of graduate students at least annually. Mentoring is an important part of predoctoral and postdoctoral education to help students reach their full potential; the faculty should be supportive of the range of career options their students might follow. For some career paths, a Master of Biomedical Science degree alone or coupled with a degree in another field (e.g., law, finance, journalism, etc.) may be appropriate and sufficient.

Summary of Recommendations

Section I: What Are the Employment Trends for Biomedical Ph.D.s?

• Data on trends in careers for biomedical Ph.D.s should be published in a timely manner by federal agencies for careful consideration by students and faculty. Professional societies should describe current employment opportunities for their members; this may identify emerging areas. To enhance career tracking, predoctoral students and postdoctoral associates should be invited by their schools to provide their social security numbers for career tracking studies. Compliance with this request should be entirely voluntary, and the resulting reports should summarize aggregate findings without identifying individual students. Submission of social security numbers should not be a condition of funding, and should temporally follow the appointment.

• Although historical trend data can be a useful guide, many changes may occur during the 10 year period from the onset of predoctoral study to the completion of a postdoctoral. Therefore, applicants for Ph.D. programs should be made aware that the job market may look quite different when they are ready to enter it than it did when they began. Research in the biomedical sciences is exciting. There is tremendous potential, but we cannot predict the impact of biomedical research advances on the job market.

Section II: How Should Admission into Biomedical Ph.D. Programs Be Regulated?

• We oppose attempts to regulate biomedical Ph.D. production at the national level. Since future employment demands for biomedical Ph.D.s cannot be determined precisely, it is inappropriate to limit Ph.D. production on the basis of guesses about the future job market. Students with strong academic qualifications and other requisite criteria and intense motivation to pursue a biomedical Ph.D. should have the opportunity to do so. It is important that only students who meet high standards be admitted and only those of proven ability be awarded the Ph.D.

• Applicant quality based on past performance and potential for success should be the essential criteria for admission to graduate school. Predoctoral applicants should not be admitted primarily to meet the institutional needs for teaching assistants or research assistants. Alternate ways to fill workforce needs should be considered.

• Students of high caliber from other nations should continue to be accepted into biomedical predoctoral programs in the U.S.A. The number of students from other nations should not be capped arbitrarily.

• Admission into graduate programs should be based on applicant quality, without gender or race discrimination. Efforts must be made to improve K-12 science education in the U.S.A. as a necessary part of expanding minority recruitment to careers in science.

Section III: Length and Type of Training: Are They Appropriate to the Spectrum of Opportunity?

• The focus of biomedical Ph.D. training is, and should continue to be, original research as embodied by the Ph.D. thesis; this training will be used for a variety of career roles in research discovery and/or transmission of knowledge. Graduate programs should develop their students' skills for critical evaluation of research results and the scientific literature, effective oral and written communication, use of modern information technology, laboratory management, and upholding the highest ethical standards in research.

• Students should be trained in depth in one specific area of biology, but also be educated broadly in areas representing different hierarchies of biological organization and exposed to the physical sciences. We commend interdisciplinary training.

• To facilitate completion of the Ph.D. in no more than 5-6 years, faculty committees should review the progress of graduate students at least annually.

• Effective mentoring is an essential part of pre- and postdoctoral education. In addition to frequent, informal contacts with the faculty mentor, we recommend that formal meetings be held at least yearly with the faculty mentor to review professional development and career aspirations.

• Faculty should be supportive of the range of career options available to their students. For some career paths, a Master of Biomedical Science degree alone or coupled with a degree in another field (e.g., law, finance, journalism, etc.) may be appropriate. Careers outside of academia, industry, and government are currently a small fraction of careers pursued by biomedical science Ph.D.s, and entire graduate programs should not be restructured to train students for these more diverse options. Nonetheless, graduate programs should be sufficiently flexible to allow individual students to broaden their education consonant with their career goals.

Section IV: How Should the Quality of Biomedical Graduate Programs Be Measured?

• The success of a predoctoral educational program should be assessed by testing the congruence of the program's missions with the career outcomes of its students. Biomedical Ph.D. programs should provide applicants with information on the career outcomes of all their predoctoral students who have completed training over the past 15 years, as this will assist students in their selection of programs matching their own goals.

• The most effective review of the success of graduate programs occurs at the local level through frequent self-study and periodic reviews by knowledgeable scientists from other academic institutions and industry. To conserve resources and valuable time of faculty and administrators, redundant reviews by national, regional, and state accreditation bodies should be eliminated wherever possible. We oppose efforts to introduce further accreditation at the national level, as it would be impractical because each program is unique; also, it might lead to homogeneity of programs, which is undesirable, at the expense of diversity, which is desirable.

• Institutions should aspire to have high quality graduate programs. For institutions seeking guidelines, the NIH and NSF training programs can be viewed as models for excellence in education in biomedical research. Since federal training grants enhance programmatic aspects of graduate education, it is desirable to spread them out to a somewhat larger number of schools, providing that their programs are high quality and merit federal support. Additional increases for pre- and postdoctoral stipends are desirable, but money for stipend increases should not compromise the federal funds for investigator-initiated research grants.

Introduction

The scientific community cares deeply about the training and career prospects of the graduate students in the biomedical sciences, and that prompted the Federation of American Societies for Experimental Biology (FASEB) to sponsor a consensus conference on this topic. Graduate education today must prepare the next generation of biomedical scientists and is the foundation upon which the continued future success of our scientific enterprise depends.

In 1974, the National Research Service Award Act was adopted by Congress to fund training in the biomedical sciences. Congress stipulated that national needs for further training in this area should be reviewed periodically; this review is carried out by the Committee on National Needs for Biomedical and Behavioral Research Personnel of the National Academy of Sciences. Thus, the topic for discussion at the 1996 FASEB Graduate Education Consensus Conference was not new, and some aspects have been deliberated for more than two decades.

One of the first groups to recognize the current salience of this topic was the Association of American Medical Colleges (AAMC) and, in response to Dr. Bill Brinkley's initiative, it formed the Graduate Research, Education and Training (GREAT) subcommittee, which has now been formalized into a permanent group with national membership and representation. Many of the issues discussed by the FASEB conferees also have been discussed at the AAMC GREAT conferences, held annually since 1994. The ongoing focus of the AAMC conferences has been on the students trained in medical schools. The FASEB Graduate Education Consensus Conference extended the discussion to the full cohort of biomedical science Ph.D. students and derived conclusions and recommendations.

Others have studied many of the same issues, 1,2 and reports with recommendations have been issued recently 3,4,5,6,7 or will be released shortly 8. Almost all the data that are the basis for the national studies derive from the Survey of Earned Doctorates and the Survey of Doctorate Recipients; summary tabulations from these data sources are published in many places including the biannual report, Science and Engineering Indicators 9. The various reports can have different recommendations, reflecting different interpretations of essentially the same database. The recommendations presented in this report are the results of deliberations by representatives from the member societies of FASEB, representing more than 52,000 scientists.

Section I: What Are the Employment Trends for Biomedical Ph.D.s?

Assessment of Employment Trends for Biomedical Ph.D.s

One way to assess trends is to ask individuals what kinds of jobs they hold. Another is to ask employers how many people they employ and whether they anticipate a staffing increase or decrease in the future. The first approach is taken by the federal government in its surveys of recent Ph.D.s. The Survey of Earned Doctorates (SED), conducted by the National Research Council (NRC) under contract to the National Science Foundation (NSF), summarizes the replies of individuals at the time they obtain their Ph.D.; when asked what their employment will be immediately after the Ph.D., the majority of biomedical scientists expect to be postdoctoral associates. Therefore, this survey is not a useful gauge of long-term employment. Another survey, a longitudinal follow-up of employment outcomes called the Survey of Doctorate Recipients (SDR), is conducted on a sample of Ph.D. recipients and is a valuable indicator of employment trends. The SDR responses are tabulated and weighed to depict employment of the entire population of Ph.D. recipients.

The statistical data tabulations presented in this report from the SED and SDR are limited to U.S. citizens. The number of non-U.S. citizens in the SDR biomedical science sample is small and estimates for U.S. citizens are virtually the same as those for the entire biomedical Ph.D. sample. Tabulations for the latter group are available upon request from FASEB. The SDR sample is representative of the entire population of U.S. Ph.D.s only, and not representative of the non-citizen Ph.D.s. The SDR sample includes only those non-citizens with Ph.D. from U.S. schools who indicated, at the time that they competed their doctoral dissertation, that they had plans to remain in the U.S.A. Not all had firm plans and it is unclear, often even to the non-citizens themselves, whether they would remain in the U.S.A. In addition, non-citizens who obtained their Ph.D. in other countries, a sizable part of the postdoctoral pool, are not captured in the SDR survey.

Information gathered by these surveys of scientists has been very useful, but it would be more helpful if it were tabulated and published in a more timely manner. Recognizing the need for additional and more timely information on the job market, the Alfred P. Sloan Foundation and NSF have funded the Commission on Professionals in Science and Technology to coordinate data collection by a group of professional associations and to monitor the supply and demand for recent graduates. Their data will be available beginning in the fall of 1997.

Individual professional scientific societies should develop descriptions and projections of employment opportunities relevant to their members and disseminate this information. It will be useful to identify emerging areas for subdisciplines.

Efforts to develop career outcomes data should be facilitated. For example, giving funding agencies access to the social security numbers of predoctoral students and postdoctorals associates will enable these agencies to track later employment more accurately and to report the aggregate employment outcomes data in a timely fashion.

Recommendation: We concur with the Committee on Science, Engineering and Public Policy 5 that the data describing trends in employment for biomedical scientists should be made available for graduate school applicants, students, postdoctoral associates, and their mentors. Data on trends in careers for biomedical Ph.D.s should be published in a timely manner by federal agencies. Professional societies should describe employment opportunities for their members; this may identify emerging areas. Predoctoral students and postdoctoral associates should be invited by their schools to provide social security numbers for internal or external (including federal) studies that track careers. Compliance with this request should be entirely voluntary, and the resulting reports should summarize aggregate findings without identifying individuals. Submission of social security numbers should not be a condition of funding and should temporally follow the appointment.

Current Data on Careers for U. S. Biomedical Ph.D.s

Survey data show both stability and change in employment of U.S. citizens with Ph.D.s in the biomedical sciences.* Unemployment for U.S. biomedical scientists has remained very low. Although the number of tenure-track plus tenured academic positions has remained fairly constant since 1981, the fraction employed in these academic positions has decreased as a percentage of the total employment opportunities. The importance of industry as a major employer of biomedical scientists has increased enormously. Employment in other permanent jobs has remained a small part of the total.

Unemployment

The percentage of U.S. citizen biomedical science Ph.D.s who are unemployed is extremely small. It was 1.0% in 1973 and 1.8% in 1995 (Table 1). Unemployment fluctuated between 0.9% and 1.9% during the intervening years with no discernable pattern.

Unemployment is only one of several labor market outcomes, and the term "underemployment" has been developed to refer to those individuals who do not find jobs appropriate to their education or experience. In an effort to measure certain aspects of underemployment, NSF has begun reporting an "involuntarily out-of-field rate" (RIOF). This is operationalized as the ratio of those who are working part time but seeking full time jobs (EPTS), or who are working outside of their degree field when a science and engineering job would be preferred (ENSP), to total employment (ET): RIOF =(EPTS +ENSP)/ET. In 1995, 3.3% of the biological and health scientists were employed "involuntarily out-of-field." This represents a slight decrease from the 1993 value of 3.5% 10. It should be noted that, unlike the majority of the statistical data contained in this report, these rates refer to all fields of biology and include both U.S. citizens and non-citizens. Individuals working full-time in their degree field in positions that are not meeting their career goals, for example senior postdoctorals who have not yet acquired permanent employment, are not indexed by this measure.

Postdoctoral Positions

Postdoctoral research allows individuals to expand their expertise; some delve into areas different from their predoctoral field. The duration of the postdoctoral experience varies with the amount of advanced training that is desired, as well as other circumstances unique to the individual. Postdoctorals also represent a pool of applicants for permanent jobs. As permanent jobs become harder to secure, individuals remain longer in postdoctoral positions. In the past two decades, the percentage of U.S. Ph.D.s working as postdoctorals has increased (Table 2).

In the early 1970s, postdoctoral positions were held by a minority of the new Ph.D.s; in 1973, one quarter of the new biomedical Ph.D.s went on to postdoctoral positions, almost always in academia. The pattern changed in the late 1970s and early 1980s; by the beginning of the 1980s, approximately half of the recent graduates were in postdoctoral positions.

Measuring the exact degree of growth is complicated by changes in survey questions and other differences in survey methodology. During the 1970s, the definition of "postdoctoral" used in the Survey of Doctorate Recipients ( SDR) was altered, and some of the change during this period may reflect differences in the respondents' interpretations of the questions asked. Moreover, between 1989 and 1991, enhancements in survey methodology increased the response rate in the survey sample. This resulted in more accurate data but changed the timing of reporting for many respondents, and may have also affected the comparability of survey statistics to those of earlier years.

These factors notwithstanding, a pattern of longer postdoctoral study is evident. In 1981, 23.6% of the employed U.S. biomedical Ph.D.s were postdoctorals 3-4 years after receiving their degree. In 1995, 32.1% of this group were postdoctorals. Increases in the percentage of postdoctorals are found at every level of experience. Whereas some of the percentages and increases are small for example, the percentage of biomedical scientists in postdoctoral positions 9-10 years post-Ph.D. rose from 0.9% in 1981 to 3.1% in 1995 the general pattern clearly indicates lengthening periods of postdoctoral study.

The postdoctoral pool is expanding even more than shown in table 2, as these data do not include the influx of graduates with doctoral degrees from other nations 11. The latter are not accounted for in the tabulations derived from the SDR, nor is it known if the length of time they spend as postdoctorals is the same as for U.S. citizens with biomedical Ph.D.s. Little is known about their career paths. 11 Further detailed analysis of this issue is needed.

In addition to more and longer postdoctorals, another trend to emerge in recent years is growth in the number of "nonacademic" postdoctorals (e.g., positions in industry). In the early 1980s, only a small fraction of the U.S. biomedical Ph.D.s held non-academic postdoctoral positions. For those 1-2 years beyond their Ph.D. in 1981, the percentage was 7.9%; for those who were 3-4 years post-Ph.D., the comparable percentage was 4.4%. By 1995, the percentage in non-academic postdoctoral positions doubled for both groups, and similar trends were recorded for more experienced individuals (Table 2).

Academic Positions

According to estimates from the SDR, the number of U.S. biomedical Ph.D.s employed in academic positions in 1981 was 35,917. This number rose to 40,176 in 1985 and remained near that level for the next 10 years (Table 3). Although there was stability in the number of academically employed U.S. biomedical Ph.D.s., there were substantial changes taking place in both the types of positions they held and in their fraction of the total number of Ph.D.s employed.

In 1981, 24,442 U.S. biomedical science Ph.D.s held tenured or tenure-track academic positions, which is comparable to the 24,082 in these positions in 1995. The number of tenure-track faculty 9-10 years post-Ph.D., however, has more than doubled since 1981. Perhaps since biomedical Ph.D.s are remaining longer as postdoctoral associates, not as many are tenured 10 years post-Ph.D. Nonetheless, the total number of tenure-track plus tenured positions has been relatively constant.

A noteworthy change in academic employment is the 65% increase in "other academic" personnel, from 7,047 in 1981 to 11,586 in 1995. This may include non tenure-track positions of teaching or research faculty and/or senior postdoctorals who are given other appointment titles.

The stability in total academic positions occurred over a period during which the total number of U.S. biomedical Ph.D.s was increasing. As a result, the percentage of employed U.S. biomedical Ph.D.s in academic (and tenured academic) positions declined (Table 4). The fraction of U.S. biomedical Ph.D.s with academic positions declined steadily from the late 1970s (when it was just under two-thirds) to the early 1990s (when it leveled off at just over half). Therefore, academia is still a major employer of U.S. biomedical Ph.D.s, representing about half of the employment, but providing jobs for a decreasing fraction of the employed Ph.D.s.

Positions in Industry

The percentage of biomedical Ph.D.s employed in industry (including the pharmaceutical and biotechnology industries) has doubled in 14 years, with 31.9% of biomedical scientists employed in this sector 9-10 years after their Ph.D. in 1995, as compared to 15.6% employed in this sector in 1981 (Table 4). A major spurt in industrial employment occurred between 1981 and 1989. In absolute numbers, 852 biomedical scientists 9-10 years post-Ph.Ds. were employed in industry in 1981, and this grew to 1,906 in 1995 12 . Thus, the importance of industry as a major employer of biomedical scientists, in both relative and absolute terms, has increased (Figure 1).

Positions in Government

The percentage of biomedical scientists employed in government has not changed substantially since 1973. Although there are some minor variations, the percentage of the employed biomedical scientists working in government jobs has remained around 10% (Table 4). These jobs include research in government laboratories, science policy positions, and science administration.

Other Positions

Employment in jobs other than those noted above has remained a small part of the total, less than 10%. For employed biomedical scientists 9-10 years after their Ph.D., it represented 6.4% of the total employment in 1973 and 7.0% of their total employment in 1995 (Table 4). This has important implications for the issue addressed later in this report of training students for careers other than in academia, industry, or government.

Future Directions for Employment of Biomedical Scientists

Projections of workforce needs are inexact, and changes in many areas affect employment trends. Therefore, although we can derive trends over the past years, we cannot predict future employment needs precisely. Some analysts believe that the increase in biomedical Ph.D. production cannot be sustained indefinitely without creating an oversupply of job seekers, and at some point it will be necessary to reach a steady state 6 . However, defining that point and determining when it is reached is difficult. Vastly different predictions can be made, depending on what assumptions are made about key variables. For example, Atkinson predicted a dire shortage in biomedical Ph.D.s 13 whereas Massy and Goldman have predicted an oversupply⁶. Key variables in the equation of supply and demand faculty retirement rates, college enrollments, migration patterns, creation of new jobs in emerging areas are notoriously difficult to forecast.

Since retirement is no longer mandatory at age 65, no one can predict when biomedical scientists will choose to retire. As longevity increases, will people opt to work longer? When people retire, will their jobs open up to be filled by younger people, or will employers tighten their belts due to economic considerations and not refill these jobs? Some groups of people are retiring earlier than in the past, and others later, making it uncertain how retirements will affect the employment pool.

If faculty retire soon after they reach 65 years of age, a significant number of jobs in academia is predicted for the next 10 years 14. Will academia need yet more faculty to accommodate the children of the baby-boomers who will reach college age in the next 10 years 13? College enrollment is estimated to increase from 14.2 million in 1995 to 16 million in 2005 15, but will this lead to increased employment of professors? Schools might opt for larger class size or computer-based learning to restrain a concomitant growth in faculty.

The rapid growth in industrial employment of the 1980s seems to have tapered off, but is this just a transient phenomenon? Hopefully, the results of the human genome project will be translated into many new medical applications, and this could produce another period of industrial expansion. With the increase of clinical, agricultural, forensic, and other applications of biological research, the need for biomedical Ph.D.s in affiliated industrial sectors will increase, as will the demand for biomedical scientists in other occupations (e.g., patent law, public policy, and science writing.) This is a very exciting time to be a biologist, and the opportunities for applications to problems of society have never been greater.

However, anxiety levels among graduate students and postdoctoral fellows have been extremely high in recent years. Some have unfulfilled expectations of securing a permanent job as a faculty member in a major research university, though they are certainly aware that for every academic position there are numerous applicants and therefore stiff competition.

Postdoctoral training has lengthened, partly because the number of tenure-track academic positions has not increased, and partly because young scientists spend more time in postdoctoral training in order to make their research records as competitive as possible.

Job expectations of predoctoral students and postdoctoral associates should be informed by the current trends in employment. Each person is different, and the kind of job that is best suited for one person may differ from the job best suited for another. Faculty mentors should encourage students to consider a wide range of employment goals. Challenging job opportunities exist in academia, industry, and other non-academic settings. Our nation thrives on the creativity of its citizens. Biomedical scientists should be encouraged to explore new areas and create new niches of employment where they can utilize their training in biomedical research.

Recommendation: Although historical trend data can be a useful guide, many changes can occur during the 10 year period from the onset of predoctoral study to the completion of a postdoctoral. Therefore, applicants for Ph.D. programs should be made aware that the job market may look quite different when they are ready to enter it than it did when they began. Research in the biomedical sciences is exciting. There is tremendous potential, but we cannot predict the impact of biomedical research advances upon the job market.

Section II: How Should Admission into Biomedical Ph.D.
Programs Be Regulated?

Trends in the Production of Biomedical Ph.D.s

The number of biomedical Ph.D.s awarded in the U.S.A. remained fairly constant during most of the 1970s and 1980s. From 1972 through 1978, the number of new Ph.D.s ranged between 3,400 and 3,600 (Table 5 and Figure 2). An increase in 1980 led to a new equilibrium ranging between 3,800 and 3,900 new Ph.D.s during the early and middle of the 1980s. In 1988, however, the number of new Ph.D.s rose to 4,369 and continued to rise to 5,878 in 1995 16,17. As a result, there was a 47.9% increase in biomedical Ph.D. production from 1987 to 1995. Between 1970 and 1994, life science Ph.D. production remained fairly constant (2,254 and 2,665, respectively) at the top 27 schools (ranked by 1970 Ph.D. production). The percentage of the total Ph.D. degrees produced by these 27 schools, however, dropped from 50.5% to 36.1% 18. Schools that historically produced fewer Ph.D. graduates had increased their output of life science Ph.D.s.

Should There Be National Regulation of Biomedical Ph.D. Production?

The increase in the number of biomedical Ph.D. graduates raises the question of whether the job market can accommodate them. If not, should regulations be imposed at the national level to dictate what the production of biomedical Ph.D.s should be? We believe that since the future job market cannot be predicted accurately, it is not advisable to limit Ph.D. production in the biomedical sciences to the number of estimated job opportunities. There are many fields (e.g., the humanities) where job opportunities are not robust, yet Ph.D. production is not capped relative to employment prospects.

Recommendation: We oppose attempts to regulate biomedical Ph.D. production at the national level. Since future employment demands for biomedical Ph.D.s cannot be determined precisely, it is inappropriate to limit Ph.D. production on the basis of guesses about the future job market. Students with strong academic qualifications and other requisite criteria and intense motivation to pursue a biomedical Ph.D. should have the opportunity to do so. It is important that only students who meet high standards be admitted and only those of proven ability be awarded the Ph.D.

Should the Workforce Needs of Schools Determine the Number of Graduate Students?

There is a symbiotic relationship between teaching, research, and graduate education. The activities of graduate assistants provide important benefits for faculty members and their institutions while offering the students unique and valuable experiences. Many schools hire graduate teaching assistants to help in undergraduate courses, which often have a laboratory component. But this cannot be a universal factor driving the number of students admitted into biomedical Ph.D. programs, since this workforce need is usually absent in medical school settings where nearly one half of the biomedical Ph.D.s are trained.

Another factor influencing the number of graduate students is the need for research assistants an important part of the workforce that carries out research supervised by faculty mentors. This apprenticeship, in turn, serves to train students in how to do research and is a mutually beneficial arrangement. The stipends and tuition for research assistants are paid either from federal sources (training grants, individual fellowships, or wages from research grants) or from the school (including state funds for publicly supported schools). Medical schools without teaching assistants may support beginning graduate students as graduate assistants and only later in their training are these students funded by their mentor's research grant. Financial arrangements reflect the synergism of the relationship the research grant pays the student's salary and tuition, while the research project obtains the benefit of the student's contributions. Finally, most applicants for faculty jobs at research universities assume there is a viable graduate program, and they expect to have graduate students helping in their research. Thus, strong graduate programs are important for recruiting the best faculty to mount active research programs and attract external funding.

The workforce needs of an institution are translated into their modes of support for graduate students. In 1994, it was estimated that 37.0% of the biomedical graduate students trained in medical schools received institutional support, 29.5% were research assistants paid from research grants, and 17.2% had traineeships from the National Institutes of Health (NIH) 1.

The forces associated with the rising production of biomedical science Ph.D.s need to be understood better. Data from NSF surveys of graduate student enrollment and support indicate that the increase in the number of biomedical Ph.D.s is correlated with growth in the number of research assistantships, suggesting that workforce needs are associated with increased production of Ph.D.s. The number of biomedical science Ph.D.s rose by 53.8% from 1980 to 1995 (Table 5). During this same period, there were declines in the number of NIH traineeship and fellowship positions and small declines in the number of biological science graduate students with teaching assistantships and "self support" (Table 6). However, the number of biology graduate students with research assistantships (from NIH, nonfederal sources, and all sources) increased dramatically from 1980 to 1995. Although the field specifications differ ("biomedical" Ph.D. data and "biological science graduate student" data on sources of support), the increase in Ph.D. production at the same time that the number of research assistantships was expanding points to the possibility of a direct connection between the two trends.

Recommendation: Applicant quality based on past performance and potential for success should be the essential criteria for admission to graduate school. Standards for admission should not be lowered and predoctoral applicants should not be admitted primarily to meet the institutional needs for teaching assistants or research assistants. Alternate ways to fill workforce needs should be considered.

How Many Students from Other Nations Should We Train?

The small rise in the number of U.S. citizens obtaining biomedical Ph.D.s appears to have been insufficient to meet the workforce needs of schools. At the same time, an increased number of citizens of other nations have come here for graduate work in the sciences and have obtained their Ph.D.s. From the early 1970s to the mid-1980s, Ph.D. production was stable, with approximately 3,500 to 3,800 degrees awarded each year (Table 5 and Figure 2). During most of this period, about 500 to 550 of the new Ph.D.s awarded each year went to individuals who identified themselves as non-U.S. citizens.*

Because much of the recent increase in biomedical Ph.D. production can be attributed to an increase in non-U.S. citizens, the question arises as to whether a cap should be imposed on the percentage of foreign students that we train. Moreover, there are some problems associated with training students from other nations. Students from other cultures frequently have to overcome difficulties with English and learn customs of a new culture. Some of the foreign students we train will return to their home countries; will they become our economic competitors? Typically, about one half of the foreign students stay in the U.S.A. 19, and they will expand the number of applicants seeking permanent employment.

Nonetheless, we believe that it is beneficial for us to attract talented citizens from other nations who wish to pursue biomedical education in U.S. institutions. In the past, scholars from abroad who work in the U.S.A. have contributed much to the advances in science and other fields. The success of our nation is based in part on the talents, creativity, and hard work of people who were at one time immigrants from other countries. We are fortunate that we can attract some of the brightest young scientists from other countries to the U.S.A. for predoctoral training, and that we can select which of these top students have the opportunity to remain for their careers.

One can ask if non-citizens have displaced U.S. citizens from graduate school in the biomedical sciences. There was a slight increase of 421 U.S. citizens obtaining the biomedical Ph.D. from 1990 to 1995, which is the same time period where the numbers of biomedical Ph.D.s. awards to non-citizens doubled (Table 5). One interpretation of these trends is that the need for biomedical predoctoral students cannot be met by U.S. citizens with sufficient qualifications, and non-citizens are recruited to fill the void. This trend continues at the postdoctoral level, where the demand for postdoctorals outstrips the supply with Ph.Ds. from U.S. institutions, so non-citizens trained in other countries are recruited. In other words, the need for predoctorals and postdoctorals in the biomedical sciences is not satisfied, and non-citizens must be recruited to satisfy this need. It may be that a greater number of qualified U.S. undergraduates do not choose to pursue predoctoral biomedical training because they hear about difficulties in obtaining academic jobs and research grants, and these considerations may be of less concern to non-citizens when weighed against other factors shaping their choices. As stated earlier, the unsatisfied need for predoctorals and postdoctrals suggests that workforce needs for research assistants are driving Ph.D. production. Creative thought should be given to whether these workforce needs can be met in other ways.

Recommendation: Students of high caliber from other nations should continue to be accepted into biomedical predoctoral programs in the U.S.A. The number of students from other nations should not be arbitrarily capped.

What Is the Status of Women and Minorities in Biomedical Science?

The percentage of women doctorates in the biomedical sciences has steadily risen, more than doubling from 19.1% in 1972 to 45.3% in 1995 (Table 7 and Figure 3). Concern remains, however, about the slow rate of progression of women through the academic ranks and the relatively small number of women reaching the highest academic positions.

Recommendation: The best qualified applicants with the most potential should be accepted into biomedical predoctoral programs without gender discrimination. Permanent employment and job advancement of biomedical Ph.D.s in the various sectors should not be affected by gender.

The percentage of underrepresented U.S. minority* Ph.D. graduates in the biomedical sciences has increased from 2.7% in 1975 to 6.4% in 1995 (Table 7 and Figure 3). Although this demonstrates improvement, it is still a very small percentage and does not reflect the proportion of these minorities in the total population of the U.S.A. Encouraging changes have taken place in the numbers of baccalaureate degrees in science and engineering earned by Blacks and Hispanics. Between 1977 and 1994, the number of Hispanics earning bachelor's degrees in science and engineering more than doubled (from 9,628 to 20,529); the number of Blacks earning bachelor's degrees in these fields rose by over one-third (from 19,552 to 26,289) 9,20. During this same period, the total number of baccalaureate degrees in science and engineering rose by only 17% (from 337,834 to 395,380).

The steepest rise has been in the most recent years, and therefore is not reflected yet in the growth of minority Ph.D.s. More needs to be known about the forces behind the rising numbers of minority bachelor's degrees in science so that these gains can be continued and expanded.

Recommendation: Admission into graduate programs should be based on applicant quality, without gender or race discrimination. There are problems in attracting minorities to the biomedical sciences. Efforts must be made to improve K-12 science education in the U.S.A. to increase the quality and quantity of science majors in college. Underrepresented minorities remain a concern, and better elementary and secondary education are a necessary part of expanding their recruitment to scientific careers. Innovative programs are needed, and efforts such as the Association of American Medical Colleges (AAMC) Health Professions Partnership Initiatives (funded by the Robert Wood Johnson and the Kellogg Foundations), and the NIH Minority Access to Research Careers (MARC) program should be continued to improve opportunities for minorities in biomedical science education.

Section III: Length and Type of Training: Are They Appropriate for the Spectrum of Opportunity?

What Is the Mission for Biomedical Graduate Education?

Although we often use the term "training," we usually mean "education." We can "train" students to use certain techniques, but we must "educate" them to have inquisitive minds, to be able to formulate meaningful biological questions, design experiments, critically analyze the results, and fathom the significance of their findings in a broad context. Most holders of the biomedical Ph.D. have careers that involve research-related activities and are part of the biomedical research enterprise. In addition, however, the skills acquired in a Ph.D. program are applicable to a variety of other careers. Biomedical Ph.D. programs can prepare students for many careers, even while maintaining their primary focus on in-depth research training.

It is crucial for biomedical Ph.D.s to be trained to communicate effectively in both oral and written formats, practice the characteristics of good teaching, be able to critically evaluate the scientific literature, use modern information technology, efficiently manage a laboratory, and uphold the highest ethical standards in research. Students also need guidance and experience in learning to function as effective members of research teams, such as those characteristic in industrial settings.

Should Predoctoral Training Be Broader and More Interdisciplinary?

It is important to be trained in depth in the techniques of a specific area of biology in order to do Ph.D. thesis research in that field. We are in an era in which many problems are dissected at the molecular and biochemical level to understand the underlying biological mechanisms. At the same time, however, biomedical research continues to investigate mechanisms at higher levels of organization that cannot be fully deduced by a simple reductionist approach. It is important to realize that certain properties from higher levels of organization are lost at the molecular level. Our students should know what the questions are at these levels of organization to be able to appreciate the significance and potential applications of their molecular findings. For biomedical Ph.D. students to be educated along the entire spectrum of biology, it is important that students be given adequate background in molecular biology, biochemistry, cell biology, systems, and whole organism (integrative) biology, and in the interaction of the organism with its environment. Exposure to the physical sciences is also important so that principles from these disciplines can be applied to biomedical research.

Recommendation: We recommend that interdisciplinary training be augmented within the subfields of biology and between biology and the physical sciences in biomedical Ph.D. programs. Biomedical Ph.D. programs that include additional clinical content (e.g., pathophysiology or pharmacology) may also benefit students seeking careers in a clinical research setting. In agreement with the Committee on Science, Engineering and Public Policy 5, we also endorse government and nongovernment sponsorship of interdisciplinary training grants to foster breadth and flexibility in biomedical Ph.D. education.

What Is the Appropriate Time Needed to Complete a Ph.D. Program in the Biomedical Sciences?

From 1973 through 1981, the median time from Bachelor's to a Ph.D. degree in the biomedical sciences remained stable at 7.0 years. It began increasing steadily in 1982, reaching a high of 8.9 years in 1995 (Table 8). Median "registered" time-to-degree is shorter than "total time," but the general trend is similar, increasing from 5.7 years in the 1970s to 6.9 years in 1995. The difference between the two "time-to-degree" measures probably reflects people taking time away from their studies after the Bachelor's degree to work in various capacities before deciding to go to graduate school.

Biomedical Ph.D. students complete their degrees in a more timely fashion at schools ranked by reputation in the top quarter than in the bottom quarter of their respective fields. Among 1987-1992 Ph.D. recipients in cell and developmental biology, the median time to degree was 1.9 years longer at schools ranked in the bottom quarter of the field than it was for schools in the top quarter 21. For molecular biology and general genetics, the comparable difference in median times to degree in the top and bottom quartile was 2.3 years. Although this differential might reflect different funding opportunities (and the greater number of students at lower-ranked schools working at other jobs to support living expenses), it might also reflect a difference between higher-ranked and lower-ranked schools in student aptitude and other factors.

Recommendation: We believe that five to six years after entry into full-time graduate study is adequate for most students to complete a biomedical Ph.D., and faculty committees should help students meet this target. Individual programs or acquisition of special skills may lengthen the time required for degree completion. The students' thesis committee should review the research progress of the graduate student at least annually. Caution should be exercised in incorporating additional requirements that might inappropriately lengthen the time to complete the Ph.D. degree, especially as additional breadth can also be obtained in postdoctoral training. The program should be carefully tailored to the needs and interests of the student. It is important for schools to exercise rigorous quality control, as it is a disservice to weaker students to let them linger on.

Importance of Mentoring in Biomedical Science Education

Efficient progress toward a Ph.D. degree benefits from the guidance of a faculty mentor and a faculty committee whose primary concern is the student's personal and professional development. In the case of postdoctorals, the faculty mentor generally assumes sole responsibility for their research guidance and help in their finding a permanent position. There should be a yearly review process (if this is not already in place) during which the faculty mentor discusses with the postdoctoral the latter's research progress, professional development, and goals for permanent employment. The counsel each student receives should be tailor-made to fit the student's unique set of skills, personality, and career aspirations.

Recommendation: Effective mentoring is an essential part of pre- and postdoctoral education. In addition to frequent, informal contacts with the faculty mentor, we recommend that formal meetings be held at least yearly with the faculty mentor to review professional development and career aspirations. In the case of predoctoral students, such review can be incorporated into their regular meeting with their faculty advisory committee.

What Are the Career Opportunities, Other than in Academia, Industry, and Government, and Shall We Educate for These Diverse Careers?

The majority of biomedical Ph.D. graduates seek careers in academia and industry, with increasing numbers finding positions in the industrial sector in recent years (Figure 1). Some also pursue careers in government, and there is increasing interest in more diverse career options. The number of biomedical Ph.D.s students in career positions outside of the more traditional careers within academia, industry, and government remains a small fraction of graduates (7.0% of biomedical Ph.D.s 9-10 years after their Ph.D.; see Table 4), and it is not reasonable to restructure entire graduate programs to train students for these more diverse options. Some schools may be ideally situated to offer dual degree programs (e.g., degrees in law or business coupled to the biomedical M.S. or Ph.D.), but this may be the exception rather than the norm. Nonetheless, faculty should be supportive of students who wish to use their biomedical Ph.D. for positions in junior colleges, patent law and regulatory affairs, business, public policy, science administration, science writing, etc. It has been suggested that there should be much greater entry of Ph.D. scientists into K-12 science education. However, K-12 teaching is very different from biomedical research and the aptitudes, education, and requirements for success are not the same. Most of the nontraditional career options require unique sets of skills, attributes, and credentials that differ substantially from those needed in a biomedical Ph.D. program.

For some of these diverse careers, a Master of Science degree with a short thesis based on bench research may be sufficient. Some schools may wish to emphasize the Master of Science degree in biomedical science alone or coupled with a degree in another field (e.g., law, finance, journalism, etc.). Consideration may be given to Master's-level degree programs directed at K-12 science teachers, with the aim of improving K-12 science education.

Recommendation: Graduate programs should expose their students to the variety of diverse career opportunities, and faculty should be supportive of biomedical graduate students who consider pursuing these diverse career possibilities. Since the number of positions outside of academia, government, and industry is currently a small percentage of positions (although more opportunities may appear in the future), graduate programs should not be totally restructured to prepare students for these diverse careers. However, graduate programs should be sufficiently flexible to allow individual students to broaden their education consonant with their career goals. Opportunities for students to explore multiple career options, such as internships in industry or exposure to teaching, should be encouraged where feasible. It is quite likely that the number of career opportunities for biomedical Ph.D.s in many diverse areas may grow in the future as spin-offs from the practical application of biomedical research; these trends should be closely monitored.

Section IV: How Should the Quality of Biomedical Graduate Programs Be Measured?

What Defines Success of a Biomedical Training Program?

The success of predoctoral educational programs can be assessed by many measures, one of which should be testing the congruence of the program's mission and goals with the career outcomes of its students. U.S. universities and academic health centers offer a broad spectrum of high-quality research training opportunities in many different disciplines. Graduate students select one or another training program because of the unique training opportunities they offer. For this reason alone, there can be no single standard or mechanism to assess program quality or success.

Although the award of a Ph.D. degree does not guarantee a job, students who have performed at a high level will have a reasonable expectation that this will make them desirable candidates for permanent employment. The success of a program's graduates in obtaining career-enhancing postdoctoral associateships or permanent positions is one measure of the program's success. There are characteristic differences in the types of employment offered to students from different programs. For example, some programs are most successful in placing their students in industry, while others characteristically place their students in academia.

Biomedical training programs should track their predoctoral graduates for 15 years, using a thorough and standardized methodology. The information on job outcomes should be made publicly available. This will be especially helpful to applicants who wish to see if their career goals match the mission of the training program they are considering, and to see the success of that program in job placement of its alumni. Programs with poor track records in career placement of their graduates may attract fewer students if such information is in the public domain.

Recommendation: Career outcomes of students are one measure of success in a biomedical graduate program. Programs sponsoring predoctoral education should provide prospective graduate students with information on the career outcomes of all predoctoral students who have completed training over the past 15 years, as this will assist students in their selection of programs matching their own goals.

How Should Graduate Programs Be Reviewed?

The development of new graduate programs, often in institutions that lacked them, and the increase in biomedical Ph.D. production has led some to propose national accreditation. Most graduate programs are already subject to multiple reviews by national, regional, and state accreditation processes. However, the most meaningful review usually occurs at the local level of the individual graduate program.

Recommendation: The most effective review of the success of graduate programs occurs at the local level through frequent self-study and periodic reviews by knowledgeable scientists from other academic institutions and industry. The process should be structured in such a way as to assure frank and full discussion of all program elements and practices. To conserve resources and valuable time of faculty and administrators, redundant reviews by national, regional, and state accreditation bodies should be eliminated wherever possible. We oppose efforts to introduce further accreditation at the national level. It would be impractical because each program is unique; and national accreditation might lead to homogeneity of programs, which is undesirable, at the expense of diversity, which is desirable.

What is the Role of Peer-Reviewed, Competitively Awarded NIH and NSF Training Grants in Setting and Maintaining Standards of Excellence in Biomedical Graduate Education?

NIH and NSF training grants set a "gold standard" against which to measure quality of students, faculty, and programmatic aspects of graduate education. Pre- and postdoctoral education programs supported by peer-reviewed and competitively awarded NIH and NSF training grants help institutions to establish and maintain standards of excellence in pre- and postdoctoral education. These programs are a model for constructive educational and institutional change.

Competitively awarded, peer-reviewed individual fellowships are a second, equally valuable mechanism for supporting pre- and postdoctoral education. They give pre- and postdoctoral fellows a measure of their self-worth and provide them with a measure of financial independence. They do not, however, have as much impact on the structure of the graduate program as do institutional training grants. On the other hand, they provide a measure of an institution's ability to attract the most outstanding individuals as predoctoral students and postdoctoral fellows.

Application for a federal training grant involves self-evaluation and extramural peer review. Such a process of self-evaluation and external review is useful for schools to undertake on their own, even if not applying for an NIH training grant. Currently, the majority of NIH predoctoral training is sponsored by NIGMS. There are 75 schools, representing about 1/3 of those that offer a Ph.D. in the biological sciences, that are recipients of one or more NIGMS training grants. Of these 75 schools, however, 20 held about 65% of the traineeships (approximately 2,300 predoctoral training positions) in 199522. Since federal training grants enhance programmatic aspects of graduate education, it is desirable to increase somewhat the number of schools that hold NIH training grants, while still maintaining rigorous standards of funding only excellent programs.

Recommendation: Institutions should aspire to have high quality graduate programs that are competitive for federal training grant support. Programs that for reasons of size, critical mass, or programmatic emphasis are not appropriate for training grant support should, nonetheless, view NIH and NSF training programs as models for designing and implementing excellence in education in biomedical research. Since federal training grants enhance programmatic aspects of graduate education, it is desirable to spread them out to a somewhat larger number of schools, providing that their programs are high quality and merit federal support. Additional increases for pre- and postdoctoral stipends are desirable, but money for stipend increases should not cause a decrease in federal funds allocated for investigator-initiated research grants. Other areas of the NIH portfolio (e.g., contracts and construction) should grow more slowly, if necessary, to maintain the priority growth of research project grants and trainee/fellowship stipends.

References

• 1. Ammons, Stanley. (1995) Association of American Medical Colleges Survey of Ph.D. Students in U.S. Medical Schools, 1994-1995. Paper presented at October, 1995 GREAT Conference, Ft. Lauderdale, FL. Washington, DC: AAMC.
  

   

• 2. National Science Foundation. (1995) Selected Data on Graduate Students and Postdoctorates in Science and Engineering: Fall 1995, Supplementary Data Release Number 2: by Enrollment Status. Washington, DC: NSF.

• 3. Bresnahan, Patricia, John Baker, Seema Kantak, Laura Mifflin, David Olson, Baerbel Rohrer, Renee Williard and Letitia Yao. (1997) Final Report for 1996 Survey of Postdoctoral Scholars. San Francisco: UCSF Postdoctal Scholars Association.

• 4. Kennedy, Thomas J. (1994) "Graduate education in the biomedical sciences: Critical observations on training for research careers." Academic Medicine 69 (October): 779-799.

• 5. National Research Council, Committee on Science, Engineering, and Public Policy (COSEPUP). (1995) Reshaping the Graduate Education of Scientists and Engineers. Washington, DC: National Academy Press.

• 6. Massy, William F., and Charles Goldman. (1995) The Production and Utilization of Science and Engineering Doctorates in the United States. Palo Alto: Stanford.

• 7. Hayes, Sharon L. (1996) "At the edge of a new frontier: a profile of the Stanford University biomedical Ph.D. class of 1996 and recommendations for the future. Report of the BioMedical Association of Stanford Students." Stanford, CA: BioMASS.

• 8. National Research Council, Committee on Trends in Early Research Careers. (In preparation, 1997) Committee Report. Washington, DC: National Research Council.

• 9. National Science Board. (1996) Science & Engineering Indicators 1996. Washington, DC: National Science Foundation. NSB-96-21.

• 10. National Science Foundation. (1996) Characteristics of Doctoral Scientists and Engineers in the United States: 1993. (NSF-96-302). Washington, DC: National Science Foundation. and National Science Foundation. (1997) Data Brief: Unemployment Among Doctoral Scientists. Washington, DC: National Science Foundation.

• 11. Perkins, John. (1996) "Are U.S. universities producing too many Ph.D.'s in the biomedical sciences: Facts and artifacts." The Pharmacologist. 38:124-128.

• 12. National Research Council. (1996) "Special Tabulations." Survey of Doctorate Recipients. Washington, DC: NRC.

• 13. Atkinson, Richard C. (1990) "Supply and demand for scientists and engineers: A national crisis in the making." Science 248: 425-432.

• 14. U.S. Department of Labor, Bureau of Labor Statistics. (1996) "Biological and Medical Scientists" and "College and University Faculty." Occupational Outlook Handbook. Washington, DC: US GPO.
  

   

• 15. Stiles, William A. (1996) "Higher education on the threshold of change." Paper presented in the session on Federal Role in Education: Effects of Changing Attitudes," AAAS Annual Meeting. Baltimore, MD.

• 16. Committee on National Needs for Biomedical and Behavioral Research Personnel. (1994) Meeting the Nation's Needs for Biomedical and Behavioral Scientists. Washington, DC: National Academy Press.

• 17. National Research Council. (1995) Survey of Earned Doctorates. Washington, DC: NRC.

• 18. Goldman, Charles A. (1996) Personal communication based upon analyses of NSF CASPAR data base. Santa Monica, CA: Rand Corporation.

• 19. Finn, Michael G., Leigh Ann Pennington, and Kathryn Hart Anderson. (1995) Foreign Nationals Who Receive Science or Engineering Ph.D.s from U.S. Universities: Stay Rates and Characteristics of Stayers. Oak Ridge Institute for Science and Education.

• 20. National Science Foundation. (1996) Science and Engineering Degrees, by Race/Ethnicity of Recipients: 1987-1994. Detailed Statistical Tables. (NSF 96-329). Washington, DC: NSF.

• 21. National Research Council. Committee for the Study of Research Doctorate Programs in the U.S. (1995) Research Doctorate Programs in the U.S.: Continuity and Change. Washington, DC: National Academy Press.

• 22. Norvell, John. (1996) Personal communication. Bethesda, MD: National Institute of General Medical Sciences.

Source: National Research Council, Survey of Doctorate Recipients

^*     Data shown are weighted estimates from sample surveys. The first line contains estimates for all biomedical scientists, while the succeeding lines present estimates for components of the population classified according to number of years since receipt of the doctorate.

Changes in survey methodology complicate some of the comparisons over time. In 1991, a significant effort was undertaken to improve accuracy of the survey estimates by increasing the response rate. Additional activities to contact non-respondents were employed to increase the reliability of the surveys, and these procedures have been maintained in the 1993 and 1995 surveys. While yielding more accurate estimates, it is not possible to determine whether changes between surveys pre- and post-1991 are due to methodological practices or substantial developments in the experiences of scientists. Trend data for this period, therefore, need to be interpreted cautiously.

As part of the more intensive follow-up program, the data collection period for the survey was pushed further back in the academic year for many survey respondents. It is possible that, for some respondents, the change in the timing of the survey had an effect on the information reported.

In addition, changes were made in the survey questionnaire in 1993. These modifications, implemented to collect more precise information, may have affected the reporting practices of the respondents, and as a result, created an additional (but unmeasured) impediment to comparisons over time.

Table 2
Percentage of Employed Biomedical Ph.D.s Who Were Postdoctoral Associates, 1973 - 1995** (U.S. Citizens Only)
	1973	1977	1981	1985	1989	1991	1993	1995
All Postdoctoral **
Total Employed	5.7	8.7	9.6	8.3	8.9	7.3	9.1	9.7
1 - 2 years post-Ph.D.	26.5	43.9	52.0	50.1	56.9	42.6	49.5	58.0
3 - 4 years post-Ph.D.	7.3	14.7	23.6	24.0	28.0	24.9	32.0	32.1
5 - 6 years post-Ph.D.	2.2	5.4	8.9	11.3	13.2	6.3	14.9	15.2
7 - 8 years post-Ph.D.	1.5	3.4	3.3	2.5	6.1	5.4	6.4	7.6
9 - 10 years post-Ph.D.	0.3	1.5	0.9	2.3	1.7	1.8	4.0	3.1
more than 10 years	0.1	0.4	0.4	0.4	0.4	0.4	0.6	0.8
Academic Postdoctoral
Total Employed	4.5	7.0	8.0	6.3	7.1	5.6	6.8	7.1
1 - 2 years post-Ph.D.	20.7	35.8	44.1	39.7	45.9	32.6	36.8	42.2
3 - 4 years post-Ph.D.	6.2	11.7	19.2	18.8	22.6	19.0	24.7	23.3
5 - 6 years post-Ph.D.	1.8	4.3	7.0	7.5	10.2	4.8	11.1	12.0
7 - 8 years post-Ph.D.	1.0	2.3	3.0	1.5	4.6	4.7	4.6	5.5
9 - 10 years post-Ph.D.	0.3	1.5	0.5	1.5	1.4	1.2	2.8	2.4
more than 10 years	0.1	0.3	0.3	0.2	0.3	0.3	0.4	0.6
Non-Academic Postdoctoral
Total Employed	1.2	1.7	1.6	2.0	1.8	1.7	2.3	2.6
1 - 2 years post-Ph.D.	5.8	8.1	7.9	10.4	11.0	10.0	12.7	15.8
3 - 4 years post-Ph.D.	1.1	3.0	4.4	5.2	5.4	5.9	7.3	8.8
5 - 6 years post-Ph.D.	0.4	1.1	1.9	3.8	3.0	1.5	3.8	3.5
7 - 8 years post-Ph.D.	0.5	1.1	0.3	1.0	1.5	0.7	1.8	2.2
9 - 10 years post-Ph.D.	n.a.	n.a.	0.4	0.8	0.3	0.6	1.2	0.7
more than 10 years	0.0	0.1	0.1	0.2	0.1	0.1	0.2	0.2

Source: National Research Council, Survey of Doctorate Recipients

*See footnote to Table 1.

** A detailed definition of "postdoctoral appointment" was provided for the first time in the 1979 Survey of Doctorate Recipients. Differences in survey results obtained before and after this date may reflect changes in questionnaires as well as temporal trends.