The Arc of the Academic Research Career: Issues and Implications for U.S. Science and Engineering Leadership: Summary of a Workshop (2014)

3

Getting Started

For young people aspiring to academic careers, the altered academic scene that Shirley Malcom and Robert Hauser described forms the context of their efforts to launch their professional lives as scientists. Amid the changes in the composition and size of the cohorts seeking to become faculty members and the increasing economic instability of both universities and research funding, something else besides the structure of the academic career has remained essentially constant: the number of faculty positions offering the possibility of tenure.

A mix of permanence and change has therefore produced “an excess supply of Ph.D.s and very uncertain demand [for their services in academe], given the challenges facing the contemporary university,” said Donna Ginther, a professor of economics at the University of Kansas, introducing the discussion of the early years of the traditional academic arc. Of course, many Ph.D.s intend to pursue careers outside of academia, but the purpose of the workshop was to look at the academic career path.

The challenges facing young people aspiring to follow that path have become increasingly demanding and their chances of attaining their goal increasingly diminished. As Ginther and Malcom suggested, however, universities continue to admit graduate students and recruit postdocs based not on the career opportunities awaiting them but on the supply of grant funding and on professors’ and departments’ resulting need for their labor in laboratories and classrooms.

Today, competition is most intense, at least numerically speaking, at the earliest stages of the career. Four or five decades ago, new Ph.D.s who studied under well-recognized professors had a reasonable assurance of landing an assistant professorship if they did well in their studies. Today, far more new Ph.D.s compete for a pool of assistant professorships that has grown much more slowly, and the chances of reaching that crucial first toehold on the tenure ladder

have shrunk. This is one reason why many young scientists who cannot launch independent academic careers accept postdoctoral positions as researchers in professors’ labs.

Ostensibly providing a year or two of additional training to equip the new Ph.D.s to better compete for an assistant professorship, Ginther noted that these appointments in fact more often become low-paid academic jobs supported by soft money that can last 5 years or more. In practical terms, postdocs are “the workhorses of the research community,” said Mary Ann Mason. Providing cheap, highly skilled labor paid for out of professors’ grants, postdocs were 43 percent of the first authors of papers in Science, according to a 1999 study, she noted (see Box 3-1).

Because of rising production of Ph.D.s and stagnant numbers of professorships, the number of postdocs has grown rapidly over the years. As Joan Girgus, professor of psychology and special assistant to the dean of the faculty on gender equity at Princeton University, noted, the postdoc population at Princeton University has doubled in a decade, for example. Indeed, as several participants noted, no one knows exactly how many postdocs there are in the country and most published figures are underestimates. Girgus noted that although Princeton limits the term of a postdoctoral appointment to three years in most fields and five years in biomedicine, many of these postdocs then move into research positions that are essentially the same but are given a different title. She added that “We do not know how many of those there are.”

As generally poorly paid temporary workers, postdocs have a very ambiguous status on most campuses, neither students nor faculty. “My sense about postdocs everywhere is that they are a pretty much forgotten population,” said Goldenberg. For most of these “trainees,” the appointment becomes either a prelude to further ambiguous soft-money academic positions or a jumping-off place for seeking work outside of academe, be it in industry, government, or elsewhere. The academic world has long regarded such work as “alternative careers,” and most graduate programs do not provide adequate preparation. As

Malcom noted, however, the so-called alternative is now the norm for the majority of Ph.D.s in most scientific fields.

THE ECONOMICS OF EARLY CAREERS

Ginther began the analysis by stressing “that each academic field is a separate labor market. You can’t combine all of science together and say, ‘This is what is happening.’ You can’t combine all academics together and say, ‘This is what is happening.’” Salary rates differ among academic fields “because there are different demands for our services, both inside and outside of academia.” Faculty members who have real opportunities to earn good incomes off campus, such as physicians, lawyers, economists, and some scientists, tend to command higher incomes on campus, too.

Ginther argued that science breaks down into four broad categories: life science, including biomedical, agricultural, and environmental sciences; physical science, including chemistry and physics, geosciences, and math and computer sciences; social science, including psychology, sociology, and economics; and engineering. These broad disciplinary categories, however, all contain large numbers of specialized fields and subfields, each of which, to some extent, constitutes a distinct labor market.

Three of these four categories have seen significant growth in the number of Ph.D.s produced in recent decades, Ginther noted. Between 1980 and 2010, annual production doubled in the life and physical sciences, from under 6,000 to almost 12,000 and from 4,000 to 8,000, respectively. Engineering Ph.D. production grew from 2,000 in 1980 to 8,000 in 2010. Growth was slower in th social sciences, from 6,000 to 8,000 Ph.D.s per year (see Figures 3-1 and 3-2). Fueling much of this growth were rapidly rising levels of federal research funding, especially in the biomedical sciences through the National Institutes of Health (NIH). Departments and professors used these funds to support the graduate students and postdocs who provided labor in their labs.

But while “Ph.D. production has been zooming,” Ginther continued, the number of tenure-track positions has grown only in engineering, where faculty ranks have increased by 50 percent over 30 years, As a result, Ginther surmised that almost half of new life science Ph.D.s and the “majority” in other fields have been finding work outside of academe. On campus, meanwhile, the “overwhelming majority” of new jobs in science has occurred in what Ginther called “non-track” positions that do not provide tenure opportunities. These include postdoc appointments as well as other kinds of soft-money research positions. Some of these jobs are essentially postdoc appointments by another, often ill-defined, title, such as staff scientist or research associate. Other positions, with names such as research assistant professor, do provide the opportunity to compete for grants as an independent investigator, but these positions are less secure because they are not eligible for tenure and rely primarily on external funding. Although these positions make it possible to

Figure 3-1 Number of doctorates in four broad fields (engineering, life sciences, physical sciences, and social science): 1980–2010.
SOURCE: 1980-2011 Survey of Earned Doctorates. National Science Foundation, Human Resources Statistics Program, National Center for Science and Engineering Statistics. Arlington, VA.

Figure 3-2 Growth rate of the number of doctorates in four broad fields (engineering, life sciences, physical sciences, and social science): 1980–2010.
SOURCE: 1980-2011 Survey of Earned Doctorates. National Science Foundation, Human Resources Statistics Program, National Center for Science and Engineering Statistics. Arlington, VA.

participate in research, many participants believed that in general they provide less prestige, less independence, less job security, less income, and fewer benefits.

Fewer than 2,000 new biomedical Ph.D.s appear to have taken postdoc appointments in 1970, as opposed to nearly 6,000 in 2008 (see Figure 3-3). The current postdocs also stay in those positions longer. The percentage of new biomedical Ph.D.s taking postdocs “has hovered around 80 percent since the 1990s,” said Ginther. As Ph.D. “numbers have increased,…you have this huge increase in the number of postdocs. You [also] see an increase in the number of behavioral and social [scientists] taking postdocs. Chemistry, [has had a] slight uptick, but not the huge increase in postdocs that you see in biomedical sciences.”

Figure 3-3 U.S. life sciences Ph.D.s with postdoctoral research plans compared to chemistry: 1970–2008.
SOURCE: Survey of Earned Doctorates, National Science Foundation, Human Resources Statistics Program, National Center for Science and Engineering Statistics. Arlington, VA; Prepared for the National Institutes of Health Advisory Committee to the Director Biomedical Workforce report (2012).

In a number of scientific fields, especially in life sciences, the prevalence of postdoc appointments means that for most students, earning a Ph.D. does not produce “a very high return” on the investment of time, effort, and opportunity involved in getting the degree, Ginther explained. Shortly after receiving the doctorate, “a Ph.D. in biological science makes 1.4 times what a BA in biological science or basic biomedical does”; this represents the lowest return of the four major science fields. This depressed income level, she said, is “driven by the postdoc,” which can pay as little as $40,000 a year. Ph.D.s who leave academe at this point “have a higher return.” “To take a postdoc,” Ginther explained, “a student is making a large financial sacrifice.” Another result of overproduction, Ginther added, is that “between 1997 and 2008, in almost every broad field, people are less likely to be using their Ph.D. to be employed in occupations that match their Ph.D. field. The one exception is math and computer sciences, and to a lesser extent…in psychology and social science.”

In many fields, however, the postdoc has become entrenched as a de facto requirement for an assistant professorship, with Ph.D.s now routinely spending several years—including, according to Ginther, a “nontrivial percentage” who “have more than 8 years in a postdoc.” Those who get to the tenure track today are therefore older than new assistant professors of past generations.

Moreover, even for the minority of scientists who do attain an assistant professorship, winning research funding has become more difficult than in past decades, in part because of the larger numbers of people, including many with non-tenure track university appointments such as research professorships, who are eligible to compete. “Since the 1990s, there has been a decoupling between the age of getting that first tenure-track job and the age of getting that first RO1,” the NIH grant that begins a scientist’s career as an independent investigator, Ginther said (see Figure 3-4).

“My research has shown that the best way to get an NIH grant is to have already had an NIH grant,” she continued. As a result, today, “the age of independence and an independent research career in biomedicine is over 40,” as opposed to the situation several decades ago when scientists in their thirties and even younger often won their own funding. The rising age of “new” researchers became “so worrying that [Elias] Zerhouni, when he was head of NIH, instituted affirmative action for young investigators…Our young scholars were being shut out of the process in favor of older scholars” as funding became increasingly scarce and competitive, she said.

Overall, Ginther concluded, “It seems [that] in the life sciences, people are hanging out in postdocs, waiting for that academic job that…isn’t going to happen… . We are taking the most productive years of our biomedical scientists’ lives, and putting them in postdocs, where the opportunity to have an independent career is highly uncertain.”

Figure 3-4 Average age at Ph.D., first tenure track position, and first R01/RPG (first major grant) for biomedical doctorates: 1980–2007.
SOURCE: Survey of Earned Doctorates, National Science Foundation, Human Resources Statistics Program, National Center for Science and Engineering Statistics. Arlington, VA; and National Institutes of Health IMPAC II; Prepared for the National Institutes of Health Advisory Committee to the Director Biomedical Workforce report (2012).

ASPIRATIONS AND REALITIES

Given these realities, what motivates new Ph.D.s to take postdoc appointments? How do they adapt to the fact that most do not emerge as assistant professors? “Do they know what they are getting into?” asked Henry Sauermann. “Do they actually know what these labor market odds are?” He and colleagues have investigated these questions in a number of surveys.

In looking at what he calls “the pre-beginning of the arc,” he explained, his research examines the supply side of the academic labor market “not in terms of numbers of pieces, but in terms of people who actually have reasons to do these things. We want to understand what these reasons are because that also speaks to…what they are expecting and what society [has] to deliver to…fulfill its contract on the other side.” Specifically, he wants to understand graduate students’ and postdocs’ “decisions to enter an academic career (or not),” their “preferences for various types of careers,” and “how these preferences change over time.” In addition, he has looked into postdocs’ level of awareness of the job market, whether they regret “ having done a postdoc,” and “how [their] ex ante preferences and postcareer outcomes match up” over time.

“We heard earlier that a lot of people…enter nonacademic careers,” he continued. But is that an “accident” and an undesired outcome, or “actually

something they planned for as they start[ed] their Ph.D.s?” An initial survey sent to 30,000 postdocs and graduate students at 40 top-tier institutions and follow-up surveys of respondents 3 years later provided some answers.

“Implicit in many of our discussions is the assumption [that] people do Ph.D.s and postdocs because they want to be academics,” Sauermann said. “A lot of advisors, from what we hear, have that assumption” and often “imposed” it on their students. But, he emphasized, expanding on Ginther’s opening point, in examining motivations as well as labor markets, “we really have to think across different fields. [The preferred career] is not the same in the life sciences as it is in engineering,” where many people choose to work in industry. “We know the actual career paths are different, and it is not surprising that the preferences are quite different, as well.”

Asked to pick their preferred career from choices that included a faculty appointment that emphasizes teaching, a faculty appointment that emphasizes research, government work, a job in an established private firm, a job in a startup company, or some other career, Ph.D.s in life sciences, chemistry, physics, and engineering expressed different preferences. By wide margins the life scientists and physicists chose faculty with a research emphasis first and faculty with a teaching emphasis as their second choice. The chemists and engineers, however, named work in an established firm as their top preference, although each group also chose faculty as their second choice, the chemists preferring a teaching emphasis and the engineers choosing a research emphasis. “The point is,” Sauermann says, that these Ph.D.s have “a very broad range of career interests. Academia is an important career interest for these people, [but] it is not the only one.”

Scientists’ career interests are not only broad, Sauermann’s research shows, but also change over time. In 2010 and again in 2013, the same group of graduate students in life sciences, chemistry, physics, and engineering rated the attractiveness of careers as university faculty with an emphasis on research, in government, in an established firm, or elsewhere. Regardless of discipline, the allure of the faculty post declined during the interval, with the steepest decline among life scientists and physicists. The desirability of working for an established firm rose for the life scientist and chemists and remained unchanged for the physicists. In all groups, “other,” which might include work for a startup firm or other endeavors altogether, also rose.

What dulled the lure of the academic life? Some insight comes from openended comments by respondents who had initially named a research-oriented professorship as their first choice, Sauermann said. He offered a few “nonrandom” examples of the most pointed comments:

• “Realizing that university faculty usually spend most of their time on activities supporting the research but not the research itself,”
• “I got tired of doing work that only matters to a handful of people, has no impact on society, and pays poorly,”

• “I don’t want to be my advisor. Ever.”
• “I’ve discovered that I’m a mediocre scientist but a really good teacher, and teaching makes me happy.”
• “The realization that very few tenure-track positions actually exist, the lack of research funding, and too many qualified applicants.”

A common theme among the graduate students’ comments, Sauermann noted, was their realization that “academic research is not actually doing academic research [but] managing, fundraising, struggling with grants, and all these issues… . A lot of these people go into science because they want to be at the bench, study stuff, discover stuff and not just tell other people what to study… . For some of these students, that seems to be one of the insights that they learn while they are watching their advisors and seeing how it plays out.” A number of people also wrote, “I want to start my own company, or I want to go work for a big pharma company, because that is where I know stuff actually happens and goes down the pipe[line], and hits people in terms of…[new] drugs…” Someone else, however, said, “I can make the biggest impact providing a really cool, important paper.” In Sauermann’s opinion, “both are…important impacts, but we have to understand that people see these things in different ways,” depending on their own values and experiences.

The career preferences that postdocs expressed were a bit different from those of the graduate students, however, perhaps because people with nonacademic bents had already taken other directions after finishing their Ph.D. Asking life sciences, chemistry, physics, and engineering postdocs to choose their most preferred career from among a faculty job emphasizing teaching, a faculty job emphasizing research, and working for government, an established firm, a start-up firm, or other, showed that faculty with a research emphasis led in all groups by very wide margins.

Given the very tight academic job market, however, do Ph.D.s enter their postdoc appointments aware that they face such low odds of ending up in the faculty job they hope for? Perhaps, Sauermann suggested, “they don’t know what the labor market conditions are. Maybe they know, but they think they are at the top end of the distribution. Maybe they actually do a postdoc for other reasons, [not intending] to become the academic researcher… . Maybe they actually want to do something else.”

Asked what percentage of Ph.D.s in their field land a tenure-track job within 5 years of earning the doctorate, the four groups of postdocs surveyed each gave estimates that on average came within 3 percentage points of the actual figures, which in these four broad categories stand only in the mid-teens. “They are smart people, and they read Science Careers,” Sauermann said. “They read the discussions; they know quite well that the odds of getting a position are not that great.”

But does this mean they’ve given up hope of getting one of those coveted posts for themselves? The majority of the postdocs—61 percent of the life

scientists, 55 percent of the chemists, 63 percent of the physicists, and 47 percent of the engineers— said that having an academic research career was indeed their motive for pursuing a postdoc. Despite their accurate knowledge of the job market, those aspiring to a tenure-track job still viewed their own chances of having one within 5 years as quite a bit better than the low averages they had quoted; they estimated a 55 percent chance for the life scientists, 68 percent for the chemists, 42 percent for the physicists, and 58 percent for the engineers.

“The labor market expectations, at the level of the global labor market, seem to be quite well calibrated,” Sauermann noted, but “maybe not so much at the level of the individual person.…Any given person could truly be the next Nobel Laureate, right? We can’t tell them ‘No, you are overestimating your chances.’ Any given person might be right, but we know that, collectively, they can’t all be right.”

Advisors, however, often have considerable influence over the aspiring scientists they teach. “We found that about 80 percent of students [say] that my advisor most strongly supports an academic research career, even though we know most of them don’t get those careers.” Young scientists, Sauermann continued, need “a broader set of information sources…to understand labor market and the different kinds of careers [available to them] and make informed career decisions.”

But society also needs to know which students end up pursuing which kinds of careers, Sauermann added. “Which students get the jobs they want? Which ones don’t?…How is it related to ability? Is it that the really, really smart people in the end get the academic positions? Is it the ones with the lowest opportunity costs, who have nothing to lose as they just stick around in the postdocs? Is it people with the biggest persistence, who are just determined to do it, and they stick around and don’t get lured away by something else?”

Beyond that, “How selective do we want the [career selection] mechanism to be? What happens to people who don’t get selected or don’t select into these different careers? How can we evaluate the career outcomes we see?”

FAMILY MATTERS

Despite the complex motives that Sauermann’s research revealed, the reason that a great many postdocs lose interest in careers as research-oriented faculty members is simple and straightforward, Mason found in her surveys of the University of California-Berkeley faculty. Her recent book is entitled Do Babies Matter?: Gender and Family in the Ivory Tower,⁶ and the answer, at least for women in the early stages of academic careers—meaning before tenure—is that the desire to start a family must often compete with the desire to have an academic research career. As Girgus also noted, “graduate students and

_____________________

⁶ Mary Ann Mason, Nicholas H. Wolfinger, and Marc Goulden. Do Babies Matter?: Gender and Family in the Ivory Tower. Rutgers University Press (2013).

postdocs, like faculty, are often trying to integrate a productive work life with family formation, something that we have come to recognize as important only recently.”

The conflicting desires are most apparent in the life sciences, where women now earn more than half the Ph.D.s. And because of overproduction of Ph.D.s fueled by NIH funding and the use of graduate students as low-cost lab workers, life sciences is also the “hardest in terms of getting a [faculty] job” and where young scientists face the greatest risk of “staying in your postdoc forever.” Life sciences is therefore the field “most hard hit by all the demographics that I am going to be telling you [about],” but the same trends exist in the other fields as well, she said.

Moving from the postdoc to the tenure track is unlikely for the great majority of scientists, but is unlikeliest of all for married mothers, who are 35 percent less likely than married fathers to make that jump, Mason said. “The major slide of women out of the academic pipeline after getting their Ph.D.s occurs before they get their tenure-track job, usually in the postdoc years, or they have changed their minds in the graduate student years,” Mason continued.

When postdocs become parents, “twice as many women [as] men are likely to change their career goal away from being a research professor… . Forty-one percent of women who had babies when they are postdocs say that they no longer want to have a research career, as opposed to 20 percent of men.” The trend is even sharper, and the difference from men is even greater, among women who have babies while in graduate school, she added. Many “babies [are] born during the postdoc years,” which, given postdocs’ ages—generally the early to middle thirties—“is natural,” especially since fewer women have babies while graduate students, at least at research universities, she said. When asked why they changed their career plans, female graduate students at the University of California-Berkeley “almost uniformly said for family reasons. ‘I want to be able to enjoy having a family, [being a] mother and wife, which are close to impossible when one chooses academia. The clock is ticking, and it doesn’t stop for anything or anyone.’ “

“The main problem for all women in academia who have babies is they just don’t have enough time,” Mason said. Berkeley postdocs who are mothers average 100 hours of work a week: 43.6 doing research, 3 teaching, 15.9 doing housework, and 36.3—almost a second full-time job—giving care. Postdoc fathers put in more time in the lab—51.6 hours—but only 2.4 on teaching. And they put in strikingly fewer hours working at home—12.6 on housework and 20.7 in child care. Regardless of gender, postdocs without children, on the other hand, work a “mere” 80 hours a week, more than 50 of them in the lab (see Figure 3-5).

Beyond a crushing workload that includes less time spent in professional competition, postdoc mothers often face another major disadvantage. At the many universities that classify them as trainees rather than employees, “they are not covered by the [Family and Medical Leave Act, which provides unpaid time

Figure 3-5 Workload by type of work (caregiving, housework, teaching, and research activities) for U.S. postdoctoral scholars with and without children, by gender: 2009.
SOURCE: Goulden, M., M. A. Mason, and K. Frasch. “UC Postdoctoral Career Life Survey”. Berkeley, CA: UC Berkeley, (2009). Available at http://ucfamilyedge.berkeley.edu/grad%20life%20survey.html.

off to care for family members, or by] Title 7,” which bans employment discrimination based on pregnancy. Some universities, however, include graduate students and postdocs in the maternity leave plans that they provide to faculty members. The University of California, for example, is among the minority of universities that give graduate students and postdocs at least 6 weeks of paid maternity leave (see Figure 3-6). Princeton provides them a range of benefits, Girgus noted, including assistance paying for child care, backup child care in case of illness, financial assistance to provide child care during professional travel, and more. Most universities, however, can afford to provide only a good deal less. Said Mason, “There is a lot of room for improvement at the graduate student and postdoc years, if we don’t want to lose women out of the pipeline.”

Figure 3-6 Percentage of institutions in the Association of American Universities (N = 62) which provide at least 6 weeks of paid maternity leave for academic populations: 2008.
SOURCE: Mason, M. A., M. Goulden, and K. Frasch. “Family Accommodation Policies for Researchers at AAU Universities Survey.” (2008).

This page intentionally left blank.