Because PATH has funded just 13 clinically trained researchers (MDs, MD/PhDs, and DVM/PhDs), we combined these groups. They finished their terminal degrees at an average age of 29.2 and, on average, landed faculty jobs at 36.5, close to the mean ages at appointment for MDs and MD/PhDs reported from the Association of American Medical Colleges (AAMC) data in the National Research Council’s Bridges to Independence report.20 Their first R01 came at an average age of 39.5, compared with the NIH’s averages of 43.4 for both MD/PhDs and MDs.22 Clinically trained PATH awardees were promoted to associate professor at an average age of 42.6. The median time to faculty promotion for all PATH awardees, 6.0 years, compares well to the AAMC’s measure of 6.2 years for PhDs, MDs, and MD/PhDs who began faculty appointments between 1987 and 1996.23
Because PATH awardees are relatively early in their careers, we looked at publication patterns across the history of our assistant professor programs. Though we had CVs from 196 awardees including the 78 supported by the PATH program, only 173 CVs representing 116 men and 57 women included full publication lists. The remaining CVs contained abbreviated publication records and were not included for full analysis.
Throughout the history of these programs, women on average have published less than men. Because the records of those who publish unusually much or unusually little can distort understanding of how the group performs, we looked at the publication productivity of the middle 50% of awardees ranked by the number of their publications. The number of awardees at each milestone decreases because newer awardees have not yet been on faculty for 10 or 15 years. At five years after faculty appointment, the middle 50% of male awardees (58 awardees) had published 28.4% more papers than the middle group of 29 female awardees. Ten years after faculty appointment, the 40 men who represent the middle 50% had published only 18.1% more papers than 22 women in the middle 50%, but by 15 years, the middle group of 30 men published an average of 35.8% more papers than the 17 women in the middle group. It is tempting to speculate that family choices affect women’s early publication rate, but we have no data on awardees’ personal lives to enable comparisons. The middle 50% of clinically trained and nonclinically trained scientists (14 and 73 people, respectively) published about the same number of articles in their first five years, with the clinically trained publishing 3% or 0.4 papers more than PhD-only awardees. By 10 years, the clinically trained group (11 people) fell behind the PhD group (51 people), publishing 14.6% fewer papers, similar to their 14.0% fewer papers by 7 clinicians and 40 nonclinicians by year 15. Clinically trained grant awardees with and without PhDs published similar numbers of papers over time. Future work will analyze the impact of awardees’ publications.
There is a striking difference in the reported average training patterns of male and female awardees. Within the complete set of 196 CVs that included earlier programs and PATH, 142 awardees (96 men and 46 women) included information on training others. On average, per year, women trained more undergraduates (1.25 for women, 1.16 for men), more master’s degree students (women, 0.42; men, 0.29), and more PhD students (women, 1.00; men 0.72), whereas men trained more dual-degree (MD/PhD or DVM/PhD) students (men, 0.19; women, 0.16) and postdocs (men, 0.73; women, 0.62) (Figure 2A).
There was considerable variation among trainees so that for each of these numbers, the standard deviation was approximately the same as the average. Among women, the 5 individuals who made up the top 10% of PhD student trainers produced 2.0 to 3.1 students per year over a range of 8 to 16 years. The 10 people who made up the top 10% of male PhD student trainers produced 1.2 to 3.5 students per year over a range of 3 to 21 years. The most prolific female trainer of PhDs had produced 40 degree holders by 2010; the most prolific man, 46. They began their faculty careers in the same year. The top 10% of women postdoc trainers produced 1.0 to 1.3 postdocs per year, and the top 10% of men, 1.3 to 2.7. The most prolific male postdoc trainer had turned out 43 postdocs by 2010, and with 7% more time as an independent investigator, the most prolific woman had trained 21.
Thirty-six women and 59 men reported training undergraduates, but awardees from the same institutions and sometimes from the same schools within institutions differed in whether they included undergraduate laboratory trainees in their CVs. Of the 4 women who made up the top 10% reporting training undergraduates, 3 were at private research universities outside their colleges of medicine; the fourth was at a land grant university without a college of medicine. The 6 men in the top 10% included 2 at public medical schools, 2 at private research universities and outside their colleges of medicine, and 2 at a land grant university outside its college of medicine. They trained a range of 2.5 to 6.2 undergraduates per year. The top 10% of men who reported training undergraduates trained a range of 2.5 to 7.7 students per year. The most prolific male and female undergraduate trainers began their faculty positions in the same year. She had trained 86 undergraduates; he, 54.
Focusing on the middle 50% of awardees ranked in terms of the numbers of each category of students they train provides a different picture (Figure 2B). Women and men differed most in rates of training graduate students and postdocs. The 23 women who made up the middle 50% of graduate student trainers produced an average of 0.8 students a year (8 PhDs trained to completion each decade), 33% more than the middle group of 48 men, who produced 0.6 PhDs per year. Training postdoctoral fellows was reported by 139 awardees: 63 women and 76 men. The middle group of 38 men trained 0.7 postdocs per year—17% more postdocs than the 0.6 trained by the middle group of 32 women (Figure 2B).
Researchers published at different rates and structured the trainee makeup of their laboratories differently. Some but not all highly successful researchers published at high rates compared with their peers. Most PATH awardees published one to three articles per year (Figure 3). We do not have sufficient data to understand whether or how training and publication rates are related. However, looking at the top 20% of PATH awardees in terms of their publication rate allowed us to at least ask whether the most frequently published awardees train more people than their peers.
Looking at publication rate, rate of graduate student training, and rate of postdoctoral fellow training for the top 20% of men and top 20% of women provides some surprising observations (Figure 4). Those who produced the most postdocs did not generally publish the most. In fact, in this sample, men and women who trained the most graduate students outpublished those who train the most postdocs. The women who published the most had, on average, labs that trained more graduate students than postdocs, and the men who published the most on average had labs that trained more postdocs. We had anticipated that a small group of awardees would have published more papers, trained more postdocs, and trained more graduate students than most other awardees in the group. To our surprise, no awardee came to the top when ranked by all three measures.
The most productive 20% of female publishers (5 individuals) had published an average of 39.8 articles since faculty appointment (range 3.4–4.8 articles per year). They averaged 9.4 years on faculty (range 8–10). Their first R01s came an average of 3.2 years after faculty appointment (range 2–4). Two did not report on number of students trained. The 3 individuals who provided information on doctoral training have educated 34 students, producing an average of 1.2 PhDs per year or 12 new PhDs graduated per decade (range 0.63–2.00 per year). The 4 individuals who reported on training postdocs advised 19 postdocs between them, an average of 0.5 postdocs per year (range 0.1–1.0).
The most productive 20% of male publishers, 11 individuals, had published an average of 57.2 articles since faculty appointment (range 4.3–12.6 papers per year, average 6.3). They averaged 8.8 years on faculty (range 5–13) and received their first R01s an average of 2 years after faculty appointment, with a range from the year before beginning the faculty appointment to 8 years after. One works in Canada, where R01 funding is rare, and has not yet received an R01. Three did not report on training. Another had not trained graduate students. Forty-six students had been trained by the 8 individuals who provided information on doctoral training, at an average of 0.6 PhDs per year (range 0.2–1.6). Two who reported training graduate students did not report training any postdocs. A total of 43 postdocs were trained by the remaining 6, an average of 0.9 postdocs per year (range 0.3–1.7).
The most productive 20% of female PhD trainers, 4 women, had published an average of 22.3 articles since faculty appointment (range 1.8–4.8 papers per year). They had been on faculty an average of 7.5 years (range 4–9), getting first R01s 1.8 years, on average, after faculty appointment (range from a year before appointment to 4 years after). At an average of 1.6 PhDs per year (range 1.2–2.0), this group has trained 48 PhDs. Three of the women reported training 10 postdocs between them at an average of 0.4 postdocs per year (range 0.1–0.6). Only 1 of them was among the most active women publishers.
The most productive 20% of male PhD trainers, 8 men, published an average of 32.4 articles since faculty appointment (range 1.3–8.9 papers per year). They averaged 8.6 years on faculty (range 7–13). Their first R01s came an average of 3.1 years after faculty appointment (range from the year of appointment to 9 years after). Eighty-three students have been trained by this group, an average of 1.4 PhDs per year (range 1.1–2.0). They have trained 44 postdocs between them, averaging 0.8 postdocs per year (range 0.3–1.7). Two of them were among the most active male publishers.
The 4 researchers in the top 20% of female postdoc trainers published an average of 21.5 articles since faculty appointment (range 1.4–4.1 papers per year). They had been faculty members 8.3 years, on average (range 7–10), and gained R01s an average of 2 years after faculty appointment (range 1–4 years). They trained 24 PhD students at an average rate of 0.7 PhDs per year (range 0.4–0.9) and 29 postdocs, at an average of 0.9 postdocs per year (range 0.7–0.9). None were among the most active PhD trainers. Only 1 was among the most active women publishers.
Eight awardees formed the most productive 20% of male postdoc trainers. They published an average of 24.3 articles since faculty appointment (range 0.9–8.9 papers per year). Their average time on faculty was 8 years (range 3–11). Seven received R01s an average of 3.1 years after faculty appointment (range from the year of appointment to 9 years after). The eighth, in Canada, has not yet had an R01. Seventy-seven postdocs had been trained by this group at an average 1.2 postdocs per year (range 1.0–1.7). They trained 63 PhD students between them, or an average of 1.0 students per year (range 0.5–2.0). Three of the most active graduate trainers among men were also in this group. Only 1 of the 8 was also a top male publisher.
Support from early faculty grants goes to researchers who are on their way to career success and gives them resources that increase their advantages. However, these awards are rare, and many excellent researchers never get one. That early success at getting funding pays off is already known. Most awardees of the PATH program outpace national statistics for time to first job, first R01, and first promotion. This is likely true for other elite grant programs aimed at outstanding early investigators. Nearly all of our awardees do well and emerge as leaders in their fields.
What is perhaps most surprising in this analysis of early-career scientists is not how extraordinary most awardees are but how ordinary. At selection, BWF’s assistant-professor-level awardees generally have had a productive training period, moved into a faculty position, and achieved independence. Although some awardees are extraordinarily productive in terms of publications and training, the high outliers are likely different from other researchers in highly individual ways. Most of our awardees’ careers progress steadily at a pace that is attainable by focused, dedicated trainees. Our observations here remind us that there is no single right way to have a scientific career.
That most emerging leaders and rising stars are doing it at a moderate, achievable pace is useful news. We can enhance the success of trainees by helping them develop the resilience needed to achieve this kind of productivity. Trainees who will someday lead academic research laboratories as primarily NIH-funded principal investigators are building their chances of success from the time they begin working in the laboratory. In the course of a few years, a person goes from graduate school, where published articles can seem like the grand prize, to the tenure track, where articles are one’s bread and butter. Timely progress through the training period and into the faculty position is essential, as is productivity demonstrated by publishing. Perfectionism and frustration must be overcome, and high levels of resilience must be developed.
We may be scaring off students and postdocs who are truly well suited for the health science professoriate by not giving them a good sense of how attainable success is. Researchers who do well, including in today’s difficult climate, are mostly not wunderkinds. Many of the factors that make a successful scientist come from within, and a researcher who cannot or will not move forward productively will not do well. But it is the responsibility of those who train graduate students and postdocs to ready them for success, including by ensuring that they truly understand how to turn their data into manuscripts and their manuscripts into publications. Judgment of what to publish, when, and where; what is enough work to move forward to the next career step; and how to build a compelling grant comes from knowledge passed from person to person in the laboratory and at the desk.
Academic departments can improve trainees’ preparation by showing them how established scientists overcome setbacks at the bench, turned-down papers, unfunded grants, and the other disappointments, annoyances, defeats, and disgruntlements that are part of professional life. There are no morbidity and mortality conferences on academic life, but perhaps there should be. Faculty should be frank about the challenge of staying productive, especially when circumstances beyond the normal workflow cause slowdowns or complete disruptions of progress. Retrospective discussions can improve trainees’ understanding of how to move from postdoc to faculty start-up to independence, how and when new lines of research are launched, how small stories unexpectedly bloom into big ones, and how collaborations begin. Students, postdocs, and junior faculty would all benefit from hearing how childbirth, parenthood, being pressed unexpectedly into substantial service or teaching roles, marriage, divorce, natural disaster, health crises, and remarkable opportunities that sometimes arise have shaped the careers, outlooks, productivity, and the science of those around them.
Here, we have focused on numbers, but we have not examined the impact of BWF awardees’ publications, the success of their trainees, their standing within their fields, or whether they are viewed as leaders. Future work will explore whether, when, and how these productive researchers drive their fields forward, and will look at how they move into institutional, national, and international leadership and service roles. That analysis may provide insight into the differences between men’s and women’s publishing rates at midcareer. Another interesting question, but not one that we are well positioned to answer, is what goes wrong with faculty careers? Although some careers implode in fascinating but not illuminating ways, there may be much to learn from looking for patterns in the career development of researchers who struggle with launching independent careers and of those who are unable to reestablish productivity after interruptions.
Acknowledgments: The authors wish to thank Ruth Reynolds for her assistance and John Burris for critical reading of the manuscript.
1. Freeman R, Weinstein E, Marincola E, Rosenbaum J, Solomon F. Careers. Competition and careers in biosciences. Science. 2001;294:2293–2294
2. Pion G, Ionescu-Pioggia M. Bridging postdoctoral training and a faculty position: Initial outcomes of the Burroughs Wellcome Fund Career Awards in the Biomedical Siences. Acad Med. 2003;78:177–186
3. Carayol N, Matt M. Individual and collective determinants of academic scientists’ productivity. Info Econ Pol. 2006;18:55–72
4. Committee for the Evaluation of the Lucille P. Markey Charitable Trust Programs in Biomedical Sciences, Board on Higher Education and Workforce, National Research Council. Funding Biomedical Research Programs: Contributions of the Markey Trust. 2006 Washington, DC National Academies Press
5. Committee for the Evaluation of the Lucille P. Markey Charitable Trust Programs in Biomedical Sciences, Board on Higher Education and Workforce, National Research Council. Evaluation of the Markey Scholars Program. 2006 Washington, DC National Academies Press
6. Mavis B, Katz M. Evaluation of a program supporting scholarly productivity for new investigators. Acad Med. 2003;78:757–765
7. Pion GM, Cordray DS. The Burroughs Wellcome Career Award in the Biomedical Sciences: Challenges to and prospects for estimating the causal effects of career development programs. Eval Health Prof. 2008;31:335–369
8. Williard R, O’Neil E. Trends among biomedical investigators at top-tier research institutions: A study of the Pew Scholars. Acad Med. 1998;73:783–789
15. Jacob BA, Lefgren L. The impact of research grant funding on scientific productivity. J Public Econ. 2011;95:1168–1177
20. Committee on Bridges to Independence: Identifying Opportunities for and Challenges to Fostering the Independence of Young Investigators in the Life Sciences, National Research Council. Bridges to Independence: Fostering the Independence of New Investigators in Biomedical Research. 2005 Washington, DC National Academies Press
© 2013 by the Association of American Medical Colleges
23. Liu CQ, Alexander H. Promotion rates for first time assistant and associate professors appointed from 1967–1997. AAMC Analysis in Brief. May 2010;9(7)