Postdoctoral training offers doctoral recipients a temporary period of mentored or scholarly experience, considered highly productive within scientific and academic communities. Such training is also ostensibly valuable for postdocs, who gain additional experience to help pursue their chosen career paths. Tenure-track faculty positions, however, are now estimated to represent a small percentage of postdoc career outcomes (~15%) (Larson et al., 2014; National Academy of Sciences, 2014). This has led to proposals that support training postdocs for additional roles beyond tenure-track faculty positions. Additional efforts by the National Institutes of Health (NIH), National Science Foundation (NSF), and National Academies of Sciences, Engineering, and Medicine also aim to increase mentor accountability (National Academy of Sciences, 2014; Institute of Medicine, 2000; The National Academies of Sciences, 2018; National Institutes of Health, 2012; Meyers et al., 2016). Persistent concerns with increasingly long periods of postdoctoral training, lack of appropriate career guidance beyond the professoriate, and comparatively low postdoctoral salaries have also led to repeated calls to reform the postdoctoral training model (National Research Council, 1969; Davis, 2005; Sauermann and Roach, 2016; Alberts et al., 2014; Gould, 2015; Schaller et al., 2017).

Despite these concerns, comprehensive data for postdocs are not routinely collected (The National Academies of Sciences, 2018). Indeed, reliable data on such basic information as the number of postdocs have been lacking, or disputed, in part due to the difficulties that a lack of job title standardization, postdoc mobility, and the ad hoc nature of institutional postdoctoral administration present to data collection efforts (National Academy of Sciences, 2014; National Institutes of Health, 2012; Schaller et al., 2017; Daniels, 2015; Pickett et al., 2017).

Possibly for these very reasons, the postdoctoral experience has not been comprehensively surveyed nationally in over a decade, following the Sigma Xi ‘Doctors without orders’ survey report in 2005, which was based on postdoc respondents from 46 participating institutions (Davis, 2005). Nevertheless, recent data collection efforts have provided insights into the postdoctoral experience (National Institutes of Health, 2012; Sauermann and Roach, 2016; Pickett et al., 2017; Heggeness et al., 2016; Gibbs et al., 2015; Ferguson et al., 2014; Phou, 2017; Kahn and Ginther, 2017). For example, the pilot phase of the NSF Early Career Doctorates Survey studied the breadth of the doctoral population at US academic institutions, including postdoctoral researchers, early career faculty, and scientists in non-postdoctoral positions such as staff scientist or administrative positions (Phou, 2017). Most of their respondents were full-time faculty (54%); only 31% of respondents were postdocs.

Comprehensive survey data that specifically targets the postdoctoral period and includes postdoctoral researchers with PhDs granted both inside and outside the US, and data regarding postdoc career plans, satisfaction with mentorship, or family demographics, are still largely lacking (National Academy of Sciences, 2014; The National Academies of Sciences, 2018). To address these gaps, and to research those postdocs who do not have clear institutional oversight, we took a grass-roots approach to conduct a postdoc-led survey of US postdoctoral researchers. We asked postdoctoral researchers a number of questions associated with professional and career development, mentoring, career choice, lifestyle, and demographics (for details see Materials and Methods and Source Data 1-3). The purpose of this work was to capture a comprehensive snapshot of the postdoctoral experience in a manner that was both broad and informative, with high diversity in the questions and topics covered, in the number and type of institutions where the postdoctoral researchers are based, and in the breadth of postdoctoral experiences included.

To collect data from institutions with a wide range of support for postdocs, we took a multi-level approach to recruit survey participants. We used publicly available contact information for university leadership, postdoctoral administrators, postdoctoral societies and associations, and asked these individuals in leadership positions to disseminate our survey to all postdocs at their institutions. In total, we contacted individuals at the 482 US institutions most likely to have postdocs, including universities, research institutions, museums and government labs. We obtained respondents from 351 institutions. In addition to direct contact with institutions, we also used a grass-roots survey dissemination approach, promoted a website describing the survey that could be freely shared on social media and by email, and contacted professional societies to encourage survey dissemination. Using these combined approaches, we collected 7,673 individual responses into a secure REDCap database (IRB Protocol Number 15–1724), which, after quality control to remove respondents from non-US institutions, provided a final dataset of 7,603 respondents (see Materials and Methods for details).

As one of our goals was to reach as many postdocs as possible, survey dissemination was not randomized to any specific subset. Responses from institutions with long-standing postdoctoral affairs offices were anticipated to be over-represented in our dataset (see Materials and Methods for more information). Nevertheless, our respondents represented all 50 states, including a large fraction of respondents from institutions without well-established offices for postdoctoral support. While the majority of respondents represented STEM disciplines, which traditionally employ the most postdocs, 8.4% reported their primary fields as humanities, psychology or social sciences (Supplementary file 1 (Table S3)).

Our postdoc respondents were 49% US citizens and 51% non-US citizens (Figure 1 and Supplementary file 1 (Table S2)). The majority were 30–34 years of age (54.5%), and 1–3 years from receipt of their doctorate (63.1%), matching their reported years of postdoctoral experience (Figure 1 and Supplementary file 1 (Table S2)). The majority of postdoc respondents (69%) were from R1 academic institutions (Carnegie classification), with the remainder from non-R1 academic institutions (16%), medical centers (10%), and national government laboratories (4%). Less than 1% were either self- or otherwise employed. Of the 6,476 postdoc respondents in academia, 57% were based at public institutions while 43% were at private institutions. While the majority of postdocs were from R1 institutions, a large number of both R1 and non-R1 institutions were represented in our data set. Respondents were based at 151 different R1 institutions, 135 non-R1 academic institutions, 42 medical centers, and 59 national government laboratories. The Carnegie classification of non-R1 academic institution includes a wide range of academic institutions, from medical and PhD degree granting institutions that do not qualify as R1, to liberal arts and historically minority serving institutions. Due to the relatively small percentage of postdoc respondents from the large number of non-R1 institutions represented in our data set, we did not parse this group further, though future work should assess whether postdocs at liberal arts and minority serving institutes have unique experiences relative to postdocs based at other non-R1 academic institutions.

Figure 1

Download asset Open asset

A majority of respondents (55%) described their primary field of study as life sciences. There were small, but significant, differences in primary field by geographic region (Figure 2; Supplementary file 1 (Table S3)). Race and ethnicity were self-reported with 60.3% White/Caucasian, 24.8% Asian/Asian American, 6.6% Hispanic/Latino, and 2.6% Black/African American (Supplementary file 1 (Table S2)). Both national and international postdocs were included in these proportions. Our respondents were 53% female, while the gender ratio of their mentors was skewed towards males (71% male; Figure 1), consistent with the most recent AAUP Gender Equity report where full-time faculty are majority (61%) male (Association of American University Professors, 2006).

Figure 2

Download asset Open asset

While the demographics of our survey respondents may differ slightly from those of the actual postdoctoral population (but see Materials and Methods for analysis suggesting lack of response bias), confirmation of a lack of response bias remains difficult as there are currently no gold-standard datasets of postdocs in the US for comparison, due to the previously mentioned broader lack of oversight and barriers to reaching postdocs. That being said, the unique characteristics of our dataset, including approximately equal representation (similar sample sizes) of men and women, as well as US citizens and non-US citizens, facilitated our comparative analyses of the postdoctoral experiences of these different groups, which we report below.

Our data indicate that gender has a significant effect on the postdoc experience (Figure 1). Men were paid more than women (Male average: $47,678.00, Female average: $46,477.43, n = 7,516, χ² = 62.337, p < 0.0001. Note that our survey answer options were categorical instead of continuous, thus Pearson χ² tests and nominal logistic regression models were used instead of ANOVAs and generalized linear regression models throughout the manuscript). Men were more likely to have a same-gender mentor, i.e. a same-gender role model (Male: 77.3%, Female: 35.4%, n = 7,459, χ² = 144.352, p < 0.0001). Men were more likely to be non-US citizens (Male: 42% US citizens, 52% temporary visas, 6% permanent residents; Female: 56% US citizens, 38% temporary visas, 6% permanent residents; n = 7,543, χ² = 169.709, p < 0.0001). In addition, a small but significantly higher proportion of male postdocs were married/partnered (Male: 68.3%, Female: 63.2%, n = 7,538, χ² = 21.693, p < 0.0001) and/or have children (Male: 31.0%, Female: 22.3%, n = 7,532, χ² = 71.561, p < 0.0001).

The gender disparity in pay was present even after male and female postdocs were matched in age, years since graduation, self-identification as white/Caucasian or other, satisfaction with mentorship, Carnegie classification of institution, public or private control of institution, whether married/partnered, or having children (nominal logistic regression model, gender effect test n = 7,280, χ² = 51.330, p < 0.0001; Supplementary file 1 (Table S4)). In other words, female postdocs earned less than male postdocs, regardless of type of institution, marital status, parental status, or majority/minority status. This gender wage gap increased with postdoc age but not with partnership status, partially supporting previous analyses of the STEM gender wage gap (Association of American University Professors, 2006; Athanasiadou et al., 2018).

Primary field of study was excluded from this analysis because field has such a large effect on salary, overshadowing most other factors, with postdocs in Engineering, Environmental Sciences, the Physical Sciences, and the Social Sciences earning significantly more money than postdocs in the Life Sciences, Humanities, Medicine, or Psychology (n = 7,542, 750.452, p < 0.0001; Supplementary file 1 (Table S5)). Carnegie classification also had a large effect on salary (Supplementary file 1 (Table S4)), as 58% of the postdocs at national government laboratories report earning more than $55,000 a year, while only 8% of postdocs at R1 institutions report earning more than $55,000 a year. Male postdocs were more likely than women to be in the primary fields of Engineering (n = 620, χ² = 76.652, p < 0.0001) or Physical Sciences (n = 846, χ² = 77.466, p < 0.0001), two fields which have historically higher salaries (Buffington et al., 2016). Interestingly, female postdocs trended towards being paid less than men in all fields except the Physical Sciences, where women trended towards being paid slightly more than men (Supplementary file 1 (Table S5)). Income, mentor gender, citizenship, and partner status are all factors that may contribute to the observed gender difference in interest in primarily research-focused academic careers (Buffington et al., 2016) (Figures 1 and 3H).

Figure 3 with 1 supplement see all

Download asset Open asset

Most postdocs reported salaries in the range of $39,000–$55,000 (median $43,750, mean $46,988, n = 7,551). In the 2014 National Postdoctoral Association’s Institutional Policy Report, 52% of the 74 institutions reported that their minimum stipend matched the current NIH NRSA minimum (Ferguson et al., 2014). At the time of this survey, the NIH minimum was $43,692 (National Institutes of Health, 2018), which matches well with $43,750, the reported median income in our study, as well as the median postdoc income reported in recent work by McDowell et al. (McDowell et al., 2018). Five percent of postdocs reported mean gross incomes of less than $39,000 and ~10% reported incomes above $55,000. Although salaries in high cost of living (COL) urban areas tend to be higher than average (Supplementary file 1 (Table S6)), when adjusted to publicly available COL data, postdocs in large metropolitan areas earn significantly less money than postdocs in college towns or rural settings (average salary when adjusted for cost of living, metropolitan: $38,045.60; non-metropolitan: $44,714.40; n = 7,551, F-ratio:12.614, p = 0.0002; Figure 2A, Supplementary file 1 (Table S6)). ‘The Postdoctoral Experience Revisited’ 2014 report recommended as a best practice that the minimum salary be set at $50,000 (National Academy of Sciences, 2014); however, this has not been enacted at most institutions, and was only enacted by the NIH in November 2018. During the months that our survey was open (February–September 2016), the effect of a proposed minimum salary update ($47,476) to the Fair Labor Standards Act (FLSA) on postdoctoral salaries was openly debated, but ultimately not federally mandated (Benderly, 2016). Our data suggest that setting a minimum salary for postdocs is particularly important for postdoctoral researchers in large metropolitan areas, where salaries are not maintaining parity with cost of living increases.

The majority of postdocs selected research-focused academic careers as their primary long-term career plan (57.7%), with industry research a distant second (17.8%; Figure 1J; Supplementary file 1 (Table S2)). Determining the ‘why’ of career choice remains the subject of much study (National Academy of Sciences, 2014; The National Academies of Sciences, 2018; National Institutes of Health, 2012; Hayter and Parker, 2018; Roach and Sauermann, 2017). To assess which factors were most influential for determining postdoc career plan in our dataset (categorized in this survey as: academia, primarily research based; academia, primarily teaching based; industry; government/non-profit; other) we conducted a nominal logistic regression model with 26 factors concerning topics considered to be important for postdoc success and career choice (Supplementary file 1 (Table S7)), which include demographics, training, productivity and mentor support matrices. The 14 significant factors in the model were (in order of effect size): 1) whether postdoc career plans had changed; 2) whether the postdoc received training in pedagogy; 3) feelings of career preparedness; 4) perceived support of career plan from mentors; 5) primary field of study; 6) residency status in the US; 7) intensity of job search; 8) postdoc gender; 9) number of first, last, or corresponding author publications; 10) number of conferences attended in the past year; 11) hours worked per week; 12) total number of publications while a postdoc; 13) mentor rank; and 14) desire to pursue a career in the US (Table 1).

Perceived mentor support, number of postdoc publications, hours worked per week, conferences attended, and postdoc feelings of career preparedness were all positively correlated with a choice to pursue a research-focused academic career (Figure 3A,B,D–F; Figure 3—figure supplement 1C). Male postdocs, and postdocs who were not US citizens, were more interested in academic research positions (Figure 3G and H). In contrast, postdocs with mentors outside of the professoriate were more likely to prefer government/non-profit positions (Figure 3—figure supplement 1A). Whether this is a cause or effect relationship is not clear from our study, though we did find that postdocs with non-academic mentors changed their career plans at the same rate as those with academic mentors (n = 7,361, χ² = 6.860, p = 0.077). In addition, postdocs actively searching for permanent positions were less interested in academic research than postdocs not yet on the job market (Figure 3—figure supplement 1B), and were more likely to have changed their career plans (n = 7,565, χ² = 224.633, p < 0.0001). These results complement recent studies suggesting that individual career choice is influenced by changing job attribute preferences and self-awareness (Buffington et al., 2016), and that academic success is influenced by mentorship during the postdoctoral period (Lienard et al., 2018).

Sixty percent of respondents were either satisfied or very satisfied with the mentorship they receive, with similar responses from both genders (Figure 1I). To assess which factors were most influential for determining satisfaction with their postdoc mentor, we conducted a nominal logistic regression model with the same 26 factors included in the model for postdoc career choice (though excluding satisfaction with mentorship as a factor, and replacing it with postdoc long-term career plan; Supplementary file 1 (Table S7)). The eight significant factors in the model (in order of effect size) were: 1) feelings of career preparedness; 2) perceived support of career plan from mentors; 3) frequency of project meetings with mentor; 4) intensity of job search; 5) whether the postdoc received training in mentorship; 6) primary field of study; 7) perception of job market; and 8) academic rank of mentor (Table 2). These factors were more important than number of postdoc publications, whether a postdoc had changed career plans, postdoc or mentor gender, residency status, or postdoc training in either grant writing or pedagogy.

Perceived mentor support had a positive effect on how satisfied a postdoc is with their mentor, as did frequency of mentor meetings, perception of preparedness for desired future career, and perception of job market (Figure 4A–D). Postdocs who received training in mentorship were more satisfied with the mentorship they received than postdocs who did not receive training in mentorship (Figure 4E). We found this to be particularly noteworthy, as mentorship training is not a common part of the postdoctoral experience, with only 26% of postdocs reporting that they have received such training. While we cannot comment further on the specific type of mentor training that postdoc respondents received, we note that several institutions in our study have mentor training programs for postdocs in place, including those that use curricula from the University of Wisconsin-Madison Center for the Improvement of Mentored Experiences in Research (CIMER) project and/or National Research Mentoring Network (NRMN), for example the Big Ten Alliance NRMN-CAN program, as well as the ‘Mentoring in Research’ program at Stanford University.

Figure 4

Download asset Open asset

Previous research on a postdoc cohort showed that high satisfaction with mentorship and perceived support correlated with increased interest in an academic research focused career (Scaffidi and Berman, 2011). In addition, in a randomized, controlled study, a different type of mentoring, ‘group career coaching,’ where a career coach works with small groups of graduate students to provide support and promote career progress, was found to increase both perceived ‘achievability’ and ‘desirability’ of academic careers in an under-represented minority student group (Williams et al., 2016). Thirdly, in a longitudinal study of PhD students interested in academic careers, their perceived ability, or self-efficacy, was a strong indicator of retaining interest in a faculty career (Roach and Sauermann, 2017).

Other studies suggest that in addition to structured oversight and professional development (Davis, 2009), provision of research mentor training improves the success of researchers-in-training at all levels (Fleming et al., 2012). These studies strongly suggest that increasing a graduate student or postdoc’s self confidence and self-efficacy increases interest in an academic faculty career. Self-efficacy is directly impacted by the primary mentor and may explain the correlation between perceived support from their mentor and a postdoc’s interest in a faculty career seen here and elsewhere (Scaffidi and Berman, 2011). While our data do not show a significant correlation between gender and satisfaction with their mentor, they do suggest that an increase in mentor support and mentorship training will increase a postdoc’s interest in academic jobs. This increase in mentor support and mentorship may be a particularly important tool for increasing female and under-represented postdocs’ pursuit of research-intensive academic careers.

In summary, our dataset represents the most comprehensive survey of the US postdoctoral population in over a decade. As such, these data may provide a benchmark for legislation and institutional policy makers, inform research questions pertaining to the evolving postdoctoral population, and serve as a precedent for understanding the important dynamics of the scientific workforce.

We found that a research-focused academic position remains the most common primary career goal for postdocs, in spite of increasing emphasis on other types of careers for doctorate holders (The National Academies of Sciences, 2018; Alberts et al., 2014; St Clair et al., 2017). Although 60% of respondents were satisfied with the mentoring that they receive, our data suggest that providing formal mentorship training for postdocs may significantly increase their satisfaction with their mentor and influence career choice (Williams et al., 2016). Our data also show that women are less interested in research-focused academic positions than men, and this may be associated with gender specific differences in postdoctoral experiences (Moss-Racusin et al., 2012).

While the data we collected allowed us to identify a number of factors influencing the postdoctoral experience, other factors, such as socioeconomic background and underrepresented status, may also play a significant role, and should be studied further. Nevertheless, our findings highlight the impact of mentoring, across all demographics, as essential to informing career choice and determining quality of postdoctoral experience.

Non-privileged data used in this study are available in supplemental tables and additional material related to this manuscript. Due to their sensitive nature, much of the raw data are privileged to prevent individual identification, in accordance with IRB protocol. However, summary data for institutions, fields, and regions with more than 50 respondents are available upon request.

1. 1. Alberts B
   2. Kirschner MW
   3. Tilghman S
   4. Varmus H

   (2014) Rescuing US biomedical research from its systemic flaws

   PNAS 111:5773–5777.

   https://doi.org/10.1073/pnas.1404402111

   • PubMed
   • Google Scholar
2. Website

   1. Association of American University Professors

   (2006) AAUP Faculty Gender Equity Indicators

   Accessed May 2017.

   https://www.aaup.org/reports-publications/publications/see-all/aaup-faculty-gender-equity-indicators-2006
3. Preprint

   1. Athanasiadou R
   2. Bankston A
   3. Carlisle M
   4. Niziolek C
   5. McDowell G

   (2018) Assessing the Landscape of U. S. Postdoc Salaries

   bioRxiv.

   https://doi.org/10.1101/227694

   • Google Scholar
4. 1. Benderly BL

   (2016) A year-end bait and switch

   Science.

   https://doi.org/10.1126/science.caredit.a1700002

   • Google Scholar
5. 1. Buffington C
   2. Harris BC
   3. Jones C
   4. Weinberg BA

   (2016) STEM training and early career outcomes of female and male graduate students: evidence from UMETRICS data linked to the 2010 Census

   American Economic Review 106:333–338.

   https://doi.org/10.1257/aer.p20161124

   • PubMed
   • Google Scholar
6. 1. Daniels RJ

   (2015) A generation at risk: young investigators and the future of the biomedical workforce

   PNAS 112:313–318.

   https://doi.org/10.1073/pnas.1418761112

   • PubMed
   • Google Scholar
7. Website

   1. Davis G

   (2005) Doctors without orders

   American Scientist. Accessed May 2017.

   http://postdoc.sigmaxi.org/results/
8. Book

   1. Davis G

   (2009)

   Improving the postdoctoral experience: an empirical approach

   In: Freeman R, Goroff D, editors. Science and Engineering Careers in the United States: An Analysis of Markets and Employment. Chicago, IL: Chicago Press. pp. 99–127.

   • Google Scholar
9. Report

   1. Ferguson K
   2. Huang B
   3. Beckman L
   4. Sinche M

   (2014)

   National Postdoctoral Association Institutional Policy Report

   National Postdoctoral Association: https://cdn.ymaws.com/npamembers.site-ym.com/resource/resmgr/docs/npa_policyreport2014_final.pdf.

   • Google Scholar
10. 1. Fleming M
    2. Burnham EL
    3. Huskins WC

    (2012) Mentoring translational science investigators

    JAMA 308:1981–1982.

    https://doi.org/10.1001/jama.2012.14367

    • PubMed
    • Google Scholar
11. 1. Gibbs KD
    2. McGready J
    3. Griffin K

    (2015) Career development among american biomedical postdocs

    CBE—Life Sciences Education 14:ar44.

    https://doi.org/10.1187/cbe.15-03-0075

    • PubMed
    • Google Scholar
12. Website

    1. Gould J

    (2015) The postdoc series: The plight of the postdoc. Nature Jobs Blog

    Accessed March 15, 2015.

    http://blogs.nature.com/naturejobs/2015/03/16/the-postdoc-series-the-plight-of-the-postdoc/
13. 1. Hayter CS
    2. Parker MA

    (2018) Factors that influence the transition of university postdocs to non-academic scientific careers: An exploratory study

    Research Policy.

    https://doi.org/10.1016/j.respol.2018.09.009

    • Google Scholar
14. Website

    1. Heggeness ML
    2. Gunsalus KT
    3. Pacas J
    4. McDowell GS

    (2016) Preparing for the 21st Century Biomedical Research Job Market: Using Census Data to Inform Policy and Career Decision-Making

    SJS. Accessed December 21, 2016.

    http://www.sjscience.org/article?id=570
15. Book

    1. Institute of Medicine

    (2000)

    Enhancing the Postdoctoral Experience for Scientists and Engineers: A Guide for Postdoctoral Scholars, Advisers, Institutions, Funding Organizations, and Disciplinary Societies

    Washington, DC: The National Academies Press.

    • Google Scholar
16. 1. Kahn S
    2. Ginther DK

    (2017) The impact of postdoctoral training on early careers in biomedicine

    Nature Biotechnology 35:90–94.

    https://doi.org/10.1038/nbt.3766

    • PubMed
    • Google Scholar
17. 1. Larson RC
    2. Ghaffarzadegan N
    3. Xue Y

    (2014) Too many PhD graduates or too few academic job openings: the basic reproductive number R₀ in academia

    Systems Research and Behavioral Science 31:745–750.

    https://doi.org/10.1002/sres.2210

    • PubMed
    • Google Scholar
18. Preprint

    1. Lienard JF
    2. Achakulvisut T
    3. Acuna DE
    4. David SV

    (2018) Intellectual synthesis in mentorship determines success in academic careers

    bioRxiv.

    https://doi.org/10.1101/273888

    • Google Scholar
19. 1. McDowell GS
    2. Athanasiadou R
    3. Bankston A
    4. Carlisle M
    5. Niziolek CA

    (2018) Assessing the landscape of US postdoc salaries

    Studies in Graduate and Postdoctoral Education 9:213–242.

    https://doi.org/10.1108/SGPE-D-17-00048

    • Google Scholar
20. 1. Meyers FJ
    2. Mathur A
    3. Fuhrmann CN
    4. O'Brien TC
    5. Wefes I
    6. Labosky PA
    7. Duncan DS
    8. August A
    9. Feig A
    10. Gould KL
    11. Friedlander MJ
    12. Schaffer CB
    13. Van Wart A
    14. Chalkley R

    (2016) The origin and implementation of the Broadening Experiences in Scientific Training programs: an NIH common fund initiative

    The FASEB Journal 30:507–514.

    https://doi.org/10.1096/fj.15-276139

    • PubMed
    • Google Scholar
21. 1. Moss-Racusin CA
    2. Dovidio JF
    3. Brescoll VL
    4. Graham MJ
    5. Handelsman J

    (2012) Science faculty's subtle gender biases favor male students

    PNAS 109:16474–16479.

    https://doi.org/10.1073/pnas.1211286109

    • PubMed
    • Google Scholar
22. Book

    1. National Academy of Sciences

    (2014)

    The Postdoctoral Experience Revisited

    Washington, DC: The National Academies Press.

    • Google Scholar
23. Report

    1. National Institutes of Health

    (2012)

    Biomedical Research Workforce Working Group Report

    National Institutes of Health: https://acd.od.nih.gov/documents/reports/bmw_report.pdf.

    • Google Scholar
24. Website

    1. National Institutes of Health

    (2018) Ruth L. Kirschstein National Research Service Awards (NRSA) postdoctoral stipends, training related expenses, institutional allowance, and tuition/fees effective for fiscal year 2017

    Accessed May 2017.

    https://grants.nih.gov/grants/guide/notice-files/NOT-OD-17-003.html
25. Book

    1. National Research Council

    (1969)

    Invisible University: Postdoctoral Education in the United States. Report of a Study Conducted Under the Auspices of the National Research Council. [Richard B. Curtis, Study Director]

    Washington, DC: The National Academies Press.

    • Google Scholar
26. Website

    1. Phou K

    (2017) Profile of early career doctorates: 2015

    Accessed June 2018.

    https://www.nsf.gov/statistics/2017/nsf17313/
27. Preprint

    1. Pickett C
    2. Bankston A
    3. McDowell GS

    (2017) The GSS is an unreliable indicator of biological sciences postdocs population trends

    bioRxiv.

    https://doi.org/10.1101/171314

    • Google Scholar
28. 1. Roach M
    2. Sauermann H

    (2017) The declining interest in an academic career

    PLoS One 12:e0184130.

    https://doi.org/10.1371/journal.pone.0184130

    • PubMed
    • Google Scholar
29. 1. Sauermann H
    2. Roach M

    (2016) Why pursue the postdoc path?

    Science 352:663–664.

    https://doi.org/10.1126/science.aaf2061

    • Google Scholar
30. 1. Scaffidi AK
    2. Berman JE

    (2011) A positive postdoctoral experience is related to quality supervision and career mentoring, collaborations, networking and a nurturing research environment

    Higher Education 62:685–698.

    https://doi.org/10.1007/s10734-011-9407-1

    • Google Scholar
31. 1. Schaller MD
    2. McDowell G
    3. Porter A
    4. Shippen D
    5. Friedman KL
    6. Gentry MS
    7. Serio TR
    8. Sundquist WI

    (2017)

    Point of view: What’s in a name?

    eLife 6:e32437.

    • Google Scholar
32. Website

    1. Select Statistical Services Limited

    (2018) Population Proportion – Sample Size calculator

    Accessed November 2017.

    https://select-statistics.co.uk/calculators/sample-size-calculator-population-proportion/
33. 1. St Clair R
    2. Hutto T
    3. MacBeth C
    4. Newstetter W
    5. McCarty NA
    6. Melkers J

    (2017) The “new normal”: Adapting doctoral trainee career preparation for broad career paths in science

    PloS One 12:e0177035.

    https://doi.org/10.1371/journal.pone.0177035

    • PubMed
    • Google Scholar
34. Book

    1. The National Academies of Sciences

    (2018)

    The Next Generation of Biomedical and Behavioral Sciences Researchers: Breaking Through

    Washington, DC: The National Academies Press.

    • Google Scholar
35. Book

    1. Tintle NL
    2. Chance BL
    3. Cobb GW
    4. Rossman A
    5. Roy S
    6. Swanson TM
    7. VanderStoep JL

    (2016)

    Introduction to Statistical Investigations

    John Wiley & Sons, Inc.

    • Google Scholar
36. Website

    1. United States Census Bureau

    (2015) Educational Attainment of the Population 18 Years and Over, by Age, Sex, Race, and Hispanic Origin; Table 1-01

    Accessed November 22, 2016.

    https://www.census.gov/data/tables/2015/demo/education-attainment/p20-578.html
37. 1. Williams SN
    2. Thakore BK
    3. McGee R

    (2016) Coaching to augment mentoring to achieve faculty diversity: a randomized controlled trial

    Academic Medicine : Journal of the Association of American Medical Colleges 91:1128–1135.

    https://doi.org/10.1097/ACM.0000000000001026

    • PubMed
    • Google Scholar

Thank you for submitting your article "United States National Postdoc Survey Results and the Interaction of Gender, Career Choice, and Mentor Impact" for consideration by eLife. Your article has been reviewed by two peer reviewers, and the evaluation has been overseen by Emma Pewsey as the Reviewing Editor and Peter Rodgers as the Senior Editor. The following individual involved in the review of your submission has agreed to reveal their identity: Jennifer Miller (Reviewer #1).

The reviewers have discussed the reviews with one another and I have drafted this decision to help you prepare a revised submission.

Summary

This study is the first nationally comprehensive survey of postdocs in over a decade. For far too long, institutions, government labs and industry teams have not kept consistent or reliable data on postdocs or equivalent positions. The study addresses a significant void and is a first step towards better understanding the experience of postdocs and how different factors impact their job satisfaction and their future career decisions. The authors should also be applauded for tying their salary recommendations to cost of living. This will be important for political feasibility, and it is also consistent with the data.

Essential revisions

1. The paper would integrate more readily with other studies of the scientific workforce if data were analyzed using regression analysis. The main advantage of a regression analysis in this case would be building a model with causal rather than descriptive framing. If you choose not to perform a regression analysis, please explain in the text why you have chosen to use chi-squared tests instead.

2. Given the importance of postdoc mentor training to the article, a more robust discussion on it is warranted. What is considered to be 'training'? Please provide more information on the goals of postdoc mentor training, the nature of mentor training or the effectiveness of the training. Please also make it clear whether only the mentors of the postdocs undergo training, or whether, in some instances, the postdocs themselves also do so.

3. The correlation between perceived mentor support and the interest in a faculty career was highly interesting, and the article would benefit from a more in-depth discussion about why this was possibly the case.

4. This study did a very good job of recruiting participants from a wide range of institutions across the country, but it would help to know more about the nature of the scientific environments that the respondents are currently in – what type of institution, and what level of funding the institutions typically hold. This sort of information would help to paint a clearer picture of the respondents and would help to better contextualize the diversity of the pool.

5. There is significant discussion about the gender gap for postdoc pay equity. It would be helpful to provide more information about the individuals experiencing the pay inequity to understand the nature of the disparity.

1. The paper would integrate more readily with other studies of the scientific workforce if data were analyzed using regression analysis. The main advantage of a regression analysis in this case would be building a model with causal rather than descriptive framing. If you choose not to perform a regression analysis, please explain in the text why you have chosen to use chi-squared tests instead.

We agree with the reviewers that regression analyses are most appropriate for these data. That is why we used regression models throughout the study. Specifically, we used nominal logistic regression models, to take into account the categorical nature of the responses and factors. We now clarify this in the text. The test statistics for nominal logistic regression models are χ², which we report in addition to p-values for whole models and effect tests. We now clarify this in both our Methods and Results sections.

2. Given the importance of postdoc mentor training to the article, a more robust discussion on it is warranted. What is considered to be 'training'? Please provide more information on the goals of postdoc mentor training, the nature of mentor training or the effectiveness of the training. Please also make it clear whether only the mentors of the postdocs undergo training, or whether, in some instances, the postdocs themselves also do so.

To clarify, we only refer to postdocs undergoing training in mentorship themselves, not their faculty mentors. We did not ask if the postdoc’s mentor had received training.

We allowed respondents to self-define their answer to the question, “Have you received mentor training?”. Thus we are unable to answer questions on the content, effectiveness, or goals of mentor training from the survey data. Of the postdoc respondents who selected “Yes”; 36% indicated in a later matrix question that the “majority of your training in [mentoring] comes from” their primary faculty mentor / PI and did not select other options. The other options (in order of most often selected) were self, other faculty, other combination, collaborators, and lab members in that order. The majority of respondents chose some combination of offered selections.

That being said, there are mentor training programs for postdocs at several institutions included in this study. The NRMN-CAN Big Ten Alliance professional development program, the UW-Madison Wisconsin CIMER Project and NRMN curricula, and “Mentoring in Research” at Stanford University are examples of such programs.

We have now revised the text (Results and Discussion section) to read:

“Postdocs who received training in mentorship were more satisfied with the mentorship they themselves received than postdocs who did not receive training in mentorship (Figure 4E). We found this to be particularly noteworthy, as mentorship training is not a common part of the postdoctoral experience, with only 26% of postdocs reporting that they have received such training. While we cannot comment further on the specific type of mentor training that postdoc respondents received, we note that several institutions included in our study have mentor training programs for postdocs in place, including those that use curricula from the UW-Madison CIMER project and/or National Research Mentoring Network (NRMN), e.g. Big Ten Alliance NRMN-CAN program, as well as “Mentoring in Research” at Stanford University.”

3. The correlation between perceived mentor support and the interest in a faculty career was highly interesting, and the article would benefit from a more in-depth discussion about why this was possibly the case.

We have now expanded our discussion of the potential causes of the correlation between perceived mentor support and interest in a faculty career, and include the following sentences (Results and Discussion section):

“Thirdly, in a longitudinal study of PhD student interest in academic careers, their perceived ability, or self-efficacy was a strong indicator of retaining interest in a faculty career (Ref 25). Taking these studies as a whole, high self-efficacy reinforces interest in an academic faculty career. In turn, self-efficacy is directly impacted by the primary mentor and may explain the correlation between perceived support from their mentor and a postdoc’s interest in a faculty career seen here and in Ref 27.”

4. This study did a very good job of recruiting participants from a wide range of institutions across the country, but it would help to know more about the nature of the scientific environments that the respondents are currently in – what type of institution, and what level of funding the institutions typically hold. This sort of information would help to paint a clearer picture of the respondents and would help to better contextualize the diversity of the pool.

We now include information on how many of our respondents were from R1 academic institutions, non-R1 academic institutions, medical centers, government labs, and other, as well as the proportion of postdocs in academia who were located at private versus public institutions.

5. There is significant discussion about the gender gap for postdoc pay equity. It would be helpful to provide more information about the individuals experiencing the pay inequity to understand the nature of the disparity.

We now expand on gender disparity in salary, and clarify that this salary disparity was independent of where postdocs worked, marital status, parental status, age, or race/ethnicity. As far as we can tell, nearly all groups of female postdocs, with the exception of those in the physical sciences, earn less than their male counterparts.

Author details

1. Sean C McConnell

   Sean C McConnell is in the Department of Science, Medicine and Public Health, American Medical Association, Chicago, United States

   Contribution

   Conceptualization, Formal analysis, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, Designed the survey and analyzed data, Disseminated the survey

   Contributed equally with

   Erica L Westerman

   Competing interests

   No competing interests declared

2. Erica L Westerman

   Erica L Westerman is in the Department of Biological Sciences, University of Arkansas, Fayetteville, United States

   Contribution

   Conceptualization, Formal analysis, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, Designed the survey and analyzed data, Performed multivariate analysis and models, Disseminated the survey

   Contributed equally with

   Sean C McConnell

   For correspondence

   ewesterm@uark.edu

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-3575-8298

3. Joseph F Pierre

   Joseph F Pierre is in the Department of Pediatrics and Department of Microbiology, Immunity, and Biochemistry, University of Tennessee Health Science Center, Memphis, United States

   Contribution

   Conceptualization, Formal analysis, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, Designed the survey and analyzed data, Disseminated the survey

   Competing interests

   No competing interests declared

4. Erin J Heckler

   Erin J Heckler is in the Office of Postdoc Affairs in the Office of the Vice Chancellor for Research, Washington University in St. Louis, St. Louis, United States

   Contribution

   Conceptualization, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing, Designed the survey and analyzed data, Disseminated the survey

   Competing interests

   No competing interests declared

5. Nancy B Schwartz

   Nancy B Schwartz is in the Department of Pediatrics and Department of Biochemistry and Molecular Biology, University of Chicago, Chicago, United States

   Contribution

   Writing—original draft, Project administration, Writing—review and editing

   Competing interests

   No competing interests declared

• 19,362

  views

• 1,039

  downloads

• 54

  citations

Views, downloads and citations are aggregated across all versions of this paper published by eLife.