Academic careers have expanded across the university sector at all career levels in recent decades, but there remain relatively few levels in the hierarchy – PhD student, postdoctoral researcher, and independent group leader/principal investigator (PI)/university lecturer. In general the system trains more PhDs and postdocs than can be employed as independent group leaders, PIs or university lecturers, so the competition for these positions can be intense. The present authors know this first-hand because we all started in such positions at universities in the UK between 2015 and 2018. Moreover, we know that new group leaders, PIs and university lecturers face a wide range of challenges as they seek to establish their research groups and undertake new responsibilities in their department or institute.

In this article we will use the term PI as a short-hand for independent group leader, principal investigator or university lecturer. There are several routes to becoming a PI in the UK. The two most frequent are: i) recruitment as a permanent lecturer (subject to passing probation) at a university; ii) recruitment as a fixed-term research fellow at a university or research institute (funded directly by a university, or externally by research councils and charities). There are pros and cons associated with both routes: lecturers typically have long-term job security, but externally-funded research fellows can often establish their research programme faster than lecturers.

To explore how new PIs are recruited, supported and evaluated, we conducted a survey of PIs appointedin the UK between 2012 and 2018. This article reports the results of this survey, discusses what it tells about the hopes and concerns of new PIs in the UK, makes recommendations to employers and funders to address some of these concerns, and concludes with advice for those hoping to become a new PI.

Our survey is described in detail in the Methods section. In summary, the survey was completed by 365 individuals who had become PIs over the past six years: 83% worked in the life sciences (Figure 1A), 57% were male (Figure 1B), 51% were from the UK (Figure 1C), and 84% classified themselves as white (Figure 1D).

Figure 1 with 2 supplements see all

Download asset Open asset

Figure 1—source data 1

Summary data for Figure 1.

https://doi.org/10.7554/eLife.46827.007

Download elife-46827-fig1-data1-v1.csv

The average respondent had spent seven years between finishing their PhD and starting their own group (Figure 1E), which suggests they had held two or three fixed-term positions before gaining independence. This seven-year period likely reflects the eligibility restrictions that were in place for some fellowships until recently, and it will be interesting to see if the recent decision by many funders to remove such restrictions has an influence on the time taken for PhDs to become PIs.

A longer time spent as a postdoctoral researcher may allow some individuals the extra time needed to complete and publish ambitious or collaborative projects. However, it might also increase the average age at which researchers become independent (which is 34 years old for the respondents to our survey; Figure 1—figure supplement 1), and this would increase the length of time during which postdocs have to balance starting a family with the pressures imposed by fixed-term contracts and the need to show mobility.

With regard to families, half of our cohort had dependents, female investigators had taken the longest career breaks and were the most likely to have dependants, and almost everyone (361/365) worked full time (Figure 1—figure supplement 2).

International mobility plays a key role in the academic career path. 51% of respondents had spent more than one year training outside of the UK as postdocs, and 67% had undertaken at least one international move as part of their career (either when moving between their PhD and postdoc, or when moving from a postdoc to a PI position in the UK; Figure 2). And when moves within the UK are included, the percentage of respondents who moved to take up their position increases to 78%: moreover, 77% had not previously worked in their current department.

Figure 2

Download asset Open asset

Figure 2—source data 1

Summary data for Figure 2.

https://doi.org/10.7554/eLife.46827.009

Download elife-46827-fig2-data1-v1.csv

Job satisfaction and well-being

While details about our cohort reveal information about the sector as a whole, it is when we ask about job satisfaction that we begin to see where problems may lie for new PIs. In general the number of respondents who replied that they were satisfied or very satisfied clearly exceeded the number who were dissatisfied or very dissatisfied about the six different aspects they were asked about: their department; their institution; space and facilities; their funder; work-life balance; and optimism (Figure 3). However, 24% (79/330) were either dissatisfied or very dissatisfied with their space and facilities, compared with 56% (185/330) who were satisfied or very satisfied. We think there is scope for funders to put pressure on institutions to address concerns about space and facilities.

Figure 3 with 1 supplement see all

Download asset Open asset

Figure 3—source data 1

Summary data for Figure 3.

https://doi.org/10.7554/eLife.46827.013

Download elife-46827-fig3-data1-v1.csv

Moreover, 33% (111/331) were either dissatisfied or very dissatisfied with their work-life balance, compared with 41% (135/331) who were satisfied or very satisfied. Flexible or part-time working is often put forward as a way to improve work-life balance but, as noted above, just four of the respondents worked part-time. Given all the challenges associated with being a new PI – find funding, build a research group, publish work, prepare and give lectures – it is not surprising that practically no one works part-time.

The pressures of being a new PI are best expressed in this direct quotation from one respondent: “I feel like I'm trying to do three separate jobs (research, management/admin, teaching) as well as be a mother... be my own postdoc (because I can't afford one), be the lab technician (because I can't afford one), be the lab manager (because I can't afford one...), be a good mentor for my students, plan strategy, write grants (constantly, I need the money), stay up to date with other research, prepare new teaching material (this takes me ages, I want to do a good job), teach, mark assessments and answer student queries etc. I could go on. No, seriously, is it even possible?”.

Clearly there is a need to find other mechanisms to improve the work-life balance besides offering part-time/flexible working patterns. However, despite elements of dissatisfaction, it is important to highlight over 50% of respondents were optimistic for the future (Figure 3). This highlights a strong resilience and positivity amongst new PIs as they tackle the various demands of their role. It was also encouraging to find that having dependants did not affect satisfaction ratings or optimism scores for new male or female PIs (Figure 3—figure supplement 1).

Career track comparison and gender disparity

As previously mentioned, there is no single route to becoming a PI in the UK. Typically, two career paths exist: (a) appointment as a lecturer, possibly followed by promotion to senior lecturer, then reader and lastly professor; (b) appointment as a research fellow, possibly followed by an advanced fellowship or a transfer to the lecturer career path. 41% of our respondents were on the lecturer career path, with 59% on the research fellow path (Figure 4). The research fellows secured funding from a range of bodies and 70% of respondents were from 24 research-focused institutions, referred to as the Russell group (Figure 4A). This puts a large amount of resources into very few institutions.

Figure 4

Download asset Open asset

Figure 4—source data 1

Summary data for Figure 4.

https://doi.org/10.7554/eLife.46827.015

Download elife-46827-fig4-data1-v1.csv

38% of our new PIs were required to successfully apply for major grants or fellowships in order to take up their position (Figure 4B), with research fellows bringing in the highest levels of funding (some 25% of research fellows secured more than £500,000 before starting their position; Figure 4C).

We also start to see a gender disparity emerge in grant funding early in the careers of new PIs: the majority of respondents (80% of men, 77% of women) had secured some additional funding within the first five years, but male respondents had secured significantly more than female respondents (Figure 5A–C). In particular, male PIs are much more likely to have secured additional funding in excess of £1 m (p=0.025), and female PIs were awarded significantly fewer grants (p=0.039). Compared to new female PIs, it looks as if new male PIs were better able to gain momentum and accelerate through continued grant success, allowing them to build critical mass expanding the numbers in their labs. We should delve deeper into this issue and ensure that new female PIs are being encouraged and supported to apply for more funding and to build their teams in the same way as new male PIs.

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Summary data for Figure 5.

https://doi.org/10.7554/eLife.46827.019

Download elife-46827-fig5-data1-v1.csv

Female PIs were also paid £3000–5000 less per year than male PIs (Figure 5D), both as lecturers and research fellows (with lecturers generally starting on lower salaries than research fellows; Figure 5—figure supplement 1A): moreover, male PIs tended to be appointed at a higher grade (eg, at grade 8 rather than grade 7 for lectureships; Figure 5—figure supplement 1B). The appointment of women to lower grades will impact their rate of career progression compared to men.

Teaching and administration

It came as a surprise to us that a majority of research fellows (57%) were expected to teach, despite their time being protected for research and a majority having their salary paid by an external funder rather than the university (Figure 6A). Although the number of contact hours was significantly less than the number for lecturers, nearly 40% of research fellows were expected to have more than 10 contact hours with undergraduates per year, and 10–15%, were expected to have more than 40 contact hours per year (Figure 6B). There is an argument that research fellows should engage with their departments, bring new material to undergraduate courses, and participate in some level of teaching early in their independent careers: moreover, if the new PI hopes to be appointed as a lecturer in the longer term, gaining teaching experience will be beneficial. However, having to give more than 40 hr of lectures and tutorials is excessive for a new research fellow, so we suggest that all funders should consider specifying a limit on teaching hours. Some funders already apply limits, but these data suggest there is a lack of enforcement by either the fellow or the institution.

Figure 6

Download asset Open asset

Figure 6—source data 1

Summary data for Figure 6.

https://doi.org/10.7554/eLife.46827.021

Download elife-46827-fig6-data1-v1.csv

In general, we would suggest that the best way for new PIs to engage with undergraduate teaching would be to focus on the supervision of undergraduate laboratory projects or literature projects, so that the teaching is directly contributing to their research programme. After 2–3 years of having their time 'protected' in order to establish their research programme, direct teaching commitments through lectures and examinations can begin.

The fact that new female PIs have higher teaching loads (Figure 6B) and higher administration loads (Figure 6C) than new male PIs may contribute to the reduced grant success (and to the smaller lab sizes discussed below). These statistics suggest female PIs are over-committing themselves and/or are more frequently tasked with non-research roles than male PIs. Efforts by universities to ensure equal numbers of men and women sit on committees could contribute to the higher administration loads on female PIs (both new and established) because there are fewer female PIs overall.

Mentorship and career development

Support for new PIs is especially important when they take on roles they have no prior experience of. We found that almost 25% of all new PIs felt that they had no mentorship: moreover, the female PIs who did not have a mentor were the least optimistic for their career progression (Figure 7). All but 3% of lecturers reported having an annual review, but almost 18% of externally funded fellows did not have an annual review. However, these fellows benefited from forms of support not available to most lecturers (for example, funders organize meetings that bring together all the fellows they fund, and these meetings can be a valuable source of peer support and career advice).

Figure 7

Download asset Open asset

Figure 7—source data 1

Summary data for Figure 7.

https://doi.org/10.7554/eLife.46827.023

Download elife-46827-fig7-data1-v1.csv

Most lecturers (88%) were provided with start-up funds, which were typically in the region of £20–60K, by their university (Figure 8). Once again there was a gender gap in favour of male PIs, who received an average of £14.6K more than female PIs.

Figure 8 with 1 supplement see all

Download asset Open asset

Figure 8—source data 1

Summary data for Figure 8.

https://doi.org/10.7554/eLife.46827.027

Download elife-46827-fig8-data1-v1.csv

A major concern for externally funded research fellows was what to do at the end of their contract. While one might assume that such a fellow would be appointed as a lecturer (or higher on the academic career ladder) at their host university when their fellowship ended, 70% of research fellows did not have an agreement for such a 'proleptic appointment' in place with their host university (Figure 8—figure supplement 1). Moreover, 58% were unaware of what they had to do to get a permanent position or how their host institution or department went about making such decisions. Since these fellows are hired by universities on contracts which are dependent on the funding source, they can be made redundant at the end of the fellowship with little consequence. These fellows do not have job security despite supporting their own salary, bringing in large grants, passing stringent external selection processes and, on the whole, being more likely to contribute to the research excellence framework (REF: this is the process through which university research is assessed in order to determine the future level of government funding). We would encourage funders to address this issue in order to protect their investment in these researchers.

Some of our respondents were extremely critical about these matters. “Career progression is very non-transparent” said one, “Vague descriptions of the areas in which excellence is required, but no idea of the level equivalent to excellence. Getting a proleptic appointment is very difficult.” A second commented as follows: “It is widely believed that if you have funded your own salary from grants for seven years then the school should take you on as a full-time lecturer. However this does not appear to be written down anywhere and may have been inconsistently applied.”

Building a research group

Although 80% of respondents had at least one PhD student (Figure 9A), both lecturers and research fellows reported difficulty in recruiting PhD students. One reason was that many PhD studentships in the UK have been gathered into large doctoral training centres, and senior researchers have been more successful in having their projects accepted by these centres: moreover, students seem to have a preference for joining established labs.

Figure 9

Download asset Open asset

Figure 9—source data 1

Summary data for Figure 9.

https://doi.org/10.7554/eLife.46827.029

Download elife-46827-fig9-data1-v1.csv

Some 80% of research fellows also had at least one postdoc in their lab, and some had three or more: however, 50% of lecturers did not have a postdoc (Figure 9B). Research fellows were also more likely to have a research assistant than lecturers (53% vs 25%; Figure 9C). However, lecturers tended to have bigger groups than research fellows (Figure 9D), which suggests high numbers of undergraduate students in the lab: in the absence of postdocs or research assistants to help train and supervise these students, this will only add to the pressure on the new PI, and may not provide the best training for the students either.

Again, the best way to convey the frustrations experienced by some respondents is to quote them directly: “To be successful as a fellow it is primordial to get a PhD student during the first year of contract,” said one. "Without hands in the lab we cannot work. This is not granted, I struggled to get my lab members. Actually I secured an external studentship, but incredibly and annoyingly my Institution does not allow me to be primary supervisor”. A second commented as follows: “I was told in no uncertain terms that the department could offer me nothing as a start-up. I am part of 2 possible PhD schemes in the university but funding only has been awarded to senior colleagues.”

Gender bias in recruitment

As reported in Figure 1B, 56.7% of respondents were men, 41.6% were women and 1.6% preferred not to say. However, when we plot the percentage of male and female respondents against year of appointment, we find a spike in the number of male researchers appointed as new PIs in 2013 (Figure 10). We feel it is likely that this spike was caused by a wave of recruitment ahead of REF2014. The next REF will be in 2021. Each REF tends to be preceded by a 'transfer window' in which new academic staff are appointed and established staff sometimes move between universities. However, it is extremely clear from our data that the wave of new appointments before REF2014 significantly favoured male applicants. It seems likely to us that this wave of male recruitment may have been due to an increase in direct head-hunting, or more informal recruitment techniques driven by networks within fields, and that these practices seem to favour men. We urge institutions to ensure that female academics are not disadvantaged in this way ahead of REF2021.

Figure 10

Download asset Open asset

Figure 10—source data 1

Summary data for Figure 10.

https://doi.org/10.7554/eLife.46827.031

Download elife-46827-fig10-data1-v1.csv

While the career trajectories of academics are diverse, our survey suggest that there are a number of overarching issues that affect most new PIs. We list below actions that could be taken by host institutions and funders to address these overarching issues and ensure that all new PIs have the opportunity to reach their full potential.

The survey was conducted using convenience sampling, with most participants finding out about it through Twitter or forwarded email invitations. The majority of those responding were in the life sciences, in part because of the networks that the survey was circulated around, and partly due to the language used in the survey (the term PI does not have the same meaning in the social sciences, for instance). We do not claim that this is a fully representative sample. However, the sample has a similar gender and nationality breakdown to those reported for the sciences elsewhere (e.g., by the Higher Education Statistics Agency in the UK), and we do feel that it allows us to say important things about at least a significant subgroup of new PIs in the UK.

A pilot version of the survey was originally run with 10–12 volunteers recruited via the UK_NewPI slack group. On the basis of this a number of questions were revised, including changing some definitions, revising categories, and changing the salary question to refer to monetary values rather than pay bands. The revised survey with full questions can be found in Supplementary file 2. The survey was distributed, through the authors networks, via the NewPI slack (both UK and US), the eLife ambassadors mailing list, LinkedIn, and Twitter. In the end, of the 365 respondents, 311 were recruited via Twitter, 11 via email links, 16 via weblinks, and 1 from LinkedIn.

Much of the analysis consists of simple descriptive statistics – that is, looking at the distribution of individual variables. However, where we were interested in the relationship between variables, we used a mix of ordinal logit regressions and chi-squared tests depending on the nature of the relationship being studied (see Supplementary file 1 for details). Ordinal regression allowed us to control for multiple factors, to be more sure that the relationship that we found was not a result of (at least measured) confounding factors, for ordinal outcome variables. Full details of these models (both model estimates and specific variables used in different models) can be found in Supplementary file 1. The models used were as follows:

1. Ordinal logit regressions, with starting salary (8 bands) as the outcome variable, and gender as the main predictor variable of interest. Control variables include ethnicity (white/non-white), nationality (UK born/not), age, year of appointment, years between PhD and appointment, number of children, and type of appointment (lecturer/research fellow). These models allow us to see what individual factors are most associated with higher and lower salaries on appointment.

2. Ordinal logit regressions, with ‘optimism for the future’ as the outcome variable (5-point likert scale). Various combinations of covariates were used including gender, ethnicity, nationality, whether respondent does teaching, starting salary (8 bands) number of grants received, value of grants received, whether respondent is in a Russell group university, whether the respondent has an RA/Tech (3 categories: Own RA, shared RA, no RA), number of PhD students, number of postdocs, years between PhD and PI appointment, year of PI appointment, and type of job (lecturer/fellow). These models allowed us to see what individual factors are most associated with feeling optimistic/pessimistic about their future careers.

3. Ordinal logit regressions, with ‘optimism for the future’ as the outcome variable (5-point likert scale). Predictor variables were satisfaction with the lab, the department and the university/institute. The purpose of these models was to show which of those three influenced the respondents’ optimism the most.

4. A Chi squared test of REF year (2013 against all other years) and gender. This revealed the increase in male (but not female) recruitment associated with the REF.

5. Chi Squared test of type of job (lecturer, senior lecturer, research fellow and other) against amount of teaching (6 categories) vi) An ordinal logit model where the outcome is grant income (5 categories) and the outcome of interest is gender, with control variables of ethnicity (white/non-white), age, year started as a PI, years of postdoctoral experience, number of children, and the year of first appointment as a PI. We also ran the same model as a standard logit model whether the outcome was dichotomised (over one million/less than one million).

Summary data analysed in this study are included in the manuscript and supporting files.

Thank you for submitting your article "The life of P.I. - Transitions to Independence in Academia" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by myself (Emma Pewsey) and the eLife Features Editor, Peter Rodgers. The following individuals involved in review of your submission have agreed to reveal their identity: Amanda N Sferruzzi-Perri (Reviewer #1).

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission. While the list of points to address may seem long and will require major revision of the article, we feel that they will greatly benefit the clarity of the article, and its ability to influence decision-makers. We hope you will be able to submit the revised version within two months.

SUMMARY OF MANUSCRIPT

This is an important and timely topic that should get major spotlight and we congratulate the authors for deciding to tackle it. The paper does a good job of identifying the major struggles of new PIs, and makes some suggestions for how PIs, host institutions and funding agencies can help to prevent these struggles. The opportunity that the findings present to influence senior academics and other decision makers therefore make it crucial that this manuscript is of excellent quality and that it leads to clear conclusions about why and how the situation should change.

LIST OF REQUIRED REVISIONS

GENERAL

1. There are a number of incorrect citations (for example, the in-text citations for the panels of Figure 10 are labelled as Figure 9, and a citation to 14F appears to be for 14G). Also, the 4th sentence in the caption to figure 13 should start: "Fellows (D) and Lecturers (E) were then.. . ". Please check and correct the text and figure captions for issues like these. Please also ensure that all in-text discussion of data that's presented in a figure cites that figure.

2. Please present the findings in a more neutral manner. For example, instead of making a broad statement such as "too many new investigators express frustration", please report on the percentages who are frustrated.

PRESENTATION OF FIGURES

3. The article will be easier to read if the number of figures is reduced. eLife allows the figures in a paper to have figure supplements; the figures appear in both the online and pdf versions; the figure supplements only appear in the online version, but the figure caption makes it clear that the figure supplement are available online. Please combine figures and convert some panels into figure supplements as suggested below.

4. Some panels are not discussed in the text (2C, 6A, 6B, 6D, 6E, 7A, 7B, 9C, 10A, 13B, 13C, 14D, 14E, 14F, 14I, 15A); please either i) delete these or ii) convert them to figure supplements (and make sure they are cited somewhere in the text or the relevant figure caption).

5. The results of yes/no questions (5A, 5B, 5C, 6C, 10B, 10C, 10D, 10E, 12A, 12B, 13A, 14A, 14B) would be better presented by stating the percentage of respondents who answered yes in the text and/or figure caption, and citing the appropriate section of your supporting data file.

6. Please include y-axes and keys for figures where these are missing (including Figures 3, 4, 11).

7. Figure panels 14G and H show the same data, so please convert one to a figure supplement. Also, for 14G, please adjust the y-axis scaling to make it possible to see the data that supports the statement in the text about start-up funds "in the region of £20-60K".

8. Please ensure that the order of the figures matches the order in which they are cited/mentioned in the text.

RESULTS

9. Please present the results of the statistical tests for all the data shown; for example, this is missing from comparisons between female and male PIs and/or lecturers and researcher fellows. Please state the types of statistical tests run for each piece of presented data, with details on how certain confounders were controlled for, as well as detailing any corresponding statistical outcomes.

10. The first sentence of the results states that '...365 respondents represents a significant proportion of new group leaders in the UK.' To support this comment, please state the number of new group leaders in the UK for the period corresponding to the results presented.

11. Please rephrase the final sentence for the second paragraph of the results ('There has been some discussion... in redressing the balance') to make its meaning clearer.

12. Please clarify what Figure 2B represents by describing it in the figure legend; you may also wish to add further clarification in the main text. Should the values for women and men add to 100%? Does this graph show only those that had dependents/caring responsibilities? Does '0' for length of break mean the individual had a dependent but took 0 months off from work? What is the mean and median for the data for each gender? How does this relate to the number of dependents per respondent? Furthermore, the panel shows possibly 1 or 2 men taking very long periods of leave, which does not appear to support the statement in the text that '... fathers are also taking significant periods of leave and sharing childcare responsibilities.' Please clarify.

13. Please provide the data supporting the following sentence: 'It was also encouraging to find that having dependents did not affect satisfaction ratings or optimism scores for either females or males'; the data presented is not separated by sex.

14. Instead of stating that the proportion of lecturers to research fellows captured in the study was roughly equal, please state the exact percentages.

15. The comments about research fellows being based at Russell group universities (bottom of page 3) seems to reflect that these universities first recruit individuals as research fellows (rather than as lecturers), as shown in Figure 5D. Please discuss this in the text.

16. Could the data presented in Fig 8 and 15 also relate to when career breaks may occur for a women (which we know are for longer periods)? Was information collected about the timing of leave with regard to the trajectory of the respondent's career paths? In regards to 8C, could this be related to men asking for more money for each grant compared to women? Could a calculation be made to determine whether the amount awarded per grant varied between women and men?

17. Based on the data presented in 9B and C, women have the larger loads for teaching and committees. Please discuss this, as these data may be related to the finding that women also secure fewer grants and have small lab sizes - as they are simply over-committing themselves to, or are more commonly tasked with, non-research activities.

18. The sentence 'New PIs should also be aware that supervision of master's students and PhD students would also count as teaching contribution', is not strictly true for all universities. There are differences in what universities classify as part of a teaching load.

19. Please check that the data presented in 10C does correspond to the sentence stating 'Nearly 18% of externally funded fellows reported not having an.. .'.

20. The statement '...participated by submitting projects to up to 5-6 in some cases (Fig 14C).' does not agree with the data shown in Figure 14C.

21. Please show the data that supports the statement that '...lecturers were the least satisfied with worklife balance compared to research fellows'.

RECOMMENDATIONS FOR IMPROVED SUPPORT OF NEW PRINCIPAL INVESTIGATORS

22. Many of the advice points are not related to issues arising from the data. It would instead be more compelling to discuss recommendations directly related to the data and/or to discuss sources that may support the potential effectiveness of these solutions. The article would also be strengthened by more recommendations that new PIs can act on regardless of host support. For example when mentioning that mentorship to new PIs is not ideal, you could direct readers to resources for effective use of mentorship.

23. For the actions by host institutions, could you recommend anything related to the observations that there are clear gender differences in salary and start-ups, as well as recruitment around the REF?

24. Please explain the rationale for suggesting different grades for research fellows and university-funded fellows (action points 7 and 8 for host institutions).

METHODS AND SUPPORTING DATA

25. Please expand the methods section to give full details of the nature of the items included in the survey and explain why these were chosen, the pilot testing and revision process, the specific definition of the sought respondents, and the recruitment strategy used. Please also refer the reader to the document of the final survey questions.

26. Please provide a supporting data file that clearly states the number and percentages of respondents to each option in the survey, broken down by gender and lecturer/fellow as appropriate.

[Editors' note: further editorial revisions were requested prior to acceptance, and these revisions resulted in changes to the numbering of the figures.]

We repeat the reviewers’ points here in italic, and include our replies point by point, as well as a description of the changes made, in Roman.

GENERAL

1. There are a number of incorrect citations (for example, the in-text citations for the panels of Figure 10 are labelled as Figure 9, and a citation to 14F appears to be for 14G). Also, the 4th sentence in the caption to figure 13 should start: "Fellows (D) and Lecturers (E) were then.. . ". Please check and correct the text and figure captions for issues like these. Please also ensure that all in-text discussion of data that's presented in a figure cites that figure.

REPLY: The figure panels and numbering have all been updated to address edit suggested below. The text has been edited to reflect all of these changes. Many panels have also been removed from the main figures to streamline the manuscript, and now form supplements for each main figure.

2. Please present the findings in a more neutral manner. For example, instead of making a broad statement such as "too many new investigators express frustration", please report on the percentages who are frustrated.

REPLY: We have edited the text with this comment in mind, all changes are tracked in the submitted text document. We have deleted statements such above, and replaced with numbers from the data collected.

PRESENTATION OF FIGURES

3. The article will be easier to read if the number of figures is reduced. eLife allows the figures in a paper to have figure supplements; the figures appear in both the online and pdf versions; the figure supplements only appear in the online version, but the figure caption makes it clear that the figure supplement are available online. Please combine figures and convert some panels into figure supplements as suggested below.

REPLY: We have rearranged all of the figures to streamline the manuscript

Please see table of rearranged figures below

4. Some panels are not discussed in the text (2C, 6A, 6B, 6D, 6E, 7A, 7B, 9C, 10A, 13B, 13C, 14D, 14E, 14F, 14I, 15A); please either i) delete these or ii) convert them to figure supplements (and make sure they are cited somewhere in the text or the relevant figure caption).

REPLY: Please see table of rearranged figures below

5. The results of yes/no questions (5A, 5B, 5C, 6C, 10B, 10C, 10D, 10E, 12A, 12B, 13A, 14A, 14B) would be better presented by stating the percentage of respondents who answered yes in the text and/or figure caption, and citing the appropriate section of your supporting data file.

REPLY: This is now changed. Please see table of rearranged figures below for updated figure numbering.

6. Please include y-axes and keys for figures where these are missing (including Figures 3, 4, 11).

REPLY: This is now changed. Please see table of rearranged figures below for updated figure numbering.

7. Figure panels 14G and H show the same data, so please convert one to a figure supplement. Also, for 14G, please adjust the y-axis scaling to make it possible to see the data that supports the statement in the text about start-up funds "in the region of £20-60K".

REPLY: Please see table of rearranged figures below

8. Please ensure that the order of the figures matches the order in which they are cited/mentioned in the text.

REPLY: Please see table of rearranged figures below

RESULTS

9. Please present the results of the statistical tests for all the data shown; for example, this is missing from comparisons between female and male PIs and/or lecturers and researcher fellows. Please state the types of statistical tests run for each piece of presented data, with details on how certain confounders were controlled for, as well as detailing any corresponding statistical outcomes.

REPLY: The methods section has been extended and updated, changes are tracked in the attached document. We now also include. xls files for each graph shown, which makes clear the total numbers for each category/response.

10. The first sentence of the results states that '...365 respondents represents a significant proportion of new group leaders in the UK.' To support this comment, please state the number of new group leaders in the UK for the period corresponding to the results presented.

REPLY: We have been unable to find a total number of PIs in the UK, this statement was based on estimates knowing the approximate number of fellowships awarded by each of the larger funding bodies per cycle. However, we would be happy to remove this statement if required.

11. Please rephrase the final sentence for the second paragraph of the results ('There has been some discussion... in redressing the balance') to make its meaning clearer.

REPLY: Please see tracked changes in the attached text.

12. Please clarify what Figure 2B represents by describing it in the figure legend; you may also wish to add further clarification in the main text. Should the values for women and men add to 100%? Does this graph show only those that had dependents/caring responsibilities? Does '0' for length of break mean the individual had a dependent but took 0 months off from work? What is the mean and median for the data for each gender? How does this relate to the number of dependents per respondent? Furthermore, the panel shows possibly 1 or 2 men taking very long periods of leave, which does not appear to support the statement in the text that '... fathers are also taking significant periods of leave and sharing childcare responsibilities.' Please clarify.

REPLY: Figure 2 B has been edited and now forms part of the supplement for figure 1. We do not have data for the number of dependents per respondent, this was sometimes supplied in the comments but we are unable to break the data down in this way.

13. Please provide the data supporting the following sentence: 'It was also encouraging to find that having dependents did not affect satisfaction ratings or optimism scores for either females or males'; the data presented is not separated by sex.

REPLY: This data is now included as a supplement to figure 3, split by gender as requested.

14. Instead of stating that the proportion of lecturers to research fellows captured in the study was roughly equal, please state the exact percentages.

REPLY: This has now been addressed

15. The comments about research fellows being based at Russell group universities (bottom of page 3) seems to reflect that these universities first recruit individuals as research fellows (rather than as lecturers), as shown in Figure 5D. Please discuss this in the text.

REPLY: This has now been addressed, please see tracked changes

16. Could the data presented in Fig 8 and 15 also relate to when career breaks may occur for a women (which we know are for longer periods)? Was information collected about the timing of leave with regard to the trajectory of the respondent's career paths? In regards to 8C, could this be related to men asking for more money for each grant compared to women? Could a calculation be made to determine whether the amount awarded per grant varied between women and men?

REPLY: We did not collect data regarding the timing of leave with the career trajectory, although this would be an interesting question. We also did not ask for grant amount per application, and since this would vary greatly depending on the type of grant it would not be possible to draw concrete conclusions. It would be very interesting to ask funders to look at the grant amount awarded split male to female.

17. Based on the data presented in 9B and C, women have the larger loads for teaching and committees. Please discuss this, as these data may be related to the finding that women also secure fewer grants and have small lab sizes - as they are simply over-committing themselves to, or are more commonly tasked with, non-research activities.

REPLY: This has now been addressed, please see tracked changes

18. The sentence 'New PIs should also be aware that supervision of master's students and PhD students would also count as teaching contribution', is not strictly true for all universities. There are differences in what universities classify as part of a teaching load.

REPLY: We have edited this please see tracked changes

19. Please check that the data presented in 10C does correspond to the sentence stating 'Nearly 18% of externally funded fellows reported not having an.. .'.

REPLY: This figure has now been removed from the manuscript

20. The statement '...participated by submitting projects to up to 5-6 in some cases (Fig 14C).' does not agree with the data shown in Figure 14C.

REPLY: This data has been move to Figure 8 - figure supplement 2, and the text has been edited to match the plots.

21. Please show the data that supports the statement that '...lecturers were the least satisfied with worklife balance compared to research fellows'.

REPLY: This is now shown in a new figure supplement - NEW Figure 9 - figure supplement 1

RECOMMENDATIONS FOR IMPROVED SUPPORT OF NEW PRINCIPAL INVESTIGATORS

22. Many of the advice points are not related to issues arising from the data. It would instead be more compelling to discuss recommendations directly related to the data and/or to discuss sources that may support the potential effectiveness of these solutions. The article would also be strengthened by more recommendations that new PIs can act on regardless of host support. For example when mentioning that mentorship to new PIs is not ideal, you could direct readers to resources for effective use of mentorship.

REPLY: Please see tracked changes, we note that peer-peer mentorship can be extremely important and an excellent source of support.

23. For the actions by host institutions, could you recommend anything related to the observations that there are clear gender differences in salary and start-ups, as well as recruitment around the REF?

REPLY: Having discussed this with our host departments this is a very difficult issue to address, as they require a degree of flexibility in starting salary in order to recruit their favoured candidates who may have other offers.

24. Please explain the rationale for suggesting different grades for research fellows and university-funded fellows (action points 7 and 8 for host institutions).

REPLY: University funded fellowships are generally a scheme to bring in potential candidates for career development fellowships, so within our peer group this is generally viewed as a stepping stone towards independence which is solidified when external funding is awarded.

METHODS AND SUPPORTING DATA

25. Please expand the methods section to give full details of the nature of the items included in the survey and explain why these were chosen, the pilot testing and revision process, the specific definition of the sought respondents, and the recruitment strategy used. Please also refer the reader to the document of the final survey questions.

REPLY: Please see tracked changes, we have added detail to this section

26. Please provide a supporting data file that clearly states the number and percentages of respondents to each option in the survey, broken down by gender and lecturer/fellow as appropriate.

REPLY: We have uploaded the supporting data for each plot now shown, broken down as presented.

Author details

1. Sophie E Acton

   Sophie E Acton is a Cancer Research UK career development fellow in the MRC Laboratory for Molecular Cell Biology, University College London, United Kingdom

   Contribution

   Conceptualization, Data curation, Investigation, Writing—original draft, Writing—review and editing

   For correspondence

   s.acton@ucl.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2704-716X

2. Andrew JD Bell

   Andrew JD Bell is a lecturer in the Sheffield Methods Institute, University of Sheffield, United Kingdom

   Contribution

   Software, Formal analysis, Validation, Methodology, Writing—review and editing

   For correspondence

   andrew.j.d.bell@sheffield.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8268-5853

3. Christopher P Toseland

   Christopher P Toseland is an MRC Career Development awardee in the Department of Oncology and Metabolism, University of Sheffield, United Kingdom

   Contribution

   Conceptualization, Data curation, Investigation, Writing—original draft, Writing—review and editing

   For correspondence

   c.toseland@sheffield.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1641-7535

4. Alison Twelvetrees

   Alison Twelvetrees is a Vice Chancellor's fellow in the Sheffield Institute for Translational Neuroscience, University of Sheffield, United Kingdom

   Contribution

   Conceptualization, Data curation, Visualization, Writing—original draft, Writing—review and editing

   For correspondence

   a.twelvetrees@sheffield.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-1796-1508

• 11,566

  views

• 653

  downloads

• 15

  citations

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