The importance of creating a positive, fair and transparent working environment in academia has become increasingly evident over the past decade. Cross-sectional surveys and institutional audits reveal high self-reported levels of stress, poor work-life balance, and disparities in pay and career progression (Acton et al., 2019; Evans et al., 2018; Guthrie et al., 2018; Kim et al., 2024; Kinman and Jones, 2008; Marck et al., 2024; Morgan et al., 2021; Susi et al., 2019; Wellcome Trust, 2020). While these studies lack demographically matched non-academic control groups, they identify systemic issues within academia that need to be addressed. Solving these problems requires intervention at all levels of the academic hierarchy. Making top-down changes at the national or institutional level – such as those proposed in the Researcher Development Concordat in the UK (RDC Strategy Group, 2019) – can take time. Conversely, bottom-up changes made by individuals – such as those proposed in this article – can be implemented more rapidly.

The research group (or “lab”) is the foundational unit throughout the academic world, and the policies of the group can determine the wellbeing of the people who work in it (Hammoudi Halat et al., 2023). The group leader is typically responsible for determining the policies of their group, and hence has the strongest influence on the lab culture. There are informative resources with advice for starting a research group (Aly, 2018; Goldstein and Avasthi, 2021; Schmidt, 2006; Somerville et al., 2019), and a number of groups have made the lab handbooks and manuals that document their various policies publicly available (Aly, 2018; Andreev et al., 2022; Martin and Stanfill, 2023; Tendler et al., 2023). In addition to being useful to people who are already part of the group, public handbooks and manuals are also helpful to individuals who are thinking of joining it.

In this article, we take a different approach with the SAFE Labs Handbook, creating a community-driven document that is applicable to any field of research within the life sciences. The handbook comprises 30 actions (or “commitments”) that can be taken to improve the culture in a research group. These commitments are all actionable items that can be directly implemented by a group leader, and they all contain well-defined, measurable outcomes. Moreover, the commitments can be implemented without institutional support. We also report the results of a survey in which more than 200 researchers from more than 20 countries – including group leaders, postdoctoral researchers, PhD students, and staff – were asked about the 30 commitments. The complete handbook, explanations of each commitment, and a growing collection of templates from international groups across disciplines, can be found here: https://safelabs.info/home/safe-labs-handbook/.

This article has its origins in the Starting Aware, Fair, and Equitable Labs (SAFE Labs) workshop that was attended by the present authors – all new group leaders in the life sciences – in Palazzone di Cortona, Italy in May 2024. The workshop was organized to explore how to foster more positive and equitable lab environments. Although cultural problems in academia are wide-ranging, the workshop focused on how group leaders could improve their lab environment without institutional support. The feedback on the workshop from the attendees was overwhelmingly positive, with everyone saying that they would make changes in their labs as a result of attending the workshop (Figure 2—figure supplement 1).

The workshop highlighted two key barriers to improving lab culture as a new group leader. First, existing resources provide either general advice or example handbooks from specific labs, rather than actionable steps that can be implemented in verifiable manner. Second, academia does not normalise the documentation of lab policy, with group leaders often relying on word-of-mouth to educate new lab members. When documentation does exist, key information is often limited to existing lab members, and is not available to prospective applicants.

Attendees at the workshop agreed that the “best” lab environment was dependent on the individual. However, they also agreed that all groups could benefit from documenting key policies to minimise expectation mismatch between lab members (existing and prospective) and the group leader. To achieve this goal, attendees agree to create a handbook to: (i) provide actionable steps for group leaders to improve their lab environments without institutional support; (ii) normalize the process of documenting key lab policies.

While the initial workshop targeted new group leaders to ensure equality of experience among attendees, and to encourage discussion, the commitments in the SAFE Labs Handbook are relevant to all research groups, irrespective of how experienced the group leader is. Moreover, subsequent workshops have been open to all applicants.

The SAFE Labs Handbook

The SAFE Labs Handbook comprises 30 commitments spanning three broad categories: Teams, Policies, and Careers (Figure 1). Each commitment is an actionable statement that asks a group leader to publicly document, internally document, or establish a policy in their laboratory. Publicly documented information should be visible to anyone (e.g. posted on the lab website). Internally documented information need only be visible to existing lab members (e.g. posted in a password-protected lab manual or wiki): this may be information that is sensitive, or only relevant to individuals who are current members of the lab. In this context “establish” indicates a commitment that requires a change at the level of lab management, such as organizing a new annual meeting. Commitments were generated through a distillation of the notes from the workshop by the 13 attendees (see Methods). Publicly vs internally documented commitments were established through anonymous vote, with the decision deferred to the community-survey if there was no consensus (10%<score < 50% in favour, Figure 2—figure supplement 1h).

Figure 1

Download asset Open asset

Commitments were chosen to be actionable and verifiable to facilitate implementation and accountability. Each commitment derived from collated documentation from the SAFE Labs Workshop. We identified prominent issues that warranted inclusion in the handbook. For example: expectations around working hours and vacation are unclear to prospective and current lab members. We then agreed on the form of the commitment that made it actionable, verifiable and minimally prescriptive. “I commit to supporting a healthy work-life balance” would not qualify because the commitment cannot be verifiably implemented. Instead, the analogous entry in the handbook states that a group leader commits to “publicly document expectations for working hours, remote working, and vacation.” This is actionable (requires specific action from the group leader), verifiable (implementation can be evidenced), and does not mandate any particular policy – it instead requires that the policy is documented so that expectations are clear to all current and prospective lab members. We therefore provide only details, suggestions, and where appropriate, a freely editable template statement for each commitment. We also encourage the growing community to add examples of their statements, to help new group leaders develop effective policies tailored to their environment.

As an illustration, here is the current entry for the above commitment:

Details

Lab rules for working hours should be clear to avoid conflicts or misunderstandings, and clear expectations for working hours can increase equity between lab members. Group leaders need to ensure that lab members feel safe to balance their work in the lab with their life outside it. Policies regarding work hours, remote working, and vacation should be explicitly included.

Suggestions

Should notice of holidays be given and how? Are there core-working hours (typically less than the full working hours)? Should lab members schedule messages if sent outside of working hours? What times are appropriate times for scheduling meetings?

Template

I am committed to creating a healthy work environment for all lab members. I anticipate all lab members taking a minimum of [institution’s] prescribed days of annual leave. “Minimum” because if experiments/conferences necessitate working on a weekend, I support lab members taking time off to compensate. I hope to schedule all meetings within [institution’s] “core” work hours, and will refrain from sending, or answering, non-urgent emails/messages outside of work hours.

Full-time lab members should work “onsite” at least four days a week. I believe regular onsite presence is important to maintain the lab community. However, I will support intermittent periods of remote work when, for example, traveling/visiting family abroad or writing up a thesis/grant.

To assess the need for each commitment, we conducted a survey in which over 200 researchers were asked to rate the importance of every commitment, and to state whether it was already implemented in their environment. The number of respondents in the “group leader” career stage (51%, n=105), suggests a strong drive to improve lab culture at the senior level. The remaining 49% of respondents were split amongst postdocs (25%, n=57), PhD students (21%, n=46), and research support staff (7%, n=16; Figure 2a). Further, more than 30% of respondents (over 40% of non-PhD student respondents) had been in their role for over four years – demonstrating that both new and experienced members of the community engaged with our survey (Figure 2b). Participants represented institutions in 24 countries and worked in a range of fields (Figure 2c and d). Although these demographics are influenced by organiser bias, no single country represented more than 25% of participants. Overall, participant demographics indicate broad engagement with the commitments in the SAFE Labs Handbook.

Figure 2 with 1 supplement see all

Download asset Open asset

When asked to rank the importance of the commitments on a scale of 1–5 – where 1 was “useless” and 5 was “critical” – the average response was around 4 for all career stages (Figure 3a and b) and countries (Figure 3—figure supplement 1a–d). However, the average implementation rate for the commitments was only about 25% (Figure 3c and d). The mean importance score amongst respondents was not significantly impacted by their years of experience (Figure 3—figure supplement 2). Across the five countries with the most respondents, the relative importance of commitments and implementation rates were broadly consistent, with the UK and Italy reporting the highest and lowest implementation rates (Figure 3—figure supplement 1a–d). There was a strong correlation (R=0.67, P<0.001) between implementation rate and importance, as rated by group leaders (Figure 3e). This suggests that group leaders prioritize commitments they perceive as more important, but it may also reflect a confirmation bias. The same correlation was also significant when separated by country (Figure 3—figure supplement 1e–I).

Figure 3 with 2 supplements see all

Download asset Open asset

The importance scores demonstrate that the commitments benefit both group leaders and lab members. Group leaders are accountable for writing and implementing the commitments, creating a framework to communicate key information to their group. For lab members, these commitments establish expectations and responsibilities for each role, define policies to uphold, and establish feedback mechanisms to initiate change or highlight failures.

Commitments had a higher importance rating when scored by lab members than group leaders (Figure 3a and b; P<0.01, Mann–Whitney U test), suggesting that group leaders systematically underestimate the importance of these commitments to their lab members. Lab members also reported lower implementation rates than group leaders (Figure 3c and d, P<0.05, Mann–Whitney U test). This may be because group leaders overestimate the completeness of their implementation, or that lab members are unaware of the policies – reinforcing the importance of communicating lab policy through a written medium. We observed a correlation between importance and implementation rate for group leaders (P<0.001, robust linear fit; Figure 3e), but not for lab members (Figure 3f–h), suggesting a disconnect between the actions taken by group leaders, and those that would be most valuable from the perspective of their lab. Indeed, the mean implementation rate across respondent groups was significantly correlated with the difference in importance score between group leaders and lab members (P<0.001, robust linear fit; Figure 3i), suggesting group leaders prioritise policies they deem important, but not necessarily those most valued by group members.

Lab members and group leaders supported internal, rather than public, documentation of lab policies. When SAFE Labs Handbook authors did not agree on whether a commitment should be publicly or internally documented Figure 2—figure supplement 1h, survey respondents were asked for their input. Group leaders indicated a broad preference against public documentation (public preference ranged from 9 to 26%), while lab members were more supportive (12–75% public preference; Figure 4—figure supplement 1). These results suggest that while public documentation is considered valuable to some lab members, it is most important the information is documented. This informed our plans for the handbook (see Discussion).

Our survey results could be biased by sampling individuals willing to volunteer their time to complete the survey. Therefore, to test the general applicability of our results, we implemented a paid survey, asking current PhD students to rate a subset of 10 commitments (see Methods). In support of our initial survey, there was a strong correlation (P<0.001, robust linear fit) with unpaid participants when considering the relative importance across commitments (Figure 4a), although importance scores were marginally lower for this paid cohort (Figure 4b).

Figure 4 with 2 supplements see all

Download asset Open asset

More than 50% of group leaders, and 70% of lab members (both unpaid and paid participants), supported full implementation of the handbook in their lab environment (Figure 4c, Figure 4—figure supplement 2). Critically, less than 6% of respondents were unsupportive of the handbook. Instead, a significant fraction of respondents indicated “maybe”—often citing institutional limitations or preferences for internal documentation (see Discussion). Given the verified importance of these commitments, this level of implementation has the potential to generate systemic change throughout the academic community.

The SAFE Labs Handbook is a tool that helps group leaders to create more positive and equitable environments. Importantly, the commitments in the handbook can be implemented by individual groups without institutional support. A cornerstone of the SAFE Labs approach is the recognition that each group leader has a distinct approach to management. Thus, commitments do not specify how a group should be managed, and different approaches are often equally valid, and appealing to different lab members. Moreover, we do not envision that the handbook will resolve all issues, or stop overt abuses of power (Moss and Mahmoudi, 2021). Rather, it will support group leaders who are committed to improving the culture in their research group.

Our survey showed that the commitments in the handbook are widely recognized as important but are implemented in a minority of labs. Despite all participant groups agreeing on the importance of the commitments, the implementation rate was notably low (<25%). We believe this percentage may even be an overestimate, as respondents included lab members from the authors of this document, where implementation is more probable. This gap between perceived importance and actual implementation demonstrates the value that the SAFE Labs Handbook can bring to the community. Group leaders are already convinced of the importance of these commitments, but the handbook will both increase awareness and ease implementation through example documentation.

Although implementation rate and importance were strongly correlated for group leaders, this correlation was not significant for lab members, suggesting that group leaders could better align their efforts with those of their lab members. Group leaders rated two commitments as marginally more important than lab members (‘Individual Feedback’ and ‘PhD Committee’). Of the remaining commitments, scored higher by lab members, the largest differences were observed for ‘Postdoc Funding’, ‘Ambitions for Roles' and 'Responsibilities'. This reveals an interesting trend, with group leaders assigning more importance to opportunities for scientific guidance, and lab members emphasizing clarity around expectations and career development.

Survey respondents reflect significant sampling biases, but correlate with the opinions of an independent cohort of paid participants. The predominance of survey respondents from the UK, France, and Italy, as well as the fields of neuroscience and cellular biology, reflect the composition of the authors of this manuscript, as well as biases in grant mechanisms and the willingness of participants to volunteer their time (see Methods). Nonetheless, our results were strongly correlated with data from a separate pool of paid participants, indicating that these results reflect opinions of the wider academic community. Nevertheless, we acknowledge that the handbook would benefit from an increased diversity of voices. Indeed, with more data, differences in the implementation rate or perceived importance across countries could serve as important indicators that changes need to be made at the institutional or national level to support group leaders.

The SAFE Labs Handbook is a dynamic resource document that will evolve through community-driven feedback and contribution. Our survey data establish the need for this tool, and the importance of the existing commitments. However, there remains significant scope for revision and expansion, both from feedback as labs implement the handbook, and from further workshops. The handbook repository includes a growing collection of real-world examples from existing implementations to help new labs develop their policies. Moreover, future SAFE Labs workshops will focus in-part on proposing changes to the handbook via the lens of a new cohort of group leaders and organisers (https://safelabs.info/).

Individual handbooks should also be considered ‘living documents’, and we expect both public and internal documentation to change as each group evolves. These changes should be communicated to existing lab members, and consulting lab members when updating the content of the handbook can be a valuable process. However, we encourage all group leaders to complete their initial documentation independently (indeed, this is the only option for new group leaders). We believe an honest assessment from the group leader regarding their preferred lab policies is the most effective starting point in producing a useful handbook. All new members should be made aware of policies before joining and asked to formally acknowledge the handbook content during onboarding (e.g. providing a signature). If a member does not adhere to these values, feedback should be communicated through the established mechanisms – which are part of the existing commitments.

In addition to quantifiable metrics, our survey presented participants with the opportunity to comment on individual commitments, and on the complete handbook. We address the four most common concerns below.

We envision the SAFE Labs Handbook as a critical tool for group leaders, as well as an opportunity to advertise their commitment to creating more positive and equitable academic environments. We acknowledge that the handbook is not sufficient to create a positive lab culture, and that this relies on sustained efforts from both group leaders and lab members. However, if expectations are clear to all lab members at the earliest opportunity, it creates a valuable foundation from which future challenges and conflicts can be addressed. We expect the handbook to be most effective in an environment that encourages open and honest bidirectional feedback, and the commitments insist on establishing dedicated meetings for this purpose. It is not possible to eliminate difficulties in lab members confronting group leaders that fail to adhere to their own handbook, but we include suggestions in the handbook for reducing this barrier (e.g. an anonymous feedback channel).

To maximise future engagement with both the handbook and future workshops, we hope to disseminate information more effectively. Avenues could include institutional welcome packages for new group leaders, early career funding schemes, and professional networks such as EMBO. Engagement is also encouraged through our dedicated mailing list and Github discussions (https://github.com/SAFE-Labs-Docs/Lab-Handbook/discussions). We maintain a list of SAFE Labs on our website (https://safelabs.info/home/the-labs/). Labs that want to be featured on this page are required to submit a request, together with a link to their publicly documented handbook or policies. These submissions are briefly scrutinised before labs are added to the community list, and the lab is encouraged to feature the lab logo on their website in recognition of their implementation. These logos are considered indicators of engagement, like open science badges, and advertise to prospective scientists that the group leader places importance on lab culture. A more formal certification system could be developed in the future for use as a key performance indicator at the institutional level and could be recognised by funders.

All data generated or analysed during this study are included in the accompanying Git repository (https://github.com/SAFE-Labs-Docs/2025_Paper; copy archived at Coen, 2025). All data, analyses, and updated versions of the handbook will remain available at the SAFE Labs Handbook website (https://safelabs.info/home/safe-labs-handbook/) and the SAFE Labs Handbook GitHub repository (https://github.com/SAFE-Labs-Docs/Lab-Handbook; commit 971f8bf at time of submission).

1. 1. Acton SE
   2. Bell AJ
   3. Toseland CP
   4. Twelvetrees A

   (2019) A survey of new PIs in the UK

   eLife 8:e46827.

   https://doi.org/10.7554/eLife.46827

   • PubMed
   • Google Scholar
2. 1. Aly M

   (2018) The key to a happy lab life is in the manual

   Nature 561:7.

   https://doi.org/10.1038/d41586-018-06167-w

   • PubMed
   • Google Scholar
3. 1. Andreev A
   2. Komatsu V
   3. Almiron P
   4. Rose K
   5. Hughes A
   6. Lee MY

   (2022) Welcome to the lab

   eLife 11:e79627.

   https://doi.org/10.7554/eLife.79627

   • PubMed
   • Google Scholar
4. Software

   1. Coen P

   (2025) 2025_Paper, version swh:1:rev:a91b1dfd787f406eeb25a514fb514f2c9292a206

   Software Heritage.

   https://archive.softwareheritage.org/swh:1:dir:b786d0b52a66d80364347425e3a5a33ef336f029;origin=https://github.com/SAFE-Labs-Docs/2025_Paper;visit=swh:1:snp:49f49cde1e62c29f313b279ee761e0fa94811fb1;anchor=swh:1:rev:a91b1dfd787f406eeb25a514fb514f2c9292a206
5. 1. Evans TM
   2. Bira L
   3. Gastelum JB
   4. Weiss LT
   5. Vanderford NL

   (2018) Evidence for a mental health crisis in graduate education

   Nature Biotechnology 36:282–284.

   https://doi.org/10.1038/nbt.4089

   • PubMed
   • Google Scholar
6. 1. Goldstein B
   2. Avasthi P

   (2021) A guide to setting up and managing a lab at a research-intensive institution

   BMC Proceedings 15:8.

   https://doi.org/10.1186/s12919-021-00214-7

   • PubMed
   • Google Scholar
7. Website

   1. Guthrie S
   2. Lichten CA
   3. Belle JV
   4. Ball S
   5. Knack A
   6. Hofman J

   (2018) Understanding Mental Health in the Research Environment: A Rapid Evidence Assessment

   Accessed October 27, 2025.

   https://www.rand.org/pubs/research_reports/RR2022.html
8. 1. Hammoudi Halat D
   2. Soltani A
   3. Dalli R
   4. Alsarraj L
   5. Malki A

   (2023) Understanding and fostering mental health and well-being among university faculty: A narrative review

   Journal of Clinical Medicine 12:4425.

   https://doi.org/10.3390/jcm12134425

   • PubMed
   • Google Scholar
9. 1. Kim L
   2. Hofstra B
   3. Galvez SMN

   (2024) A persistent gender pay gap among faculty in a public university system

   Scientific Reports 14:22212.

   https://doi.org/10.1038/s41598-024-72871-5

   • PubMed
   • Google Scholar
10. 1. Kinman G
    2. Jones F

    (2008) A life beyond work? Job demands, work-life balance, and wellbeing in UK Academics

    Journal of Human Behavior in the Social Environment 17:41–60.

    https://doi.org/10.1080/10911350802165478

    • Google Scholar
11. 1. Marck CH
    2. Ayton D
    3. Steward T
    4. Koay HF
    5. Wiley JF
    6. Taiaroa G
    7. Walton CC
    8. Weld-Blundell I
    9. Greaves MD
    10. Singh A

    (2024) The workplace culture, mental health and wellbeing of early- and mid-career health academics: A cross-sectional analysis

    BMC Public Health 24:1122.

    https://doi.org/10.1186/s12889-024-18556-0

    • PubMed
    • Google Scholar
12. 1. Martin MY
    2. Stanfill AG

    (2023) On the road to achieving work-life balance in academia

    Clinical and Translational Science 16:553–556.

    https://doi.org/10.1111/cts.13485

    • PubMed
    • Google Scholar
13. 1. Morgan AC
    2. Way SF
    3. Hoefer MJD
    4. Larremore DB
    5. Galesic M
    6. Clauset A

    (2021) The unequal impact of parenthood in academia

    Science Advances 7:eabd1996.

    https://doi.org/10.1126/sciadv.abd1996

    • PubMed
    • Google Scholar
14. 1. Moss SE
    2. Mahmoudi M

    (2021) STEM the bullying: An empirical investigation of abusive supervision in academic science

    EClinicalMedicine 40:101121.

    https://doi.org/10.1016/j.eclinm.2021.101121

    • PubMed
    • Google Scholar
15. Website

    1. RDC Strategy Group

    (2019) Concordat to Support the Career Development of Researchers

    Accessed October 27, 2025.

    https://researcherdevelopmentconcordat.ac.uk/
16. 1. Schmidt EW

    (2006) Making the right moves [A review of "A Practical Guide to Scientific Management for Postdocs and New Faculty"]

    Journal of Natural Products 69:309.

    https://doi.org/10.1021/np0582731

    • Google Scholar
17. 1. Somerville LH
    2. Cunningham WA
    3. Gruber J
    4. Bavel JJV

    (2019) Three keys to launching your own lab

    Science.

    https://doi.org/10.1126/science.caredit.aaz1327

    • Google Scholar
18. 1. Susi T
    2. Shalvi S
    3. Srinivas M

    (2019) ‘I’ll work on it over the weekend’: high workload and other pressures faced by early-career researchers

    Nature.

    https://doi.org/10.1038/d41586-019-01914-z

    • Google Scholar
19. 1. Tendler BC
    2. Welland M
    3. Miller KL
    4. The WIN Handbook Team

    (2023) Why every lab needs a handbook

    eLife 12:e88853.

    https://doi.org/10.7554/eLife.88853

    • PubMed
    • Google Scholar
20. Website

    1. Wellcome Trust

    (2020) What Researchers Think About The Culture They Work In

    Accessed October 27, 2025.

    https://wellcome.org/insights/reports/what-researchers-think-about-research-culture

Decision letter after peer review:

Thank you for submitting your article "The SAFE Labs Handbook: Community-Driven Commitments for Group Leaders to Improve Lab Culture" to eLife for consideration as a Feature Article. Your article has been reviewed by three peer reviewers, and there reports are below. The following reviewer has agreed to reveal their identity: Benjamin Tendler (Reviewer #3).

General Assessments

Reviewer #1:

This is an important article that provides concrete action items and resources for group leaders who are seeking to document their lab policies in a handbook. This is a much-needed initiative with the potential to improve transparency in academia, which in turn is important for improving lab culture and mental well-being. Its strengths include: (i) data on the perceived importance and implementation of individual commitments and on willingness to implement lab handbooks; (ii) concrete advice, templates, and an easily accessible starting point to what can be a daunting task of creating a handbook; (iii) excellent acknowledgement of limitations in the approach and biases inherent to the sampling of workshop and survey participants; and (iv) a clearly written article with links to resources that the community can use. The main limitations revolve around an increased need for discussion of several issues, summarized below.

Reviewer #2:

I had come across the SAFE labs initiative before, and read this work with great pleasure. I strongly agree with the author that bottom-up, community initiatives are a key ingredient of academic culture change, and this work is a fantastic step in that direction. Given eLife's commitment to community and research culture, I think this would be the ideal outlet for publication.

This article is a bit of a mix between a general feature (to showcase the workshop and handbook itself) and research (surveys to draw more quantitative conclusions). While I think the latter is great to have, it is the former that I believe will have the greatest impact through wider communication about the existence of the handbook. I have a few suggestions that the authors may consider (see below).Reviewer #3:

In this work, Donà and colleagues present the SAFE Labs Handbook, an evidence-based research culture initiative actionable by research groups. The SAFE Labs Handbooks initiative proposes the adoption of a document comprising of 30 community-driven 'commitments' designed to improve the way research groups operate. The recommendations are backed up by community evidence (taken in the form of surveys), designed to be easily actionable, and are non-prescriptive to incorporate different viewpoints.

Establishing the 30 commitments utilised a carefully considered methodology. The researchers have gone to considerable effort to quantitatively evaluate their recommendations, demonstrating that their commitments are perceived as important at all levels of a research group whilst simultaneously poorly described/documented in lab policy. The number of survey respondents (particularly from group leaders) is impressive, and also helps extract differences in how researchers at different career stages interpret the relative importance and implementation of different policies. From my perspective, I frame the commitments as a distillation of previous lab handbook initiatives designed to improve research culture, sourcing evidence from communities (as opposed to an individual group), and describing a direct set of recommendations that can be more rapidly implemented without institutional support.

Whilst the authors go to great length to describe how they established their commitments, the aim of their work is not simply to establish a template, but to launch an initiative promoting the adoption of SAFE handbooks in the wider research community. A key area that should be addressed is how to approach writing the handbook, who it is for, and how it is incorporated into a research lab. Below I provide some comments focused on these themes.

Essential Revisions:

Reviewer #1:

[1] The Introduction states that "academic environments are associated with impaired mental health, poor work-life balance, and disparities in pay and career progression". It is important to clarify what the comparison group is. Impaired mental health compared to what group? Poor work-life balance compared to what group? Is the comparison group matched on important demographic features, and differ only in the presence vs. absence of an academic work environment? Or does this claim rest on self-reported ratings of mental health and work-life balance without a matched comparison group? Likewise, it seems important to clarify whether the disparities in pay and career progression are larger in academia vs. other settings.

[2] It is stated that "existing resources provide general advice rather than actionable steps that can be verifiably implemented". Yet, some of the cited papers include links to publicly available lab policies that touch on many of the SAFE Labs Handbook Commitments, and these papers invite group leaders to follow their example by creating their own lab policies using those publicly available ones as a template. I definitely see the value of the SAFE Labs Handbook and the list of commitments, but the statements that prior work in this space has been more general advice rather than actionable steps does not seem accurate.

[3] The paper notes a disconnect between actions taken by group leaders and those perceived to be most important by lab members. It further notes that increased divergence between group leaders and lab members was correlated with a higher global importance score (Figure 3i). But Figure 3i seems to suggest that the actions with the most divergence (-0.5 or lower on the y-axis) are relatively *less* important. Actions that are more important (>4.5 on the x-axis) have values closer to 0 on the y-axis (e.g., -0.2), indicating less divergence between group leaders and lab members. Because most of the points in that figure lie below 0 on the y-axis, more divergence reflects *lower*, more negative, y-axis values, which are related to *lower* importance scores on the x-axis. This seems to contradict the statement in the paper that increased divergence is correlated with higher global importance. Can the authors clarify where the misunderstanding is?

[4] There are two additional discussion points that I think are worth including in the paper. First is the difference between implementing the suggested practices and actually having a positive lab culture. Transparently communicating expectations does not necessarily lead to a positive lab environment. In addition to fair and equitable expectations, a positive lab environment requires sustained efforts on the part of the group leader and the lab members. It might be advisable to mention that this lab handbook approach should be the first step, and that it may be necessary but is certainly not sufficient for a supportive environment. A second, related, issue is that of accountability. I appreciate the recommendation for annual lab-wide feedback sessions and mechanisms for sharing concerns and conflicts. But it can be very difficult for lab members to confront their group leader when the group leader is not following their own stated policies. This challenge is almost inevitable because of power dynamics and how much lab members depend on their group leader for advancement. It may be worth discussing how it can be very difficult to hold group leaders accountable and ensure that they follow their stated policies. It would also be valuable to include recommendations for how this can be approached in a way that protects lab members and eases the discomfort and anxieties associated with giving their group leader negative feedback.Reviewer #2:

[5] My first one is to speculate about future ways in which the potential of this resource could be maximized. The website will spread through subfields by word-of-mouth, some people will see the article/preprint, and a low number may attend the workshop itself as a new PI. Would there be other ways to include this resource in e.g. welcome packages at institutions? Through funders who award early career grants? Through e.g. FENS, EMBO, SfN, NeuroMatch, NewPI Slack? I do not expect the authors to carry out all this work in community building, but a vision for the way forward would help in identifying where new cohorts of the workshop could build on further anchoring these practices in the community.

[6] The second one is to zoom in a bit to the items that have the highest discrepancy between PIs and lab members, from figure 3i. Which topics would need most attention or additional discussion?Reviewer #3:

[7] Writing a SAFE Handbook

The survey findings identify an increased interpreted level of policy implementation for group leaders versus members (Figure 3c), suggesting that policies are being miscommunicated by group leaders to group members. In addition, the findings identify increased perceived importance of several topics from group members versus leaders (Figure 3a), suggesting that the voices of group members are not being effectively heard. This evidence suggests that a SAFE Handbook should be written collectively as a group exercise rather than a top-down approach, in agreement with previous handbook initiatives that strongly advocated for writing as a group exercise (Tendler et al. 2023), or findings highlighting the gap of interpretation of expectations (Wellcome Trust 2020).

The manuscript does not present any information about how to approach writing a SAFE handbook beyond hinting as a group writing exercise in 'Concern 4'. The authors should add a clear description of this to ensure that the benefits of the document are maximised, and do not risk parts of the initiative boiling down to a template copy-and-paste box ticking exercise by a group leader without broader thought and consideration of lab viewpoints.

[8] Who is the Handbook for?

The writing of the manuscript and SAFE Handbook website suggests that the thirty commitments are for group leaders. However, I would argue that many of the themes (e.g. code of conduct) are the responsibility of everybody in the lab, and require an explicit commitment from individual group members. Please describe and provide justification as to who is accountable for the contents of the lab handbook. If the expectation is that ensuring the commitments are adhered to is the responsibility of the entire group, recommendations for achieving group buy-in.

[9] How is the Handbook integrated into the lab?

Please describe more explicitly what the expectations are for the established handbook in the lab. Specifically, (1) the expectations surrounding group members reading and committing to the handbook, (2) the recommended process if somebody is not living up to the values expressed in the handbook, (3) expectations for the handbook content to be reviewed and updated (e.g. on an annual basis as a 'living document').

Essential Revisions:

Reviewer #1:

[1] The Introduction states that "academic environments are associated with impaired mental health, poor work-life balance, and disparities in pay and career progression". It is important to clarify what the comparison group is. Impaired mental health compared to what group? Poor work-life balance compared to what group? Is the comparison group matched on important demographic features, and differ only in the presence vs. absence of an academic work environment? Or does this claim rest on self-reported ratings of mental health and work-life balance without a matched comparison group? Likewise, it seems important to clarify whether the disparities in pay and career progression are larger in academia vs. other settings.

We agree with the reviewer and have now corrected this omission in the manuscript (Intro, ∼ Para 1). We now make it clear that these studies lack comparisons to demographically matched control groups, but that they still suggest systemic issues within academia that need to be addressed.

[2] It is stated that "existing resources provide general advice rather than actionable steps that can be verifiably implemented". Yet, some of the cited papers include links to publicly available lab policies that touch on many of the SAFE Labs Handbook Commitments, and these papers invite group leaders to follow their example by creating their own lab policies using those publicly available ones as a template. I definitely see the value of the SAFE Labs Handbook and the list of commitments, but the statements that prior work in this space has been more general advice rather than actionable steps does not seem accurate.

We agree with the reviewer that there are existing resources providing both general advice and specific examples. These resources are incredibly valuable, and we agree that our previous statement unintentionally minimised this contribution. However, as the reviewer points out, these resources typically consist of a single, actionable step: using an example handbook to help create a personalised one. We feel that the breakdown of this process into separate, actionable, commitments will make this process more manageable and more universally applicable. We now specify this distinction in the manuscript (Intro, ∼Para 2).

[3] The paper notes a disconnect between actions taken by group leaders and those perceived to be most important by lab members. It further notes that increased divergence between group leaders and lab members was correlated with a higher global importance score (Figure 3i). But Figure 3i seems to suggest that the actions with the most divergence (-0.5 or lower on the y-axis) are relatively *less* important. Actions that are more important (>4.5 on the x-axis) have values closer to 0 on the y-axis (e.g., -0.2), indicating less divergence between group leaders and lab members. Because most of the points in that figure lie below 0 on the y-axis, more divergence reflects *lower*, more negative, y-axis values, which are related to *lower* importance scores on the x-axis. This seems to contradict the statement in the paper that increased divergence is correlated with higher global importance. Can the authors clarify where the misunderstanding is?

We thank the reviewer for raising this issue: our description of the correlation in the previous figure was mis-stated. Further, the effect was small and not significant when evaluated with Pearson correlation, the metric used in panels 3e-h (see minor point, Reviewer 1.c), we have therefore replaced Figure3i with a more informative analysis, capturing a more substantial effect.

The new Figure 3i shows that the difference in importance-score between group leaders and group members correlates with the mean implementation rate across groups. Thus, implementation is lower for commitments that group members deem more important than group leaders; in other words, group leaders prioritise policies they deem important over those that may be more valued by their group members. We have corrected the figure, legend, and text (Results of a Community Survey, ∼ Para 4).

[4] There are two additional discussion points that I think are worth including in the paper. First is the difference between implementing the suggested practices and actually having a positive lab culture. Transparently communicating expectations does not necessarily lead to a positive lab environment. In addition to fair and equitable expectations, a positive lab environment requires sustained efforts on the part of the group leader and the lab members. It might be advisable to mention that this lab handbook approach should be the first step, and that it may be necessary but is certainly not sufficient for a supportive environment. A second, related, issue is that of accountability. I appreciate the recommendation for annual lab-wide feedback sessions and mechanisms for sharing concerns and conflicts. But it can be very difficult for lab members to confront their group leader when the group leader is not following their own stated policies. This challenge is almost inevitable because of power dynamics and how much lab members depend on their group leader for advancement. It may be worth discussing how it can be very difficult to hold group leaders accountable and ensure that they follow their stated policies. It would also be valuable to include recommendations for how this can be approached in a way that protects lab members and eases the discomfort and anxieties associated with giving their group leader negative feedback.

We agree with the reviewer on both these points. We have now re-emphasised the limits of the handbook, including the difficulties in providing feedback to group leaders (Discussion, ∼ Para 12). We also highlight that the handbook includes suggestions for reducing the barrier to providing feedback to group leaders, and include one example in the paper (maintaining an anonymous feedback channel, Discussion, ∼ Para 12).

Reviewer #2:

[5] My first one is to speculate about future ways in which the potential of this resource could be maximized. The website will spread through subfields by word-of-mouth, some people will see the article/preprint, and a low number may attend the workshop itself as a new PI. Would there be other ways to include this resource in e.g. welcome packages at institutions? Through funders who award early career grants? Through e.g. FENS, EMBO, SfN, NeuroMatch, NewPI Slack? I do not expect the authors to carry out all this work in community building, but a vision for the way forward would help in identifying where new cohorts of the workshop could build on further anchoring these practices in the community.

We agree with these points. We have expanded our discussion to touch on these points (∼ Discussion, ∼ Final Para). We further highlight our future plans, and the available mailing list and github discussions for community engagement. We are committed to continuing the SAFE Labs Website and repository as this initiative evolves.

[6] The second one is to zoom in a bit to the items that have the highest discrepancy between PIs and lab members, from figure 3i. Which topics would need most attention or additional discussion?

This is an excellent suggestion, and we now highlight some points of discrepancy in the discussion. This revealed a suggestive difference, where group leaders felt scientific guidance was more important while lab members highlighted the importance of clear expectations and career-development goals (Discussion, ∼Para 3).

Reviewer #3:

[7] Writing a SAFE Handbook

The survey findings identify an increased interpreted level of policy implementation for group leaders versus members (Figure 3c), suggesting that policies are being miscommunicated by group leaders to group members. In addition, the findings identify increased perceived importance of several topics from group members versus leaders (Figure 3a), suggesting that the voices of group members are not being effectively heard. This evidence suggests that a SAFE Handbook should be written collectively as a group exercise rather than a top-down approach, in agreement with previous handbook initiatives that strongly advocated for writing as a group exercise (Tendler et al. 2023), or findings highlighting the gap of interpretation of expectations (Wellcome Trust 2020).

The manuscript does not present any information about how to approach writing a SAFE handbook beyond hinting as a group writing exercise in 'Concern 4'. The authors should add a clear description of this to ensure that the benefits of the document are maximised, and do not risk parts of the initiative boiling down to a template copy-and-paste box ticking exercise by a group leader without broader thought and consideration of lab viewpoints.

We thank the reviewer for highlighting a significant omission from the manuscript. We agree that feedback from lab members is critical, and that the Handbook should be regularly reviewed and edited in conversation with lab members (Discussion, ∼ Para 6). However, we do not believe the initial handbook should be written as a group exercise for the following reasons:

New group leaders do not have the facility to collectively write their handbook, but we do not want to intimate that this process should be delayed until the lab is larger. Handbooks should be completed before the first lab member joins to provide them with a clear impression of the expectations of the group leader.

Commitments in the Handbook are designed to minimise expectation mismatch between lab members (both prospective and current) and group leaders, with the acknowledgement that each group leader is different. We believe an honest assessment from the group leader regarding their preferred lab policies is the most effective starting point in producing a useful Handbook, rather than deferring this responsibility to a collective effort. After the Handbook is written, feedback from lab members is extremely valuable, and if suggestions are not incorporated, the reasoning behind this should be explained.

The group leader is (usually) the only “constant” presence in the lab, while the other personalities will change over time. Therefore, we believe it is appropriate for the group leader to have a driving role in setting the expectations for their group rather than relying on lab members (likely the more vocal members).

We therefore suggest that documentation be written by the group leader in the first instance (Discussion, ∼ Para 6).

[8] Who is the Handbook for?

The writing of the manuscript and SAFE Handbook website suggests that the thirty commitments are for group leaders. However, I would argue that many of the themes (e.g. code of conduct) are the responsibility of everybody in the lab, and require an explicit commitment from individual group members. Please describe and provide justification as to who is accountable for the contents of the lab handbook. If the expectation is that ensuring the commitments are adhered to is the responsibility of the entire group, recommendations for achieving group buy-in.

The reviewer makes an excellent point, echoing feedback that we also received from our survey, which we should have addressed more plainly. We now clarify (Results of a Community Survey, ∼ Para 3) that the Handbook is for the benefit of both group leaders and lab members. Group leaders are accountable for writing an implementing the commitments, creating a framework to communicate key information to their group. For lab members, these commitments establish expectations and responsibilities for each lab role, define policies to uphold, and establish feedback mechanisms to initiate change or highlight failures. To promote this framework, we now specify that the onboarding procedure needs to include reading and acknowledging the lab handbook by the lab member.

[9] How is the Handbook integrated into the lab?

Please describe more explicitly what the expectations are for the established handbook in the lab. Specifically, (1) the expectations surrounding group members reading and committing to the handbook, (2) the recommended process if somebody is not living up to the values expressed in the handbook, (3) expectations for the handbook content to be reviewed and updated (e.g. on an annual basis as a 'living document').

These are excellent suggestions, and we now specifically address them in the revision (Discussion, ∼Para 6).

Author details

1. Erika Donà

   Erika Donà is in the Institute of Neuroscience, Consiglio Nazionale delle Ricerche (CNR), Vedano al Lambro, Italy

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   James M Gahan, Petrina Lau, Jana Jeschke, Torben Ott, Katja Reinhard, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-9268-3619

2. James M Gahan

   James M Gahan is in the Centre for Chromosome Biology, School of Biological and Chemical Sciences, University of Galway, Galway, Ireland

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, Petrina Lau, Jana Jeschke, Torben Ott, Katja Reinhard, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0001-7043-4873

3. Petrina Lau

   Petrina Lau is in the Department of Psychiatry and the Gerald Choa Neuroscience Institute, The Chinese University of Hong Kong, Shatin, Hong Kong

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Jana Jeschke, Torben Ott, Katja Reinhard, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2257-2344

4. Jana Jeschke

   Jana Jeschke is in the Institut Jules Bordet, Université libre de Bruxelles, Brussels, Belgium

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Torben Ott, Katja Reinhard, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

5. Torben Ott

   Torben Ott is in the Institute of Biology and Bernstein Center for Computational Neuroscience Berlin, Humboldt University of Berlin, Berlin, Germany

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Jana Jeschke, Katja Reinhard, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4168-2832

6. Katja Reinhard

   Katja Reinhard is in the Neuroscience Department, Scuola Internazionale Superiore di Studi Avanzati, Trieste, Italy

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Jana Jeschke, Torben Ott, Chiara Sinigaglia, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-8719-7445

7. Chiara Sinigaglia

   Chiara Sinigaglia is at the Biologie Intégrative des Organismes Marins (BIOM), CNRS and Sorbonne Université, Banyuls-sur-Mer, France

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Jana Jeschke, Torben Ott, Katja Reinhard, Jorien L Treur and Thomas Vogl

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7195-7091

8. Jorien L Treur

   Jorien L Treur is in the Department of Psychiatry, Amsterdam UMC, University of Amsterdam, Amsterdam, Netherlands

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Jana Jeschke, Torben Ott, Katja Reinhard, Chiara Sinigaglia and Thomas Vogl

   Competing interests

   No competing interests declared

9. Thomas Vogl

   Thomas Vogl is in the Center for Cancer Research, Medical University of Vienna, Vienna, Austria

   Contribution

   Conceptualization, Writing – review and editing

   Contributed equally with

   Erika Donà, James M Gahan, Petrina Lau, Jana Jeschke, Torben Ott, Katja Reinhard, Chiara Sinigaglia and Jorien L Treur

   Competing interests

   No competing interests declared

10. Stephane Bugeon

    Stephane Bugeon is at the Institut de Neurobiologie de la Méditerranée (INMED), INSERM and Aix Marseille Université, Marseille, France

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

    Contributed equally with

    Letizia Mariotti, L Federico Rossi and Philip Coen

    For correspondence

    stephane.bugeon@inserm.fr

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-5703-4848

11. Letizia Mariotti

    Letizia Mariotti is in the Institute of Neuroscience, Consiglio Nazionale delle Ricerche (CNR), Padova, Italy

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

    Contributed equally with

    Stephane Bugeon, L Federico Rossi and Philip Coen

    For correspondence

    letizia.mariotti@cnr.it

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0002-6749-1006

12. L Federico Rossi

    L Federico Rossi is in the Center for Neuroscience and Cognitive Systems, Istituto Italiano di Tecnologia, Rovereto, Italy

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

    Contributed equally with

    Stephane Bugeon, Letizia Mariotti and Philip Coen

    For correspondence

    federico.rossi@iit.it

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0001-5831-4860

13. Philip Coen

    Philip Coen is in the Cell and Developmental Biology Department, University College London, London, United Kingdom

    Contribution

    Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Visualization, Methodology, Writing – original draft, Project administration, Writing – review and editing

    Contributed equally with

    Stephane Bugeon, Letizia Mariotti and L Federico Rossi

    For correspondence

    p.coen@ucl.ac.uk

    Competing interests

    No competing interests declared

    "This ORCID iD identifies the author of this article:" 0000-0003-1495-1061

• 0

  citations

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