F1000Research F1000Research 2046-1402 F1000 Research Limited London, UK 10.12688/f1000research.161819.2 Research Article Articles The most prevalent perceived barriers to sharing research data at the point of manuscript acceptance in health and life sciences: a survey at the University of Bristol

[version 2; peer review: 2 approved, 2 approved with reservations]

 Bajraktari Rrita Formal Analysis Investigation Methodology Visualization Writing – Original Draft Preparation 1 Booth Fiona Conceptualization Funding Acquisition Methodology Supervision Writing – Review & Editing 2 Munafo Marcus Funding Acquisition Writing – Review & Editing 2 Beazley-Long Nicholas Conceptualization Data Curation Formal Analysis Investigation Methodology Supervision Writing – Original Draft Preparation Writing – Review & Editing https://orcid.org/0000-0001-7342-2771 a 2 Universite Libre de Bruxelles, Brussels, Brussels, Belgium University of Bristol Faculty of Life Sciences, Bristol, England, UK nick.beazley-long@bristol.ac.uk

No competing interests were disclosed.

 15 8 2025 2025 14 482 13 8 2025 Copyright: © 2025 Bajraktari R et al. 2025 This is an open access article distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Background

Sharing research data is critical for study validation and reuse, yet challenges persist across disciplines, such as psychology 1 and biomedical science 2. While global initiatives promote open science, understanding localized barriers in specific academic contexts is vital to implementing effective solutions.

 Objective

To investigate the most prevalent perceived barriers and reasons that prevent the sharing of research data underlying manuscripts, at the point of their acceptance for publication, within the Faculty of Health and Life Sciences (FHLS) at the University of Bristol, a research-intensive university in the UK.

 Methods

We distributed a comprehensive survey to FHLS researchers, addressing logistical, technical, and cultural challenges. A total of 143 participants provided insights into their experiences with data sharing.

 Results

The primary obstacles identified were time constraints and the complexity of the preparation process, with 34% reporting they “usually” or “always” lack sufficient time to adequately prepare their data for sharing. Additional barriers included not having the rights to share (27%), insufficient technical support (15%), and limited incentives within research teams. Moreover, qualitative responses highlighted a lack of confidence in data sharing infrastructure and guidance.

 Conclusions

These findings highlight the importance of targeted interventions to enhance data-sharing practices. Solutions should prioritize data preparation processes, clarify data ownership policies, and offer tailored training programs. Integrating data-sharing requirements into research workflows from the outset could significantly alleviate these challenges. Our study provides actionable recommendations to inform the development of resources and infrastructure that support a culture of open science within the FHLS at the University of Bristol.

 Health Sciences Life Sciences data sharing open data Open Science reproducibility reuse barriers to sharing. Philanthropic donation from Professor John Climax This research was supported by a philanthropic donation from Professor John Climax to the University of Bristol. No funding organization has had any role in the survey’s design, implementation or analysis. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Amendments from Version 1

The title has been revised to reflect better the objective of the study, clarifying that respondents were instructed to consider barriers to sharing research data underpinning manuscripts at the point of manuscript acceptance for publication. Minor revisions to the text in the Abstract, Introduction and Study Design were made throughout for consistency and clarity on the respondents' instructions.   A sentence has been added to Survey Design explaining that the question in Part 2 of the survey was framed as “researchers around you”. This wording was intentionally chosen to capture participants’ perceptions of the data sharing practices and barriers surrounding them — not objective knowledge of others’ behaviours. This framing has been shown to reduce potential social desirability bias. In Survey Design we have also included the exact text used in the survey instructing the respondents on how to rate the statements.

 Introduction

Scientific research is inherently a collective enterprise. The era of solitary discoveries is long past; today, advancing knowledge requires collaboration, both within local research groups and across broader networks. Knowledge sharing has become essential to drive scientific progress and enhance our understanding through community efforts. For centuries, researchers have relied on publishing in peer-reviewed journals as the primary means of sharing their findings. However, accumulating evidence suggests that this traditional approach alone no longer meets the standards required by modern research for example in clinical trials 3 and societal challenges. 4 Poor replicability of research across disciplines in health and life science 5 underscores the need for greater openness and transparency in the research process. 3

Open research promotes openness beyond merely sharing final research findings. Key elements of openness include making datasets available when feasible and ethical, providing detailed documentation of methodologies, including metadata, and sharing research materials and code. Increased transparency not only allows for critical validation but also facilitates replication and enhances the credibility of both the research and researchers. 6

While data sharing was historically not standard practice, it is now an expectation, driven by shifts in research culture and mandates from funders and publishers. 1 However, transitioning to data sharing practices is far from straightforward, encountering challenges in definitions, 7 implementation, 8 ecological concerns (i.e. the energy needed to store the data being too high 9 ) and ensuring equitable practices. 10

For the purpose of this study, we define data sharing at the point of manuscript acceptance, as: the practice of making research data as open as possible online, allowing other researchers and the public to freely access, reuse, and analyse the data in a format which is openly available.

While the importance of data sharing is widely recognized, our study focuses on the practical and perceived barriers that researchers face when manuscripts are accepted for publication—a critical stage when data is finalized. By exploring these perceived barriers and reasons, we aim to help researchers recognize and overcome those challenges, where possible, to enable future data sharing. Our study highlights the main reasons for not sharing data within the Faculty of Health and Life Sciences (FHLS) at the University of Bristol, a research-intensive university in the UK, and reinforces findings from previous studies elsewhere, adding further evidence to support and expand on existing research and teaching in this area.

 Methods Study ethics approval

The survey design was approved by the University of Bristol’s School of Psychological Science Review Ethics Committee (Ethic approval: 18486, approved 05/06/2024). Informed consent was provided by each participant prior to beginning the survey, research data was anonymous at point of collection and no personally identifiable information was collected. The Participant Information Sheet and consent form are available with the other materials from this study.

 Survey design

To better understand the reasons and perceived barriers preventing the sharing of research data underlying manuscripts following their acceptance for publication, participants from across the Faculty of Health and Life Sciences were asked to rate statements related to data sharing. The survey consisted of two sections. In the first section, statements were framed as self-referential (e.g., “I don’t share data because it is too complicated”). In the second section, the same statements were rephrased to describe researchers’ behaviour around them (e.g., “They don’t share because it is too complicated”). The question was framed as “researchers around you”, to reduce social desirability bias – see survey statement development.

Participants were asked how often some barriers were encountered when sharing data at point of manuscript acceptance. They rated their responses using a four-point scale: “never,” “sometimes,” “usually,” and “every time.” For comparison, these responses were converted into numeric values ranging from 0 (never) to 3 (every time), and mean scores were compared across the two groups. Participants had the option to provide additional comments through a free-text box included at the end of each section. The survey was hosted on Microsoft Forms for online administration. Participants were asked to sign in using their university credentials to prevent entry duplication and external access by unauthorized individuals.

Participant instructions preceding the statements (see Extended Data for the full list of statements):

“On a scale from Never to Every time, can you rate how often each statement below applies to you when thinking about sharing the research data underlying your manuscript following its acceptance for publication:”

“On a scale from Never to Every time, can you rate on how often each statement below applies to others around you when thinking about sharing the research data underlying your manuscript following its acceptance for publication:”

 Survey advertising and incentive

The survey targeted health and life science researchers and was advertised through faculty research mailing lists and displayed on digital screens across the faculty. The survey remained open for one month. To encourage participation, three £50 prepaid gift cards were offered as incentives, with winners selected randomly using Google’s random number generator immediately following survey closure.

 Survey statement development

The survey items were developed based on themes from four key studies addressing data-sharing practices and barriers. Perrier and colleagues 11 identified issues such as data integrity, research conduct, and feasibility challenges, while Gomes et al. 12 emphasized barriers like process complexity, lack of incentives, and reuse concerns, and Gownaris et al. 13 highlighted early-career researchers’ concerns, including fear of misuse and career implications. Toelch and Ostwald 14 provided best practices for transparent research, shaping the phrasing of survey items.

A total of 21 items were selected from these themes. In the first section of the survey, the items were framed as personal statements (e.g., “I’m unsure about the process”) to foster self-reflection and participant engagement, consistent with findings from D’Ailly et al. 15 and Brenner. 16 The second section, where the statements applied to colleagues (e.g., “They are unsure about the process”), introduced psychological distance, which can reduce social desirability bias. 17, 18

Thematic analysis of comments

After rating the statements, participants had the opportunity to leave comments, which were subjected to a thematic analysis. This analysis was conducted independently by two researchers using a systematic approach: read through all comments, then identify themes on second read-through, and categorize these themes into broader categories. Once both researchers completed their independent analyses, they shared and discussed their identified themes and categorizations to reach a consensus. In cases where disagreements persisted, a third researcher was consulted to resolve the differences and finalize the thematic categorization. This collaborative process helped to improve the reliability and validity of the thematic analysis. 19

Results Participant demographics

A total of 143 active researchers completed the survey, composed of 114 research staff, 28 postgraduate research (PGR) students and 1 undergraduate student. The estimated response rate across the faculty research staff was 9.1% (Supplementary Figure S1) and researchers from all career stages contributed to the survey (Supplementary Figure S2).

 General survey findings

The primary perceived barrier to data sharing was a lack of time, identified as the top-ranked barrier in both sections of the survey. The statement “I/they don’t have enough time to prepare my/their data for sharing” had the highest mean scores in part 1 (personal framing: 1.19) and part 2 (colleague framing: 1.42).

Figure 1 (self-referential) and Figure 2 (colleague referential) rank the statements by mean score. The top ranked statements were:

The lack of time to prepare data for sharing

The complexity of the process (“too complicated”)

Managing and sharing large datasets (“too many data files” and “difficulty sharing large datasets”)

Lack of rights to share data

Inadequate infrastructure

Lack of team support (“my team doesn’t do it”)

Lack of knowledge on how to share data

Results for perceived barriers to data sharing with personal framing.

This figure illustrates participant responses to statements relating to perceived barriers to data sharing, framed as applying to themselves. The white numbers within each bar represent the absolute number of participants who selected “Never,” “Sometimes,” “Usually,” or “Every Time,” as indicated in the legend. The bold numbers on the right of each bar show the mean score for each statement. Mean scores were calculated by assigning numerical values to responses (0 = Never, 1 = Sometimes, 2 = Usually, 3 = Every Time), multiplying these values by the number of responses, and dividing the total by 143 (the number of participants).

 Results for perceived barriers to data sharing with colleagues framing (‘They’).

This figure illustrates participant responses to statements relating to perceived barriers to data sharing, framed as applying to colleagues. The white numbers within each bar represent the absolute number of participants who selected “Never,” “Sometimes,” “Usually,” or “Every Time,” as indicated in the legend. The bold numbers on the right of each bar show the mean score for each statement. Mean scores were calculated by assigning numerical values to responses (0 = Never, 1 = Sometimes, 2 = Usually, 3 = Every Time), multiplying these values by the number of responses, and dividing the total by 143 (the number of participants).

These results suggest that barriers are not only logistical but also cultural and knowledge-based. In part 2, similar barriers emerged but in a different order, with “Their team doesn’t do it” rising to second place. It is also interesting to observe that all mean scores for each statement were higher and that the top 8 statements in part 1 match those in part 2. This positive difference could be explained by a potential desirability bias or self-deception.

 Career stage and departmental differences

When breaking down the responses to the different career stages a similar general pattern to the combined responses was observed (Supplementary Figure S3), with “not enough time” being the top reason in two thirds of the career stages and the complexity around sharing (“too complicated”, “too many data files”) scoring highly. However, PGR students expressed more concerns about:

Managing large datasets

Risks of plagiarism

Sharing intellectual property

Differences between departments (Supplementary Figure S4), aligning with other published findings, 2 also reflected disciplinary nuances:

Population Health Sciences: The top barrier was the lack of rights to share sensitive data, reflecting the challenges and ethical concerns when handling health-related datasets.

Psychological Sciences: Responses highlighted concerns about anonymising and sharing human-centred data.

Biochemistry: The primary barrier was managing large and complex datasets.

These subtle differences between the schools and career stages will help us to tailor our internal training appropriately for difference audiences.

 Survey free text comments

Of the 143 participants, 57 left detailed comments, which were analysed thematically to provide additional insights. This process led to the identification of 14 themes consisting of some that were linked to the pre-defined statements (e.g. a lack of time), also others that were not covered by the statements (e.g. fear of failure to replicate, fear of error detection). This analysis has provided additional context enriching our understanding of the challenges and perceived barriers. The full thematic analysis is available in the available extended data. Below are examples of the identified themes and additional context.

Data organization and clarity: disorganized datasets not designed for sharing can lead to misinterpretation.

Training and support gaps: a lack of accessible guidance and resources on data sharing was frequently mentioned especially guidance on sharing qualitative data.

Challenges in anonymizing qualitative data were highlighted by several participants.

Fear of error detection and reproducibility issues: new concerns for our study emerged about criticism over errors or failed replication when data is shared.

Participants also highlighted distinctions within certain barriers, such as the “lack of benefit” theme. For example, 6 comments differentiated between a lack of personal benefit and a broader perception of the practice being “useless” for the field. This rich qualitative data will inform our training resources and workshops and enable the refinement of the statements for further iterations of the survey. The comments in full and the identified themes are available with the extended data. 20

Discussion

This survey aligns with previous findings on the barriers to data sharing. 11– 14 For instance, the predominant challenges identified, such as time constraints, complexity of data sharing, and lack of incentives, mirror those reported in the literature. But while our results reveal consistent tendencies with these earlier studies, no significance tests were conducted.

Our findings highlight several areas that could be of some interest for further investigation. First, the analysis of participants’ comments suggests that greater specificity in survey items addressing complexity would be valuable. For example, the item “it is too complicated” could be refined to distinguish between technical challenges, organizational issues, and insufficient infrastructure. This distinction would enable future research to better identify and address specific barriers. Similarly, the item “I don’t see any benefits” should differentiate between personal perceptions (e.g., “no benefit for me”) and broader views of utility to the scientific enterprise. This level of detail could inform tailored interventions, such as workshops to raise awareness of data-sharing benefits or institutional policies offering incentives for researchers.

Another important consideration for future research is the impact of question phrasing on responses. Our survey’s dual framing of items (“I” vs. “They”) provides preliminary evidence of a potential social desirability bias. As data sharing becomes increasingly common, it is possible that open science practices are perceived as socially desirable behaviors, influencing self-reported attitudes and behaviors. Investigating this bias further could provide valuable insights into the evolving perception of open practices within the scientific community. To our knowledge, no prior survey has employed this dual framing approach, making our study a novel contribution to the evaluation of scientific practices.

Finally, the comments from our participants revealed an additional barrier not initially considered: fear of error detection or lack of reproducibility. This concern underscores the importance of fostering a culture of transparency that minimizes stigma around errors. An initiative to achieve this was recently launched with the Estimating The Reliability & Robustness Of Research (ERROR) project. 21 Future surveys should include items explicitly addressing this issue to capture its prevalence and impact on data-sharing behavior.

Limitations and recommendations for future surveys

This study has several limitations that should be considered when interpreting the findings. First, participants were not given the option to skip statements they felt uncomfortable answering or did not know how to respond to, which may have affected the reliability of some responses. Second, we did not collect information about the primary research data type (e.g., quantitative or qualitative) for each participant. Including this information in future surveys could provide important context for understanding the specific barriers faced by researchers working with different types of data. Third, the sample size was limited for certain schools, which may have introduced bias in the representation of discipline-specific challenges. Conducting future surveys during periods when more researchers are available, such as outside of the summer months, could help improve response rates and ensure a more representative sample.

Despite these limitations, the survey has provided direct evidence of the barriers to data sharing at the University of Bristol and we consider these findings to be generalizable to other similar research-intensive institutions. These valuable insights have informed in-house training and guidance initiatives. Specifically, these results have shaped resources aimed at improving data management skills for data sharing and reproducible research practices. Additionally, this survey serves as a benchmark for future iterations, allowing for the evaluation of progress in addressing the barriers identified and refining strategies to support data sharing within the academic community.

 Ethical considerations and consent to participate

The survey design was approved by the University of Bristol’s School of Psychological Science Review Ethics Committee (Ethic approval: 18486, approved 05/06/2024). A Participation Information Sheet (PIS) was made available to the participants introducing the background and objectives of the study, that their research data (composed of their ratings to the statements and any comments) was anonymous at the point of collection, therefore not containing any personally identifiable information, and that the research data would be made open following the completion of the study. Written informed consent was collected digitally via Microsoft Forms before the survey began. Participants provided consent by ticking a checkbox. If consent was not given, the survey could not be started. Following the completion of the survey participants had the opportunity to enter a prize draw to win one of three £50 gift cards by submitting their university email address. These email addresses were stored separately, and not linked to, the anonymous research data and following the prize draw the email addresses were immediately deleted.

 Data availability

Study data, code, figures and supplementary figures are openly available on the University of Bristol’s open repository ( https://doi.org/10.5523/bris.qldqe39bi5yd22hkg5q3zce3i) 22 Data and figures are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0) ( https://creativecommons.org/licenses/by/4.0/). The deposit includes the following:

Readme.txt - A readme file including descriptive project metadata and the computing requirements to reuse the code

UoB_FHLS_2024DataAvailabilitySurvey.csv – the raw data in csv

UoB FHLS 2024DataAvailabilitySurvey.xls – raw data in xlsx

Excel_to_Figures_DATA_survey.R – code to generate figures

Reporting guidelines

This research followed the Consensus-based Checklist for Reporting of Survey Studies, CROSS. 23 The checklist accompanying this article is available. 20

Extended data

Additional data and study materials are available in the project OSF deposit 20 ( https://doi.org/10.17605/OSF.IO/V2XP3) available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0). This deposit includes the following:

ParticipantInformationSheet.docx – the PIS

OnlineConsentStatementBarriersToDA.docx – consent statement

CrossSurveyGuidelines.docx – guidelines followed for this survey

F1000ResDAQualiThemeExtraction.xlxs – theme analysis of participants’ comments

Acknowledgements

We would like to thank all the participants that took the time to complete the survey.

 References Hardwicke TE Mathur MB MacDonald K : Data availability, reusability, and analytic reproducibility: Evaluating the impact of a mandatory open data policy at the journal Cognition. R. Soc. Open Sci. 2018;5(8):180448. 30225032 10.1098/rsos.180448 PMC6124055 Tedersoo L Küngas R Oras E : Data sharing practices and data availability upon request differ across scientific disciplines. Sci. Data. 2021;8(1):192. 34315906 10.1038/s41597-021-00981-0 PMC8381906 Bauchner H Golub RM Fontanarosa PB : Data Sharing: An Ethical and Scientific Imperative. JAMA. 2016;315(12):1238. 10.1001/jama.2016.2420 Gelenbe E Brasseur G Chefneux L : On sharing knowledge and fostering ‘open science’. Ubiquity. 2021;2021(May):1–13. 10.1145/3462221 Świątkowski W Dompnier B : Replicability Crisis in Social Psychology: Looking at the Past to Find New Pathways for the Future. Int. Rev. Soc. Psychol. 2017;30(1):111. 10.5334/irsp.66 Nosek BA Hardwicke TE Moshontz H : Replicability, Robustness, and Reproducibility in Psychological Science. Annu. Rev. Psychol. 2022;73(1):719–748. 34665669 10.1146/annurev-psych-020821-114157 Elliott KC Resnik DB : Making Open Science Work for Science and Society. Environ. Health Perspect. 2019;127(7):75002. 31353949 10.1289/EHP4808 PMC6792383 Clark MP Luce CH AghaKouchak A : Open Science: Open Data, Open Models, … and Open Publications?. Water Resour. Res. 2021;57(4):e2020WR029480. 10.1029/2020WR029480 Farley SS Dawson A Goring SJ : Situating Ecology as a Big-Data Science: Current Advances, Challenges, and Solutions. Bioscience. 2018;68(8):563–576. 10.1093/biosci/biy068 How to share data—Not just equally, but equitably. Nature. 2023;622(7983):431–432. 10.1038/d41586-023-03239-4 Perrier L Blondal E MacDonald H : The views, perspectives, and experiences of academic researchers with data sharing and reuse: A meta-synthesis. PLOS ONE. 2020;15(2):e0229182. 32106224 10.1371/journal.pone.0229182 PMC7046208 Gomes DGE Pottier P Crystal-Ornelas R : Why don’t we share data and code? Perceived barriers and benefits to public archiving practices. Proc. R. Soc. B Biol. Sci. 2022;289(1987):20221113. 36416041 10.1098/rspb.2022.1113 PMC9682438 Gownaris NJ Vermeir K Bittner M-I : Barriers to Full Participation in the Open Science Life Cycle among Early Career Researchers. Data Sci. J. 2022;21:2. 10.5334/dsj-2022-002 Toelch U Ostwald D : Digital open science—Teaching digital tools for reproducible and transparent research. PLoS Biol. 2018;16(7):e2006022. 30048447 10.1371/journal.pbio.2006022 PMC6095603 D’Ailly HH Murray HG Corkill A : Cognitive Effects of Self-Referencing. Contemp. Educ. Psychol. 1995;20(1):88–113. 10.1006/ceps.1995.1005 Brenner PS : Advancing Theories of Socially Desirable Responding: How Identity Processes Influence Answers to “Sensitive Questions”. Brenner PS , editor. Understanding Survey Methodology. Springer International Publishing;2020; Vol.4: pp.45–65. 10.1007/978-3-030-47256-6_3 Krumpal I : Determinants of social desirability bias in sensitive surveys: A literature review. Qual. Quant. 2013;47(4):2025–2047. 10.1007/s11135-011-9640-9 Larson RB : Controlling social desirability bias. Int. J. Mark. Res. 2019;61(5):534–547. 10.1177/1470785318805305 Nowell LS Norris JM White DE : Thematic Analysis: Striving to Meet the Trustworthiness Criteria. Int. J. Qual. Methods. 2017;16(1). 10.1177/1609406917733847 Bajraktari R Booth F Beazley-Long N : Barriers to Data availability at the University of Bristol. OSF. 2025. 10.17605/OSF.IO/V2XP3 ERROR Project Team: Estimating The Reliability & Robustness Of Research (ERROR) project. Cited 2025 Apr 15. Reference Source Bajraktari R Booth F Beazley-Long N : The perceived barriers to data sharing in health and life sciences: a survey at the University of Bristol. 2025. 10.5523/bris.qldqe39bi5yd22hkg5q3zce3i Sharma A Minh Duc N Thang LL : A Consensus-Based Checklist for Reporting of Survey Studies (CROSS). J. Gen. Intern. Med. 2021;36:3179–3187. 33886027 10.1007/s11606-021-06737-1 PMC8481359 10.5256/f1000research.186377.r407252 Reviewer response for version 2 Amaral Olavo B. 1 Referee https://orcid.org/0000-0002-4299-8978 Costa Gabriel G. 1 Co-referee Federal University of Rio de Janeiro, Rio de Janeiro, Brazil

Competing interests: No competing interests were disclosed.

 10 10 2025 Copyright: © 2025 Amaral OB and Costa GG 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

Summary:

The study aims to investigate the most prevalent perceived barriers and underlying reasons that prevent the sharing of research data underpinning manuscripts at the point of acceptance for publication across Health and Life Sciences at the University of Bristol. The manuscript is clear and well-written, and the research question is relevant, although generalizability is limited by the specificity of the sample.

Major points:

- There is a data inconsistency between the results section, which mentions 114 research staff responses and Supplementary Table S1, which mentions 115 responses. Please revise.

- In spite of the statement “see Extended Data for the full list of statements”, we could not find the actual survey instrument in the OSF page, and could only find the statements in the raw data spreadsheet. Also, it would be much more useful to provide a direct link to such materials, rather than pointing to the Extended Data section, which then points to the general OSF page. The same holds for supplementary figures, which would be more easily accessible if they were made part of the paper (or via a direct link to each figure in the OSF).

- We don’t think that the difference in mean scores between the “I” and “people around me” responses is evidence for social desirability bias. First, I’d assume that, due to responder bias, people answering such a survey will have above-average interest in data sharing and may be genuinely more prone to doing it than the those around them. Second, as the answer options are “never/sometimes/usually/always”, it is expected that the “never” and “always” categories should shrink if one is talking about a large group of people rather than about a single person (as variance in the larger group will lead “sometimes” and “usually” to be more likely). From the figure, this indeed seems to be the case. Reduction in “never” responses (which are the most prevalent responses in the personal section) would thus by itself lead to a greater average in the third-party condition, even with no bias present.

- As the authors make some comparisons between response categories, it might be helpful to include 95% confidence intervals for mean scores to allow the reader to assess how likely these differences would be expected to happen by chance.

- Given the low response rate, responder bias is an important issue here and could indicate skepticism about data sharing among non-responders. The potential consequences on the findings should be discussed explicitly in the limitations section.

- Greater context about the University of Bristol is warranted if the authors want to imply that these findings are generalizable to other research-intensive institutions, so the reader can form an opinion about the scope of this generalizability. In terms of context, it would also be useful if the authors could discuss the implications of this data for potential policy changes in the institution.

Minor points

- In the first paragraph, the phrase "Poor replicability of research across disciplines in health and life science" is supported by reference number five (5), which only addresses replicability in social psychology. It would be useful to include references on replicability in other areas as well.

- In the second paragraph under survey design, it is written that "mean scores were compared across the two groups", but it is not clear which groups are being talked about (e.g. the groups of responses).

- It is not clear from which schools the PGRs and the single undergraduate respondent came, and how they were contacted. Was the survey distributed exclusively via active research staff, or was it sent directly to PGRs and undergraduates as well?

- It would also help to state more clearly in the methodology that the survey was conducted in the summer months, as this may have contributed to the low response rate

- In the legends of Figure 1 and 2, the full meaning of GDPR should be stated.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Partly

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Meta-research applied to research reproducibility

We confirm that we have read this submission and believe that we have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however we have significant reservations, as outlined above.

Bajraktari Rrita Social psychology, Universite Libre de Bruxelles, Brussels, Brussels, Belgium

Competing interests: No competing interests were disclosed.

 16 12 2025

We are grateful to Reviewer 4 (and 5 :')) for providing thoughtful and constructive feedback on our manuscript. We appreciate the time and care taken in reviewing this work, and we address the comments below. As the article has already been accepted as Version 2, we respond here and through updates to the extended materials.

Data inconsistency (114 vs 115 participants)

Thank you for pointing this out. The additional response was an undergraduate response – this has been corrected in the Supplementary Figure S1

Survey instrument

We appreciate the suggestion and the full survey instrument is now available in the Extended Data on OSF.

Interpretation of social desirability bias

We sincerely thank the reviewer for providing an alternative interpretation. In the manuscript, social desirability was presented as a possible hypothesis, not a definitive conclusion. We agree that alternative explanations exist and appreciate this thoughtful interpretation. Given the open peer review format, readers will be able to consider the reviewer’s perspective alongside ours.

We agree that responder bias is an important limitation, and this is why it was stated as one in the section ‘Limitations and recommendations for future surveys’.

Institutional context and policy implications

Thank you for highlighting this! We have now added the relevant University of Bristol policy on Open Data to the Extended Data in OSF. This allows readers to better assess the institutional context, generalisability, and potential implications for future policy development.

Minor comments

We thank the reviewer for the suggestions. Where appropriate, clarifications have been added in the extended materials. Regarding GDPR, we consider the acronym to be sufficiently well recognised in this context and does not need to be spelled out accordingly to the F1000Research author guidelines.

We once again thank Reviewer 4 for their constructive and helpful suggestions. We believe these comments have further improved the transparency and contextualisation of the study.

 10.5256/f1000research.186377.r407248 Reviewer response for version 2 Vieira Armond Anna Catharina 1 Referee https://orcid.org/0000-0002-7121-5354 University of Ottawa Heart Institute, Ottawa, Canada

Competing interests: No competing interests were disclosed.

 19 9 2025 Copyright: © 2025 Vieira Armond AC 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve-with-reservations

The study aims to investigate the barriers to research data sharing within the Faculty of Health and Life Sciences at the University of Bristol. The manuscript is clear and well-written, and the question is relevant. However, important clarifications and reporting gaps should be addressed.

First, the manuscript would benefit from some clarifications about the survey development and administration:

The full survey instrument is not included. I was only able to locate the survey statements by opening the dataset. The complete instrument, including instructions and questions, should be made available in the supplementary material or extended data.

Briefly describing this process (or noting that it was not piloted) would be helpful.

The timing of the survey is not specified (month and year), and it is not clear whether reminders were sent after the initial invitation. Adding this information would help readers interpret the reported response rates.

 Results

The authors report the response rate for staff, but not for the other groups. Response rates for all groups should be provided.

Study participants are not fully characterized. While some demographics are reported in the main text, a more detailed description is needed, including career stage and department.

Supplementary Figures 3 and 4 are important to the results and discussion. It would improve the paper if these were presented in the main text rather than as supplementary material.

 Discussion and limitations

Several points are missing or underdeveloped in the discussion. The study was conducted in a single faculty. While the results are useful for developing tailored interventions, the findings are not generalizable given the uneven representation across groups, the limited sample size, and the cultural specificity of data-sharing practices. This should be discussed with caution.

The interpretation of the findings would also benefit from more institutional context. Data sharing is influenced by both cultural and institutional factors. Therefore, the authors may also wish to add more details about whether the University of Bristol has mandates or recommendations for data sharing, whether infrastructure, training, and resources are available, and whether local funders have relevant requirements.

Finally, the definition of data sharing used in the study is restrictive, excluding controlled-access data, for example. This decision should be discussed, as biomedical researchers often face challenges in balancing de-identification with data usability, and controlled-access repositories are often the only viable model for sensitive health data.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

Partly

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Partly

Are sufficient details of methods and analysis provided to allow replication by others?

Partly

Reviewer Expertise:

Health sciences with a focus on metaresearch, research integrity, open science practices, and research data management.

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Bajraktari Rrita Social psychology, Universite Libre de Bruxelles, Brussels, Brussels, Belgium

Competing interests: NA

 16 12 2025

We would like to thank you for their thoughtful and constructive feedback. We appreciate the time dedicated to reviewing our manuscript and for the helpful suggestions, which have enabled us to further improve the transparency of our work. As the manuscript has already been accepted by two reviewers as Version 2, we address these points here and by clarifying the extended materials, rather than submitting a new version.

Survey instrument availability

Thank you! The full survey instrument, including instructions and all questions, is now available in OSF.

Survey timing and reminders

We agree that providing timing information helps contextualise the response rate. Month/year of data collection and information regarding reminders have now been added to the extended materials on OSF.

Response rate for all groups

We appreciate the comment. While the number of active research staff was known and therefore a response rate could be calculated, this information was not available for postgraduate researchers at the time of the survey, and therefore a response rate for this group could not be determined. We have clarified this point in the extended data.

Participant characterisation

The demographic detail is available in the supplementary figures. Supplementary Figures S1&2 display the department and career stage breakdowns of research staff. Supplementary Figures S3&4 include the total number of responses (staff and PGRs) per department and career stage in the column heading.

Generalisability and institutional context

We thank the reviewer for this insightful suggestion. We have now uploaded the University of Bristol’s Open Data policy at the time of the survey to the Extended Data on OSF. This provides institutional context on OSF that readers may better assess the extent to which findings relate to local institutional conditions.

Definition of data sharing

We believe that our definition was appropriate in the context of our project. Our aim was to assess current practices within our faculty, with a specific focus on open access data sharing. Importantly, the third part of the survey allowed participants to freely explain why they might not be able to share certain data, which we report in the Results and discuss in the Discussion section. The definition was explicitly provided to participants at the start of the survey. While we acknowledge that alternative definitions are possible, and that our definition may be considered restrictive, we needed to make a clear methodological decision and ensure that all respondents answered based on the same definition. For this reason, we do not consider it appropriate to modify the manuscript retrospectively, as it reflects the study procedures as they were conducted.

We again thank Reviewer 3 for their valuable and constructive comments, which we believe further strengthen the clarity and transparency of our study.

 10.5256/f1000research.186377.r405743 Reviewer response for version 2 Reichmann Stefan 1 Referee https://orcid.org/0000-0003-1544-5064 Graz University of Technology, Graz, Austria

Competing interests: No competing interests were disclosed.

 23 8 2025 Copyright: © 2025 Reichmann S 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve

Many thanks to the authors for carefully revising the paper. With the additional information on survey questions, the authors significantly improve the transparency of their methodology and the soundness of their argumentation. In the version that I reviewed, it was relatively unclear (at least to me) what questions had been used in they survey. Now that this has been appropriately addressed and included in the text, I will be happy to recommend the work for indexing.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

No

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

No

Are the conclusions drawn adequately supported by the results?

No

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

NA

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.

 10.5256/f1000research.177903.r386103 Reviewer response for version 1 Reichmann Stefan 1 Referee https://orcid.org/0000-0003-1544-5064 Graz University of Technology, Graz, Austria

Competing interests: No competing interests were disclosed.

 16 6 2025 Copyright: © 2025 Reichmann S 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation reject

Article: The perceived barriers to data sharing in health and life sciences: A survey at the University of Bristol

Reviewer: Dr. Stefan Reichmann

Article Synopsis

The article investigates the perceived barriers to data sharing in the Faculty of Health and Life Sciences at the University of Bristol. A survey (n=143) was distributed addressing 3 types of perceived barriers: cultural, logistical, and technical. The survey found the primary obstacles to be time constraints and the complexity of the preparation process, with 34% reporting they “usually” or “always” lack sufficient time to adequately prepare their data for sharing. Additional barriers included not having the rights to share (27%), insufficient technical support (15%), and limited incentives within research teams. Moreover, qualitative responses highlighted a lack of confidence in data sharing infrastructure and guidance.

Reviewer Comments

This is a carefully executed study with a timely research question. Nevertheless, there are some problems with the work that should be addressed.

The part of the work that I take issue with the most is the definition of data sharing given, which seems to me to be unnecessarily restrictive. The authors define data sharing as “the practice of making research data as open as possible online, allowing other researchers and the public to freely access, reuse, and analyze in a format which is openly available”. There are many other ways that data sharing happens, and by choosing such a restricted definition, the study risks missing quite a bit of the practice that may (or may not – this would be an interesting finding in its own right) be happening at FHLS. For instance, FHLS might have invested in data sharing infrastructure (an institutional repository, say), but respondents predominantly choose to share their data via other means or not share them at all – this would be undetectable with this survey. Christine Borgman (and others) have documented data sharing practices extensively, finding large variance across research settings in the ways research data are defined, produced, shared, reused, etc. I believe that defining data sharing as stated above restricts the domain of study too much – the phenomenon of data sharing is simply much broader than the study admits to. As is probably evident, this critique is written from the standpoint of a qualitative researcher who is mindful that the phenomenon of data sharing is richer and much more varied than the definition given.

The narrowness of the definition has ramifications for the study design. Since the concept has not been defined vis-à-vis respondents (at least, this is not mentioned in the study), we do not know which definition of data sharing respondents subscribe to (whether a narrow definition, as is given by the authors, or a wide definition as I suggested above). In that sense, the responses given would be very difficult to interpret adequately.

The scale used is not fit for the way the questions are phrased, at least in some cases (whenever the Likert scale is interpreted as “Never” to “Every time”): For example, “I don’t share data because it’s too complicated” seems to me to be a yes or no question. Applying the Likert scale in this (and other) cases seems forced, to say the least. In this particular example, a different question formulation (and scale) would have been more appropriate. In particular, the question combines two variables (“I don’t share data” and “Data sharing is complicated”) which make interpretation difficult and risks losing information. Additionally, it creates ambiguity in interpretation, because respondents could be assenting to either one of the two statements in their response. Because the question combines them, it becomes impossible to interpret meaningfully. A better way to approach this would have been to ask, in a separate question, whether respondents share their data, how often, with whom, and by what means, and then to ask about barriers.

The gulf between the way the questions are phrased and the response options offered becomes even more apparent when considering the second framing of the questions (“they” instead of “I”). There, it seems that very few respondents chose the option “Every time” for all of the survey items. I suspect that this is due to the way the questions are phrased and does not necessarily reflect how respondents view their colleagues. At the very least, a considerable number of statements would have to be assessed based on an “Completely agree-Completely disagree”-Likert Scale. Also, many of the questions logically imply an “I don’t know”-option. That respondents did not have the option to skip questions either (a limitation the authors acknowledge), in my view, hampers the results further. This could have been solved relatively easily by including an “I don’t know/Cannot answer” option. In general, while a survey should be constructed so as to yield generalizable, quantifiable results, the questions should still be as close as possible to our normal way of speaking, i.e. even if responses are collected on a five-(four, three)-point-scale, the questions should be as close as possible to the way respondents normally speak about the issues in question. Consequently, either the statements or the response options need to be rephrased (where in fact, I would opt for the latter in most cases). In addition, I would need to see the questions that respondents were actually asked – at present, the article merely presents the statements that respondents rated but I suspect that there would have been some kind of prompt.

Further, I am having trouble understanding the rationale of the two question framings. While I do think that reducing social desirability bias is laudable, I am hesitant to concur with the way this has been achieved because the resulting survey questions are difficult to interpret at face value – “they” is highly ambiguous, and the quality of a response changes considerably depending on how a respondent interprets “they” (colleagues can be people in the same department, field, even university – we don’t really know what group this is referring to). For instance, “They have difficulty sharing very large datasets” could be taken to refer to colleagues in the same field, which suggests a lack of skills, or it could be referring to colleagues in other disciplines, for which an explanation seems a little more complicated. In addition, some of the statements make considerably less sense when substituting “I” for “They”, for instance “Their team doesn’t do it” seems to me to be very difficult to interpret – whose team?

In terms of the findings, I would expect them to mirror those found in the literature since the survey items have been developed based on landmark studies, but I would caution against interpreting these findings as simply confirming previous work; rather, because of the survey design, I would argue that the work reproduces the assumptions built into the work that the present article builds upon.

In order to be of some help, I would like to end with some suggestions for how to improve the paper: The simplest (but probably most time-consuming) would be to follow-up with respondents across the disciplines represented in the sample and ask them for an interview. This would give respondents space to clarify how they interpret the core ideas used in the survey (such as “data sharing”, “reproducibility”, etc.), and to clarify whom they regard as in- resp. outgroup. Other than that, I think the most sensible course would be to revise the questions as indicated (I would be happy to offer my help there).

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

No

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

No

Are the conclusions drawn adequately supported by the results?

No

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Sociology of Science, Science and Technology Studies, Open Science, Information Science

I confirm that I have read this submission and believe that I have an appropriate level of expertise to state that I do not consider it to be of an acceptable scientific standard, for reasons outlined above.

Beazley-Long Nicholas Psychology, University of Bristol, Bristol, Bristol, UK

Competing interests: No competing interests were disclosed.

 8 8 2025

We thank the reviewer for the constructive criticism. We have made several changes and additions to the article for clarity. We believe we have addressed all the reviewers concerns below. 

Addressing the concern about the definition of data sharing used in this study. Our aim was to investigate the most prevalent perceived barriers and underlying reasons that prevent the sharing of research data underpinning research manuscripts, at the point of acceptance for publication across Health and Life Sciences at Bristol. To this end, we defined data sharing at the point of manuscript acceptance as: the practice of making research data as open as possible online, allowing other researchers and the public to freely access, reuse, and analyse the data in a format which is openly available. We have adjusted the text in the introduction to clarify this definition. To clarify that respondents were instructed to consider barriers to sharing data underlying manuscripts’ findings at the point of publication acceptance we have revised the title accordingly. We specifically chose this time point in the research lifecycle because at this point the data underlying manuscript findings are - in the vast majority - finalised and this time represents a critical decision-making moment when researchers are often required to consider data availability for publication. In addition, journal and funder policies around data sharing typically come into effect at this time. Focusing on this stage allows us to assess perceived and experienced barriers at the point when research data sharing is most actionable and policy-relevant. We have revised the article to include the text that instructed respondents how to rate the statements in part 1 and part 2. See below. These instructions help to define the ‘what’ (data underlying the manuscript) and the ‘when’ (at point of acceptance) for the respondents to consider. We did not want to state the exact types of data to consider (i.e raw, processed, summary, primary/secondary, methodological, unstructured etc) as this could have been restrictive. 

The instructing text in the survey.

Part 1. “On a scale from Never to Every time, can you rate on how often each statement below applies to you when thinking about sharing the research data underlying your manuscript following its acceptance for publication:” 

Part 2. “On a scale from Never to Every time, can you rate on how often each statement below applies to others around you when thinking about sharing the research data underlying your manuscript following its acceptance for publication:” 

Clarification of the Framing in Part 2 ("They" Framing) 

The reviewer raised a concern that participants may not know what their colleagues do or don’t do in respect to sharing research data. However, the survey did not ask about colleagues specifically. The actual question posed was: “How often each statement below applies to others around you...?” This wording was intentionally chosen to capture participants’ perceptions of the data sharing practices and barriers surrounding them — not objective knowledge of others’ behaviours. This framing has been shown to reduce potential social desirability bias. 

Clarification of the Question Format 

Reviewer 2 also questioned the framing of some items, such as “I don’t share data because it’s too complicated,” suggesting these are yes/no questions. However, the survey asked participants to indicate how often each statement applied, not whether it applied or not. This frequency-based approach was deliberate: researchers may encounter certain barriers (e.g., complexity, time constraints) in some projects but not others. A binary yes/no format would have been overly restrictive and failed to capture this variability. By asking about frequency, we aimed to better understand the prevalence of different barriers across contexts. 

Reproduction of assumptions argument 

The reviewer makes the argument that our survey design reproduces assumptions built into the study on which this article builds. However, our intention was to assess the prevalence of barriers and underlying reasons for data sharing challenges in Health and Life Sciences research. To this end, we developed 21 statements based on existing literature, complemented by free-text boxes to allow respondents to share additional thoughts, including barriers not captured by the predefined items. This approach builds on established knowledge in the field, which we believe is essential rather than limiting and it would have been remiss of us to ignore. We acknowledge, that if our primary aim had been to uncover entirely new or undocumented barriers, a different methodological approach would have been needed. However, our focus was on identifying the most prevalent barriers in order to inform practical strategies for addressing them.

 10.5256/f1000research.177903.r384382 Reviewer response for version 1 Boté-Vericad Juan-José 1 Referee https://orcid.org/0000-0001-9815-6190 Universitat de Barcelona, Barcelona, Catalonia, Spain

Competing interests: No competing interests were disclosed.

 9 6 2025 Copyright: © 2025 Boté-Vericad JJ 2025 This is an open access peer review report distributed under the terms of the Creative Commons Attribution Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. recommendation approve

The article is well-organized and clearly communicates its objective.

Current literature is cited across the introduction and methods sections.

Some introductory sentences are slightly redundant or overly long. Introduction needs only minor stylistic refinements for clarity.

Survey design is rigorous and includes ethical approval documentation.

Dual framing (I/They) is a thoughtful method reducing bias.Thematic analysis methodology is clear, collaborative, and appropriately validated.

Framing encourages reflection and distinguishes self versus group perceptions.

Survey questions are described clearly and linked to past research.

All source materials, survey documents, and code are openly available.

Analysis steps are explained and data structure is well described.

Qualitative coding steps include inter-rater review and consensus process.

Raw data is shared via the University of Bristol repository.

Files include survey results, R code, and metadata readme.

Repository uses Creative Commons license ensuring open data access.

Documentation enables re-use and verification by external researchers.

No file or description appears to be missing or unclear.

This article is suitable for publication after minor edits.

Language in introduction and discussion could be slightly condensed.

Consider clarifying broad terms like “too complicated” in future work.

The study is robust, timely, and methodologically well executed.

It contributes valuable evidence for improving data sharing culture.

I recommend acceptance with very minor stylistic improvements.

This is a useful model for open research infrastructure surveys.

Is the work clearly and accurately presented and does it cite the current literature?

Yes

If applicable, is the statistical analysis and its interpretation appropriate?

Not applicable

Are all the source data underlying the results available to ensure full reproducibility?

Yes

Is the study design appropriate and is the work technically sound?

Yes

Are the conclusions drawn adequately supported by the results?

Yes

Are sufficient details of methods and analysis provided to allow replication by others?

Yes

Reviewer Expertise:

Open Science

I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard.