Exploratory Insights into Barriers and Open Practices for Computational Reproducibility in Scientific Research
Abstract
Background:Rapid adoption of digital technologies across research disciplines underlines the need for accessible and reusable computational data and code.Methods:An anonymous, multidisciplinary survey examined researchers’ perceptions, needs, barriers, and self-reported practices concerning open science, data and code publishing and reuse.Results:Of 254 respondents who initiated the survey, 133 completed it, mostly from Europe. Registered reports, replication studies and pre-registration were among the least frequently reported practices (52%, 38% and 42%, reported as Never applied), while open software and OA publishing demonstrated widespread adoption (83% and 69%) of the respondents, respectively. The main perceived barriers to data sharing were lack of time (60%) and insufficient funding (44%). For code sharing, they were lack of time to prepare documentation (65%), publication pressure (51%), and insufficient funding (42%). Journal requirements (score: 482) and institutional incentives and rewards (score: 439) were the highest-ranked supporting measures. 28% of respondents indicated that they never tried to reproduce a study, and when replication was attempted, researchers often found that open data (70%), open code (71%), and metadata (86%) were never, rarely, or only sometimes available in the publications they read.. Open-ended responses emphasized training, career-stage guidelines, and basic programming skills.Conclusions:As survey was disseminated through open-science channels using volunteer sampling, no response rate could be calculated, and the sample was self-selected toward researchers already engaged with reproducibility. Findings should therefore not be generalized to the wider research community. Within this group, the central finding is a gap between awareness and implementation. High endorsement of open and reproducible practices coexists with lower self-reported adoption. Responses also emphasized the need for structural incentives and institutional support, reflecting perceived limitations in time, resources, expertise, and professional recognition.
Related
- Ten simple rules for achieving computational reproducibility in cognitive neuroscience
- Clinical trial data transparency should match data access to scientific purpose: a tiered framework for reproducibility and reuse
- hpci improves reproducibility of computational biology research by integrating Continuous Integration with High Performance Computing clusters.
- Antimicrobial practices for Mycoplasma infections in cattle: A survey of veterinary practitioners in light of current scientific evidence and national guidelines
- Antibiotic prescribing practices of community health extension workers in East Central Uganda: an exploratory qualitative study
BibTeX
@article{gelsleichter2026exploratory,
title = {{Exploratory Insights into Barriers and Open Practices for Computational Reproducibility in Scientific Research}},
author = {Yuri Andrei Gelsleichter and Rita Banzi and Florian Naudet and Constant Vinatier and István Kertész and Monika Varga},
journal = {F1000Research},
year = {2026},
month = aug,
doi = {10.12688/f1000research.172013.3},
url = {https://hal.science/hal-05742638v1},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Exploratory Insights into Barriers and Open Practices for Computational Reproducibility in Scientific Research
AU - Yuri Andrei Gelsleichter
AU - Rita Banzi
AU - Florian Naudet
AU - Constant Vinatier
AU - István Kertész
AU - Monika Varga
PY - 2026
DA - 2026/08/11
JO - F1000Research
DO - 10.12688/f1000research.172013.3
UR - https://hal.science/hal-05742638v1
AB - Background:Rapid adoption of digital technologies across research disciplines underlines the need for accessible and reusable computational data and code.Methods:An anonymous, multidisciplinary survey examined researchers’ perceptions, needs, barriers, and self-reported practices concerning open science, data and code publishing and reuse.Results:Of 254 respondents who initiated the survey, 133 completed it, mostly from Europe. Registered reports, replication studies and pre-registration were among the least frequently reported practices (52%, 38% and 42%, reported as Never applied), while open software and OA publishing demonstrated widespread adoption (83% and 69%) of the respondents, respectively. The main perceived barriers to data sharing were lack of time (60%) and insufficient funding (44%). For code sharing, they were lack of time to prepare documentation (65%), publication pressure (51%), and insufficient funding (42%). Journal requirements (score: 482) and institutional incentives and rewards (score: 439) were the highest-ranked supporting measures. 28% of respondents indicated that they never tried to reproduce a study, and when replication was attempted, researchers often found that open data (70%), open code (71%), and metadata (86%) were never, rarely, or only sometimes available in the publications they read.. Open-ended responses emphasized training, career-stage guidelines, and basic programming skills.Conclusions:As survey was disseminated through open-science channels using volunteer sampling, no response rate could be calculated, and the sample was self-selected toward researchers already engaged with reproducibility. Findings should therefore not be generalized to the wider research community. Within this group, the central finding is a gap between awareness and implementation. High endorsement of open and reproducible practices coexists with lower self-reported adoption. Responses also emphasized the need for structural incentives and institutional support, reflecting perceived limitations in time, resources, expertise, and professional recognition.
SN - 2046-1402
ER -