科研速览 · Science Skim继续刷下去 · Keep skimming →
◆ F1000Research2025-01-01

Exploratory Insights into Barriers and Open Practices for Computational Reproducibility in Scientific Research.

Yuri Andrei Gelsleichter, Rita Banzi, Florian Naudet, Constant Vinatier, István Kertész, Monika Varga

一句话结论 · In one sentence

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.

原始摘要(英文原文)· Original 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.
读原文 · Read the paper ↗

AI 追问PRO

登录后使用 AI 追问

讨论区

登录后参与讨论

相关论文 · Related

Exploratory Insights into Barriers and Open Practices for Computational Reproducibility in Scientific Research. — 科研速览 Science Skim