PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
November 18, 2025F1000Research0 citationsOpen Access

Survey about Barriers and Solutions for Enhancing Computational Reproducibility in Scientific Research

View Full Paper
YGYuri Andrei GelsleichterRBRita BanziFNFlorian Naudet

Key Points

  • Survey highlights the crucial role of institutional support in enhancing data sharing and reproducibility.
  • A significant 70% of respondents noted missing open data and 71% reported unavailable open code during replication attempts.
  • Training and proper resources are essential for researchers to tackle challenges in reproducibility and data sharing.
  • Addressing barriers like lack of time and insufficient funding is vital for improving computational reproducibility.

Abstract

Background Rapid development and adoption of digital technologies across all research disciplines underlines the need for accessible and reusable computational data and code. Methods An anonymous, multidisciplinary survey covering open science, data publishing and reuse, as well as code publishing and reuse was conducted to gather insights into researchers’ practices, needs, and barriers. Results A total of 254 people initiated the survey, with 133 complete responses (mostly from Europe, equally distributed among scientific fields). Survey revealed that registered reports, replication studies and pre-registration are the least applied practices (52%, 38% and 42%), while open software and OA publishing demonstrated widespread adoption (83% and 69%) of the respondents, respectively. Data sharing is hindered mostly by lack of time (60%) and sufficient funding (44%). Among the predefined obstacles of code sharing, again, the lack of time to build proper documentation (65%), pressure to publish (51%), and the insufficient funding (42%) are the most mentioned reasons. On the other hand, most stimulating factors are the requirement of journals to share data/codes (score: 482), followed by incentives and rewards by institutions (score: 439). The survey showed that 28% of researchers 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 missing or incomplete. The analysis of open-ended responses highlighted the need for training, career-stage guidelines, and basic programming skills for researchers. Conclusions Although the likely low response rate may limit its representativeness, this study provides an up-to-date snapshot. A recurrent theme throughout the responses is the need for structural incentives and institutional support. Researchers claim that making work reproducible requires time, resources, and expertise; however, these efforts are rarely rewarded in conventional academic evaluation systems, highlighting the need for a systemic cultural shift.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Gelsleichter et al. (2025) studied this question.

synapsesocial.com/papers/6924fee6c0ce034ddc3517c5https://doi.org/10.12688/f1000research.172013.1
Ask AI
Helpful
Bookmark
Share
View Full Paper