Researchers' attitudes toward sharing code: Findings from a survey and policy implementation

Whilst many funders and publishers require or encourage researchers to share data underlying publications, the same cannot be said for code, which is also vital for reproducibility and transparency of research. We surveyed researchers in computational biology disciplines to understand if the community would support a mandatory code sharing policy, and their experiences with code sharing. Respondents (n=214) reported that, on average, 71% of their research articles have associated code, and that on average, code has not been shared for 32% of these articles. The most common reasons for not sharing code were practical issues (e.g. lack of time) and intellectual property concerns were cited by 22%. Overall, respondents reported they were on average more likely to submit to the journal if it had a mandatory code sharing policy, although Medicine and Health scientists and senior researchers viewed the proposed policy less favourably. A stronger code sharing policy has since been introduced at PLOS Computational Biology and this talk will discuss the learnings from the survey and the impact of the stronger code sharing policy implemented at the journal.