Can I Haz Data? Analyzing DAS Gap in Journals Publishing Papers Funded by NWO and ZonMw

Summary
Research shows that there is often a mismatch between Data Availability Statements (DAS) and actual data availability. Here, we estimate how frequent DAS mismatch occurs based existing, automatically scraped data about DAS and actual data availability from 8,826 Dutch-funded articles published across a wide variety of journals. We found that 45% of articles have a DAS, from which a large majority states that data are available to some extent. Of articles that promise data, 45% of the articles contained a link to a dataset in a repositories. We present DAS mismatch for a subset of journals and publishers separately. The DAS mismatch estimates contain uncertainty, but hint at a non-trivial degree of DAS mismatch in the literature. We discuss limitations and provide recommendations for journals, authors, and data reusers.