A quantitative account of how limited information sharing undermines comprehensive literature syntheses

Crucial aspects of reproducible, replicable and reusable science include the responsiveness of study authors for clarifications and the availability of research data and analysis results. Getting in contact with authors and obtaining information or results is, however, not always straight-forward. Here we report and discuss the issues and obstacles we faced when contacting authors of scientific papers with such requests. Our investigation rests on the results of a retrospective quantitative analysis of research data requests sent to authors of neuroimaging studies for a series of meta-analyses. Overall, only 52% of the requests received a reply, and only 29% contributed data or information that was relevant for the respective meta-analysis. Obtaining a response was less likely if (i) the request was sent to the contact e-mail address provided in the publication, (ii) behavioral data was requested, (iii) reminders had to be sent, or (iv) there was no personal acquaintance with the contacted author. As expected, obtaining unpublished data or information from older publications was significantly more difficult than for more recent ones. We discuss possible reasons for the observed low response rates and limited sharing of information and conclude our account by providing suggestions to improve open-science practices and by pointing to a need for change in the academic system to foster better research data management for transparency and efficient reuse of results.