Introducing FAIR Scores in a Global Agricultural Science Reporting Service: An Analysis of the First Reporting Period

Abstract

Communicating the meaning and value of the FAIR (Findable, Accessible, Interoperable, and Reusable) principles and suitable implementation strategies to research communities and their broader ecosystem is a crucial challenge. Automated FAIR scoring algorithms are being developed to provide immediate machine-driven insight into (meta)data compliance. They are also expected to return instructions to improve FAIR compliance where applicable. However, we still lack a systematic understanding of how automated FAIR scoring impacts adoption in repositories that implement them. Accordingly, we are excited to share findings from the first large-scale machine-driven FAIR scoring of global agricultural research at One CGIAR, a global partnership of international organizations dedicated to sustainable food production. Characteristics that make this analysis particularly interesting include the organization-wide visibility of the FAIR scoring and the mandatory reporting of all scientific resources in the preceding calendar year. We find that 39% of the 418 records received updates. Out of those, 7.3% resulted in improved FAIR scores. Those improved records outperformed the mean FAIR scores of the complete 2022 reporting data, while falling behind them before the update. We further show differences according to the knowledge product type and outline how we gather additional data on users’ perceptions for the OR2023 presentation.