A comprehensive evaluation of life sciences data resources reveals significant accessibility barriers

Individuals with disabilities participate notably less in the scientific workforce. While the reasons for this discrepancy are multifaceted, accessibility of knowledge is likely a factor. In the life sciences, digital resources play an important role in gaining new knowledge and conducting data-driven research. However, there is little data on how accessible essential life sciences resources are for people with disabilities. Our work is the first to comprehensively evaluate the accessibility of life sciences resources. To understand the current state of accessibility of digital data resources in the life sciences, we pose three research questions: (1) What are the most common accessibility issues?; (2) What factors may have contributed to the current state of accessibility?; and (3) What is the potential impact of accessibility issues in real-world use cases? To answer these questions, we collected large-scale accessibility data about two essential resources: data portals (n = 3,112) and journal websites (n = 5,099). Our analysis shows that many life sciences resources contain severe accessibility issues (74.8% of data portals and 69.1% of journal websites) and are significantly less accessible than US government websites, which we used as a baseline. Focusing on visual impairment, we further conducted a preliminary study to evaluate three data portals in-depth with a blind user, unveiling the practical impact of the identified accessibility issues on common tasks (53.3% success rate), such as data discovery tasks. Based on our results, we find that simply implementing accessibility standards does not guarantee real-world accessibility of life sciences data resources. We believe that our data and analysis results bring insights into how the scientific community can address critical accessibility barriers and increase awareness of accessibility, leading to more inclusive life sciences research and education. Our analysis results are publicly available at http://inscidar.org/ .