Evaluating institutional open access performance: Methodology, challenges and assessment

The role of open access has rapidly become central to the global research community (and beyond).
This is increasingly so with organisations and governments mandating serious considerations for
publishing in open access forms (e.g., Plan S). However, tracking the progress of open access at the
institutional level remains a difficult problem, with most bibliographic databases lacking exhaustive
open access and affiliation labelling, using non-standardised metadata formats, and having potentially
significant differences in coverage. This necessitates methodologies in integrating diverse
data sources to provide more complete evidence for open access performance at the institutional
level. In this study, we build the first comprehensive and reproducible data workflow in combining
digital object identifiers (DOIs) affiliated to individual institutions from Microsoft Academic, Web
of Science, and Scopus. Subsequently, we normalise publication dates using Crossref metadata and
query each DOI’s OA status through Unpaywall. We use this methodology to produce various open
access scores for the Top 1000 universities in the Times Higher Education World University Rankings
2019, and supplement this list with additional universities from countries with a low presence.
Analyses of the resulting data highlights the existence of different open access paths that universities
take on and regional differences across the globe due to varying policies and infrastructure. We
also present the top 100 performing universities in each of the categories of total open access, open
access (gold) publishing and repository-mediated (green) open access percentages. The presence of
African, Asian and Latin American universities in the top 100 shows encouraging progress in those
regions.