Abstract
This paper examines research output quality evaluation in the 2021 national research evaluation exercise (REF 2021), which reportedly have not relied on algorithmic rankings and, instead, used peer-reviews for all individual research outputs. In our study, we use Hole’s (2017) methodology for inferring a journal ranking from the REF outputs’ data for unit of assessment 17 (business and management studies). The analysis finds evidence of high correlation between REF results and CABS journal ranking. When we introduce institution-level heterogeneity, we show that journal rankings appear not to be both statistically and quantitatively different for DORA signers and non-signers. The implications of these findings are significant: they enhance our understanding of responsible use of metrics in research evaluations and pave the way to a more nuanced employment of DORA and DORA-like principles in future assessment exercises.