Does Open Access Matter for Citations? A Topic-Stratified Analysis of Library and Information Science Discipline

Open access publishing aims to make research accessible to a wide range of readers, with broader accessibility offering a potential pathway to increased citation exposure. Citation impact therefore remains an important justification for OA policies and motivates the question examined in this study. Using BERTopic, 52 research topics were derived from 14,720 Scopus-indexed Library and Information Science (LIS) publications published between 2010 and 2024. This study examined the OA-citation association overall, within individual research topics, and across progressively controlled regression models. OA publications showed a modest overall citation association (r = .141, p < .001). This association was not uniform across topics: among the 50 topics included in the topic-level analysis, 40 (80%) showed no detectable OA association, while 10 topics, concentrated in health-sciences-adjacent LIS subfields, accounted for most of the aggregate association. In the fully specified regression model, the journal-level citation baseline was the strongest predictor of citation rate (β = .235, p < .001), exceeding the coefficients for international collaboration, document type, and author count. Its inclusion reduced the OA coefficient by more than half, from β = .114 to β = .054 (both p < .001). There was no statistically detectable temporal change in the OA-citation association from 2010 to 2022. Findings indicate that the citation association of OA is topic-dependent and conditional on the journal-level citation environment. This study uses an observational design and therefore estimates statistical associations rather than causal effects.