How research funding weakens the OA citation premium: Evidence from Chinese LIS publications

This study investigates whether and how the research funding of a focal paper moderates the citation impact of Open Access (OA) for the publications in Chinese Library and Information Science (LIS), within an institutional context where OA is not mandatorily linked to research funding. Using a dataset of 30,386 Chinese LIS journal articles published between 2018 and 2022 and indexed in the China National Knowledge Infrastructure (CNKI), we alleviate selection bias via propensity score matching and estimate negative binomial regression models to assess the main effects of OA and funding as well as their interaction on citation counts. We further conduct heterogeneity analyses across stratified journal impact factor (JIF) levels, funding hierarchy, subfields within Library and Information Science, and geographic regions of the first author’s affiliated institution. Results indicate that (1) under the assumption of variable exogeneity, a robust open access citation advantage (OACA) is confirmed in the LIS field; (2) research funding exerts a significantly negative moderating effect on the citation effect of OA at the overall level; (3) this negative moderating effect is observed across most dimensions, particularly in medium-impact journals, interdisciplinary fields, national-level and other funding tiers, and North China. In contrast, a significantly positive moderating effect is observed in Southwest China. The effectiveness of OA as a citation-enhancing strategy is context-dependent. OA policies and interventions may yield larger marginal benefits in settings with limited funding support or weaker disciplinary and regional research infrastructures. This study advances OA scholarship by explicitly modeling funding as a moderator of OACA in a social-science disciplinary context where funding and OA are not institutionally coupled. It provides large-scale evidence from China and demonstrates that the OA-citation relationship varies systematically across resource environments, offering nuanced implications for open science policy design and research evaluation.