AIMS: Over the last two decades, the existence of an open access citation advantage (OACA)-increased citation of articles made available open access (OA)-has been the topic of much discussion. While there has been substantial research to address this question, findings have been contradictory and inconclusive. We conducted a systematic review to compare studies of citations to OA and non-OA articles.
METHODS: A systematic search of 17 databases attempted to capture all relevant studies authored since 2001. The protocol was registered in Open Science Framework. We included studies with a direct comparison between OA and non-OA items and reported article-level citation as an outcome. Both randomized and non-randomized studies were included. No limitations were placed on study design, language, or publication type.
RESULTS: A total of 5,744 items were retrieved. Ultimately, 134 items were identified for inclusion. 64 studies (47.8%) confirmed the existence of OACA, while 37 (27.6%) found that it did not exist, 32 (23.9%) found OACA only in subsets of their sample, and 1 study (0.8%) was inconclusive. Studies with a focus on multiple disciplines were significantly positively associated with finding that OACA exists in subsets, and are less associated with finding that OACA did not exist. In the critical appraisal of the included studies, 3 were found to have an overall low risk of bias. Of these, one found that an OACA existed, one found that it did not, and one found that an OACA occurred in subsets.
CONCLUSIONS: As seen through the large number of studies identified for this review, OACA is a topic of continuing interest. Quality and heterogeneity of the component studies pose challenges for generalization. The results suggest the need for reporting guidelines for bibliometrics studies.
64 studies (47.8%) confirmed the existence of OACA, while 37 (27.6%) found that it did not exist, 32 (23.9%) found OACA only in subsets of their sample, and 1 study (0.8%) was inconclusive.
direct descriptive tally of all 134 included studies, consistently reported across the paper
Studies with a focus on multiple disciplines were significantly positively associated with finding that OACA exists in subsets, and are less associated with finding that OACA did not exist.
reported chi-square supports association, though the moderator analysis is exploratory
Of the three studies that had low overall risk of bias, one found that an OACA existed, one found that it did not, and one found that an OACA occurred in subsets.
direct descriptive result from the independently conducted risk-of-bias appraisal
Quality and heterogeneity of the component studies pose challenges for generalization.
I²=90% and pervasive high risk of bias directly support the qualified conclusion
Derived from the full evaluation — not a separate score.
Strengths
The Methods describe searches across 17 multidisciplinary databases, reference-list handsearching, translation of non-English records, and an OSF-registered protocol.
↳ Methods, search strategy and protocol paragraphs
Only three of 134 studies were rated overall low risk of bias, and that result is carried into the explicit refusal to draw definitive publication recommendations.
↳ Results, Risk of Bias; Discussion
The Introduction engages confounding, selection bias, early-view effects, and disciplinary variation from earlier reviews, while the Discussion connects these issues to the observed evidence limitations.
↳ Introduction, prior-review discussion; Discussion
Limitations
Nearly 98% of included studies were judged high risk of bias, with only two randomized studies among 134. This confines the contribution to diagnosing the literature rather than settling the substantive question.
↳ Results, Risk of Bias; Discussion
Risk of bias and heterogeneity are integrated narratively, but the review does not provide a formal body-of-evidence certainty assessment comparable to GRADE.
↳ Methods; Results, Risk of Bias; Discussion
The paper explicitly cannot recommend whether authors should choose OA to increase citations. Its recommendations instead require further methodological development and stronger primary evidence.
↳ Discussion; Limitations and Future Work
The review's principal contribution is its systematic characterization and appraisal of 134 OACA studies, not a directional estimate of the citation advantage. Its 17-database search, dual independent procedures, and decision not to pool at I²=90% support a strong methods assessment. Scores remain below the highest band because the body of evidence lacks a formal certainty rating, the moderator analyses are exploratory, and the underlying literature is overwhelmingly high risk of bias. The clearest practical value lies in the proposed reporting guidelines and future living review rather than immediate advice to authors choosing publication venues.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Strong4.1
Confidence highThe review meaningfully extends prior narrative work by systematically characterizing 134 comparative studies and formally assessing their risk of bias. Its advance concerns the condition of the evidence base rather than a definitive answer about OACA.
Quality and heterogeneity of the component studies pose challenges for generalization.
The methods combine a registered 17-database search with dual screening, extraction, and risk-of-bias assessment, and the authors appropriately declined pooling at I²=90%. The absence of a formal body-of-evidence certainty framework and the exploratory chi-square analyses limit a higher score.
An I2 of 90% was found among studies with sufficient data for pooling.
The argument proceeds coherently from a contested question to mixed findings and a deliberately non-definitive conclusion. Dense percentage reporting in Results and the extended Discussion modestly reduce accessibility without changing the paper's meaning.
it is not possible to draw definitive conclusions and recommendations
The paper engages prior reviews addressing confounding, selection bias, early-view effects, and disciplinary variation, then carries those concerns into its interpretation. Its limitations are linked to reduced generalizability rather than acknowledged and ignored.
Non-English language content and publishers from outside of North America, the UK, and western Europe are underrepresented
Caveats4 of 4 checks
Retrieval, screening, outcome-frequency, and risk-of-bias counts reconcile across the Abstract and Results. The reported discipline chi-square is plausible but close to the stated significance threshold and was not independently recomputed.
The review reports OSF protocol registration, dual independent procedures, and an itemized study dataset, with no text-visible conduct inconsistency. Ethics approval is not applicable to this review of published records.
Flags: 2 declared / 5 total
138 of 138 checkable references verified
171 references in manuscript 33 have no canonical index record — counted, but not index-checkable citation details diverge from the cited records for 5 references 16 references confirmed by manual review
References not verified automatically are resolved manually within 24 hours.
No retraction or expression of concern found in Crossref (includes Retraction Watch).
Sources: Crossref (incl. Retraction Watch) ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
The synthesis is not ready to guide OA publication choices because nearly all included studies were high risk of bias and pooling was inappropriate. Its more immediate use is improving research and reporting practice.
We found high risk of bias in nearly 98% of the included studies
AI-generated, human-governed. Something look off? Contact us to request a review.
Keep reading
Other papers Nabu has evaluated blind, starting with the same field.
Baptiste Scancar, Jennifer Anne Byrne, David Causeur, et al. · 2026 · Media and communications
These findings suggest that paper mills inflate the citation metrics of supported publications and affected journals at scale.
TentativeChao Ma, Yiwei Li, Feng Guo, et al. · 2019 · Economics
papers published between October and December on average get as much as 18.5% fewer citations than those published in the other months in the year
TentativeMasashi Kawai, Koichi Ojiro, Makoto Asaoka, et al. · 2022 · Other social sciences
日本の機関リポジトリによるオープンアクセス論文の数は限られるものの，その被引用数と様々なタイプのオープンアクセス論文の被引用数の比較からは，日本の機関リポジトリによるオープンアクセスのメリットは確認できなかった．
Tentative