Crunching the numbers. Responsibly.
Abstract Research suggests an open-access citation advantage, in which open-access articles are associated with greater citations compared to articles behind paywalls. In this paper, we examined whether publicly accessible articles are associated with increased citations in the field of special education. We reviewed 1678 articles in 43 special education journals published in 2022 and identified (i) whether and how each paper was publicly accessible and (ii) the number of citations in Web of Science Core Collection (WoS) and Google Scholar (GS) for each article. Results of a series of multi-level linear models indicated publicly accessible articles were associated with 26% (β = 0.26, p < 0.001) and 41% (β = 0.41, p < 0.001) more citations than closed articles in WoS and GS, respectively, after controlling for other variables potentially associated with citations (i.e. funding support, reporting an empirical study, reporting a review or meta-analysis, part of a special issue, number of authors, publication date and journal impact factor). A significant positive association was found across individual accessibility categories (e.g. gold OA, green OA, bronze). Limitations and implications are discussed.
publicly accessible articles were associated with 26% (β = 0.26, p < 0.001) and 41% (β = 0.41, p < 0.001) more citations than closed articles in WoS and GS, respectively, after controlling for other variables
observational models support the association, though access timing and printed coefficient labels remain unresolved
all four types of public accessibility were associated with higher citations in both WoS and GS. Gold OA was associated with the largest percentage change in both WoS (35%) and GS (46%)
cross-sectional category estimates are positive, though access timing and one WoS significance display remain uncertain
Spearman’s correlation (ρ = 0.83, p < 0.001) and Pearson’s correlation (r = 0.84, 95% CI [0.82, 0.85], p < 0.001) indicated a strong, positive relation between WoS and GS citations
two directly reported correlation estimates consistently support a strong positive database relationship
Derived from the full evaluation — not a separate score.
Strengths
The analysis addresses a documented gap by estimating accessibility–citation associations across 1,678 articles from 43 special-education journals and distinguishes four accessibility categories.
↳ Introduction §§1.3–1.4; Methods §2.1; Results §3.1
The paper justifies article-within-journal modelling using reported ICCs, compares OLS with Poisson and negative-binomial alternatives, and describes residual diagnostics and model selection.
↳ Methods §2.3; Results §3.1
The discussion identifies quality, topical relevance, prestige, and selection into accessibility as alternative explanations and repeatedly frames the findings as associations rather than causal effects.
↳ Discussion §4; Limitations §4.2
Limitations
Accessibility was assessed in Q2 2024, while several access forms can emerge after publication and after citations begin accumulating. The analysis does not establish when exposure preceded citation accrual for those articles.
↳ Methods §§2.1–2.2; Limitations §4.2
The Abstract and Results print β values of 0.26 and 0.41, whereas Table 1 reports 0.23 and 0.34; the stated percentages align with the table after back-transformation.
↳ Abstract; Results §3.1; Table 1
Article quality, topic, author prestige, and authors’ choices about making papers accessible may jointly affect accessibility and citations, but no sensitivity or quantitative bias analysis bounds this uncertainty.
↳ Discussion §4; Limitations §4.2
The paper makes a clear but incremental contribution by extending the open-access citation literature to a large, defined sample of special-education journals. Its multilevel modelling, diagnostic comparisons, coding checks, and associative language support the reported direction of the findings. Methodological Rigour is constrained by the fact that accessibility status was observed after publication and may postdate part of the citation accumulation period, particularly for green, bronze, and shared access. The headline percentages remain interpretable from Table 1, but the accompanying β labels and one Table 2 significance presentation should be corrected.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.3
Confidence highThe study provides a clearly delineated field-specific extension of open-access citation research using 1,678 articles, two citation databases, and separate accessibility categories. Its advance is meaningful but incremental because it applies an established question to one discipline and publication year.
“Given this variability, it is not clear whether and how previous findings...generalize to special education.”
Multilevel models, clustering diagnostics, model comparisons, covariate adjustment, and coding reliability support the analysis, but access status was assessed alongside 2024 citation counts and could have arisen after some citations. Selection into accessibility is also acknowledged without sensitivity analysis, leaving meaningful uncertainty around the association.
“the timing of public accessibility may affect citation trends”
Research questions, analyses, and associative conclusions are easy to follow, and the paper explicitly disclaims causal inference. Precision is reduced by the mismatch between the β labels in the Abstract and Results and the estimates displayed in Table 1.
“we did not design the study to permit causal inferences”
The paper engages mixed prior evidence, compares its estimates with earlier work, and traces sampling, confounding, timing, and database limitations to interpretation. The contextual treatment is thorough, though it does not quantitatively bound the selection and timing mechanisms it identifies.
“this relationship should be interpreted with caution”
Lower confidence on Methodological Rigour — domain match limited.
Caveats4 of 4 checks
The direction and percentage magnitude of the headline association are supported by Table 1, but the β labels printed in the Abstract and Results do not match the table. A secondary significance inconsistency affects the shared-journal WoS estimate.
Required declarations are present, and the paper identifies an OSF project containing materials, data, code, and preregistration. The preregistration's completion after data collection but before cleaning and analysis is explicitly disclosed.
Flags: 5 declared / 5 total
38 of 38 checkable references verified
43 references in manuscript 5 are books, websites or datasets — counted, but not index-checkable
No retraction notice found in Retraction Watch.
Sources: Retraction Watch ✓
Where this paper’s evidence sits on the path from initial observation to real-world use.
The analysis uses real publication and citation records, but it characterises an association rather than testing an accessibility intervention. It can inform choices at the margin but is not sufficient for a causal publishing-policy recommendation.
“we cannot infer that public accessibility caused an increase in article citations”
AI-generated, human-governed. Something look off? Contact us to request a review.