Crunching the numbers. Responsibly.
there is no generalizable gold OA citation advantage, neither at article nor at journal level
broad unadjusted category comparisons support the descriptive pattern but cannot identify the effect of OA itself
the percentage of OA articles has increased in the time period 2009-2014 in all three indexes
direct 2009 and 2014 prevalence measurements consistently increase across SCIE, SSCI, and AHCI
Although there are scientific disciplines where the average impact of an article is higher in the case of OA (36 categories), in most disciplines the opposite happens (173 categories)
Table 4 directly reports the category counts, though category aggregation limits interpretation of underlying article-level effects
Derived from the full evaluation — not a separate score.
Strengths
The analysis includes 1,137,634 articles from the 2009 Web of Science cohort and reports results across 249 subject categories, substantially broadening the evidence base beyond individual fields.
↳ Abstract; Methodology; Table 1
The paper defines fully gold OA, fixes a six-year citation window, provides the OACA formula, and illustrates its interpretation with a worked example.
↳ Methodology, pp. 3–4; Table 3
The Introduction discusses studies both supporting and questioning the OA citation advantage and explicitly distinguishes open-access, early-view, and selection-bias explanations.
↳ Introduction, pp. 2–3
Limitations
The article compares unadjusted OA and non-OA averages without controlling for journal prestige, article quality, author characteristics, or selection into fully OA journals, although selection bias is identified as necessary to consider.
↳ Introduction, p. 3; Methodology, p. 4
Results are aggregated at subject-category level, and the Conclusions acknowledge that opposing journal- or article-level effects may create misleading similarity without quantifying that risk.
↳ Conclusions, final paragraph
The paper identifies no specific policy process, institutional decision, or implementation pathway, and its descriptive design cannot isolate the effect of OA itself.
↳ Methodology; Conclusions
The paper's principal value lies in its scale: Table 1 covers all 249 Web of Science subject categories, while Figure 4 adds a journal-level comparison. The descriptive calculations are transparent, and the conclusion that a uniform advantage is not visible in these aggregates follows from Tables 4 and 5. Confidence in interpreting this as an OA effect is substantially lower because the Methodology provides no adjustment for selection bias or journal visibility, despite the Introduction identifying these mechanisms as central. The literature framing and exposition are stronger than the causal identification, placing the paper near the middle of the quality scale overall.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.0
Confidence mediumThe paper adds a broad descriptive comparison across 249 Web of Science categories and examines both article citations and journal impact factors. Its advance is empirical coverage rather than resolution of why OA and non-OA outcomes differ.
this paper is the first multidisciplinary study that includes all scientific disciplines
Gold OA, the citation window, and the OACA calculation are transparently defined, but comparisons rely on unadjusted category averages. Selection bias and journal visibility are acknowledged without matching, adjustment, or sensitivity analysis, binding the observational-methods ceiling.
any analysis must take into account the effect of time and selection bias
The paper follows a linear progression from the disputed postulate through data construction, article-level results, journal-level results, and conclusions. The OACA formula and worked example aid interpretation, while the main conclusion is expressly framed as what cannot be concluded.
it cannot be concluded that gold OA has increased its impact
The Introduction covers both supportive and sceptical OA-citation literature and identifies early-view and selection-bias explanations. The Conclusions disclose journal visibility, APC, and aggregation limitations, but do not quantify how they alter the reported pattern.
there may be advantageous and no-advantageous journals/articles which may counterpart their effects
Caveats4 of 4 checks
The principal totals and reported correlations are internally coherent, but one large unexplained category discontinuity reduces confidence in comparisons of subject-category prevalence over time. It does not directly affect the 2009 citation comparison underlying the main descriptive result.
The bibliometric design does not require human- or animal-subject approval, funding is disclosed, and no visible conflict, registration, or availability claim creates a conduct concern.
Flags: 0 declared / 5 total
20 of 20 checkable references verified
21 references in manuscript 1 have no canonical index record — counted, but not index-checkable 1 reference confirmed by manual review
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 evidence comes from operational publication and citation records, but the analysis is retrospective and descriptive rather than an evaluated policy or intervention. Substantial causal and context-specific work would be needed before acting on it.
AI-generated, human-governed. Something look off? Contact us to request a review.