Using matching and regression analyses, we measure the difference in citations between articles posted to Academia.edu and other articles from similar journals, controlling for field, impact factor, and other variables. Based on a sample size of 31,216 papers, we find that a paper in a median impact factor journal uploaded to Academia.edu receives 16% more citations after one year than a similar article not available online, 51% more citations after three years, and 69% after five years. We also found that articles also posted to Academia.edu had 58% more citations than articles only posted to other online venues, such as personal and departmental home pages, after five years.
a paper in a median impact factor journal uploaded to Academia.edu receives 16% more citations after one year than a similar article not available online, 51% more citations after three years, and 69% after five years
modeled citation differences are consistent, but observational matching leaves author- and article-level confounding unresolved
articles also posted to Academia.edu had 58% more citations than articles only posted to other online venues, such as personal and departmental home pages, after five years
the observational comparison is limited and the reported five-year magnitude conflicts with the Conclusions
Derived from the full evaluation — not a separate score.
Strengths
The study compares Academia.edu posting with both articles unavailable online and articles available through other online venues, allowing a more specific platform-level question than a generic open-access comparison.
↳ Introduction; Results, Predicted Citation Advantages, Tables 13–14
The P=U restriction limits ex post uploading of already successful articles, while conservative MTurk classification removes non-research content and matching accounts for publication year and journal impact factor.
↳ Materials and Methods, On-Academia and Off-Academia Articles; Document Types; Results, Matching by Impact Factor
The Background addresses early-view and selection-bias postulates as well as randomized and instrumental-variable studies that challenge a causal open-access advantage.
↳ Background, Critiques of the Citation Advantage
Limitations
The paper identifies author reputation, promotional behavior, and selective uploading of promising articles as possible confounders, but the models contain no direct author- or article-quality controls or sensitivity bounds.
↳ Background, Sources of Selection Bias; Results, Regression Analysis—Covariates
The Abstract and predicted-results discussion report a 58% five-year advantage over articles available only elsewhere online, whereas the Conclusions report 49% for the corresponding comparison.
↳ Abstract; Results, Predicted Citation Advantages; Conclusions
The platform comparison does not isolate feeds, notifications, or another specific discovery feature; the Discussion presents these as hypotheses requiring further study.
↳ Discussion, Issues and Topics for Further Research
The study’s strongest feature is its large platform-specific comparison, supported by impact-factor matching, the P=U temporal restriction, online-availability controls, and broadly consistent linear and count-model results. Those methods support an association but do not resolve the author- and article-quality selection mechanisms explicitly described in the Background. The literature review is balanced, although the implications of residual confounding are not fully carried into the certainty of the headline percentages. The conflicting 58% and 49% figures further reduce confidence in the exact magnitude of the secondary comparison.
Nabu’s assessment, alongside the field’s view.
Are you an author of this paper?
Sound3.4
Confidence mediumThe study meaningfully extends open-access citation research by comparing a platform combining access and discovery features with both unavailable and other-online articles. The advance remains bounded because the analysis does not directly isolate feeds, notifications, or another discoverability mechanism.
“Further work could be done to measure the effect of these distributional properties.”
The score reflects impact-factor matching, a publication-year upload restriction, field and availability controls, and convergence across three model families. Author reputation, promotional behavior, and article-quality selection remain unmeasured and are not bounded through sensitivity analysis.
“Many of these factors cannot be observed directly or completely.”
The progression from literature and bias mechanisms through cohort construction, modeling, and predicted contrasts is readily followable. Precision is reduced by the unreconciled 58% estimate in the Abstract and Results versus 49% in the Conclusions, alongside occasional use of causal “effect” terminology.
“a typical article that is also posted to Academia.edu has 49% more citations”
The Background engages both supportive open-access studies and randomized, quasi-experimental, and selection-bias critiques. Residual author and article-quality selection is acknowledged, but its implications for the certainty and magnitude of the headline estimates are not fully traced in the Discussion.
“the true effect of open access on citations remains debated in the literature”
Caveats4 of 4 checks
The direction of the central association is consistent, but one secondary headline magnitude differs across sections. The discrepancy affects the exact estimate rather than reversing the result.
The paper states that data and code are available at a named GitHub repository, but it contains no conflict-of-interest disclosure despite using platform data and describing motivation from reports to the platform team.
Flags: 2 declared / 5 total
14 of 14 checkable references verified
19 references in manuscript 5 are books, websites or datasets — counted, but not index-checkable
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 evidence comes from real publication and citation behavior, but it remains an observational association rather than a tested platform intervention. The paper offers no implementation protocol and states that further work is needed to identify the responsible features.
“Further work could be done to measure the effect of these distributional properties.”
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