TL;DR
Open Access (OA) isn't just a boon for researchers—it’s the engine driving public scientific literacy. This study reveals that OA articles are 64.7% more likely to be cited on Wikipedia than paywalled research, with the effect being most pronounced for high-impact, recently published papers.
What Problem Does This Paper Address?
Wikipedia is the world’s most accessible repository of knowledge, yet its reliability depends on citations to primary scientific literature. While the "Open Access Citation Advantage" (OACA) is well-documented within academia (researchers citing other researchers), its impact on the public dissemination of science via platforms like Wikipedia remained poorly understood.
Previous studies were often limited by small sample sizes or focused on journal-level data rather than individual articles. This paper addresses the gap by asking: Does making an article free to read actually increase the likelihood that a Wikipedia editor will find, read, and use it as a source?
Methods
The authors employed a robust, article-level regression analysis to isolate the "OA effect" from other confounding variables. Key steps included:
- Data Enrichment: They merged Wikipedia citation data with metadata from OpenAlex (for OA status and academic impact) and Scimago (for journal rankings).
- Stratified Sampling: To create a fair comparison, they built a balanced dataset of ~522,000 articles. For every article cited on Wikipedia, they randomly sampled a "negative" (non-cited) article from the same field, publication year, and journal.
- Logistic Regression: They used two models to calculate the odds ratio of an article being cited. Model 1 looked at the total effect of OA, while Model 2 controlled for academic citation counts to see if OA has a "direct" influence on Wikipedia editors beyond just being a "famous" paper in its field.
Datasets
The study utilized several massive, open-science datasets:
- Wikipedia Citations: An updated version of the dataset containing over 45 million citations extracted from the February 2024 English Wikipedia dump.
- OpenAlex: A free platform providing data on over 240 million academic works, used here to identify OA status (Gold, Green, Hybrid, Bronze, or Closed) and "concepts" (fields of study).
- Scimago: Used to obtain the SJR (SCImago Journal Rank) and H-index for over 40,000 journals.
Key Insights and Ideas
- The "Bronze" Dominance: The most common type of OA cited on Wikipedia is "Bronze" (articles free to read on a publisher's site but without a formal open license). This suggests that simple "readability" is the primary driver for editors, rather than legal re-use rights.
- The Recency Preference: Wikipedia editors have a strong bias toward "novel results." The OA advantage is significantly stronger for articles published in the last four years.
- The Impact Multiplier: While high academic citation counts (impact) naturally increase the odds of a Wikipedia mention, being OA acts as a force multiplier. The "gap" in citation probability between OA and closed articles actually widens as the paper's academic impact increases.
- Disciplinary Divergence: STEM fields (Biology, Physics, Medicine) show the highest proportions of OA citations. Conversely, the Humanities (Art, History) have lower OA rates but show a high reliance on the few OA sources available.
Key Results
- The 64.7% Boost: After controlling for field, year, and journal rank, OA articles have a 64.7% higher likelihood of being cited on Wikipedia compared to closed-access articles.
- Crossing the Rubicon: For articles published after 2015, the percentage of Wikipedia citations that are Open Access has surpassed 50%.
- The "Long Tail" of Journals: Interestingly, nearly 90% of journals cited on Wikipedia have a relatively low SJR score (less than 10). This suggests Wikipedia editors don't just stick to Nature or Science; they explore a vast "long tail" of literature, provided they can get past the paywall.
Limitations and Future Work
- Scope: The study focused exclusively on the English Wikipedia. Citation patterns may differ significantly in other languages or regional contexts.
- Format Bias: The researchers excluded conference papers and books, which are critical in fields like Computer Science and the Social Sciences.
- Causal Timing: The current model doesn't account for when a citation was added. Future work could analyze Wikipedia’s edit history to see if an article becoming OA (e.g., via a Green OA repository) triggers a subsequent spike in Wikipedia citations.
- Readability: The study doesn't measure the "lay-friendliness" of the prose, which likely influences whether a volunteer editor chooses to cite a specific paper.
Summary updated