Retracted scientific papers cited on Wikipedia tend to linger on the popular website for years, according to a study examining nearly 1,200 citations.
The study’s authors, led by Ph.D. candidate Haohan Shi from the Media, Technology, and Society Program at Northwestern University, used the Retraction Watch Database to compile a list of retracted papers and cross-referenced that list with Wikipedia citations. Of the 1,181 retracted citations identified, just over half were added to Wikipedia before the paper was retracted; a fifth were added after retraction but without any reader warning; and just over a quarter explicitly noted the retraction.
They also measured how long it took the Wikipedia community to correct citations added before the paper was retracted. The team found that while many corrections occurred swiftly, the median time for a correction was 3.68 years.
“While Wikipedia is often recognized for its agility in responding to breaking news and high-salience events, our findings reveal a much slower and uneven response to post-publication knowledge failures,” the authors wrote in a preprint posted last September on arXiv.org and accepted at the 2026 ACM Conference on Computer-supported Cooperative Work and Social Computing.
These problematic citations could “quietly misinform millions of people,” Shi told Retraction Watch.
Other studies have shown retractions don’t keep studies from getting cited in the scientific literature. Journalists who cover studies rarely cover the retractions, and retracted studies turn up in policy documents and in health-related reviews.
Given the size of Wikipedia — which contains over 66 million articles and receives about 508 million views per day — the presence of retracted citations isn’t surprising, said Laura Kurek, a Ph.D. candidate at the University of Michigan’s School of Information who has researched how the Wikipedia community maintains articles about the Russia-Ukraine war.
“Wikipedia eventually gets it right,” Kurek said, but noted the median correction time of nearly four years was substantial.
Many of the retracted papers the researchers found were cited in specialized Wikipedia pages, Shi said, but some did appear on general-interest pages. “Ordinary readers who read those pages without knowing that the paper is retracted are likely to just believe in the claim told by the paper,” Shi said.
The time it takes Wikipedia editors to remove or edit a citation to a retracted paper varies based on the academic field, the researchers found. Physical science citations are corrected the fastest, with a median correction time of 228 days. Corrections to life sciences and health sciences citations take significantly longer, with median correction times of 2,720 days and 1,846 days, respectively.
The study points to a “fundamental disconnect” between retraction data and the Wikipedia community’s workflow, the researchers wrote. Shi acknowledged, however, that expecting Wikipedia editors to continuously monitor every citation for new retractions is unrealistic.
Instead, he suggested Wikipedia editors use automated tools such as RetractionBot to accelerate the correction process. The bot downloads information from the Retraction Watch Database, crosschecks those studies against citations found on Wikipedia and flags any retracted citations. Browser extensions like LibKey Nomad will also flag links to retracted papers for users who have them installed.
Of the 1.6 million scientific papers cited on Wikipedia, the number of retracted articles is less than 0.1% – a datapoint worth noting, Kurek said. “People keep tending to [Wikipedia] like a garden, and over time things do get caught,” she said.
But Shi said the total number of cited scientific papers across the website shouldn’t be compared to the total number of retracted citations. “An overall rate of 0.034% doesn’t mean Wikipedia is successfully avoiding bad science — it simply reflects that retractions are incredibly rare in the scientific world at large.”
Like Retraction Watch? You can make a tax-deductible contribution to support our work, follow us on X or Bluesky, like us on Facebook, follow us on LinkedIn, add us to your RSS reader, or subscribe to our daily digest. If you find a retraction that’s not in our database, you can let us know here. For comments or feedback, email us at [email protected].