The number of retractions issued for research articles in 2023 has passed 10,000 — smashing annual records — as publishers struggle to clean up a slew of sham papers and peer-review fraud. Among large research-producing nations, Saudi Arabia, Pakistan, Russia and China have the highest retraction rates over the past two decades, a Nature analysis has found.

The bulk of 2023’s retractions were from journals owned by Hindawi, a London-based subsidiary of the publisher Wiley (see ‘A bumper year for retractions’). So far this year, Hindawi journals have pulled more than 8,000 articles, citing factors such as “concerns that the peer review process has been compromised” and “systematic manipulation of the publication and peer-review process”, after investigations prompted by internal editors and by research-integrity sleuths who raised questions about incoherent text and irrelevant references in thousands of papers.

A BUMPER YEAR FOR RETRACTIONS. Chart shows retraction notices in 2023 have passed 10,000.

 

Most of the Hindawi retractions are from special issues: collections of articles that are often overseen by guest editors and that have become notorious for being exploited by scammers to rapidly publish low-quality or sham papers.

On 6 December, Wiley announced on an earnings call that it would stop using the Hindawi brand name altogether, having previously shuttered four Hindawi titles and, towards the end of 2022, temporarily paused special-issue publication. Wiley will fold existing titles back into its own brand. As a result of the problems, said Wiley’s interim chief executive, Matthew Kissner, the publisher expects to lose out on between US$35 million and $40 million in revenue this fiscal year.

A Wiley spokesperson said that the publisher anticipated further retractions — they did not say how many — but that the company takes the view that “special issues continue to play a valuable role in serving the research community”. The spokesperson added that Wiley had put in place more rigorous processes to confirm the identity of guest editors and oversee manuscripts, removed ‘hundreds’ of bad actors — some of whom had held guest editor roles — from its systems, and scaled up its research-integrity team. It is also “pursuing legal means” to share data about the bad actors with other publishers and providers of tools and databases.

Hindawi’s retracted papers might have been mostly sham articles, but they were still collectively cited more than 35,000 times, says Guillaume Cabanac, a computer scientist at the University of Toulouse in France who tracks problems in papers, including ‘tortured phrases’ — strange wording choices used in efforts to evade plagiarism detectors — and signs of undisclosed use of artificial intelligence. “These problematic papers get cited,” he says.

RISING RETRACTION RATES. Graphic shows the ratio of retracted papers to articles published has risen to above 0.2%

 

Retractions are rising at a rate that outstrips the growth of scientific papers (see ‘Rising retraction rates’), and this year’s deluge means that the total number of retractions issued so far has passed 50,000. Although analyses have previously shown that the majority of retractions are due to misconduct, this is not always the case: some are led by authors who discover honest errors in their work.

The world’s largest database to track retractions, collated by the media organization Retraction Watch, does not yet include all of 2023’s withdrawn papers. To analyse trends, Nature combined the roughly 45,000 retractions detailed in that data set — which in September was acquired for public distribution by Crossref, a non-profit organization that indexes publishing data — with another 5,000 retractions from Hindawi and other publishers, with the aid of the Dimensions database.

Rising rates

Nature’s analysis suggests that the retraction rate — the proportion of papers published in any given year that go on to be retracted — has more than trebled in the past decade. In 2022, it exceeded 0.2%.

Among countries that have published more than 100,000 articles in the past two decades, Nature’s analysis suggests that Saudi Arabia has the highest retraction rate, of 30 per 10,000 articles, excluding retractions based on conference papers. (This analysis counts an article for a country if at least one co-author has an affiliation in that country.) If conference papers are included, withdrawals from the Institute of Electrical and Electronics Engineers (IEEE) in New York City put China in the lead, with a retraction rate above 30 per 10,000 articles (see ‘Countries with highest retraction rates’).

COUNTRIES WITH HIGHEST RETRACTION RATES. Chart shows the top 8 countries with the highest retraction rates over the past two decades.

 

The analysis shows that around one-quarter of the total number of retractions are conference papers — and the bulk of those comprise withdrawals by the IEEE, which has pulled more than 10,000 such papers in the past two decades. The IEEE was the publisher with the highest number of retractions. It does not record when it retracts papers, but most of those removed were published between 2010 and 2011.

Preventive measures

Monika Stickel, director of corporate communications at the IEEE, says that the institute thinks its preventive measures and efforts identify almost all submitted papers that do not meet the organization’s standards.

However, Cabanac and Kendra Albert, a technology lawyer at Harvard Law School in Cambridge, Massachusetts, have found issues, including tortured phrases, citation fraud and plagiarism, in hundreds of IEEE papers published in the past few years, Retraction Watch reported earlier this year. Stickel says that the IEEE has evaluated those papers and found fewer than 60 that didn’t conform to its publication standards, with 39 retracted so far.

The 50,000 or so retractions recorded around the world thus far are only the tip of the iceberg of work that should be retracted, integrity sleuths say. The number of articles produced by ‘paper mills’ — businesses that sell bogus work and authorships to scientists — is estimated to be in the hundreds of thousands alone, quite apart from genuine papers that might be scientifically flawed. “Paper-mill products are a problem even if no-one reads them, because they get aggregated with others into review articles and laundered into the mainstream literature,” says David Bimler, a New Zealand-based research-integrity sleuth also known by the pseudonym Smut Clyde.