An analysis identified studies that contained duplicated images or tortured phrases.Credit: bernie_photo/GettyA biotechnology journal that was inundated with paper-mill submissions in 2021 — and claimed in 2023 that it had tackled the problem — still harbours hundreds of dubious papers, an analysis by research-integrity sleuths has found.At least 226 studies on rodents published in the open-access journal Bioengineered between 2010 and 2023 contain manipulated or duplicated images, says the analysis, which was posted on the arXiv preprint server on 28 March1. These are hallmarks of papers produced by paper mills — companies that churn out fake scientific manuscripts to order. Most of these 226 studies have not been retracted.Science’s fake-paper problem: high-profile effort will tackle paper millsBioengineered and its publisher, Taylor & Francis, said in a 2023 blog post that they became suspicious about paper-mill submissions in early 2021. In response, the publisher said, it had introduced policies to screen manuscripts more rigorously. However, the preprint’s authors say that efforts to clean up published papers have been too slow.“The problem isn’t that they were targeted by paper mills, because that can happen. It’s just the way that you solve the problem afterwards. And I think that this hasn’t happened correctly,” says René Aquarius, a neurosurgery researcher at Radboud University Medical Center in the Netherlands and a co-author of the preprint. “It’s been two to four years since most of these bad papers have been published,” he adds.Taylor & Francis says that it is investigating “a large number of articles” in Bioengineered that it suspects were produced by paper mills. “We will cross-check that list with papers of concern identified by the preprint’s authors,” a spokesperson for the UK-based publisher told Nature. “Alongside this work, we have made a range of editorial and process changes to the journal, which, as noted in the preprint, have prevented the publication of further paper-mill content.” The spokesperson declined to answer specific questions about the preprint’s findings.Suspicious activityFrom its launch in 2010 until 2019, Bioengineered published around 70 papers a year. This rose to 131 papers in 2020, then soared to more than 1,000 per year in 2021 and 2022.“Our previous experience on other journals that had been targeted by paper mills signaled the need for a thorough investigation,” said Todd Hummel, a global publishing director at Taylor & Francis, in the 2023 blog post. He added that the journal had brought in measures to root out fraudulent submissions. These included asking authors to provide raw data with their manuscripts, training editors on how to spot paper-mill submissions and applying closer scrutiny to requests to change author lists. In 2023, the number of papers the journal published dropped back down to 64.Science journals crack down on image manipulation“I think it’s admirable that they have been transparent about this,” says Aquarius. “It’s great that [they] stopped the influx of paper-mill papers, or at least the acceptance of them.”“But why leave some of those papers for multiple years?” he adds. “I think cleaning up the literature is also an important task.”Jana Christopher, an image-integrity analyst at FEBS Press, who is based in Heidelberg, Germany, agrees. “It’s important that the effort of correcting the literature takes place efficiently, effectively, thoroughly and at a reasonable pace,” she says.Manipulated imagesIn their analysis, Aquarius and his colleagues decided to screen studies that mentioned mice or rats in their titles or abstracts. The number of articles that met these criteria had increased sharply in 2021 — before 2019, they made up less than 15% of Bioengineered papers, but in 2021 and 2022, they accounted for more than 35%.
Enjoying our latest content?
Login or create an account to continue
Access the most recent journalism from Nature's award-winning team
Explore the latest features & opinion covering groundbreaking research
Access through your institution
or
Sign in or create an account
Continue with Google
Continue with ORCiD