Imprimir Republish

GOOD PRACTICES

Misconduct on display

Website that allows users to make anonymous comments on scientific articles has established itself as a platform for reporting ethical misconduct

Hulton Archive / Getty Images

Created 11 years ago as a forum for discussing scientific articles, the PubPeer website has established itself as a channel for reporting scientific misconduct. Its influence was evident in the dismissal of Stanford University president Marc Tessier-Lavigne in mid-July. The first claims that several articles published by Tessier-Lavigne’s team contained manipulated images were made in comments on PubPeer in 2015, gaining traction after a reporter from a Stanford student newspaper compiled the allegations in a report last year. An investigation by the university concluded that the president did not participate in the fraud and was not aware of it before the articles were published, but investigators criticized him for not asking for the articles to be retracted as soon as he learned of the problems — his inaction ultimately sealed his fate. “I agree that in some instances I should have been more diligent when seeking corrections, and I regret that I was not,” Tessier-Lavigne wrote in his letter of resignation.

Anyone following PubPeer’s trajectory was not surprised by the outcome at Stanford. In 2016, software capable of identifying statistical errors in scientific articles was used to analyze 50,000 papers published in psychology journals, with the results shared on PubPeer to the embarrassment of thousands of authors. The following year, the journal Nature decided to revoke an award it had given to Spanish biochemist Carlos López-Otín in recognition of his work as a mentor for young researchers after nine articles by his research group at the University of Oviedo were retracted for image manipulation — the evidence for which was found on PubPeer. More recently, seven articles by Gregg Semenza of Johns Hopkins School of Medicine, winner of the 2019 Nobel Prize in Physiology or Medicine, were retracted due to image manipulation. Once again, the problem was brought to light on PubPeer.

Any article published in a scientific journal is subject to comments on the website. In practice, the platform promotes a similar type of scrutiny to peer review, but after studies have been published, with readers able to point out errors or signs of misconduct. Reviewers can choose to identify themselves, but they do not have to do so. According to the website’s creators, around 90% of the reviews are anonymous and PubPeer does not deliberately store any information about its users. “It’s only a matter of time before user information is leaked in some way. The only sure means of protection is never to record any information we do not want to see exposed,” explained the website when announcing its new technical regulations in 2017.

The anonymous nature of the complaints has already caused at least one major legal headache for PubPeer, but the website has emerged stronger. In 2016, Fazlul Sarkar, an oncologist at Wayne State University, USA, who later had 13 articles retracted for image manipulation, sued PubPeer, demanding that the identity of the individuals who highlighted the problems in his work be revealed. He claimed that a job offer from the University of Mississippi was revoked as a result of the allegations. A Michigan state court found that anonymous reports were protected by law.

In PubPeer’s early years, its own creators remained anonymous and the platform was shrouded in mystery. The founders only revealed their identities in 2015 as the website’s reputation blossomed and they decided it was time to seek funding to expand operations. American neuroscientist Brandon Stell of Paris Descartes University, France, announced that he was behind the initiative, supported by two brothers: his former student Richard Smith, and Richard’s brother George Smith, who works in web development. Boris Barbour, another neuroscientist from the same university, and Gabor Brasnjo, a lawyer based in San Francisco, USA, also joined the team after PubPeer was created.

At the time, Stell told the journal Science that he originally had the idea during his undergraduate days at the University of Colorado Boulder, USA, where he attended meetings with researchers and students at which articles published in journals were discussed and dissected. The experience inspired him to create a virtual discussion club, open to all, with the ambition of accelerating the process of correcting science. The website grew and led to the creation of the PubPeer Foundation, based in California and funded by a philanthropic organization backed by billionaires Laura and John Arnold.

The platform asks commenters to only post concrete information that can be publicly verified. It expressly prohibits rumors. “Allegations of misconduct are forbidden on PubPeer,” the website informs users. “They are anyway unnecessary. Your audience on the site is mostly composed of highly intelligent researchers and scientists. They are quite capable of drawing their own conclusions if the facts are clearly presented.”

This does not mean that all comments posted on the site are reliable. Denise-Marie Ordway, who works for the website The Journalist’s Resource, interviewed three journalists who use PubPeer as a source of information and asked them for recommendations on how to responsibly handle claims made on PubPeer. The most important thing is to understand that the information is merely evidence that needs to be confirmed through detailed investigation. “You need to treat anonymous postings on PubPeer the same as you would treat anonymous postings of any kind,” said Charles Piller, one of the interviewees, who writes for the journal Science and is one of the founders of American nonprofit investigative journalism organization The Center for Public Integrity. The first precaution Piller takes is to ask scientific integrity experts to verify any suspicions raised in the comments: “You have no way of knowing if that person has some sort of ulterior motive.” He raises concerns about the damage such complaints can cause to reputations. “Show evidence,” he advises. “It’s better to let the evidence speak for itself most of the time.”

Stephanie Lee, a reporter for the news website The Chronicle of Higher Education, says she uses the platform whenever writing about a scientist or group whose work interests her. “PubPeer is the first place I’ll go to see if questions have been raised,” she explained.

Julia Belluz, a correspondent for the website Vox, highlighted the role PubPeer plays in the process of correcting the scientific record. “We know peer review can fail to catch errors or even outright fraud in research before it’s published. PubPeer is a place where scientists go to whistleblow about problematic research.”

Republish