Peer review…and by marketing

It is often repeated that scientific publications are first evaluated by peers, experts in the field, ensuring, as much as possible, the validity of the published results. Sometimes this is done singly (only the identity of the reviewers is anonymized), sometimes in pairs (the names of the article's authors are also removed). To push doubts even further, some reviews (I've tried it!) have even gone triple-blind! Therefore, the editor responsible for the decision to accept or reject the article does not know the names of the authors and reviewers.

But we forget a crucial question: Who decides in light of contradictory evaluation reports? Are we looking for new reviewers or does the magazine director decide for himself? A recent scandal involving Ranga Dias from the University of Rochester in the United States sheds some light on what is happening at a prestigious magazine, nature. An expert in superconductivity, the physicist made headlines in 2020 thanks to a publication in this journal announcing the development of the first superconducting material at room temperature. The startling result attracted the attention of experts, and as doubts mounted about the veracity of the data, the magazine was forced to retract the article two years later, while the fraud investigation began. But in March 2023, almost a year later, Ranga Dias returned to the charge and published in the same journal, albeit a very selective one, another exciting article announcing a room-temperature superconducting material!

Many physicists after that He objected to the magazine's decision They demanded to see the evaluation reports to no avail. but Journalists had access to it. All four reports were extremely critical. Someone even mentioned the danger of this case becoming another “Shaun case”! Jan Hendrik Schön, an employee of Bell Laboratories in the United States, had published exceptional results in the field of molecular transistors, which turned out to be completely invented. These scams, which were discovered in 2002, led to the retraction of several articles, eight of which appeared in 2001 in science And nature.

Two reviewers from Ranga Dias noted that there were serious doubts about the 2020 article — which was still under investigation at the time — and that important details were missing about the procedures for compiling the material. After five publications in which the physicist attempted to defend his results, only one reviewer concluded that the results were robust, while another slightly supported the publication. The other two were against the explanations provided and remained dissatisfied with them or demanded further action.

In the face of this division, the magazine decided to publish the article. Knowing that the former was under investigation for fraud and that the publisher and reviewers knew the authors' names, why not ask for the opinions of other reviewers? There is no doubt that the race for scientific leadership was in favor of publishing. After all, once the excitement is over, it's always possible to withdraw the item without much fuss. On the other hand, refusal to publish it may lead to it being published in another journal, which may then gain fame associated with a potentially major discovery… so the benefit seems greater than the risk, especially since this is not the first time this has been done. Extraordinary results are retracted in prestigious journals. Sean's case reminds us that this is quickly forgotten. Ranga Dias' new article was finally retracted in November 2023.

In short, peer review is necessary, but not sufficient. Magazine managers still need to put science before marketing…