Should I tell anyone that I suspect inappropriate behavior in the article I'm viewing?

The cancer researcher is suspicious of the article he was asked to review. What steps should they take?

Council

Paper retreat is growing, along with awareness of the problem of scientific misconduct. From plagiarism to text created by artificial intelligence to authorship for sale, many articles are unfairly submitted to magazines. If these articles are published later, it may weaken the areas of research and potentially have tragic consequences if, for example, oncologists rely on unreliable research in the treatment of people with cancer.

But part of the verification process is also the verification of errors made in good faith. Not all marked documents are actually fraudulent. As you assume, errors can penetrate into articles, even if the authors did not intend to use labels. Unfounded accusations can prevent the publication of important studies, damage the career and potentially retaliate against criticism.

Reviewers play a crucial role, together with editors and publishers of magazines, helping to ensure the quality and reliability of published journal articles. But lack of time and review recommendations may limit the ability of reviewers to express concern.

Doubt

There may be a reasonable explanation of the similarity of images. Jennifer Byrne, an oncologist researcher at the University of Sydney, Australia, offers to check two things. First, do the images in the previous article refer to the same things (e.g. proteins and cell lines) that are described in a later article? (You can also review the annotations around the image, for example, whether new descriptions have been added.) Secondly, do the two articles have common authors? If both conditions are met, the authors can reuse their previous data. It may be a justified or honest mistake, says Byrne, who also researches scientific integrity.

Even if the author's commands are different, there may be a good reason to reuse the images. "Some images can be provided by repositories in the form of data that will be reused for various research questions," explains Byrne. Or the authors submitting the application could obtain permission from the copyright holders to reuse the images from the previous study for comparison purposes. In any case, says Byrne, "any intentional reuse of the image, however, must be clearly described in the manuscript". Therefore, it should also be cited in the links, following the recommendations of almost all publishers.

Such problems should be mentioned in the reviewers' comments to the authors. Nevertheless, Byrne says that it would be more alarming if the images in the article in question referred to things other than those in the published article.

In addition to apparently duplicated images, the reviewer can consider other problems in the submitted manuscript. For example, Brian Jones, a biologist at Murdoch University in Perth, Australia, checks the selection of references to make sure they support the statements made in the submission. Again, the discrepancy may be an honest mistake of the author in a hurry. Or it may be a sign of manipulation of citations, in which irrelevant links are deliberately inserted into the document to increase citation rates.

Talk to the editor

Questions should be raised confidentially to the magazine editor, for example, by sending an email to the editor or using the "contact the editor" function in online application systems. "It is important to recognize that manuscripts are provided to reviewers on a confidential basis, so reviewers should not report problems to anyone outside the review process," Byrne emphasizes.

This message should include the evidence you have collected, including a reference to a previous publication and figures that seem to be duplicated, says Jillian Goldfarb, a researcher at Cornell University in Ithaca, New York. In this first warning to the editor, she recommends that you "ask for recommendations on whether you should continue viewing the article or whether they would want to suspend the review process to investigate this issue".

Check your tone

Whether in communication with the editor or in comments for authors, it is risky to assume that offenses have been committed. Goldfarb saw with his own eyes the political difficulties of publishing. In 2023, she resigned from the post of magazine editor, partly because of ethical concerns about suspicious documents.

Advisors emphasize that in order to avoid a lawsuit, it is not a good idea to use words such as "fraud". Instead, it is best to express the reasons for concern briefly and objectively. "You will get into trouble if you assume that the mistakes were intentional," Jones warns. Instead of mentioning plagiarism, for example, he suggests writing something like "paragraph 6 seems to be identical to what is in article XXX". For detailed legal advice, however, you should contact a professional.

This advice is applicable even in seemingly blatant cases of misconduct. Jones gives an example of the presented article, in which the author, who died five years earlier, participated. Nevertheless, he covered his foundations: "I checked the email in this case, and the institute confirmed the death and expressed a lack of knowledge of the document, so it was not a posthumous submission."

Goldfarb advises that "accusing someone with a firearm is not the most professional tactic in these situations. Rather, stick to a professional, curious, but worried tone."

Balance your responsibilities and restrictions

Ideally, "it's enough to warn people about the problem to solve it," Goldfarb comments. "I've seen cases when the best angels of the authors appear, and suddenly new figures appeared in the revised project, and it became even better for such changes." She hopes that raising concerns in this way can also contribute to improving the laboratory's research culture. "I hope this means that you have caused cultural changes in the research group, that such actions are unacceptable."

Further options are available if the article is published despite convincing evidence of misconduct, but they bear their own risks. You can write a letter to the editor for publication, says Jones, but only "if you are very confident in your facts, because some authors threaten a lawsuit". For potentially anonymous options, he suggests contacting Retraction Watch, a research integrity blog or other press organizations, or expressing concern on a review site after publishing PubPeer.

Given this sensitivity, reviewers may be tempted to simply ignore the problem. But this affects science as a whole: some organizations claim that reviewers are obliged to report their problems. As Goldfarb says, "failure to report suspicions raises ethical questions and threatens the reliability of science itself".