New retraction vocabulary
One of the topics discussed at the 5th World Conference on Research Integrity, held from May 28 to 31, 2017, in Amsterdam, the Netherlands, was the need to establish alternatives to the concept of retraction, which is the cancelation of scientific papers when fraud or errors are discovered after the papers are published. The main argument in favor of the change is that, today, errors made in good faith are confused with cases of fraud and are stigmatized; for that reason, a different term should be used.
One case that occurred in 2012 was mentioned at the event. Mathematician Richard Mann, of the University of Leeds in the United Kingdom, was informed that a paper he published in PLOS Computational Biology, on the collective motion of glass shrimp, was seriously flawed. Due to a coding error, Mann included only 1/100th of his data in the computational modeling. While under investigation, he feared that retracting the paper could be construed as research misconduct. “I became worried about public shaming,” he said during the conference.
Unlike Mann, who had the opportunity to correct and republish the paper, many other researchers make innocent mistakes and their careers are called into question after they need to retract an article. Nicholas Steneck, director of the Research Ethics and Integrity Program at the University of Michigan, told Science “You have to change the language.”
There are precedents for such cases. The Embo Journal adopted the term “partial retraction” for cases in which a minor error was found in a paper, but with no effect on the conclusions. In 2015, JAMA Psychiatry introduced the concept of “retracting and replacing” for clinical trials that had pervasive errors, but once corrected, were still worth publishing. In both situations, the paper is not withdrawn from the journal, but a corrected version is published.
Some proposals go beyond that. In an article published in March 2017 in the bioRxiv repository, Virginia Barbour, a professor at the Queensland University of Technology in Australia, suggested replacing the term “retraction” with “amendment,” and there would be three categories: insubstantial for minor errors, such as typos; substantial, for errors involving data or figures that nonetheless do not jeopardize the work as a whole; and wholesale, when the entire paper is considered unreliable and must be canceled, corrected and republished.