When it comes to correcting the scientific record, chemists prefer to have a quiet word

A survey of chemists’ attitudes to correcting errors in scientific publications shows most prefer to discuss them informally with colleagues and students or to write privately to editors. However, the researcher that conducted the study suggests using online platforms and repositories to encourage more corrections to the public record.

Frédérique Bordignon, a researcher at the LISIS laboratory and bibliometrician and research integrity officer at Ecole des Ponts, France, contacted 50,000 corresponding authors of at least two publications between 2020–2023 in American Chemical Society and/or Royal Society of Chemistry journals. Almost a thousand replied, a response rate of 2%.

Half of respondents said they’d found an error in their own work and, among those, 86% said they had taken action. A large majority (88%) reported finding an error in someone else’s publication and almost four-fifths said that they had done something about it. Some wrote to the journal (13%), others tried to replicate the work (4%), published a refutation (4%) or posted on PubPeer (2%), an online post-publication peer-review platform. But most (42%) brought up errors in private discussions; 32% did this for errors in their own work. Around a third – 16% in the case of their own errors – discussed the problem with their students. Reasons for ignoring errors included time considerations, fear of offence and authors’ seniority or status.

While Bordignon heard from respondents about errors avoided by editors, reviewers and authors cooperating closely, she heard more examples of correction failures ranging from reluctant editors to authors giving in to reviewers and introducing changes that were ‘wrong’.

‘Authors sometimes feel pressured to modify their manuscripts to satisfy reviewers and, without falsifying their data, they may end up writing things they know aren’t quite accurate or even are wrong based on their own study,’ Bordignon says. ‘This is obviously very damaging and concerning for the quality of peer review.’

Concerning findings

Overall, however, she stresses the study shows that researchers are actively correcting errors, even if this doesn’t always result in corrections to the scholarly record. ‘It would be beneficial if researchers used post-publication peer-review platforms such as PubPeer or those dedicated to commenting on preprints to report errors. This leaves a public written record that other researchers can use, for instance to avoid embarking on experiments doomed to fail, or to prevent errors from spreading.’

Chris Brewin, a clinical psychologist at UCL, says the survey’s most concerning finding is that authors might be introducing errors into their papers to satisfy reviewers. ‘They are under a lot of pressure to publish papers, but they should be free to challenge reviewers’ comments.’

Jennifer Byrne, a molecular oncologist at the University of Sydney, was also surprised by this. ‘It’s unclear [from the paper] whether the requested changes had affected the manuscript’s conclusions, or whether they pertained to more minor issues. The idea that researchers might make changes that they believe are incorrect is worrying. While it is possible that some of these “errors” (as viewed by authors), were actually correctly highlighted by reviewers, it seems unlikely that all such cases could be explained this way. It would be interesting to know whether [these] authors are predominantly from particular groups, such as early-career researchers.’

Given the low rates of published corrections, Byrne agrees with Bordignon that more opportunities are needed to flag errors outside the formal post-publication corrections offered by journals and publishers. She thinks PubPeer notices, which describe verifiable errors, should be directly linked with publications. ‘However, these results suggest that PubPeer has a low rate of use by the chemistry community. It would be interesting to understand whether this reflects low awareness of PubPeer, or other factors.’

However, David Peterson, a sociologist at Purdue University, takes issue with the author’s suggestions for reform. He argues that science self-corrects not only through formal mechanisms like retractions and corrections but also through ‘organic’ ones which ‘reformers’ have overlooked or downplayed.

‘The author acknowledges these but treats them as inefficient or problematic because they don’t correct the published record,’ he notes. ‘I’m not convinced this is a big problem. The published literature is one source of information, but [there are] many others including colleagues. There is a narrative among reformers that errors corrupt or pollute the literature, but I find this metaphor problematic. The scientific ecosystem has mechanisms for self-protection. If an article makes big claims but generates little follow-up and few citations, you learn it was probably a dead end.’

He thinks it comes down to priorities. ‘You can imagine fields in which self-correction has grown so lax that errors have proliferated out of control. Is there any reason to suggest that chemistry is in this boat? I’m sceptical. If not, the goal is more about making science more efficient. But, again, I don’t see any evidence that chemists devoting time to correcting the published literature of – in many cases – small errors, would be a better use of their time.’