Better Bored Than Confused: Why Consistency Is Key in Biomedical and Scientific Writing

ABSTRACT

Formal corrections of research publications should be clearly communicated to users of the scholarly literature, especially when these corrections have consequences for reported knowledge. However, the information that a publication was corrected is not always shared and indicated clearly. To test correction communication, we formed a corpus of publications with corrections most likely consequential to their reported knowledge. To this end, we selected 3177 publications that were recently corrected according to Crossref and that received at least two comments on PubPeer. We used this corpus as a probe to assess correction communication on a range of scholarly platforms. The majority of these platforms does not indicate the presence of corrections together with the entry for these publications: Web of Science, OpenAlex, Dimensions, Google Scholar, Lens, Mendeley, Zotero, Endnote, and Elicit. Those that do indicate the existence of corrections, do so for nearly all corrected publications: Scopus 90.8%, PubMed 98.7%, EuropePMC 98.8%, Scite.ai 99.7%, and the journal's webpage 94.4%. Even when the presence of a correction is communicated, this is not always indicated clearly. Readers of the scholarly literature might miss corrections when they are communicated under a generic header such as ‘referred to by’, or only indicated by a small line of text in the same colours as the rest of the page. We provide suggestions for improved communication of corrections through better integration of information across platforms, improved labelling on platforms, and the development of correction communication standards.