The Existence of Stealth Corrections in Scientific Literature—A Threat to Scientific Integrity

Abstract

One of the cornerstones of publication integrity is the thorough maintenance of the scientific record to ensure the trustworthiness of its content. This includes strict and transparent record-keeping when implementing post-publication changes through a clearly visible corrigendum or erratum, which provides details of the changes and the reasons for them (ICMJE 2024).

However, such record-keeping is not always practised as stealth changes, post-publication changes to the scientific literature without any accompanying note, have been observed. One notable kind of a stealth change is a stealth retraction: published papers simply disappearing from the website of a journal without a formal retraction notice (Teixeira da Silva 2016; Teixeira da Silva and Daly 2024).

Besides stealth retractions, a second problematic type of stealth change exists in the scientific literature: stealth corrections. We define a stealth correction as at least one post-publication change made to a scientific article, without providing a correction note or any other indicator that the publication was temporarily or permanently altered.

Stealth corrections in the scientific literature have occasionally been described by online blogs (Bimler 2021; Schneider 2020), but the scientific literature itself has yet to address the numerous stealth corrections in its own corpus. In this article, we provide examples of such stealth corrections.

We found a total of 131 published articles that were affected by stealth corrections, across a variety of large and small publishers (Table 1). In most cases, the content of the articles was changed (Table 2). An overview of all stealth corrections (# 1-131) can be found in Table S1, which also contains the links to all accompanying PubPeer posts for additional detail. Five articles with stealth corrections eventually received an official correction notice and 17 articles were reverted to their original version (Table S1). Nine expressions of concern were published, and 11 articles were eventually retracted (Table S1). Additionally, seven book chapters were removed (Table S1).

The stealth corrections presented in this paper demonstrate a fundamental and mostly ignored problem in the scientific literature. Correct documentation and transparency are of the utmost importance to uphold scientific integrity and the trustworthiness of science. Post-publication changes need to be clear for readers to understand if, and why, changes have been made. However, little attention is paid to post-publication alterations. For example, the Committee on Publication Ethics (COPE) provides clear flowcharts indicating that a published correction is needed in case of: post-publication plagiarism (# 20) (COPE 2021b), post-publication author removal (# 2-9) (COPE 2021a) or inappropriate image manipulation in a published article (# 10-11, 18-19, 21-101) (COPE 2024). However, no specific flowcharts exist on post-publication changes in author affiliations (# 1–2), consent or ethics statements (# 129-130), handling editor (# 102-103), or allocation to journal issue (# 104-128). Post-publication amendments that are made silently, without a visible correction note, will give rise to questions regarding the ethics and integrity of the specific journal, editors and publisher, and might undermine the validity of the published literature as a whole.

It is particularly concerning when publications with serious problems such as image duplication or data duplication receive stealth corrections. In cases of serious issues, a stealth correction could mask the need for a formal retraction, which could potentially benefit journals, publishers and authors engaging in misconduct. Future readers are likely to be unaware that such problems were ever present, which might lead to less critical assessment of these publications. Readers of our findings might even get the impression that editors or publishers are actively trying to hide problems, such as in one case where the journal's Editor-in-chief was listed as an author of the stealthily corrected paper (# 19, the article received an official correction 7 months after the stealth correction was discovered). Or the cases where the publisher moved articles authored by editors of a special issue from that special issue to a “section”, which is a different part of the MDPI infrastructure and not linked to a special issue, possibly in order to meet the criteria of the Directory of Open Access Journals (DOAJ) (# 104-128).

The question then becomes how to incentivise transparency; how can we encourage journals and publishers to issue formal notices instead of resorting to stealth corrections? We need clear definitions of the term “correction.” Situations can become complex very quickly as publishers often post preliminary or pre-publication versions of papers while they are undergoing type setting or even pre-prints on their own hosting platform. What is the precise moment of publication in these cases? If anything changes during these stages, will these changes count as corrections? Should undocumented, fundamental changes such as the replacing of images or results during this phase of the publishing process still be regarded as stealth corrections or not?

Another complicating factor is that publishers might use internal guidelines on handling corrections. These guidelines might leave room for interpretation by their reader, as is the case with Elsevier: “If […] new material should replace the original content of the accepted article, the editor may consider the publication of an erratum or a corrigendum” (Elsevier 2024). Guidelines might also lack the information needed for editors to make a decision on the topic, as is the case with MDPI's guidelines, which do not describe what to do when removing articles from a special issue (MDPI 2024). The abovementioned COPE flowcharts might help to circumvent these problems and could offer publisher-independent guidance on handling post-publication corrections. However, additional flowcharts are needed to adequately address all types of post-publication corrections.

Stealth corrections add yet another threat to the integrity of scientific publishing, besides previously described phenomena such as paper mills (Bishop and Abalkina 2023; Candal-Pedreira et al. 2022; Else and Van Noorden 2021), fake peer review (Oviedo-García 2024) and abuse of special issues (Mills, Mertkan, and Onurkan Aliusta 2024). Authors are often at the heart of these problems, fuelled by the harsh publish-or-perish pressure of academia (Al-Adawi, Ali, and Al-Zakwani 2016; Vasconez-Gonzalez et al. 2024). However, adequate handling of corrections is exclusively the responsibility of publishers, journals and editorial boards. Academic editors are often seen as the safeguards of the scientific literature and its integrity (Desai and Shortell 2011; Marusic 2010), but in the case of stealth corrections, this integrity can come into question.

We strongly recommend that every post-publication change, even in a preliminary or pre-publication version, be publicly logged in the interest of transparency. In the long term, this should ideally become an automatic process, built into journal content management systems, such as Editorial Manager; any time an article is updated, the change would be recorded in the article's public-facing metadata. This fully transparent approach would allow readers to make their own determinations as to whether a change is consequential or not, rather than relying on the editors' discretion. Automating the logging of changes would make stealth changes impossible, giving readers an assurance that all changes are transparently declared. This would also enable indexing services, such as Web of Science or PubMed, to track a journal's post-publication corrections and incorporate this information into their journal evaluation processes. In the short term, indexing services should document stealth changes when they are identified and consider sanctions for the journal in question, with removal from the index for repeat offenders.

Stealth corrections are extremely difficult to detect and we ask the scientific community to stay vigilant in order to detect and report them. Reporting should preferably take place on a public platform such as PubPeer to provide the necessary transparency and long-term documentation.

In order to add new cases and monitor the cases we have described in this article, we have created an online spreadsheet listing stealth corrections (find it here). We invite interested readers to forward any PubPeer link detailing stealth corrections to us, so we can add them to the spreadsheet.

R.A.: conceptualization, data curation, formal analysis, investigation, methodology, supervision, validation, writing – original draft, writing – review and editing. F.S.: conceptualization, data curation, investigation, validation, writing – original draft, writing – review and editing. N.W.: conceptualization, investigation, writing – review and editing. A.G.: conceptualization, investigation, writing – review and editing. G.C.: conceptualization, investigation, supervision, writing – review and editing.

Nick Wise is a research integrity manager at Taylor & Francis. The final version of this article was submitted before he started working for Taylor & Francis.