RETRACTION NOTICES: THE GATEWAY TO UNDERSTANDING RETRACTIONS

Retractions, primarily of the biomedical literature, continue to rise, including COVID-19 literature ¹. Many of those retractions have arisen from post-publication peer review (PPPR) ², leading to negative criticisms and the exposure of errors, problems, or fraud in papers ³. Retraction notices (RNs) have generally become more informative, but a large body of RNs remain incomplete and opaque about details leading to or surrounding the retraction procedure and its background and thus have limited scholarly use ⁴. Consequently, opaque, superficial, or incomplete RNs are poor vessels of communication with the broader academic population. Linguistic tactics are frequently used in RNs to euphemistically assign responsibility, thereby often not allowing the identification of those agents responsible for drafting the RN. This legally and ethically evasive tactic reduces editors' and publisher's accountability ⁵.

To the author's knowledge, and based on experience through observations of cases over the years, there appear to be two broad patterns that would lead to the final wording of a RN: 1) In the case where an author retracts their own paper, author-suggested reasons, explanations or apologies might be incorporated into the RN, but rarely are an author's suggested wording 100 % incorporated into a RN, and these are often moderated by the editors or publisher; 2) more commonly, independent of who calls for the retraction of a paper, the final wording is agreed upon by the editor in conjunction with the publisher. In the latter, the teams tend to be individuals who form part of an "ethics" or integrity team, one or more of whom may be legal specialists. Consequently, it is not uncommon to find only short and succinct statements in RNs that may carry specific legalese or legal jargon that may be difficult for academics to appreciate or interpret with clarity.

This opinion paper describes and discusses a closely related phenomenon, namely, authors' disagreement with retractions, as is sometimes explicitly noted in RNs.

DISAGREEMENTS WITH RETRACTIONS: POSSIBLE REASONS

Since statements that record such a disagreement tend to be short and uninformative, for example, "The author/authors does/do not agree with the retraction," other academics and the public are somewhat left in the dark regarding the reason for such a disagreement. An academic paper-at least in scholarly, indexed peer-reviewed journals-usually results from rigorous peer review and strict editorial handling and may follow an equally long research period, except for pay-on-demand paper mill-derived papers ⁶. Thus, it makes sense to transparently explain in the RN the reason for that disagreement to respect authors' rights to have their opinions properly heard and expressed ⁷ and to duly inform the academic community of those reasons.

In principle, it is possible to have fundamental disagreements with one or more of the following aspects of a retraction and/or RN: a) its findings; b) the procedure; c) the wording of the RN; d) the outcome of the ethical investigation; e) the editorial or publisher's handling; f) the conclusions drawn; g) other reasons not listed here.

EXEMPLIFYING OBJECTIONS: SIX RETRACTION NOTICES

Six examples are shown next to highlight issues with author-stated objections, either to the retraction or the RN. In the first (Springer Nature), 10/11 authors explicitly disagreed with the retraction, but the RN does not indicate the reason(s) for their disagreement ⁸. In the second (BMC, part of Springer Nature), the RN indicates that "authors agree with the retraction but disagree with the wording of the retraction notice," although it is unclear which wording precisely they disagree with ⁹. In the third (Wiley Periodicals LLC), the RN provides apparently contradictory information, namely agreement between the authors, editor-in-chief, and publisher ("has been retracted by agreement between the authors, the journal's Editor in Chief, Prof. Dr. Christian Behl, and Wiley Periodicals LLC"), but then a statement indicated that the authors disagree ("[t]he authors disagree with this decision") ¹⁰. In the fourth (Portland Press Ltd.), the statement that points to the authors' disagreement carries a grammatical error ("[t]he authors disagree to [sic] the Retraction") ¹¹. In the fifth (Karger Publishers) ¹² and sixth (Taylor & Francis) ¹³ RNs, the authors' disagreement is straightforward and stated simply.

In all these cases, the journals and/or publishers are Committee on Publication Ethics (COPE) members. Despite this, only the sixth RN states a direct link to, or association with, COPE: "We have been informed in our decision-making by our policy on publishing ethics and integrity and the COPE guidelines on retractions." It is thus unclear why there is such variation in the wording and content of RNs among these COPE members and to what extent there is editorial freedom in determining the final content and wording of an RN. In all of these RNs, aspects of the background (e.g., people or procedure involved) are shrouded in "secrecy" (or confidentiality), and the explanation provided by the RN for the entire process is, for most of these cases, incomplete and/ or opaque. In one or more of the cited cases, which are apparently derived from paper mills ¹⁴, it is unclear if the authors' response and statement in the RN represent the voice, opinion, and decision of the actual authors or of a "representative" of the paper mill, particularly where the authenticity of the corresponding author's email is in question.

In all RNs, blame appears to be placed squarely on authors' shoulders, completely ignoring a failure in peer review and thus by peer reviewers and editors, with potential underlying aspects of oversight, peer or editorial incompetence, or lack of professionalism, in detecting the issues exposed during PPPR that led to these retractions. In other words, in all RNs, only the authors' integrity is questioned, but not that of the peer reviewers, editors, and/or publishers. This skewed position suggests bias. Similarly, in all six RNs that were profiled, no apology was offered to the public, academia, or clients who may have used, read, cited, or purchased fraudulent, flawed, and/or erroneous research. The latter point raises a critical yet distasteful reality, namely whether journals-via boosted metrics such as citations-or publishers-via boosted rankings, sales, subscriptions, or open access article processing charges-may have benefitted unfairly from the publication and/or sale of retracted research ¹⁵. Finally, these retracted papers were indexed in PubMed (and/or Medline or PubMed Commons), an apparently respectable public biomedical database and search platform. However, with rising cases of retractions of PubMed-indexed literature, the scholarly legitimacy and reliability of PubMed as a source of trustworthy biomedical information is being increasingly questioned ¹⁶.

It is incompatible for COPE member journals and publishers to claim that they have employed open and transparent, fair and principled procedures prior to issuing a retraction and then be opaque regarding details of the retraction's history, procedure, and statements in the RN, including reasons for authors' disagreements with the retraction of their papers. Claimed transparency, on the one hand (related to post-publication discovery and investigation), followed by actual opacity, on the other (related to peer review failure), is an incompatible virtue, even more so when applied as stated policies.

There are also several unpalatable questions that most academics most likely do not wish to ask openly. To what extent are publishers' legal counsel and/or legal departments controlling the wording of RNs? Is there any editorial leverage, and what is the extent of editorial independence throughout the investigative process leading up to a RN? Do authors have an opportunity to decide the wording of the RN, even if they are at fault or disagree with the retraction? These are three key questions that require urgent answers as academia witnesses not only a rise in retractions ¹⁷ but also in the number of statements claiming authors' disagreement.

Evidently and importantly, interviewing the authors of retracted papers would be ideal to better appreciate their reasons for disagreement. However, such affidavits or declarations would likely not be easy to publish in academic journals, as they might trigger legal threats or take-down notices by publishers who may be the subject of criticism. Ultimately, deservedly or not, retractions can hurt individuals, careers, and reputations because they remain symbols of stigmatization despite their corrective nature ¹⁸. Moreover, some are advocating for a solidification of this shaming culture ^19,20.

Curiously, an extensive analysis of 7,650 RNs only considered reasons for retraction, but seemingly failed to consider the crucial issue of authors' disagreements ²¹, suggesting that such analyses are incomplete, or portray a skewed perspective regarding RNs. This discussion thus provides an opportunity for researchers interested and invested in this topic to reassess their methodological approach to understanding RNs by placing greater emphasis on authors' opinions and perspectives rather than relying blindly or exclusively on editor and/or publisher-controlled statements in RNs.