Cheating in scientific publishing: the scourge must stop!

I opine it isn't moral For a writer-man to cheat, And despicable to wear a laurel As was gotten by deceit Here's a message to those who would cheat in one form or another in the publishing world: Hey, wake up! We are living in the 21st century, in the electronic age, and the scientific world has indeed turned into the ‘global village’ envisioned by the Canadian writer/visionary Marshall McLuhan almost five decades ago. While it may have been possible, even probable, to get away with cheating activities such as duplicate publication and plagiarism many decades ago, in today's internet age, with the wide availability of search engines for publications, cheating seems incomprehensible, pointless and stupid. It is almost certain to be detected, and usually very quickly. Despite this, cheating continues. A PubMed search using the term ‘duplicate publication’ generated approximately 1300 citations, many of these notices of duplicate publication. A 2005 anonymous survey of more than 3000 American NIH-funded researchers, found that 1.4% admitted to plagiarism and 4.7% to duplicate publication (1)! Recently, a vigilant reader notified us that a Liver International letter by Fang S. H. and colleagues (2) was a duplicate publication (3). Examining the dates of submission of both items, it became clear that they had been simultaneously submitted. When confronted, the authors replied, somewhat disingenuously, that they thought that a ‘Letter to the Editor’ was ‘informal format and was not a paper’. Adding to the egregiousness, their report had been selected as the cover image for that issue of our journal. Several months ago, an even more insidious attempt at multiple publication was thwarted by the vigilance of Hartmut Jaeschke, an Associate Editor handling a manuscript by S. L. Yang, F. Hong and Y. J. Lou, from Nanchang and Hangzhou, China. A PubMed search followed by reading the papers revealed that Yang and Lou, by themselves or with assorted coauthors, had already published essentially the same work three times in 2007 (Biochem Pharmacol 2007; 73: 724–35. J Surg Res 2007; 140: 36–44. Pharmacology 2007; 80: 11–20). Simply reading the abstracts of these papers and the submitted manuscript failed to detect the cheating – they were worded sufficiently differently. However, on closely reading the papers, it became clear that the same body of work had been cleverly ‘disguised’ and presented slightly differently. There are types of ‘duplicate publication’ which are not so clear-cut. As an international journal, we are occasionally faced with an English-language manuscript reporting material some or all of which has been previously published in another language by the same authors. Is this truly duplicate publication? The answer, with a nod to the ‘global village’, is again ‘yes’. This issue has been thoughtfully addressed by Farrell (4) in a recent commentary, and we agree with his view that the English-language manuscript must differ from the ‘version originale’ by at least two-thirds or it will be rejected as a duplicate. We also recently rejected a manuscript which overlapped somewhat with a previous ‘Letter to the Editor’ by the same authors. The authors had no malicious intent as they cited their previous letter. The submitted full paper showed some of the same data but added much more. However, on careful inspection, it appeared that approximately a third of the work had already been reported in the letter and the increment in new knowledge provided by the remainder was not sufficient to warrant publication (due to a large increase in submissions over the past year, our rejection rates have significantly increased). It was rather unfortunate that the authors had chosen to report their preliminary data in a letter, as it is likely that the overall complete study, if previously unpublished, would have been accepted for publication. This sad story raises the issue of another prevalent practice, which while not unethical is certainly inappropriate and a bane of journals: the ‘least publishable unit’ (LPU) method. This refers to the practice of doing a study and then splitting it into several small papers instead of a single large manuscript. Thus in the ‘publish or perish’ academic world, it increases the authors' ‘paper count’. We, like all other journals, strongly discourage this practice and urge you to submit a single, complete paper rather than less-impactful LPU papers. In other words, quality should trump quantity. As more academic centres and research-granting agencies now ask for a small number of the most significant papers rather than a simple list of total publications for promotion or grant applications, the LPU practice will hopefully diminish. I have not directly addressed the plagiarism issue in this editorial, but this is as unethical and unacceptable as multiple publication. What can be done to reduce or eliminate these scourges of the publication process? Most reputable journals have policies in place to punish the guilty. At Liver International, culprits will suffer the following consequences: (i) the other journal (or other authors in cases of plagiarism) will be notified. (ii) the dean or head of the authors' institutions will be notified. (iii) authors will be proscribed from submitting manuscripts to Liver International for 3 years. How can we improve our ability to catch the culprits? Of course we are grateful to our readers for alerting us to instances of wrongdoing, but we hope to detect cheating before it appears in print. In that respect, there are several software programs that detect multiple publication and plagiarism in the literature. Two such programs have been made available as ‘freeware’ (without charge) by the UT-Southwestern group (5). These programs look for text similarity in the abstract as well as the body of the paper. etblast, an automated citation matching tool, and Déjà vu, the duplicate citation database, are available at http://invention.swmed.edu/ and http://spore.swmed.edu/dejavu. Cheaters beware – and stay away from Liver International!

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.