A journal was alerted to potential image manipulation in four papers published over the course of twelve years by the same corresponding author. The journal contacted the corresponding author who provided some raw data for some of the papers but not all of them, and was not able to explain the apparent manipulation (which included, in one paper, a duplicate image from a paper published in another journal).
Because of the complexity of the case and the unsatisfactory responses from the author, the journal decided to refer the matter to the author's institute. The institute completed its investigation and concluded that the concerns highlighted in each paper were valid. In one case, the committee agreed that correction was reasonable; in another, earlier, paper they felt that a correction was more 'borderline'. For the remaining two papers, the committee did not find any justification for 'eventual corrections'. However, they believed that the scientific results and conclusions generated by all four papers were solid and were not qualitatively invalidated by the manipulated images. They suggested that this fact be mentioned in any eventual retractions.
The unexplained manipulation of the data makes the journal doubt the reliability of the findings despite what the institution's investigation found. They are, however, conflicted about what do to, because ICMJE guidelines say '..if the investigation proves scientific misconduct, publish a retraction of the article'. But COPE guidelines say that 'Retraction should usually be reserved for publications that are so seriously flawed (for whatever reason) that their findings or conclusions should not be relied upon'.
Question for COPE Council
- Does COPE think that it is acceptable to retract an article because of scientific misconduct, even if the results and conclusions of a paper are sound?
Advice on this case is from a small number of COPE Council Members. Most cases on the COPE website are presented to the COPE Forum where advice is offered by a wider group of COPE Members and COPE Council Members. Advice on individual cases is not formal COPE guidance.
Each of the four papers contained an image that, at best, cannot be relied upon, and at worst, is fraudulent. It is therefore a disservice to the readers of the journal not to mention the issues with the image for the sake of transparency and reproducibility. Hence something needs to be posted—correction, retraction or expression of concern—to alert readers to the finding. The editor, taking into consideration the findings of the university, needs to determine how much of the article is called into question. Did the image in question support the main findings of the article? Does this call into question the conclusions? Or was the image somewhat supplementary, meaning the conclusions are still valid?
To the specific question asked, this goes to the COPE retraction guidelines. The point of a retraction is to correct the literature, not to punish the authors. If the science is sound, then a retraction is likely not appropriate. An exception would be if the conclusions fit the data but the misconduct affected the research integrity (ie, lack of ethics approval or informed consent from participants or lack of permission from a patient to publish a case report). Hence a statement of concern could be published which clearly expresses the issues, or a correction alerting readers to disregard the images in question. Once the questionable figures are identified, the reader will be able to determine for themselves if the results are believable.