A few months we were contacted by a dean of an institution who informed us about misconduct of one of the senior scientists in that institution. An investigation launched by the institution showed that author A and coauthors reused the same images to show controls in many figures in their different publications. This problem was found in three publications in our journal.
We decided to treat the three articles on a case by case basis. We launched our own investigation and contacted the original referees. One of the advisors felt that the authors could not be trusted anymore and we should retract the articles because of the authors’ misconduct. The original referee R1 confirmed that the images were indeed duplicated but did not feel that the main findings of the papers and their general merit depended on those images.
Based on that advice, we decided to give the authors an opportunity to provide corrected images and publish corrections rather than retract the articles. The authors provided new control images and corresponding results for most of the figures in question. However, two of the three articles had some images for which only the control panel was new, and the image showing the corresponding results was the same (ie, the entire experiment had not been repeated). The original referee R1, approached for further advice, felt that the corrected articles could be published and the fact that some results were not repeated was acceptable because it would involve a huge amount of work and it was not necessary.
We also sought a second opinion from an editorial board member (R2) who, in contrast to R1, felt that the three articles should be retracted due to fraudulent behaviour and the authors’ results were not credible.
Since allowing the authors to provide potential corrections on specific gel images, the case of author A has received extensive publicity in various journals, which has made the decision in this case difficult and may have influenced our advisors. Also, the past behaviour of the author in several publications casts doubt over whether the revisions are trustable. We now feel we have three options on how to proceed:
1) Retract the articles because the authors have lost credibility; we have confirmed their misconduct and scientific dishonesty in duplicating control images and the authors have not repeated all the experiments that were questioned in two out of three articles.
2) Publish a correction notice on each of the articles to warn the readers that we have found that some of the control images have been duplicated and the results of specific figures should be treated with caution but state that the main message of each article is not changed. This would have to be issued as a ‘correction’ but the text would amount to an ‘expression of concern’.
3) Publish the corrections supplied by the authors (or at least the one with all the images corrected).
We are leaning towards the second option because we have somewhat conflicting advice on the value of the corrections supplied by the authors and we would like to warn the readers that the control images were duplicated.
Would the Forum agree with this course of action or does it have any other views?
The Forum agreed that the articles should be reviewed on a case by case basis. All agreed that the editor needs clarification of what is going on, perhaps by obtaining further expert opinion on the validity of the work. Is the duplication a result of laziness? The editor needs to establish whether the scientific record needs to be corrected. At the moment, the editor does not have enough evidence. The advice was to go back to the institution and ask them for the actual findings of the investigation. Why did the authors use duplicated images? Does the institution believe it was misconduct on the part of the authors? Depending on the findings of the institution, the Forum agreed that the editor should follow option (1) or (2).
The editors contacted the authors’ institution with a request for more detail on the actual findings of their investigation. Unfortunately we have not heard back from the institution despite sending several chases. We also sought further and more detailed advice from the editorial board member (R2), who still felt that all three articles should be retracted because of the authors’ misconduct and loss of credibility. After further discussion of the available advice we decided to retract the two articles where not all of the questioned experiments have been repeated, and publish a correction for the article where the authors provided new results and controls for all the questioned experiments.