A research article, published several years ago, was alleged to have integrity issues relating to some of the figures. Following detailed checking of the figures, the editors confirmed these issues and found more issues. The authors provided the raw data, some of which also had integrity issues. An institutional investigation concluded that the data were sound but the editor still feels uneasy about the manuscript.
The journal asked the authors to provide an explanation for all of the issues, as well as the raw data underlying the experimental data. The authors replied that there were errors during the processing of their files, have provided corrections for all of the issues, and also indicated that their findings have been reproduced by other laboratories.
The raw data that the authors provided match all of the corrected files, but the journal has received no original raw data files for two figures. Also, in the replicate data provided by the authors for one figure, duplications across replicates were found as well as a duplication across two different experimental conditions. There were no duplications in the corrected files.
In parallel, the journal contacted the institute director, who initiated an institutional investigation. The institutional committee said that they checked all of the original data and found no signs of manipulation. The institution also confirmed that the corrected files were in order.
Question for COPE Council
- The editors believe the paper should be retracted based on the serious issues with the figures and raw data. Can the results of the institutional investigation be ignored?
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.
A journal has the right to retract an article, even when an institutional investigation does not find misconduct. The editor is responsible for the ethical standards of their journal, and if the editor has independent evidence that warrants correction or retraction, it is appropriate for the editor to express concern and retract, regardless of the findings of the institutional investigation. While an institutional investigation should not be "ignored", the recommendation from an institution is just that—a recommendation. Unfortunately, some institutional investigations are conflicted or inadequate. The journal cannot assess the adequacy of the institutional investigation because each university has different standards for investigations of misconduct and so the editor cannot verify how rigorous the investigation was. A statement along with the retraction can explain that there were additional concerns not addressed by the results of the investigation.
The journal might choose to: highlight to the institution that this laboratory needs training and improved processes in data handling; check whether the right people conducted the investigation; and escalate the case to a national oversight body. The editor could also consider informing the other authors' institutions.
Retraction does not require evidence of misconduct, but rather major error. The journal can state in the notice the factual issues with the figures, replicates, and raw data. The journal could note that the investigation found no misconduct, and the authors provided data to the journal and provided a correction, but the editor was not satisfied by the revisions and decided to retract, although the institution and authors disagree with this course of action.
The editor may wish to consult the relevant extracts from the COPE Retraction Guidelines:
"clear evidence that the findings are unreliable, either as a result of major error (eg, miscalculation or experimental error), or as a result of fabrication (eg, of data) or falsification (eg, image manipulation"
"When editors or journals have credible grounds to suspect misconduct, this should be brought to the attention of the authors’ institutions as early as possible, but the decision to correct or retract an article should be made by the journal and does not necessarily depend on an institutional finding of misconduct".
"When possible, the outcome of institutional investigations should be quoted from and cited in the notice, and any findings of misconduct should be appropriately attributed to the institution who made the finding."
There is also a relevant case in the COPE database:
Editors, representing the readers, are stewards of the literature. If the editorial team believes an article is fatally flawed, for scientific or ethical reasons, then it should not be published, or if published, it should be retracted. While it is desirable to have the institutions on board, it is not a precondition.