A paper was published in July 2012. The author was told by their institution that one of the figures had to be replaced, in the interests of national security. Failure to do this would result in imprisonment. The editor checked with one of his reviewers who said that replacing the figure will not affect the results or conclusions of the paper.
So, can we replace the published version directly in order to avoid further dissemination of this figure or should we republish this paper? Or should we withdraw the paper? Is it possible to block the paper to avoid further dissemination and then republish this paper with the new figure?
As we were unable to contact this editor on the day, it was agreed that COPE council would provide advice and forward it to the editor.
Council advice was as follows.
This is a confusing cases and several council members were concerned that they were not clear what the whole story was and suggested that the editor needed to be really sure that they agreed the figure needed to be removed. The suggestions below are mostly about process therefore.
There are several options to that the editor could consider.
Most council members agreed that once a paper is published, even if the first publication is online, it should not be changed without a clear notice of a correction as this undermines the integrity of the publishing record. If something subsequently needs to be changed, a corrigendum must be submitted to address an inaccuracy, omission.
Another suggestion was to withdraw the current paper and publish the new one after the manuscript has been peer reviewed. But all of authors on the original paper would need to agree. However, the problem with “withdrawing” a paper and publishing a new one is that the publication record becomes rather confused. Will the new version have the same DOI/citation or a different one? If the same, how will readers know that they are not looking at the same version as the one someone else perhaps saw and referenced last week? Therefore COPE council does not recommend this action.
In this specific instance the editor could replace or remove the figure provided that the overall conclusions are not affected (this is really critical). Two possible processes are outlined below based on what different journals do in correcting errors
- Some journals institute an erratum process that involves changing the online version so as to eliminate the error. At the time the corrected version goes live, publish an erratum stating what the error was and that the online version is being corrected. The corrected version of the article itself also carries a statement that it has been corrected and when.
- Other journals would remove the figure with a corrigendum, without replacing the full paper.
The editor agreed to follow the advice of the COPE Forum. He will replace the published paper with a corrected figure, and also include a note explaining why this has been replaced. The editor also plans to publish a separate correction notice.