A journal received an allegation of scientific misconduct from an anonymous individual stating they were from the group that had written the paper (Institution-1, there are two institutions involved in this research). The email stated that the scientific bases of the article were unreliable. The paper was currently with the authors who were revising the paper after the first round of review, and additional experiments were required.
The editors followed-up with the whistleblower requesting more information and their identity. The whistleblower emailed back, concealing their identity, but provided additional information, highlighting a specific component of the research as unreliable.
Institution-2 (the one that was not claimed by the whistleblower) was informed of the whistleblower. Institution-2 responded by saying that the authors believed there was an initial problem with the data used, but these had been updated and were not fabricated. Institution-2, however, was not the institution that carried out the experiments in question.
The editors made the decision to obtain more information. On resubmission of the paper, the three original reviewers looked at the manuscript but were not told about the whistleblower. All three were satisfied with the changes made to the paper and approved publication. A fourth reviewer was asked to look at the paper and told about the anonymous whistleblower. This reviewer found no clear evidence of fraud, but he could not assess the experiments in question. This reviewer did, however, raise new concerns about technical deficiencies in the work. Aside from fraud, these new issues made it unsuitable for publication in the journal.
The editors requested outside assessment from researchers with knowledge of the work to redo the informatics analyses to see if the raw data (included with the paper) gave the same results as the processed data. Again, there was no clear evidence of fraud, but there was difficulty in complete reproducibility due to poor methods descriptions and lack of access to all of the data.
At this time the whistleblower sent an email recanting his/her original statement and saying they have assessed the work and the authors have made the appropriate changes to fix everything. This was an odd email as there had been no change in the manuscript since the resubmission.
The editors ultimately decided to reject the paper based on remaining concerns of misconduct and the heavy criticism of the fourth reviewer.
The authors from Institution-1 requested a meeting with the editors. At this meeting the authors: expressly denied misconduct; showed the editors pages of data from the experiments of concern; and provided the editor with a copy of an email from an anonymous whistleblower that the leader of the group had received 2 days before the journal editor had received their anonymous email. This anonymous individual claimed to be from Institution-1, and stated that one of the authors (from Institution-1) had visited their institution and removed data.
An additional oddity is that the email addresses from the whistleblower to the editor and to the authors were created to indicate they came from the institutions claimed. Investigation of the email origins showed that the two anonymous whistleblower emails came from the same individual, not two different individuals as claimed.
Question(s) for the COPE Forum
• How should editors act on a tip from an anonymous whistleblower, where there is uncertainty about the unknown person's goals and that, should institutions take punitive action without investigation of the whistleblower’s intent, careers could be heavily impacted.
• If the institutions, when informed, decide to take no action, do the editors have a responsibility to investigate to get a better sense of whether they should further push the institution given the authors can simply submit a paper elsewhere where those editors will not know about the potential for misconduct.
• Due to the importance for the career of individuals on this paper, can the editors aid the authors submitting elsewhere, given that the whistleblower lied, but there is no way to disprove fraud. Or should the editors provide the information about the whistleblower to the institution.
This was a very complicated situation and the editors have no clear means to further investigate beyond what they have done already. The editors had gone far beyond what is normally assumed to be the role of the journal in such cases. The advice from the Forum was for the journal to contact the head of the institutions and the ethics board. The editor can use the Office for Research Integrity (ORI) website to check affiliations if the institutions are outside the US (there are some listings of cooperating national ethics approval boards). The Forum advised contacting the authors first before informing the institution. The institutions should also be informed of the whistleblower’s behaviour and the apparent falsification of email addresses to create the appearance of two whistleblowers.
Suggestions for another journal and assistance with revision might be appropriate in some cases. Because of the issues raised by review #4, the article was not suitable for this journal but there are other journals in the field that might accept a revised article. The suggestions for revision already provided to the authors will help them modify the article and correct any errors.
The editor informed the authors of the overall discussions at the COPE Forum, saying that they had done all they could based on COPE guidelines in terms of trying to assess the veracity of the claims (and detailed the standard COPE guidelines). The editor told the authors that he felt that at this point, further investigation into the situation had to move to their institution if they wanted to pursue that avenue and gain some resolution. Because of the serious nature of this issue, the journal told the authors that they had an additional referee look at the work, one who was very familiar with the journal. He/she raised additional concerns about the work from the standpoint of suitability to the journal, and hence the decision was made that the work was not suited to the journal.
The authors accepted the decision.