Editing of reviewer comments: Forum discussion topic March 2020

We instruct our Editors (mostly in-house staff) that they should only ever edit a reviewers comment if there are inappropriate comments (e.g. defamatory or personal comments) or they need anonymising. If there's something in the report that's unclear, we ask Editors to advise/explain in their own comments, rather than edit the comments of the reviewer.

I think it is ethical and at times necessary to edit reviewer comments. I do think that the editor should provide clarity that the original comment has been modified. For example, if "the reviewer recommends inclusion of their own work in the reference list without clear reasons." [Ed note: Portions of this review were redacted to exclude extraneous recommendations that the authors cite the reviewer's own work.]

I think an Editor always has the right to contextualize reviewer comments. Another example you gave: The reviewer's comments are very different from those of the other reviewers and it seems that the reviewer did not understand the paper. The Editor can certainly include context: "While Reviewer 3's comments are different from the other 2 reviewers, I felt it may be beneficial to allow you to read them. While I am not convinced that all of these criticisms are legitimate, it is important to show that the article was interpreted in a much different way, so you may want to review some of the areas where the confusion may be caused by your language choice or framing."

It is best practice for an author to respond to all reviewer comments, if even to say "I disagree with this." So, the editor could also give a pass on responding to reviewer 3 or ask to have a call or further discussion on the full suite of comments. Alternately, if the editor modifies some of reviewer 3's criticisms, that should be noted. [Ed note: I have edited reviewer 3's comments because some of the points raised were not relevant to your submission. Please respond to all reviewer comments provided.]

This has come up recently at my organization, and I feel that editorial oversight includes an obligation to ensure that the feedback an author is receiving will be helpful to revision. When (if) we decide to allow reviewers to post their reviews after an article is accepted, then the full suite of comments would be available for all to read. Presumably, those authors keen to share their reviews would not be the ones with vitriol or flagrant self-promotion. I think the authors would still appreciate the curated touch over full transparency.

Firstly, my declaration of interest: I am a non-member, neither publisher nor editor, but frequent author and peer reviewer of research papers over the past 40+ years, with views on the changing landscape of scientific publishing.

Addressing the questions - I think it is not only acceptable but imperative for an editor to edit the content of a peer review when it a) does not follow guidelines provided for such reviewing, or (where those guidelines, if they exist, fall short) b) includes criticism of the article that is based on unsubstantiated opinion rather than evidence-based fact.  Criticism needs to be constructive; without example it is useless. It can be harsh but needs to be evidence based; thus “the authors show a poor understanding of current research given their emphasis on the findings in ref 5 which have subsequently been shown erroneous (Smith et al 2018)” might be acceptable, but “the authors show a poor understanding of current research” without elaboration is not.  Unfortunately, it seems that there is an increasing over-reliance on peer reviews, collected and administered without regard to quality by a professional editorial assistant with minimal editorial input.

Let me provide an extreme example. With foreign colleagues I was involved in work written up for a Springer journal. After a very long delay (another story) we received a rejection based on a single two-sentence review, of which the second was “Unfortunately the data reported are not new and do not add more to the actual knowledge.”  Our appeal addressed several points which can be condensed to the following: the review was inadequate, in both content and number, did not follow the publisher’s guidelines for peer review (which existed, although were hard to find), and was demonstrably wrong. However, given an adequate editorial process, we should never have received such a review or to have had to waste time responding to it.

This was indeed an extreme example, but it would be valuable to consider why it happened.  My unconfirmed belief is that it was assigned to a deputy editor who had difficulty in identifying peer reviewers and did not attempt to assess the merit or otherwise of the one review (s)he received.  Our efforts to respond were frustrated by the fact it was impossible to contact the editor directly - emails to the editorial assistant rarely generated the promised response from the chief Editor.

I see all these problems as being a consequence of the formalisation of peer review in which procedures aimed at simplifying the review process discourage the handling editor from becoming more involved.  Some editors appear de facto to devolve responsibility for a final decision to peer reviewers, rather than making an informed decision based on a critical assessment of the peer reviews and their own reading of the manuscript.  In this environment a single negative review frequently trumps multiple positive reviews. Of course, there is some pressure to find reasons to reject in journals with limited space, but this is not the way to do it.

This all reflects a wider issue: COPE may have an impact on preventing a deterioration in authorial responsibility; however there is little attempt to counter a deterioration in editorial responsibility to authors.

I think it's right for editors to correct spelling errors and typos in reviews without notice. Sometimes, when reviewers use language that's intemperate, it makes sense to revise words that might give offense while keeping the remainder of a review intact - and some reviewers can be quite rude, while making sensible points elsewhere.

But I would never think it right to change a review which is contrary to other reviews. Rather, as the commenter above suggests, it is appropriate to provide guidance as to the revisions that are necessary, by (for example) suggesting that an author makes spelling and grammar corrections suggested by one reviewer (usually linked to detailed page numbers, etc), before making more serious structural changes - or vice versa.

As editors, we have to make sure that the instructions given to authors by reviewers are reasonably easy to follow. And when we receive contrary reviews - 'minor corrections' from one reviewer, and 'reject and resubmit' from another is not uncommon - it is up to us as editors to give guidance, without changing the sense of the reviews.

Edits: I do occasionally edit a reviewer comment, if it has many typo's or the sentences are badly constructed (non-English speaking reviewers). If I really do not understand the review, I ask the reviewer for clarification. This editing happens perhaps in two-four reviews annually of the > 200 that pass my desk.

I may consider not to send a review to an author, if a reviewer's comment is inappropriate from my point of view (not constructive, a one-liner). In that case we seek another reviewer for the manuscript, because we make sure that we always have at least two reviews per manuscript.

If the review is inappropriate, I may even suggest the reviewer to do an on-line reviewing course (SpringerNature, Publons) to improve his/her skills. This happens perhaps once a year.

I was Reviewer Selection Editor with Taylor and Francis. I think the peer review reports are intellectual property of the reviewers; so the editor will not have any rights to make corrections. Ethically the reviewers need to be informed, even when the review reports are corrected for defamation and abusive contents and sent back to the reviewers for consent before sharing the corrected review reports. When they don’t consent the reports can be paraphrased by the editor and the editorial opinion can be shared with other reports or a third report could be sent. Sometimes I have come across single sentence review reports that carries no valuable comments. These reviewers can be asked to add points to justify their decisions and recommendations about the manuscript, and when they do not reply the reports can be quashed. Similar stance could be taken on reports when the reviewer who used abusive and defamations as well. However, the editor can draw opinion from such reports and seek another report. I think reviewer consent is highly recommended to edit and use a report; and this make sense in these times when open peer review reports and credits to review reports are being given on par with a published manuscript. These opinions are my personal views. 

It's not just ethical, it's required if a journal is to maintain a sense of fairness, humanity, and professionalism. Reviewers' comments represent the journal. If they are ignorant, mean-spirited, or hurtful, then this is how the journal will be perceived. Sometimes, they simply are inaccurate or incorrect. Many journals have thousands of reviewers, some with much more experience than others in different areas; just as not every article is perfect, not every review is. We seek to provide a good author experience; unprofessional or inaccurate reviewer comments have the opposite effect. This still leaves plenty of room for reviewers to politely and professionally find fault with an article, and to make constructive suggestions. We ask all of our editors who work with reviewers to read reviews carefully and make sure that they present each reviewer's comments in a way that is compassionate, accurate, and helpful. Sometimes we fall short in this task. But such is our aspirational standard, and we try hard to achieve it.

My goal in the verdict letter to the authors is to 1) explain to the author why the paper is being rejected OR ii) offer to potentially publish the paper and provide the author with optimal suggestions to turn their manuscript into a masterpiece. I do not think that it is wrong or evil to delete comments from the reviewer that I do not think will contribute to the latter. The other option is to add into the verdict letter that I do not suggest that the author heed that particular piece of advice but  that seems more disrespectful to the reviewer. I also often add "Editor's comments" to the end of the verdict letter if I think that the reviewers missed points that are important to me. 

As a journal editor, I have always encouraged in house and academic editors to resist the temptation to edit reviewer reports. If a reviewer's report contains hostile language, ad hominem attacks, unethical requests, or in some other way contravenes the ethics and responsibilities of a reviewer, I have usually returned the report to a reviewer with an explanation of why the journal cannot use their report and offer them an opportunity to amend their comments so that the journal can use the review. This is particularly the case when sound scientific comments are lost in hostile language. In my experience, reviewers usually take the opportunity to amend their report, often apologise for the inappropriate tone, and hopefully improve their understanding of what is acceptable or not to include in a future report.

For reviewer reports that contain typos and language errors, providing the scientific points being made are clear and unambiguous I would not edit a report. If the language issues have hampered my' understanding of the points being made, I would return the report to the reviewer for clarification.

In general, I do not think that editors should edit reviewers' reports. It opens up opportunities for inappropriate changes to be made to reviewers' comments and it blurs an ethical boundary. If the content of a reviewer's report is inappropriate, I believe it is better practice to use the situation as an opportunity to educate a reviewer, by returning their report to them, and giving them the opportunity to remove the hostile language or mistakes included in their report, and to restate their comments more appropriately and in their own words.

The editor's job is to use the peer review reports to help improve the paper and help in the editor's decision-making about the paper, rather than just to forward reviews to authors. If there's a way to make the reviews more professional and helpful, then a policy can be set and made clear in the journal website and reviewers' instruction letter. A journal could say it routinely corrects spelling and grammar mistakes and deletes recommendations/comments meant as notes for the editor. Potentially libelous comments, confusing advice, offensive remarks, factual errors, potential coercive citation, etc could be highlighted and sent back to the reviewer for revision by a deadline, after which if there is no reply, those portions would be deleted or the review not used. If any advice conflicts with that of another reviewer, whose advice the editor prefers, the editor's letter to the authors can state which item to ignore. The item in the review could also be footnoted with an Editor's Note explaining why the advice should be disregarded. If the reviewer is asked to review the revision, they can receive a copy of the editor's letter to the authors as well as the footnoted review, to understand why one item was ignored (they might also receive the other reviewer's report as a reference). Alternatively, the editor could delete that item from the review before sending it to the authors, and also send a tracked copy to the reviewer, thanking them for the review and just notifying them of the edit with a reason (eg, editor favours another approach, or the advice goes against journal style).