22

I was asked to review a paper which was submitted to a conference. The author of the paper was one of my former professors* and of course I declined to review the submission due to conflict interest.

I was wondering however, would it be unethical to write an email to my professor give him a feedback on the paper and perhaps, provide him some kind of unofficial review?

Of course it would not change the fate of the paper but if the paper gets accepted, they can use my pseudo-review when doing their final editing. I was not sure if this is ethical and/or according to the etiquette of academia.

Edit: Perhaps I should clarify what do I mean by feedback. What I had in my mind wasn't a complete review but rather some suggestions. Something in the lines of:

Very interesting stuff bla bla... however, I think it would be better if you had written sec. IV in this fashion and you add data-field foo to figure bar.


*I have a good relationship with professor but we never had any projects or publications together.

  • 3
    Why is it conflict of interest ? – seteropere Feb 27 '15 at 5:33
  • 4
    @seteropere, My huge respect and interest for this professor may blind me to see possible flaws in his paper. Is not categorize as COI? – Pouya Feb 27 '15 at 14:26
  • 1
    @Pouya While I cannot speak from personal experience, I believe it is not entirely uncommon for people to be asked to referee papers by people they are closer to than you describe, for example their students, colleagues or collaborators. – Jessica B Feb 27 '15 at 21:49
  • 1
    @Pouya: Did you explicitly agree to confidentiality, or did the email that asked you to review explicitly say you should keep this confidential? – cbeleites supports Monica Feb 28 '15 at 10:24
8

As user1258240 pointed out, if the paper is publicly available you can avoid the entire issue of how it came to your attention. But if it isn't, then you shouldn't get in touch with the authors directly:

  1. There's at least a mild ethical issue, namely why you read the paper at all, rather than discarding it once you recognized the conflict of interest. You would have had a legitimate reason to read it as a reviewer, but now you've put yourself in the same potentially awkward position as a reviewer (where you have to be careful not to take advantage of having early access to the paper's contents) without the justification. It may have been simple curiosity on your part, but it could also look a little suspicious, like you hope to derive some benefit from having read the paper.

  2. Identifying yourself to the authors could make this impression worse. The illegitimate benefit you're deriving could be getting to show off in front of the authors or establishing that they are in your debt. (On the other hand, sending anonymous e-mail with unofficial feedback feels creepy to me.)

  3. Even aside from whether it's actually unethical, you don't want to get a reputation as someone who pushes ethical boundaries. That can leave people with an uncomfortable impression, like you're a loose cannon who might cause trouble even while you're trying to behave ethically.

Instead of contacting the authors yourself, you could write to the editor and ask whether you could supply unofficial feedback, which the editor wouldn't use for deciding whether to accept the paper (but would provide to the authors). It can't hurt to ask, and the editor might agree. However, I'd avoid saying that you've already read the paper, since I think there's a good chance the editor will say "No, if you have a conflict of interest you shouldn't even read the paper in the first place."

28

There might be an ethical problem with telling the professor that you were asked to review his paper, but I think it's a minor issue: the anonymity of the reviewers is mostly there to allow them to give their opinion freely. If you are eventually not reviewing the paper then I don't think you are required to preserve that anonymity (even if you do review the paper, it's not totally clear that you must stay quiet about it).

And if the paper is available online on some archive or on the authors' web pages, then it is definitely ok to give the authors your opinion based on the online version if there is a plausible way in which you could know about the paper had you not been asked about it.

I feel that you main concern is that the fact that the paper was sent to you in error would give the authors an unfair advantage by giving them the opportunity to hear your opinion about it. Well, it is unfair, as are many things in life, but not un-ethical. Ethics is there to keep science working well and making progress, and by letting the authors know your opinion you are doing exactly that. Fairness is only relevant when there is competition for competition's sake, and science should not be that.

  • 24
    "would give the authors an unfair advantage" - over whom? And do I also get an unfair advantage by showing my paper to friends first for feedback? Do I get an unfair advantage if I upload it at arXiv and get comments that way? Let's not overthink things :) – xLeitix Feb 26 '15 at 15:53
  • 4
    Yes @xLeitix, that's my point... – user1258240 Feb 26 '15 at 17:33
  • 1
    "And if the paper is available online on some archive or on the authors' web pages": this looks to me like the most important point. If you can find the preprint online, just say you are sending feedback because you had a look at the preprint online. If you cannot find the preprint but can find information about the existence of the paper online (e.g., its title), you could ask your former prof for a copy, and then send comments. – a3nm Feb 26 '15 at 23:57
4

The question covers a wider issue than what you explicitly ask. To contact an author about a manuscript that you have not reviewed is not unethical as such but carries with it several other issues. You will essentially provide comments as any other colleague of the authors would except the comments would be unsolicited. If you contacted the authors as a reviewer outside of the review process the matter would be very different, of course.

First, there is the personal issue. How well do you know your former professor, do you think he/she will react positively to your comments "out of the blue"? I think it is worth thinking twice about providing comments on something the author would not expect you would necessarily see other than as a reviewer of course. If this happened to me and the colleague was a close friend with whom I feel comfortable discussing our respective science freely it would not be a problem. But, somewhere is a line where the comments would be come uncomfortable due to an uncertainty about the reaction.

Second, as an author, I would not expect the journal/conference to send out the entire article when asking for reviews. That behaviour is closer to unethical than what your question concerns. It is more appropriate to provide the title and abstract to the prospective reviewer so here I think the conference, in this case, should make their routines more strict. As you can see from this perspective, it is not your fault you have gained access to the entire manuscript so at least you are not, in my opinion, breaking any confidence, that has already been broken by "the conference".

So while I do not see any strictly ethical problems with providing comments, you should perhaps consider the action twice before continuing for the first reason above. I am sure there are differing opinion about this but the main problem here lies in the fact that you were given access to the entire manuscript even though you have not agreed to be a reviewer. In an open review system this would be perfectly fine but I would be quite irritated if I found my manuscript has been distributed to large numbers of prospective reviewers.

  • 7
    Second, as an author, I would not expect the journal/conference to send out the entire article when asking for reviews.I would. Sending the full paper with a review request (or at least making it available to the prospective reviewer) is standard practice in some subfields of computer science, including my own. – JeffE Feb 27 '15 at 15:34
3

That depends upon the guildelines of your review board. Consult the person or group that organized the peer review, and ask them if it would be acceptable practice. If you strongly wish to send this feedback, mention to the group that you will not mention the names of the other peer reviewers (if you know them) or any other facts about the review, and will only be addressing the content itself.

From there, it becomes an issue of whether or not the group will allow it, and act on the assumption that they will not until proven otherwise, to avoid being placed in a sticky spot.


You can also ask if the paper will be made public after the conference, and if so then you should be able to offer your critique afterwards without any ethical qualms, except to mention to the professor why you withheld the comments (to avoid compromising ethical guidelines).

I think you should be safe giving your professor this feedback as long as you don't breach the anonymity of the peer reviewers that are on his review board - but don't assume that is the case.

2

My gut instinct is to avoid doing this. Typically reviews are supposed to be anonymous and letting the professor know you've seen it breaks the anonymity of it all. You may be breaking the ethical guidelines of the conference.

Best case scenario: the professor appreciates your comments, and nothing else happens. Worst case scenario: it becomes known what you did and people (editors, other researchers, etc.) stop trusting you.

  • Thank you for interesting answer. Do you think your answer is the same if you consider my edit of the original post? – Pouya Feb 26 '15 at 15:19
  • 17
    I am trying to figure out what the breach of trust here actually is. If @Pouya emails the author, he has learned nothing more than that Pouya was originally planned to be a reviewer but declined. That does not sound like sensitive information to me - the potential reviewers are typically either well-known anyway (e.g., PC of a conference) or at least pretty obvious in case of most journals. Specifically, the author does not know more about who actually reviews his paper than before the mail. – xLeitix Feb 26 '15 at 15:50
  • @xLeitix, You have mentioned out very good points and some of my doubts in this comment and the comment you posted on the other answer. Would you kindly elaborate more by providing an answer? – Pouya Feb 26 '15 at 17:19
  • 3
    @xLeitix: in information theory terms I'm sure it tells the author something, but very little. At least, that you're on that editor's list of possible reviewers in that field (telling any anecdote about reviewing a paper would do the same), and that the actual reviews came from others on the list (which the author knows being just as aware of the conflict of interest). If the anonymity were a top secret matter of critical importance then best practice might be to say nothing just in case multiple information leaks add up to a pattern someone can analyse, but this isn't the secret service ;-) – Steve Jessop Feb 26 '15 at 18:09
  • 2
    @cbeleites Yes, this is also true for my field. Unfortunately, the abstract and invitation specifically does not contain the names of the authors, so I am also occasionally in the situation that I accept a review, find out that I have a conflict with one of the authors, and have to back out of the review again. It wouldn't call this a breach of trust of anyone, just an unfortunate practice that is not really one person's fault. – xLeitix Feb 27 '15 at 10:45
1

Summary: I don't see any particular problem in discussing the paper with the authors. But I do have a problem with an editor sending out full manuscripts before confidentiality is agreed upon. As a reviewer and researcher, I want to be able to regain my full freedom of speaking to the authors by declining to review.

Here, a diplomatic approach may be

  • to let the authors know you were asked and have declined to review, and
  • ask them whether they would like to discuss the paper
  • suggest to the editor to switch to a process where reviewers accept or decline based on title, abstract (and author names if not double blinded) only and do not get the full manuscript until they accepted to review.

  • In my field, agreeing to treat the manuscript confidential happens when accepting to review.

  • I assume that there was no process where the OP signed up as reviewer and at that point agreed to confidentiality. (Otherwise I suppose this question wouldn't have been asked)

  • I also assume that this question wouldn't have been asked if the manuscript was attached to an email that clearly says "keep this confidential".
    Although I object to this situation: I prefer to agree to confidentiality on a case by case basis rather than being bound to it by an email I cannot influence. The extreme case of what makes me uneasy here would the ridiculous situation that I'm bound to not talk to a close colleague because a 3rd party sent me some email.
    Being "spoiled" by the practice in my field, I'd now answer the editor in question that I do not want to receive confidential information such as the full manuscript before having agreed to confidentiality and I'm not available as reviewer in a procedure where even declining to review binds me to not talk to the authors about the manuscript. I'd also point out that the decide-by-abstract procedure practically avoids the problem. Yes, I do keep the abstract confidential but that is no practical limitation for my work as researcher: title and abstract typically do not contain information (in my field) that wasn't available already from conference proceedings/talks (I'm in a proceedings-are-unimportant field). There's typically nothing to talk about in the abstract, at least nothing that cannot wait until the paper is out officially.

  • Last but not least, IANAL, but from the editor's perspective I would not rely on a one-sided confidentiality "agreement" in unsolicited email (full manuscript attached to "do you accept for review?"-email as legally valid across all kinds of legislations. Explicit consent to confidentiality is IMHO legally much safer, ethically much better, and not so difficult to implement:

  • In my field, the decision to accept or decline to do a review is done based on title + abstract (+ author names). That way, I don't get particularly sensitive/confidential information before having agreed to treat it as such and the problem is avoided.
    Maybe you could suggest this to the editor.

  • Also from this point of view, it was the editor who violated the confidentiality. After all: there's no way to say how many other people got the manuscript without agreeing to confidentiality. Blaming the OP for discussing the paper with the author is killing the messenger of the bad news.

  • The major ethical problem on the OP's side was solved appropriately by declining to review a paper of a close colleague.

  • I then don't see any problem giving feedback to a close colleague. Discussing science and papers and manuscripts is what you do with close colleagues. Otherwise they aren't close collegues IMHO.

  • Just to be clear: I do make a rather large distinction between confidentiality towards third parties and being restricted in talking to the authors.

  • If I need to decline the review in order to be free to discuss with the authors whatever I want, that's fine. But I need the ethical and legal possibility to keep this freedom, so collaborations are not strained by reviews I didn't even do.
    While close colleagues shouldn't be invited to review at least for a couple of years, I know of several instances where this happened (oversight by overworked editor I assume).
  • 5
    You're assuming that somebody can only agree to confidentiality when they agree to review a paper. This is incorrect. It is perfectly possible to say "Here is a manuscript. You must keep it confidential. Do you agree to review it?" Indeed, this is exactly what happens in my field (theoretical computer science). Especially if this is done via a website ("By downloading this article, you agree to keep it confidential"), the editor is in no way breaching confidentiality. – David Richerby Feb 27 '15 at 12:45
  • 2
    What David said. I assume that any manuscript I receive, for any purpose whatsoever, is confidential by default. Don't you? – JeffE Feb 27 '15 at 15:35
  • 3
    Clearly, the conventions among fields differ. I normally wouldn’t even consider accepting a request to review a paper without seeing the full manuscript first (treating it as confidential, of course), as that makes it impossible to estimate whether it will be a complete waste of my time, and whether I can realistically meet the review deadline. – Emil Jeřábek supports Monica Feb 27 '15 at 17:12
  • 2
    To me, two things are confidential: the fact that person A reviewed the paper; the contents of the paper (assuming it's not already been made public, e.g., on ArXiv). If one has declined to review the paper, the first kind of confidentiality is moot (we agree on this). Confidentiality of the paper contents mean that they shouldn't be revealed to unauthorized persons. But it is obvious that the authors are authorized to know what is in their own paper. So, if you decline to review, there is no breach in confidentiality in discussing the paper with its authors. – David Richerby Feb 28 '15 at 13:13
  • 2
    The point about deadlines was that if I a get a 50-page paper full of technical combinatorial proofs, I know it will take an awful lot of time to review even if the paper is of high quality, and I (probably) wouldn’t know this just by reading the abstract. Problematic papers that call for a major revision or rejection can also take a lot of time and effort, however this is typically something that only transpires on closer scrutiny, so I can’t use it for the initial decision whether to accept a review request anyway. – Emil Jeřábek supports Monica Feb 28 '15 at 17:08
0

I've had anonymous reviewers disclose to me and I've always found it very uncomfortable. Almost as if they were asking for a tit-for-tat (an unspoken "I wrote a positive review of your paper, so I hope you return the favor in the future").

Unless the issues you found would be something that only you with your specialized expertise could provide insight into -- or there is something damning that needs to be said or people will die, I'd just hold your tongue. Wait until the paper comes out and then comment on it as if you've never seen it before.

Alternately, if the prof has posted a line in his CV with the paper's title and a note "under review" -- you could ask him for a copy of the draft and then send back comments as if you had just seen it de novo.

  • 2
    But the situation here is IMHO very different from what you describe: the OP declined to review exactly because there may be a conflict of interest. I agree with you in that if I am reviewer, and disclosure may seem a good option for whatever reason I'd always offer to the editor that they may disclose my identity (so I can discuss with the authors) The editor would then need to look for another reviewer. – cbeleites supports Monica Feb 27 '15 at 11:26
  • Although I find the rest of your answer interesting, as cbeleites has mentioned, the firs paragraph of your answer doesn't apply to me. I have already rejected the review so there is nothing to bargain for future submissions. – Pouya Feb 27 '15 at 16:39
  • I agree, it's just the preamble as to why disclosure in general is a bad idea. – RoboKaren Feb 27 '15 at 17:16
0

Of course you could send your feedback anonymously to the mentioned professor, using an anonymous email service. In that case you would not breach any guidelines as you would not have revealed yourself. In the email you could explain that you declined to review due to conflict of interest, but want to send some suggestions anyway.

  • 3
    What guidelines say that it's unacceptable to reveal that you were asked to review a paper but declined? – David Richerby Feb 27 '15 at 12:41
  • I am not saying that I personally think that as a reviewer you should stay anonymous. I just want to point out that if you are worried to breach anonymity (for whatever reason), there is also a way to provide feedback while staying anonymous. – Danny Ruijters Feb 28 '15 at 8:52
  • If you declined to review the paper, you're not a reviewer! – David Richerby Feb 28 '15 at 11:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.