16
$\begingroup$

Suppose I am refereeing a paper and wish to remain anonymous as referee. Suppose I had trouble understanding some of the proofs in the paper and made a post on MathOverflow and received a very helpful answer, that in my opinion will be useful not just to me as a referee, but also to the author in order to improve their paper. It seems reasonable that when I point out the argument I received from MO I should acknowledge where I got it from. But in doing so I will need to point to my post, and therefore reveal my identity. Would it be appropriate in this scenario to make the MO post using an anonymous login, specially made just for that purpose? I would be interested in hearing what other ways to proceed may be.

$\endgroup$
12
  • 19
    $\begingroup$ Your suggestion to post using an anonymous account in such situation seems reasonable to me. $\endgroup$ – Stefan Kohl Jun 3 at 20:45
  • 4
    $\begingroup$ Ask me, will be happy to play sockpuppet for such occasion :) $\endgroup$ – მამუკა ჯიბლაძე Jun 4 at 2:02
  • 4
    $\begingroup$ In this older post Dmitri Pavlov raised another issue related to refereeing: Privacy of software-generated links to questions and answers. $\endgroup$ – Martin Sleziak Jun 4 at 4:19
  • 3
    $\begingroup$ I was thinking about tags suitable for this question. Maybe citation - since it involves citing a MO post - and maybe user-accounts - since creation of an alternative account is discussed. $\endgroup$ – Martin Sleziak Jun 4 at 4:20
  • 2
    $\begingroup$ To avoid duplicates in the future it would be useful to improve the title of this meta-post (so that it reflects more faithfully the question) $\endgroup$ – YCor Jun 4 at 9:04
  • 24
    $\begingroup$ You'll want to be careful to disclose contents from the paper on a public forum, as a referee you are bound to confidentiality; discussing with a colleague is permitted, a more public discussion is not. $\endgroup$ – Carlo Beenakker Jun 4 at 9:57
  • $\begingroup$ @Ycor I have edited the title to a better one $\endgroup$ – Valerio Jun 4 at 10:48
  • $\begingroup$ @Martin Sleziak I edited the tags as you suggested $\endgroup$ – Valerio Jun 4 at 10:49
  • 2
    $\begingroup$ @Carlo Beenakker Yes your point is well made. The question in my post is general enough that only addresses a step in a proof involving standard and well-known facts, so I don't think it reveals much of the result itself. $\endgroup$ – Valerio Jun 4 at 10:53
  • $\begingroup$ @StefanKohl, it's definitely an appropriate situation to use a sockpuppet account. The one caveat is that any references should be to the best answer and not to the question, lest one be accused of anonymously pointing directly to one's own item. $\endgroup$ – Ray Butterworth Jun 4 at 13:19
  • 7
    $\begingroup$ @Carlo depending on whether the paper is already public, for instance on the arXiv. It's then perfectly natural for anyone interested in the paper to ask a specific question about a specific step if it'sunclear (though clearly not ok to just outright ask if the paper/theorem at hand is correct). $\endgroup$ – David Roberts Jun 5 at 4:03
  • 2
    $\begingroup$ @Valerio If the previous comment addressed to me was made by you - but from the other account - maybe it would be better to remove that comment, so that the connection between the two accounts is not visible here. (After all, if you want keep the comment, you can post it again from the main account.) $\endgroup$ – Martin Sleziak Jun 5 at 4:30
14
$\begingroup$

I don't think there is anything inappropriate about using alternate accounts to ask questions. I have asked many questions using alternate accounts over the years (for all kinds of reasons, e.g. shame at not knowing something basic!). As long as your different accounts don't interact (and especially don't vote for each other!), this does not violate any rule that I am aware of.

That being said, I think it is a little weird to initiate an anonymous online discussion of a paper you're refereeing. I'm not sure I would go so far as to say it is unethical, but it's definitely not standard practice, and I think that many editors would get annoyed at one of their referees doing it.

$\endgroup$
1
  • $\begingroup$ Thank you for your comments. Of course my question was hypothetical. I would not initiate a discussion on a paper I am refereeing, whether anonymously or not. $\endgroup$ – Valerio Jun 4 at 21:33
9
$\begingroup$

On my opinion, if you have a question on the paper that you referee, you may ask only the author. A referee is not only anonymous but also is supposed to be confidential. Which means you are not supposed to discuss anything in the paper that you referee with the third parties, anonymously or not.

$\endgroup$
4
  • 5
    $\begingroup$ I'm not sure I would go quite this far. Several times when refereeing papers, I have asked colleagues/collaborators of mine for help understanding certain technical points. For many papers, there is no single person other than the author who deeply understands every input to the proofs! I don't think that you should do this in public, but for people whose discretion you trust it can be fine in the right circumstances. $\endgroup$ – Andy Putman Jun 8 at 20:12
  • 3
    $\begingroup$ With preprints being readily available, I don't see a problem with referees discussing papers (discreetly) with others. I can imagine there being a problem if the preprint were not freely available though. Any mechanism that helps the referee understand more thoroughly whether an argument is correct or not, I think it is to be welcomed. $\endgroup$ – Lucia Jun 9 at 0:52
  • 3
    $\begingroup$ I agree with both remarks. IF a preprint is available, of course one can discuss it with anyone. Discussion with a colleague you trust is different from a discussion n MO:-) $\endgroup$ – Alexandre Eremenko Jun 9 at 4:30
  • $\begingroup$ An anonymous account on MO is anonymous only for ordinary users of MO. Moderators know your IP, email and so on. Thus by asking a question on MO violates the privacy of refereeing. $\endgroup$ – Mark Sapir Jun 16 at 0:48

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .