9

My question is about submitting a paper to a venue (journal or conference) when some of the authors of the paper are also members of the program committee (PC) of the venue (i.e., they will review papers for the journal or conference). I call this a PC paper.

In any modern conference or journal management system that I have seen, PC papers don't seem to pose any particular issue: when logged in as a PC member, you do not see pending submissions of which you are an author, and of course you are never given one of your own papers to review. I have never participated to a PC board meeting (to decide on which papers to accept or reject), but I assume that PC members would be excluded from discussions about their own papers (as is routinely done for conflicts of interest). In conferences where the accept/reject decision is taken by a PC chair or editor, of course they are not allowed to submit papers, and this of course I can understand.

However, in my field (theoretical computer science), I have heard that submitting a PC paper is generally frowned upon. Further, there appears to be a tradition that, on borderline papers, papers by PC members will be rejected in priority.

Hence my question: Is there indeed a bias against PC papers in academia? If yes, is there any ethical or practical justification for this?

  • 1
    btw is there any difference between submitting papers to a conference where the author is PC member and submitting to a journal where the author is Editor? I mean on the ethical level of the question, not the practical. If it is common in journals, I do not see why would it be different in conferences. – PsySp Mar 8 '17 at 9:21
  • I think you discuss everything that is to say about the topic yourself. Yes, there are some communities were submitting as a PC member is frowned upon. In most communities, it's no issue, and some conferences even specifically encourage their PC to submit themselves. Academia varies. – xLeitix Mar 8 '17 at 11:21
  • @xLeitix: But in the communities where this is frowned upon, what is the justification for it? There has to be a reason, right? Or is it a completely gratuitous custom? – a3nm Mar 8 '17 at 11:46
  • @PsySp, in my experience it is fine for a journal to handle submissions from members of the editorial board -- they just get processed by other editorial board members. But editors of a special issue, or the editor-in-chief, don't get to submit. – Fred Douglis Mar 10 '17 at 17:14
3

No is not frowned upon, at least in our field (TCS). I can provide numerous examples for that. I saw it many times. PC members submit their papers, but they are handled of course by other members and, afaik, the whole process is not visible to them until the notification date.

Note that, some conferences (like IPCO) have strict rules. They explicitly say that no PC member shall submit a paper, alone or as a co-author. But if the Call for Papers does not mention, I guess a PC member is free to submit the paper.

What I have heard though is that even if a PC member submits a paper, it is claimed that higher standards apply to that paper (but I am not sure).

  • Thanks for your explanations! Do you know why higher standards would apply to PC submissions even for conferences that do not disallow them? For conferences where the call for papers disallow PC submissions, do you know why could be the reason for the conferences to disallow such submissions? – a3nm Mar 8 '17 at 9:27
  • For the first part, it is just an unproved claim. Personally, I do not believe it because at the end of the day it is just reviewers who make the review, comments, scoring and the paper is sorted analogously. I am guessing it is there to satisfy people who disagree with such practices. For the 2nd part, I cannot answer whey some conferences have such strict rules. I am guessing it's related to the very limited number of accepted papers (IPCO has 30) and they want to show complete fairness. – PsySp Mar 8 '17 at 9:31
  • 1
    @a3nm Also, probably, because in such (small, but very prestigious) conferences the whole PC committee discusses about all potentially accepted papers. Something like unanimity vote? – PsySp Mar 8 '17 at 9:33
1

If yes, is there any ethical or practical justification for this?

Especially if the venue only attracts of small number of submissions, one justification is reviewer anonymity.

In some conference systems, PC members can see all reviews except for those of their own papers. Based on this information, it would be easy for a PC member with an own submission to infer the identities of the reviewers of her paper.

  • I am not sure how this can be done. Could you please elaborate? – PsySp Mar 8 '17 at 20:26
  • Let's take an extreme example: If there are 10 submissions, 15 PC members and each paper should have three reviews, then usually each PC member reviews exactly two papers. So, as a PC member with a submitted paper, I can look at the reviewers of the other papers and just identify those reviewers who only have one review. – lighthouse keeper Mar 8 '17 at 20:45
  • You know that PC members send the papers for review to external reviewers right? – PsySp Mar 8 '17 at 20:48
  • Sure, but despite that, the PC member identities are normally not revealed to the authors of a regular submission. In this case, they are. – lighthouse keeper Mar 8 '17 at 21:06
  • 1
    I once raised this issue with the HotCRP review system, but I didn't convince its author to do anything about it. The problem was that even if every reviewer gets, say, 15 papers, if you see that some reviewers have 15 papers but 1 reviewer has only 12, and you're conflicted on 3 papers, you have a good idea that reviewer saw your papers. But it's all inexact. for instance, maybe that reviewer only was assigned 12, or assigned 14 but not 15. I agree that other PC members shouldn't have access to the number of reviews others have, at least if they have conflicts. – Fred Douglis Mar 10 '17 at 17:05
1

Adding my own answers to add some other reasonable justifications I heard from colleagues about why PC papers are discouraged:

  • If a conference has a significant proportion of PC papers among the accepted papers, it gives to outsiders the impression of being a closed "mutual gratification" community, where the PC is accepting their own papers. Of course, this perception is also tied to other signs of how healthy a community is (e.g., does the PC rotate sufficiently, etc.), but a simple way to avoid giving this impression is to reject PC submissions entirely or at least discourage them. This seems to me like a reasonable argument.

  • If the conference is not double-blind, reviewers of a PC paper may be influenced by the fact that authors are members of PC, and may also be influenced from their interactions while discussing reviews of other papers (where you may get a positive or negative impression of other reviewers). This problem can be patched, of course, by making the conference double-blind.

  • I have seen this scenario. In the first 5-10 or so years of the conference, almost all the papers were written by PC members or their Ph.D. students. (Reviewing was not anonymous.) Gradually as the conference intellectually matured, larger numbers of papers by non-PC members began to be accepted. The same was true of a new journal in that field. – E. Douglas Jensen Dec 21 '17 at 20:24
1

It does seem to vary depending on subdiscipline. In systems conferences, it used to be acceptable for PC members to submit, but they were often discouraged from submitting too many. But I remember one conference where the chair said specifically that PC members shouldn't be discouraged or discriminated against -- hurts young faculty in particular.

Modern review systems do a good job of handling anonymity, but any conflicts make any face-to-face meeting more challenging. HotCRP is way better than easychair for this by the way.

With cochairs, even a program chair can potentially submit to their own conference, though it still can look bad when it gets in. But it does help with institutional conflicts.

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.