49

I have got a rebuttal for the conference and require to respond the questions reviewers asked. However, the response section says "Your response must contain at most 500 words." How can I answer almost 20 reviewers questions in 500 words? Any suggestions?

  • 4
    Can you group the questions so that a single answer will cover all questions in a subset? – Patricia Shanahan Feb 21 at 10:14
  • 2
    @PatriciaShanahan The reviewers comments itself tooooooo lengthy and require a detailed reply. – Skipper07 Feb 21 at 10:21
  • 4
    @Skipper07 if you say the task is impossible, then that is your answer. If you want us to help you, you need to be open to suggestions. – Maarten Buis Feb 21 at 10:27
  • 8
    Contact the convenor or bite the bullet. If you really can't group some questions, then that leaves 25 words (or two medium-length sentences) per question. That doesn't sound impossible. – henning -- reinstate Monica Feb 21 at 10:31
  • 4
    Depending on the questions, some might be answered with 'This has been resolved in the paper' or similar which uses few words. That is you don't put the actual answer into your answers but rather into the paper. – quarague Feb 21 at 11:52
106

I suspect I know which conference you are talking about (I am in the Program Committee) :)

Answering questions in the rebuttal phase of conferences is a newly emerging skill. Almost any paper faces the issue that you are facing - rebuttals are strictly word-limited, and the reviews / questions are waaaaay too long to clarify everything.

A good strategy for dealing with this depends on you understanding how the rebuttal will actually be used in the final phase of reviewing. Essentially, after your rebuttal, there will be a final phase of (online or physical) discussion among the reviewers and PC chairs to decide if the merits of your paper outweight its limitations, if the paper is important enough to warrant presentation at the conference, and so on. Many papers have nobody really willing to fight for it (i.e., everybody mostly sees limitations), and for these papers your rebuttal will not change anything. A handful of papers will only have people positive about the work, and in these rather rare cases the paper will be accepted no matter what you write. Some papers will have one or more quite positive people ("champions") and one or more people who are sceptical or downright negative. In these cases your rebuttal can play a major role in how the discussion goes.

Your rebuttal should be written for this specific case, and it should be written to give the champion(s) the best possible ammunition to counter the arguments of the sceptical crowd. That means:

  • Answer questions where you have a really convincing answer, especially if they come from the most positive reviewers.
  • If you identify factual mistakes or misunderstandings, clarify them directly and without the polite hollering one tends to do during journal reviews.
  • Focus on questions or concerns that multiple reviewers have, especially if you have a good argument.
  • Focus on concerns regarding methodology, validity, or importance of the work. Don't focus on presentation-level issues (the champion can always make an argument that these can be fixed in the camera-ready version, you don't have to point this out).
  • Remember that you don't get an actual revision for most conferences. Do not agree that some additional work will be needed (even if it's small), this will work against you.
  • Do not get into discussions that are a matter of taste or subjectivity (you won't change anybody's mind on those).
  • Do not feel like you need to address everything.
  • Skip polite phraseology ("Dear Reviewers, we are very grateful that you took the time to ..."). Get to the point.

Importantly, it's not similar to a response letter for a journal. You do not have to convince every single reviewer. Counter-intuitively, you need to focus your attention on the reviewers that may fight for your paper, and the ones that could realistically change their mind (in that order). Don't worry too much about a single reviewer that will never love your paper (especially one that does not like the kind of work you do or doesn't think it's important). You won't change their mind, and if the others champion your paper the PC chairs will likely side with the positive reviewers. Whereas in a response letter you will often "agree" with certain reviewer comments and fix them in the revision (even if you don't really agree), your goal here is to convince at least a subset of the reviewers that your current version is good as it is. Agreeing with the reviewer's concern may or may not get you sympathy points, but it will definitely not help you get the paper accepted.

| improve this answer | |
  • 8
    What does "PC" mean? I assume the C is for committee, but what's the P? Paper? Is this something specific to those fields that treat conferences as major publications (math and physics, I think), because I've never heard of the term in my field (biology) where conferences aren't considered that important. – terdon Feb 22 at 19:40
  • 3
    @terdon Program committee maybe? – curiousdannii Feb 22 at 22:49
  • 2
    Yes, that's Program Committee. – Clement C. Feb 23 at 8:00
  • @Clement C Yes it is program committee and the paper rejected as usual...One reviewer accepted and Other rejected and third was a quite indecisive... – Skipper07 Mar 3 at 7:24
8

You need to be strategic about the issues (including questions) you comment and don't comment on. Identify and address those issues that seem significant enough to sway a reviewer. Possible indicators for these issues are:

  • The issue is mentioned in a context which suggests a possible major limitation/flaw of your work.
  • The issue is mentioned repeatedly and/or with many words in the same review.
  • The issue is pointed out by several reviewers.

If one of the reviewers sticks out as more negative than the others, give the others some hints to help them "disarm" the negative reviewer in the discussion. Point out any factual mistakes and misunderstandings in the review, possibly by pointing to sections in the paper they might have missed.

| improve this answer | |
  • Thanks...I am answering the similar questions in a one go and then the other important ones. In the end, I am saying the remaining concerns will be addressed in the next submission...something like that – Skipper07 Feb 21 at 10:42
  • Sounds like a good response. – lighthouse keeper Feb 21 at 10:44
  • @ lighthouse keeper Thanks a lot for helping... – Skipper07 Feb 21 at 10:47
2

Is the 500-word limit suggested, or strictly enforced?

If the former, my experience is that you can simply ignore the limit, and include all answers to explicit questions where your answer is likely to be helpful. Of course, do not include answers that are not helpful. Since the reviewers are not required to read past the first 500 words, make sure to include the most important questions or points at the top, and more optional / unconvincing answers lower down.

This is the prevalent culture in my area of CS, where the limits in current top conferences are soft, rather than hard. I don't know about all other areas.

| improve this answer | |
  • definitely enforced...See the text it says 'at most' – Skipper07 Feb 22 at 4:52
  • 1
    @Skipper07 Well, in my field the limits sound serious, but they are really not (it says "your response should contain at most 500 words", or whatever, but most people don't actually pay attention to what it says). It sounds like for your case, the limit is hard. I'll leave this up in case someone else comes across your question and finds this answer useful. – 6005 Feb 22 at 4:55
  • @6005 It's quite common that one has to type the rebuttal into a web-page form, and then press a submit button, which will refuse to work if the response is over the limit, i.e. not unlike how comments work here. – Robert Furber Feb 23 at 3:48
  • @RobertFurber Thanks for the info. It appears to vary by field / conference. – 6005 Feb 23 at 3:55
2

Is there an editor/convenor who could have potentially backed this 500-word idea and whom you can address with this question? They will possibly answer vaguely in the sense of shifting the burden of decision on you. Or possibly not, understanding the impossibility. At a minimum, if they are sensitive organisers, they may reconsider this system for future offerings of the conference (for example, by capping the number of remarks reviewers can rise). Probably you are not the only one to complain, and the feedback can reach critical mass.

In your approach to the editor you might want to underline that the unreasonable and unfair overhead comes precisely from selecting what is important to react to: so "pick up the most important points" just identifies where the solution cannot be implemented at a reasonable cost, and not a solution in itself.

You are certainly not bothered by improving the quality of your work, but this has to be proportionate to time normally allocated for a event of similar kind. Twenty (!!) reviewers are not two or three, quantity changes quality. Work has to be workable after all.

| improve this answer | |
1

Some questions are just questions.

Some questions are suggestions for further work.

Some questions imply that reviewers may want to reject the work.

Address the last questions in your rebuttal and leave the others for now.

If there is an opportunity to give a revised work, then address the first group in the revision.

But it is really only the ones that seem to suggest rejection that are essential to address. If there are so many of those that they cannot be addressed within the limits then the paper may be doomed for now, but you can still give it your best effort.

| improve this answer | |
  • 2
    and some questions imply that the reviewers are not competent enough to review the paper... – Skipper07 Feb 22 at 4:51
  • @Skipper07 Seconding Buffy, it is indeed unfortunate that implications ought not to be part of a genuine reviewing process, and they do creep into it, because they are not recognised and/or held back by the reviewers themselves – XavierStuvw Feb 22 at 8:03
1

Here are some responses that are under 25 words, each.

  1. Fixed.

  2. Fixed, except for subpoint 2.e on transmission spectroscopy, where we disagree (see ref 43, page 36 for details).

  3. Not fixed. We are stating this as a possible interpretation of the data, with that caveat well stated.

You should be able to do the same. At least try. And of course if some answers are shorter than 25 words, than that allows for detailed discussion elsewhere. If they are minor issues, you may want to use a lot of "1", rather than debating not important things.

If things are still unreasonable, than (a) contact the editor/convenor to ask for relief. If that doesn't work, (b) pull the paper, ditch the conference, submit elsewhere. (Duh.) But something tells me from your initial squawks in the question and the comments below it, that your paper and you may need a little kick in the ass, that the reviewer is not being unreasonable.

| improve this answer | |
  • 2
    I think you're mixing this up the revision phase for a journal paper, where it's usually expected to answer every referee comment, but without a word limit. In the case of a conference rebuttal, it's usually not expected that every single comment is addressed. Doing so might actually be a waste of space. – lighthouse keeper Feb 21 at 13:22
  • +1 for concrete examples. Sentences can be abbreviated further: "We state this". Bullet point style. – Captain Emacs Feb 21 at 13:22
  • 4
    These are really bad things to write in a conference rebuttal. A conference rebuttal is not a response letter for a journal. – xLeitix Feb 21 at 14:32
  • Done is even shorter then Fixed. Also there is no need to add a "." ;) – quant Feb 21 at 23:42
  • 1
    @quant It's a word limit, not a character limit - so punctuation and word length have no bearing here. We're not golfing here, fortunately ;) – Klaycon Feb 23 at 0:03

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.