The reviewers are always right…even when they are wrong!-

Science has a culture that is inherently cautious and that is normally not a bad thing. You could even say conservative, because of the peer review process and because the scientific method prizes uncertainty and penalises anyone who goes out on any sort of a limb that is not held in place by abundant and well-documented evidence.
– Al Gore

One important aspect of scientific research is dissemination of the results through peer-reviewed publications. In a previous post, I discussed the choice of venue and the relative (un)importance of the journal impact factor. In this post, I address what happen in-between that first submission and the actual publication.

Revision as we call it, or the process of making modifications to your manuscript based on the comments made by two or three reviewers (and the editorial team) is an important step for the key reason that your own work is obvious to you and it must become fully understandable to all of those who are not familiar with it. Usually a central aspect of most reviewers comments is the issue of the missing details (which again are obvious to you but not to those outside your immediate team).

The famous rebuttal letter is a document that is almost as important as the manuscript itself. It must be crafted with care, using a polite language. All of the reviewers comments should be addressed one by one. Furthermore, it is entirely possible to add materials to the rebuttal letter that is not part of the manuscript to convince a reviewer that your results are sounds (it could also happen that a reviewer ask you in return to make those results available in a supplemental document). Once, in my time as a postdoc, the rebuttal letter was over 8 pages longs for a 4 pages Physical Review Letter so that the reviewer could see all of the intermediate results supporting those selected for the PRL document. 

There is but one golden rule that I tell all of my students in replying to reviewers: the reviewer is always right…even when he/she is wrong. 

That does not mean you cannot disagree with a reviewer, it happen quite often actually, but the fact that the reviewer made a certain comment or took a certain stance with regards to part of your work (or sometimes all of it!), even if you think it is wrong in the first place, rest entirely on your shoulder as the author(s) of the manuscript. 

Let me explain myself:

1. After all of the internal reviews, discussions with your advisor and colleagues, it is entirely possible that you (collectively) have missed something. Thus it is important to go carefully over the argument put forth by the reviewer.
2. It is entirely possible that the reviewer did not understand correctly the message, data or the explanation put forth in your written document. While the reviewer in this case, is indeed wrong, the problem still rest on your shoulder to provide a clear and understandable written document. 

In the latter case, you need to thank the reviewer for pointing out a part of your manuscript that needed to be better written, explained or substantiated, basically making your manuscript stronger.

Note that reviewers do this task for free as a service to the community. I would content that 99+% of the time it is done with good intent and really in the spirit of making the document read better and easier to understand the main message. 

If for any reason, you think a reviewer is having a non-ethical behavior (especially unsubstantiated rejection of your work with a one liner review or invoking arguments that cannot be argued scientifically – i.e. a belief, you should report it to the editorial board. Most serious journals would have pick it up before you anyhow.

In my scientific life, I have a bit less than 8% of my manuscripts that failed after submission and never got published for one reason or another (too early for the field, too much work to satisfy the reviewers, included as part of a larger subsequent body of work…) for about 225 published one. Of those published manuscripts, about 4 or 5 needed an appeal to the editorial board of the journal a some point of the review process for the exact reasons given above…and, at least in my case, they all got published.