Somebody once told me: “if a reviewer asks for certain changes in your manuscript, you must follow those instructions or your article will get rejected.” Is that indeed the case?
I think that it depends. Very often, once we have overcome our justified anger that this nitwit of a reviewer dares assume that our carefully written manuscript is anything less than perfect, it actually turns out that the reviewer has a good point and that incorporating the suggestion will, in fact, improve the manuscript. So that’s the easy case: do it! (And say “thank you” in your cover letter).
Then there’s the case where a reviewer wants some inconsequential change, for reasons that move them but otherwise remain obscure. Well, if it matters that little, humor them. After all, they just spend some valuable hours of their lives going through your manuscript. (Yes, I know it was flawless, but they had to read it anyway before they could acknowledge your brilliance).
Finally, there’s a case where the reviewer wants you to make some changes, but you strongly disagree with what they say. Some people will argue that you just have to bite the bullet. “The referee said it, so we must do it otherwise the editor will reject it”.
I beg to differ. A good editor does not simply count reviewers’ votes. (Which is why I prefer to call them “reviewers” and not “referees”). A reviewer advises the editor. Nothing less than that, but nothing more either. Sometimes you see a reviewer write something like “I cannot accept this for publication” (or even “I reject this for publication”). Well, that’s a reviewer who doesn’t understand the editing process… Reviewers give advice, editors make decisions.
Anyway, so a reviewer makes a comment and you disagree. Again, several cases are possible. To start with, “that idiot does not understand what I am talking about”! Calm down. Yes, your manuscript is flawless, but if a reviewer who is committed to read your manuscript attentively misunderstands something that you wrote, then the risk that a more casual reader misunderstands you in a similar way is probably quite considerable. In this case, I gingerly suggest that the possibility exist that it might be feasible to reformulate your text a teeny tiny bit, just to make it clearer for that dumb reviewer. My advise to my graduate students (and to authors in general) always is that your writing must be so clear that it is impossible to misunderstand, even by a hypothetical somebody who actively wants to misunderstand you. Hence: revise and tell the reviewer (and editor) that you regret not being clear enough and have clarified your text.
Another case is when a reviewer disagrees with your interpretation of your data. You might then consider to address this not only in your (detailed) cover letter to the editor outlining your responses, but actually incorporate this in your discussion. Again, if the reviewer comes up with this, some reader might do the same. So the best course of action is to address the issue and explain clearly in your manuscript (and without telling the reviewer how stupid their suggestion is, for some reason they often don’t react well to that) why you disagree with their point of view.
And then there is the case when a reviewer really has an opinion with which you strongly disagree and which you think is clearly wrong. In that case, it is perfectly acceptable to state in your cover letter: “Not done”. Of course, you then dispassionately explain why you disagree and why you think this part of your manuscript should not be changed.
Once your revised manuscript and its cover letter have been submitted, the editor may decide to send it out for re-review, most likely to the same reviewers, or they may decide for themselves whether or not they agree with your reasoning and make an editorial decision of accept or reject based on their own judgment. This is why a clear, dispassionate, and well-reasoned cover letter is so important when revising your manuscript.
Of course, very often you have to go through multiple iterations of this process. And while this whole process is often quite painful, it helps to keep in mind that almost always, editors and reviewers are trying to help you improve your manuscript. Sure, I have also heard all the horror stories about incompetent and/or malevolent reviewers, but in my own experience as an editor, I have only once or twice encountered a bad-faith reviewer. At worst, a reviewer writes something like “good article, publish it”. That’s not an advice that is very helpful for either the authors or the editor (who based on this brief comment cannot even be sure that the reviewer actually read the manuscript) and is likely to be ignored completely.