When peer review leads to invitations to revise-and-resubmit, editors commonly request that a revision memo accompany the returned paper. Faculty increasingly request revision memos as a result of student peer review processes as well. The revision memo specifies how the author(s) responded to the reviewers’ and editor’s feedback. Even if a revision memo is not explicitly requested, submitting one with any R&R is good professional form. This is especially true if the reviewers’ recommendations included conflicting feedback or other significant stumbling blocks such as conceptually or methodologically impossible suggestions. You will get more of these than you might expect, simply because not all of us are experts in any particular methodology or conceptual field.
Most revision memos briefly discuss any global adjustments then respond in bullet form to specific recommendations from individual reviewers. I recommend addressing anything the editor explicitly required first, then consider the remarks of each reviewer. Reviews are anonymous, but the files are usually identified as “Rev1, Rev2,” etc., and you can reference these in your memo. Focus on significant changes, not correcting typos or otherwise line-editing; that said, reducing the manuscript length by a significant amount is worth noting. Some journals send revision memos with the revised manuscript to the second-round reviewers; at other journals, they’re for editorial use only.
You should address each significant reviewer recommendation in your memo, but remember that this is your paper. You do not have to do anything to it that you do not want to, but accommodating reviewer requests where feasible is both expected and a good way to increase the chances of publication. If you choose not to act on a particular recommendation, you should explain your justification in your revision memo, and in many cases, you should add a footnote to the paper containing the suggestion and why you chose not to do it. Remember that the editor and reviewers had good reasons to request what they did; you need to have equally good reasons to not accommodate their requests. No such publicly available data exist, the requested options are not possible with the model employed (i.e., you can’t employ clustered standard errors in seemingly unrelated regression models or use logit as the first stage of a selection model), no additional cases meeting those criteria exist, or similar matters constitute appropriate justifications. An editor may have explicitly required certain revisions as conditions of publication, but even there, if you have significant reasons not to do these – as in, the suggestion is methodologically impossible or conceptually invalid – you can choose not to do these things if you provide sufficient justification. I suggest contacting the editor directly to discuss these situations. She may request that instead of you doing the required revisions, you discuss in the paper or in a footnote why this is not possible.
At the student level, insufficient time to conduct requested additional research is usually an acceptable explanation, particularly if it requests the addition of several variables (in quantitative or qualitative work) or new cases (in qualitative work). At the professional level, this is generally not an appropriate response. If the feedback asks you to essentially start over or collect massive amounts of new data simply for a secondary robustness check, you can usually express this (politely) to the editor and indicate that while it is possible, it is not feasible in the time allowed for the R&R (if a deadline is expressed), that you have other support and robustness checks for the questioned variable, or similar things. If the editor explicitly required that you address or act on these particular suggestions, you will want to contact the editor directly and discuss the situation. It may be possible to consider acceptance on the basis of the existing data, or with limited additional data collection in the timeframe allowed for revision, but these will generally require negotiation. Be prepared for the editor to say no; depending on your timeline, you may choose to withdraw the paper from consideration at the initial journal and begin the review process anew at a different journal. You’re essentially betting that the review process at the second journal will take less time than doing the requested additional work, but sometimes that’s a risk you may need to take.