"Reviewer 2" is shorthand for the reviewer whose report is the hardest to act on — not necessarily the most negative one, but the one whose comments are underspecified, inconsistent with the other reviewers, or delivered in a way that reads as dismissive. Responding well starts before you write anything: read the full report twice, then sort every numbered comment into one of four buckets.
Sort every comment first
- Valid and specific. The reviewer names a real gap — a missing baseline, an unclear method step, a citation you should have included. Fix it and say so plainly.
- Valid but vague. The reviewer is pointing at a real problem but hasn't said exactly what they want changed ("the related work section is weak"). Your job is to figure out the most likely intended fix and respond to that interpretation explicitly, so the reviewer can correct you in the next round if you guessed wrong.
- Misunderstanding. The reviewer's comment suggests they missed something already in the paper. Don't say "as we already stated" — that reads as defensive. Point to the exact section, and consider whether the fact that a careful reader missed it means it needs to be more prominent.
- Out of scope or unreasonable. The reviewer is asking for a different paper — a new dataset, a different research question, an experiment outside your resources. You can decline, but you need a substantive reason, not just "this is out of scope."
Respond to each bucket differently
For bucket 1, state the change and where it landed in the revision ("Added in Section 4.2, lines 210–224"). For bucket 2, state your interpretation and the change you made based on it. For bucket 3, correct politely and point to the source. For bucket 4, explain the tradeoff: why the suggested change is out of scope, what it would take, and — if there's a smaller version of the request you can accommodate — offer that instead of a flat no.
Tone rules that actually matter
- Never use the reviewer's exact critical phrasing back at them ("as the reviewer incorrectly claims"). Restate neutrally instead.
- Don't apologize repeatedly. One acknowledgment per substantive point is enough; more reads as filler, not humility.
- Match specificity to specificity. A one-line comment gets a proportionate response — a paragraph of defensive justification for a minor point looks worse than a two-sentence fix.
- If two reviewers disagree with each other, say so in your response to the editor's summary, and explain which direction you took and why. Don't silently pick a side in each individual response.
A short checklist before you submit the response
- Every numbered comment has a numbered response — no comment left unanswered.
- Every claimed change has a location in the revised manuscript (page, line, or section number).
- No response is purely emotional — each one states a fact, a change, or a reasoned tradeoff.
- The cover letter to the editor summarizes the biggest changes in three sentences or fewer, for a reader who won't read the full point-by-point.
If you want a second pass before you send it, Response to Reviewers runs a three-persona panel — Skeptical Reviewer, Tone Editor, Editor-in-Chief — against your draft response and flags tone problems, missing responses, and comments you hand-waved instead of actually addressing.