Peer review is the part of the research cycle most people learn entirely by absorption, usually from whichever reviews their supervisor forwarded. It has two halves — getting reviewed, and reviewing — and the second is the faster way to get good at the first.
On the receiving end, the practical skill is separating the review’s substance from its tone. A blunt report often contains a fixable objection; a polite one can contain a request that would take six months and should be pushed back on. Response letters have a form: address every point, state plainly what changed and where, and where you disagree, say so with a reason rather than quietly not doing it. Reviewers notice.
On the giving end, the standard worth holding is that a review should let the authors improve the paper. That means separating what is wrong from what is merely different from how you would have done it, being specific enough to act on, and being explicit about your confidence — flagging that you cannot evaluate the harmonization procedure is more useful than a vague reservation about the methods.
Preprints change the timing of all of this. Posting to bioRxiv or arXiv before submission gets the work cited and read while review runs, and is expected in this field.
Related¶
Reporting guidelines — what reviewers will check
Pre-registration — registered reports