The invitation arrives: a journal editor wants you to review a manuscript. It is a genuine professional compliment — and, if it is your first time, a little daunting. Peer review has no formal training, so most reviewers learn by imitation. This guide gives you the shortcuts: when to say yes, how to read a paper as a reviewer rather than a reader, and a report structure you can reuse every time. It also spells out the etiquette that separates a helpful reviewer from a dreaded one.
Before you accept
Say yes only if all three of these are true:
- Competence — the manuscript is genuinely within your area. You do not need to be the world expert, but you should be able to judge the methods.
- No conflict of interest — you are not a collaborator, competitor with a direct stake, or close colleague of the authors. If in doubt, declare it to the editor and let them decide.
- Time — you can deliver a considered report by the deadline. If not, decline quickly; suggesting an alternative reviewer is one of the most useful things you can do.
And one rule that overrides everything: the manuscript is confidential. Do not share it, do not use its findings before publication, and do not contact the authors.
How to read the manuscript
Read it twice. The first pass is fast, just to understand what the authors did and what they claim to have found. The second pass is slow and sceptical, and this is where you ask the questions that matter:
- Is the research question clear, and does it matter?
- Do the methods actually answer that question? Could the study design support the conclusions?
- Are the results presented honestly — do the figures and tables match the text?
- Do the conclusions follow from the data, or do they overreach?
- Is there anything that looks like an ethics or integrity problem — missing consent, image manipulation, text that reads as plagiarised?
A review-report template
Editors read dozens of reports; a structured one is a gift. Use this every time:
Summary — Two or three sentences describing what the paper does and its main finding, in your own words. This shows the editor (and authors) that you understood it.
Major points — Numbered. The issues that must be addressed before the paper could be published: flaws in design, unsupported claims, missing analyses.
Minor points — Numbered. Smaller fixes: unclear sentences, mislabelled figures, missing references, typos.
Confidential comments to the editor — Anything you would not put to the authors, and your overall recommendation.
Number your points so authors can respond to each one — it makes their revision, and the next round of review, far easier.
Tone: be the reviewer you would want
Critique the work, never the person. "The sample size is too small to support this conclusion" is useful; "the authors clearly do not understand statistics" is not. Be specific — point to the line or figure — and where you can, say how to fix a problem, not just that it exists. A firm, respectful report that a stressed author can act on is worth more than a brilliant one that only wounds.
What the editor actually wants
You are not the judge; you are the expert witness. The editor makes the decision and weighs your report against others. So give them evidence, not a verdict: explain your reasoning, flag your uncertainty honestly, and make a clear recommendation (accept, minor revision, major revision, or reject) knowing it is advice, not the ruling. Do that well and you will be invited back — which, in the quiet economy of academia, is exactly the point.