A note on the examples below: these are illustrative, composite comments written to demonstrate the pattern reviewers are asked to follow — they are not quotations from any real manuscript, journal, or reviewer, and no real paper, author, or institution is referenced. Use them as models to adapt to your own manuscript, not as text to paste in unchanged. For the full process — how to read a manuscript, structure a report, and decide on a recommendation — see CASRAI’s How to Conduct a Peer Review guide, which this page extends with a category-by-category example library.
What makes a peer review comment constructive
Editors, COPE’s Ethical Guidelines for Peer Reviewers, and the ICMJE’s guidance on the peer-review process converge on the same underlying test: a comment is constructive when it gives the author something specific and actionable, and destructive when it only asserts a judgment. The difference isn’t politeness — a blunt comment can still be actionable, and a warmly worded one can still be useless. Three things separate the two:
- Specificity. Point to a line, table, figure, or section rather than the manuscript as a whole.
- A reason. Explain why something is a problem, not just that it is one.
- A path forward, where one exists. Say what would resolve the concern, even if you can’t be certain it will.
Every example below follows a consistent format: a version that only asserts a judgment, and a rewritten version that applies these three elements.
Major revision comments
Major comments address issues that affect whether the manuscript’s central claims hold up. They’re typically numbered so authors and editors can track them through a revision cycle.
Example: unsupported causal claim
Weak: “The authors overstate their conclusions.”
Constructive: “The Discussion (paragraph 2) describes the intervention as ‘causing’ the observed reduction in outcome Y, but the study design is cross-sectional and cannot establish causation. The conclusions should be reframed in terms of association, or the authors should explain what design feature supports a causal claim.”
Example: missing comparison condition
Weak: “The experimental design is inadequate.”
Constructive: “Experiment 2 (Section 3.1) reports the treatment condition’s results but doesn’t include a control or baseline comparison, so it’s not possible to attribute the reported effect to the intervention rather than to other factors. Either add a comparison condition or explain why one wasn’t feasible, and adjust the strength of the conclusions accordingly.”
Example: gap in prior-work coverage
Weak: “The literature review is incomplete.”
Constructive: “The introduction presents this approach as novel, but doesn’t address prior work reporting a similar effect using a related method. The manuscript should either distinguish its contribution from that line of work or reconcile the differing results — as written, a reader familiar with that literature may not see what’s new here.”
Minor revision comments
Minor comments flag issues that should be fixed but don’t threaten the paper’s validity — unclear labeling, small inconsistencies, missing detail that doesn’t change the interpretation of results.
Example: figure clarity
Weak: “Figure 3 is confusing.”
Constructive: “Figure 3’s y-axis isn’t labeled with units, and the legend doesn’t distinguish the two shaded regions described in the caption. Adding units and a one-line legend key would make the figure interpretable without needing to cross-reference the text.”
Example: text/table inconsistency
Weak: “The numbers don’t match.”
Constructive: “The Results text (paragraph 4) reports n=42 for the treatment group, but Table 2 lists n=45. Please confirm which figure is correct and correct the other, since a reader trying to reproduce the analysis will hit this discrepancy immediately.”
Example: missing citation
Weak: “Add more references.”
Constructive: “The claim in the introduction that this measurement approach is ‘widely adopted’ would benefit from one or two citations to studies that have used it, so a reader can verify the claim independently.”
Methodological critique
Methodology comments assess whether the study design is actually capable of answering the question the authors pose — sample size, controls, procedural detail, and instrument or protocol choice.
Example: sample size justification
Weak: “The sample size is too small.”
Constructive: “The Methods section doesn’t report a power calculation or other justification for the sample size (n=18 per group). Given the effect size discussed in the Discussion, it would help to know whether the study was adequately powered to detect it, or whether the modest sample should temper how strongly the results are interpreted.”
Example: procedural detail missing
Weak: “Methods need more detail.”
Constructive: “Section 2.3 describes the survey instrument as ‘adapted from prior work’ without specifying which items were changed or why. Given that the instrument’s validity is central to the paper’s claims, please include the adapted instrument (or the changes made) as supplementary material.”
Example: unmatched comparison groups
Weak: “The groups aren’t comparable.”
Constructive: “Table 1 shows the control group has a substantially different age distribution from the treatment group, but age isn’t included as a covariate in the primary analysis (Section 3.2). Please either match on age, include it as a covariate, or discuss why it’s unlikely to confound the reported effect.”
Statistical critique
Statistical comments check whether the chosen analysis actually matches the data and design — a category where vague criticism is particularly unhelpful, since “the stats are wrong” gives an author nothing to act on.
Example: test/data mismatch
Weak: “The statistics are wrong.”
Constructive: “The paired t-test reported in Table 2 assumes normally distributed differences, which isn’t demonstrated for this sample size (n=8). A Wilcoxon signed-rank test, or an explicit normality check reported alongside the t-test, would strengthen this result.”
Example: missing correction for multiple comparisons
Weak: “Too many comparisons were run.”
Constructive: “Section 3.4 reports p-values for eleven pairwise comparisons without a stated correction for multiple testing. Please apply an appropriate correction (e.g. Bonferroni or false discovery rate) or explain why an uncorrected threshold is appropriate here, since as reported some of these findings may not survive correction.”
Example: effect size not reported
Weak: “The result isn’t meaningful.”
Constructive: “The abstract reports statistical significance (p=0.03) but the manuscript doesn’t report an effect size or confidence interval anywhere. Adding these would let readers judge the practical magnitude of the effect, not just whether it’s distinguishable from zero.”
Ethical and disclosure comments
These flag gaps in required disclosures rather than the substance of the research itself — consistent with CASRAI’s Types of Conflict of Interest guide and COPE’s guidance on reviewer conduct.
Example: missing ethics statement
Weak: “Ethics approval isn’t mentioned.”
Constructive: “The Methods section describes recruiting human participants but doesn’t state whether the study received IRB/ethics committee approval or an exemption. Please add this statement, including the approving body and approval or exemption reference, consistent with the journal’s requirements.”
Example: undisclosed funding
Weak: “There might be a conflict of interest here.”
Constructive: “The manuscript doesn’t include a funding or competing-interests statement. Given that the study evaluates a commercial product, please add a disclosure of the funding source and any author relationships with the manufacturer, even if there is none to report.”
A concern that rises to suspected fabrication, falsification, plagiarism, or image manipulation should be flagged to the editor directly, not investigated or accused in the review itself — see CASRAI’s How to Conduct a Peer Review guide for how COPE and ICMJE expect this to be handled.
Positive and strength comments
A review that only lists problems can be as unhelpful as one with no substance — naming what the manuscript does well, specifically, helps authors know what to preserve through revision and gives the editor a fuller picture than a list of complaints alone.
Example: acknowledging a strong design
“The stepped-wedge design in Section 2.1 is a strong fit for the operational constraints described, and the pre-registration referenced in the Methods gives real confidence that the primary outcome wasn’t chosen post hoc.”
Example: acknowledging clear writing
“The introduction does an unusually clear job of motivating why the three conditions were compared rather than just describing that they were — this made the rest of the manuscript easier to evaluate.”
Overall recommendation framing
The recommendation (accept, minor revision, major revision, reject) should read as the logical conclusion of the comments above it, not a separate judgment layered on top. Two examples of that connective framing:
Example: major revision
“This manuscript addresses a question of clear relevance to the field, but the central claim in Section 4 currently rests on an analysis (Major Comment 1) that doesn’t yet support it, and the missing power calculation (Major Comment 2) makes the sample size difficult to evaluate. I recommend major revision: if these two points are addressed, along with the minor comments below, I’d expect to be able to support publication.”
Example: reject, written constructively
“The research question is worthwhile, but the study design (Major Comments 1-3) can’t currently support the causal claims made in the Discussion, and I don’t see a revision path within the scope of a resubmission that would resolve this without substantially new data collection. I’d encourage the authors to consider a revised design for a future submission, and the writing itself is clear enough that I’d be glad to see this question addressed again.”
Even a reject recommendation benefits from specific, professional comments — authors often revise and submit elsewhere, and a report that’s all major concerns but delivered destructively teaches them nothing they can use.
Confidential comments to the editor
Some journals offer a field for comments the authors won’t see. COPE’s guidance is explicit that this field isn’t a place to say something harsher than you were willing to put in the comments the authors do see — it’s reserved for things they genuinely shouldn’t see.
Appropriate use
“I want to flag for the editor, separate from the comments to the authors: two paragraphs in the Discussion (page 9) are very close in phrasing to a passage in [a source I recognize but haven’t cited here since I can’t be certain without a formal similarity check] — worth a plagiarism-detection pass before this goes further.”
Inappropriate use (avoid this pattern)
“Reject — this is clearly the same weak group that’s been submitting to this journal for years and won’t ever get it right.” This kind of comment isn’t actionable, isn’t about the manuscript in front of the reviewer, and is exactly what COPE’s guidance on reviewer conduct exists to prevent.
A short manuscript peer review template
Most journals provide their own submission form, but the underlying structure is consistent and worth using even when a journal gives you a blank box. For the full walkthrough of each section, see CASRAI’s How to Conduct a Peer Review guide; here is the skeleton with placeholder text showing what belongs in each field:
- Summary: “This manuscript reports [one-sentence restatement of the central claim and method, in your own words].”
- Major concerns: Numbered list, one issue per number, each following the weak-vs-constructive pattern above.
- Minor concerns: Numbered list of everything else worth flagging that doesn’t threaten validity.
- Recommendation: Accept / minor revision / major revision / reject, with one sentence connecting it to the comments above.
- Confidential comments to the editor (if offered): Reserved for conflicts, suspected misconduct, or process notes — not a harsher version of the comments above.
Common questions
Can I copy these examples directly into my review?
Only as structural models. The specifics — line numbers, table references, the actual statistical or methodological issue — need to come from the manuscript in front of you. A comment copied verbatim without matching the real manuscript detail will read as generic and won’t hold up to an editor or author checking it against the text.
What’s the difference between this page and CASRAI’s guide on how to conduct a peer review?
The How to Conduct a Peer Review guide covers the full process end to end: reading the manuscript, deciding what to assess, and structuring a report. This page is the example library for one specific part of that process — what constructive comment language actually looks like across categories — for reviewers who already know the process and want models to adapt.
How many major vs. minor comments is normal?
There’s no fixed ratio. A report is well-calibrated when major comments are reserved for issues that actually affect whether the paper’s conclusions hold up, and everything else — however numerous — is filed as minor rather than inflated to seem more serious than it is.
Should feedback tone differ between a reject and a minor-revision recommendation?
The specificity and professionalism shouldn’t differ at all; only the number and severity of the underlying issues does. A reject recommendation written destructively teaches an author nothing, while one written constructively — even while recommending against publication in its current form — still gives them something to act on for a future submission.
Is it acceptable to praise a manuscript in a review?
Yes, and it’s good practice. Specific positive comments help authors know what to preserve through revision and give the editor a more complete picture than a list of problems alone — see the Positive and strength comments section above for examples.
Related CASRAI resources
- How to Conduct a Peer Review: A Practical Guide for First-Time Reviewers — the full process this page’s examples support
- Peer Review: What It Is, How It Works, and Why It Matters
- Peer Reviewer: Role, Responsibilities, and How to Become One
- COPE Guidelines Explained
- How to Decline a Peer Review Invitation
- Types of Conflict of Interest
- Single-anonymized peer review and double-anonymized peer review
