Writing a peer review is one of the most important contributions you can make to your academic or professional field. When done well, a peer review strengthens a manuscript, helps a fellow researcher improve their work, and maintains the standards that make published research trustworthy. When done poorly, it can be discouraging, unhelpful, or even harmful to the author's career and the field's progress.
• An estimated 2.8 million peer-reviewed articles are published annually across over 28,000 academic journals worldwide (STM Global Brief, 2024).
• The average peer review takes 5-8 hours to complete, yet a 2023 Publons survey found that most reviewers receive no compensation beyond acknowledgment.
• Research published in PLOS ONE found that reviews providing specific, actionable feedback lead to manuscripts that are cited 23% more frequently after publication.
• Early-career researchers who participate in peer review report 40% faster improvement in their own writing and methodology compared to those who do not review (Nature survey, 2024).
Whether you are reviewing your first manuscript or your fiftieth, this guide provides a systematic framework for producing reviews that are thorough, fair, and genuinely useful to the author. It covers preparation, analysis, writing, and the ethical considerations that distinguish a good reviewer from a great one.
Understanding the Purpose of Peer Review
Peer review serves the scholarly community at three levels, and understanding these levels helps you write a review that contributes meaningfully at each one.
At the manuscript level, your review helps the author improve their specific paper. You identify errors in methodology, gaps in argumentation, unclear writing, and opportunities for strengthening the work. This is the most direct and tangible function of your review.
At the journal level, your review helps the editor decide whether to accept, revise, or reject the manuscript. Your recommendation carries significant weight, and the editor relies on your expertise to evaluate aspects of the paper that they may not be qualified to assess independently.
At the field level, peer review serves as a quality control mechanism for the entire body of published knowledge. When you catch a methodological flaw or identify an unsupported claim, you are protecting the integrity of the literature that other researchers will cite and build upon.
Keeping all three levels in mind helps you calibrate your feedback. Minor stylistic issues matter at the manuscript level but not at the field level. A fundamental methodological error matters at all three levels.
Preparing to Write Your Review
First Read: Get the Big Picture
Read the manuscript once through without taking notes. The goal is to understand the overall argument, methodology, and contribution before evaluating specifics. Pay attention to your gut reactions: where do you feel confused? Where do you feel convinced? Where does the argument feel thin? These initial impressions often point to the most significant issues.
Second Read: Detailed Analysis
Read again with a pen, highlighter, or commenting tool. This time, evaluate each section systematically:
- Abstract and Introduction: Are the research question and rationale clearly stated? Does the introduction provide sufficient context?
- Literature Review: Is relevant prior work cited? Are there significant omissions? Does the review fairly represent the state of the field?
- Methodology: Is the research design appropriate for the question? Are methods described in enough detail to be replicated? Are there threats to validity?
- Results: Are findings presented clearly? Do the statistics (if applicable) support the conclusions? Are tables and figures informative and necessary?
- Discussion: Do the authors interpret their results accurately? Are limitations acknowledged? Are claims proportional to the evidence?
- References: Are citations current and relevant? Are any key sources missing?
Third Read: Synthesis
After your detailed analysis, step back and ask: Does this paper make a meaningful contribution to the field? Is the contribution supported by the evidence presented? These questions guide your overall recommendation and help you prioritize your feedback.
Structure of a Peer Review
Most journals provide specific review forms, but the underlying structure is consistent. A well-organized review makes your feedback easy to understand and act upon.
Summary (1-2 paragraphs)
Begin with a brief, objective summary of the manuscript's main argument, methodology, and findings. This demonstrates that you have understood the paper and gives the author confidence that your subsequent feedback is informed. Keep this factual; do not evaluate yet.
Major Strengths (2-4 points)
Identify what the paper does well. This is not a courtesy; it is essential feedback. Authors need to know which elements to preserve as they revise. Strengths might include an innovative research design, a particularly thorough literature review, clear writing, or novel findings that advance the field.
Major Concerns (2-5 points)
Address the most significant issues that affect the paper's validity, clarity, or contribution. These are issues that must be resolved before publication. Number each concern and provide specific evidence and suggestions for resolution. For example, rather than "the methodology is weak," write "the sample size of 12 participants is insufficient to support the generalized claims in the discussion. I recommend either expanding the sample or qualifying the conclusions as preliminary findings."
Minor Concerns (5-10 points)
Address smaller issues: unclear sentences, formatting inconsistencies, missing references, minor data presentation issues, and stylistic suggestions. Reference specific page numbers, paragraphs, or line numbers so the author can locate each issue easily.
Recommendation
Most journals ask for one of four recommendations: Accept, Minor Revisions, Major Revisions, or Reject. Base your recommendation on the totality of your assessment. A paper with minor concerns and strong contributions warrants Minor Revisions. A paper with fundamental methodological issues that may not be resolvable warrants Reject.
1. Summary
This manuscript examines [research question] using [methodology] with [sample/data]. The authors find that [main findings]. The paper contributes to the literature on [topic] by [specific contribution].
2. Strengths
S1. [Specific strength with evidence from the paper]
S2. [Specific strength with evidence from the paper]
S3. [Specific strength with evidence from the paper]
3. Major Concerns
C1. [Specific concern]. The authors state [quote or paraphrase] on page [X], but [explain why this is problematic]. I recommend [specific suggestion for resolution].
C2. [Specific concern with the same structure]
C3. [Specific concern with the same structure]
4. Minor Concerns
m1. Page [X], paragraph [Y]: [specific issue and suggestion]
m2. Table [X]: [specific issue and suggestion]
m3. [Additional minor points as needed]
5. Recommendation
Based on the above, I recommend [Accept / Minor Revisions / Major Revisions / Reject] because [1-2 sentence justification].
Writing Constructive Feedback
The language of your review matters as much as its content. Constructive feedback is specific, actionable, and respectful. Here are the principles that separate helpful reviews from harmful ones.
Be Specific, Not Vague
Vague: "The methodology section needs work."
Specific: "The methodology section does not specify how participants were recruited or what inclusion/exclusion criteria were used. Adding a subsection on participant selection would strengthen the paper's replicability."
Specific feedback tells the author exactly what to fix and how. Vague feedback leaves them guessing, which often leads to revisions that still do not address the underlying problem.
Separate Observation from Recommendation
First describe what you observed, then suggest what the author might do about it. "The sample includes only participants from urban areas (observation), which limits the generalizability of findings to rural populations. Consider either including rural participants in a follow-up study or adding this as an explicit limitation in the discussion (recommendation)."
Use Professional, Respectful Language
Write as you would want to be written to. Avoid dismissive language ("obviously," "clearly the authors failed to," "this is a fundamental misunderstanding"), condescending tone, or personal criticism. Focus on the work, not the authors. "The analysis would benefit from" is more productive than "The authors should have known to."
Balance Criticism with Acknowledgment
Even deeply flawed manuscripts usually contain some genuine effort and occasional strengths. Acknowledging these is not about being nice; it is about being accurate. A review that identifies only flaws is as incomplete as one that identifies only strengths. Both distort the picture.
Distinguish Essential from Optional
Clearly differentiate between changes the author must make for the paper to be publishable and suggestions that would improve but are not strictly necessary. This helps both the author and the editor prioritize the revision process.
"A good peer review is an act of generosity. You are giving an author the gift of your expertise, and the quality of that gift depends on the care you bring to it."
-- Dr. Irene Hames, author of "Peer Review and Manuscript Management in Scientific Journals" and former editor at The Plant Journal
Evaluating Different Components of a Manuscript
Assessing Methodology
Methodology is where most peer reviews have the greatest impact. Ask yourself: Could another researcher replicate this study based solely on the methods section? Are the chosen methods appropriate for the research question? Are there alternative approaches that would be stronger?
For quantitative research, evaluate sample size, sampling method, statistical tests, and potential confounds. For qualitative research, assess the appropriateness of the chosen approach (grounded theory, phenomenology, case study, etc.), data saturation, and analytical rigor.
Assessing Data and Results
Look for consistency between the reported methods and the presented results. Do the tables and figures support the claims in the text? Are effect sizes reported alongside p-values? Are negative or unexpected results addressed or selectively omitted?
Assessing Claims and Conclusions
This is where intellectual honesty is most important. Are the conclusions supported by the data, or do the authors over-interpret their findings? Watch for the gap between what the data shows (correlation, limited sample) and what the authors claim (causation, generalizability). Strong papers acknowledge this gap; weak papers pretend it does not exist.
Assessing Writing Quality
While you are not a copy editor, significant writing issues affect a paper's clarity and therefore its contribution. Note structural problems (poor organization, missing transitions), logical gaps (conclusions that do not follow from premises), and jargon that excludes rather than clarifies. Minor grammar issues can be noted briefly without detailed correction.
Common Mistakes to Avoid
These five errors undermine the effectiveness and fairness of peer reviews. Being aware of them helps you produce reviews that genuinely serve the author and the field.
1. Reviewing the paper you wish the author had written instead of the paper they actually wrote. This is the most subtle and common mistake. You may have a different theoretical framework, prefer different methods, or disagree with the research question itself. None of these are valid grounds for a negative review. Evaluate whether the paper achieves what it sets out to do, not whether you would have approached the topic differently. Reserve your criticisms for actual flaws in logic, methodology, or evidence.
2. Being vague in your criticism. "The paper needs significant improvement" tells the author nothing actionable. Every criticism should identify a specific problem and, wherever possible, suggest a specific solution. "The discussion section on pages 14-16 makes causal claims that the correlational design cannot support. I recommend qualifying these statements or providing additional evidence" gives the author a clear path forward.
3. Letting personal bias influence your evaluation. Bias can be conscious (you disagree with the author's conclusions and are harder on the paper as a result) or unconscious (you evaluate work from prestigious institutions more favorably). Actively check yourself for these tendencies. Ask: Would I make the same criticism if this paper came from a different lab, country, or theoretical tradition?
4. Providing only negative feedback. A review that lists only weaknesses is incomplete. Every manuscript has strengths, and identifying them serves two purposes: it helps the author preserve what works during revision, and it demonstrates to the editor that you evaluated the paper fairly rather than approaching it with a predetermined verdict.
5. Exceeding your expertise without disclosure. If the manuscript uses statistical methods, theoretical frameworks, or domain knowledge that are outside your expertise, say so in your review. "I am not qualified to evaluate the Bayesian analysis in Section 4 and recommend the editor seek a statistical reviewer for that component" is far more helpful than attempting an evaluation you are not equipped to make.
Ethical Considerations in Peer Review
Confidentiality: The manuscript you are reviewing is unpublished work. Do not share it, discuss it, or use its ideas in your own research before publication. Violations of confidentiality are among the most serious breaches of academic ethics.
Conflicts of interest: If you are a direct competitor, a close collaborator, or have a personal relationship with the authors, disclose this to the editor and, if necessary, decline the review. Even the appearance of conflict can undermine trust in the review process.
Timeliness: Complete your review within the requested timeframe, typically 2-4 weeks. Delayed reviews hold up the author's work and the publication process. If you cannot meet the deadline, notify the editor immediately so they can find an alternative reviewer.
Honesty: Your review should reflect your genuine assessment. Do not inflate your praise to avoid conflict or deflate your evaluation to serve personal agendas. Honest, fair, and constructive feedback is the foundation of the peer review system.
Using AI to Assist with Peer Review
AI tools can help you organize your thoughts, identify areas you may have overlooked, and draft portions of your review. They should never replace your expert judgment, but they can enhance your efficiency and thoroughness.
I am writing a peer review for a manuscript on [topic] in [field/journal]. Here are my raw notes from reading the paper:
[paste your notes]
Help me organize these into a structured peer review with:
1. A 2-sentence summary of the paper
2. 3 major strengths (numbered S1-S3)
3. Major concerns ranked by importance (numbered C1-C5)
4. Minor concerns with page/section references (numbered m1-m10)
5. A recommendation (accept, minor revisions, major revisions, or reject) with justification
Make sure each concern is specific and includes a suggested resolution.
Here is my draft peer review:
[paste your review]
Check whether I have addressed all standard components:
- Summary of the manuscript
- At least 3 strengths
- Methodological evaluation
- Assessment of results and conclusions
- Writing quality notes
- Clear recommendation with justification
- Respectful, professional tone throughout
Identify any gaps or areas where my feedback could be more specific or constructive.
Review the following peer review feedback I wrote and flag any language that could be perceived as:
- Dismissive or condescending
- Vague or non-actionable
- Personal rather than focused on the work
- Overly harsh without being constructive
[paste your feedback]
For each flagged item, suggest a revised version that communicates the same concern more constructively. Maintain the substance of the criticism while improving the delivery.
I am reviewing a [quantitative/qualitative/mixed methods] study in [field] that uses [describe methodology briefly]. Generate a checklist of methodological elements I should evaluate, including:
- Design appropriateness for the research question
- Sampling and participant selection
- Data collection procedures
- Analytical methods
- Threats to validity/trustworthiness
- Ethical considerations
For each element, provide 1-2 questions I should ask while reading the methods section.
Important ethical note: Many journals have specific policies about AI use in peer review. Some prohibit uploading manuscript content to AI tools due to confidentiality concerns. Always check the journal's reviewer guidelines before using AI, and never upload the full manuscript text to any AI tool unless the journal explicitly permits it.
Building Your Skills as a Reviewer
Peer review is a skill that improves with practice and reflection. Read published reviews (some journals make them publicly available), seek mentorship from experienced reviewers, and treat each review as an opportunity to deepen your own understanding of research methodology and scientific communication.
Many professional organizations offer peer review training workshops, including COPE (Committee on Publication Ethics), APA, and individual journal publishers. These resources can accelerate your development and introduce you to best practices you might not discover on your own.
Frequently Asked Questions
How long should a peer review be?
A thorough review typically runs 500-1,500 words. The length should be proportional to the manuscript's complexity and the significance of the issues you identify. A strong paper with minor issues may warrant a shorter review. A paper with fundamental concerns requires more detailed feedback.
What if I am asked to review a paper outside my expertise?
Decline the invitation or accept with the caveat that certain aspects fall outside your expertise. It is better to provide a partial review within your competence than a complete review that includes uninformed evaluations. Most editors appreciate honesty about the limits of your expertise.
How do I handle a paper that I think should be rejected?
Write the review with the same thoroughness and respect you would give a paper you admire. Provide specific, evidence-based reasons for your recommendation. Focus on fundamental issues (methodology, logic, contribution) rather than accumulating minor complaints. A well-reasoned rejection review can be one of the most valuable pieces of feedback an author receives.
Should I recommend references to my own work?
Only if your work is genuinely relevant and the omission represents a meaningful gap in the literature review. Recommending your own papers purely to increase your citation count is a form of ethical violation known as "coercive citation." If in doubt, suggest the relevant finding without specifying the source.
What if I disagree with the other reviewer?
Disagreement between reviewers is common and expected. Provide your honest assessment independently. The editor's job is to synthesize different perspectives and make a final decision. You are not responsible for aligning with other reviewers.
How do I review a revised manuscript?
Compare the revision against your original concerns point by point. Acknowledge changes that satisfactorily address your feedback, identify concerns that remain unresolved, and evaluate any new issues introduced by the revision. A revised manuscript review should be shorter and more focused than the original.
