Peer Reviewing a Manuscript Efficiently and Fairly

Peer review sustains the scholarly record, yet most researchers receive little training in how to do it. Invitations arrive with short deadlines, manuscripts vary enormously in quality and the work is unpaid. The result is reviews that range from meticulous to perfunctory, and authors who often feel their work was misunderstood. A structured approach makes reviewing faster for the reviewer and more useful for authors and editors. This article sets out a practical workflow for reviewing a manuscript, from deciding whether to accept the invitation to writing the report, with notes on how a source-linked assistant such as Yila can help verify claims and citations without taking over the reviewer's judgement.

Decide whether to accept

Before accepting, check three things. First, expertise: can you assess the core methods and the subject matter, or at least a substantial part of them? If not, decline or tell the editor which parts you can assess. Second, time: can you complete a careful review by the deadline? A late or rushed review helps no one. Third, conflicts: do you have a personal, professional or financial relationship with the authors, or a competing project? Declare any potential conflict to the editor.

Declining promptly, ideally with suggestions of alternative reviewers, is a service in itself.

First pass: the big picture

Read the manuscript once through without taking detailed notes. Aim to understand the question, the approach, the main findings and the claimed contribution. At the end, write a short paragraph summarising the paper in your own words. If you cannot, that is already useful information about clarity.

Then ask the fundamental questions. Is the question important and clearly stated? Is the design appropriate to answer it? Do the conclusions follow from the results? Is the contribution novel relative to existing work? Serious problems at this level determine the recommendation; there is little point in detailed comments on a paper whose design cannot answer its question.

Second pass: the detail

On the second reading, work section by section. In the introduction, check that the literature is represented fairly and the gap is genuine. In the methods, check that the design, sample, measures and analysis are described in enough detail to evaluate and reproduce. Look for registration or a protocol and compare planned with reported outcomes. In the results, check that numbers are consistent across text, tables and figures, that uncertainty is reported and that all planned analyses are presented. In the discussion, check that limitations are acknowledged and that claims are proportionate to the evidence.

Take notes as you go, referencing page and line numbers so authors can find each point.

Check key citations

Citation errors are common: papers cited for claims they do not make, or findings misreported. You cannot check every reference, but you can check those that support the central claims, particularly where the cited finding seems surprising or crucial to the argument.

A source-linked assistant helps here. Asking Yila whether a cited paper supports a particular statement returns the relevant passages with links, so you can confirm or question the citation quickly. This is especially useful when the cited work is outside your immediate area. The judgement about whether the citation is appropriate remains yours, but the search for the relevant passage is much faster.

Assess data and code where available

Increasingly, journals expect data and code to be shared. Where they are, look at them, even briefly. Check that the files exist, are documented and correspond to the analyses described. You are not usually expected to reproduce every analysis, but obvious mismatches are worth noting. Where data and code are not shared and no reason is given, it is reasonable to ask why.

Write the report

A good report has a clear structure:

1. Summary. Two or three sentences restating the paper's aim, approach and main findings, showing the authors you understood their work.
2. Overall assessment. Your view of the paper's strengths and its most important weaknesses, and whether the conclusions are supported.
3. Major comments. Numbered points addressing issues that affect the validity or interpretation of the findings. For each, explain the problem, why it matters and what the authors could do.
4. Minor comments. Numbered points on clarity, presentation, typographical errors and small inconsistencies, with page and line references.

Separate comments for the editor, including your recommendation and any concerns you prefer not to share with the authors, belong in the confidential section.

Tone matters

Write as you would wish to be reviewed. Be direct about problems but respectful about the people. Comment on the work, not the authors. Acknowledge strengths genuinely. Phrase requests constructively: instead of "the analysis is wrong", explain what concerns you and what alternative would address it. Avoid demanding citations of your own work unless they are genuinely essential.

Make a clear recommendation

Editors need a recommendation they can act on. Choose the category that matches your assessment and make sure your comments justify it. A recommendation of minor revision accompanied by fundamental objections confuses everyone. If you are uncertain, say so in the confidential comments and explain why.

Manage your time

A careful review of a typical research article takes several hours. Block that time in your calendar when you accept, rather than hoping to fit it in. Splitting the work into two sessions, one for the first pass and one for the detailed second pass and report, often produces better reviews than a single long sitting, because the interval allows your thoughts on the big picture to settle. Keep a simple template for reports so you are not starting from a blank page each time, and a record of reviews completed for your own career documentation.

Protect confidentiality

Manuscripts under review are confidential. Do not share them, discuss them with colleagues without the editor's permission, or use their ideas or data in your own work. Be cautious with online tools: uploading a confidential manuscript to an external service may breach the journal's policy. Check the journal's guidance before using any software on the manuscript itself, and prefer tools that work only with published sources you are checking against.

Reviewing revisions

When a revised manuscript returns, check each of your original comments against the authors' response and the revised text. Do not introduce new major issues unless the revisions have created them or you missed something genuinely important. Be efficient: a revision review is usually much shorter than the original.

Common pitfalls

• Accepting reviews outside your expertise without saying so.
• Detailed copy-editing of a paper with fundamental flaws.
• Vague comments without page references.
• Hostile or dismissive tone.
• Recommendations inconsistent with comments.
• Raising new major issues in a revision review without good reason.

Conclusion

Efficient, fair peer review follows a simple pattern: accept only what you can do well, read first for the big picture and then for detail, check the citations that matter, write a structured report with a constructive tone and make a clear recommendation. Tools that link claims to their sources make citation checking practical, helping reviewers protect the accuracy of the scholarly record with less effort.