Reviewer Information

This page provides basic practical information to reviewers. For more details see our Editorial Policy and Terms & Conditions. For any further queries please email hello@peer-reviews.co.uk.

Writing Your Report

The review and publication process on this platform is different to the traditional model. Please fully read this section before writing your report to make sure it’s in line with this platform’s requirements.

Two types of content are accepted i) scholarly – work that is academic in nature and would normally be in scope for a traditional journal, for example a report of original research or a perspective and ii) technical – all other work on which it is possible to provide a meaningful review e.g. a confidential internal report, a new standard operating procedure or a public white paper.

All work is accepted for publication as submitted unless misconduct is detected. On publication, reviewer’s comments and scores are published alongside the work. Review scores are made against a standard rubric and are intended to provide a indication of the quality of the work to readers. The thinking behind this approach is outlined in the White Paper.

General Guidelines

The review should be your honest and objective assessment of the work as it was submitted. It should be written in a clear and factual manner with detailed justifications for every point, citing references as needed. For technical work remember that authors may not be knowledgeable about the state of the art so jargon or unclear references should be avoided. Plain statements of opinion are encouraged; reviews are not unfair simply because they are stated robustly. However, you should not seek to cause offence and be mindful that reviews will be published and read by a wide audience.

The text of the review should address all aspects of the work and be structured into the seven sections described below. The scores on each of these sections should be a summary of your review and congruent with the text. You may provide guidance on further work, suggested changes etc. but remember that work is published as submitted so it is not a requirement for authors to revise. The work must therefore be assessed ‘as is’.

Misconduct, Plagiarism & Confidential Comments

You should flag any indication of misconduct, which would typically include any:

• Fabrication (e.g. making up data)
• Falsification (e.g. dropping outliers or manipulating images)
• Misrepresentation (e.g. of people’s involvement, conflicts of interest)
• Breach of duty of care (e.g. failing to obtain ethical approval)

You are also asked to highlight any instances of plagiarism, broadly defined as ‘using other people’s ideas, intellectual property or work (written or otherwise) without acknowledgement or permission’. ¹ Plagiarism is a particular concern for scholarly work that is published and different standards may apply to some of the technical work reviewed, e.g. standard operating procedures for internal use -however all concerns should be flagged regardless of content type. Note that the policy of this platform is to be agnostic in the use of AI tools and we do not require their use to be disclosed however, authors are fully and solely responsible for ensuring the appropriate acknowledgement of primary sources. The use of AI tools is not an excuse for plagiarism in the submitted work.

Comments on any topic can also be supplied confidentially. These will not be published or provided to the authors.

General Comments

In this section provide a brief summary of the main points of the work, and the overall conclusion of your review. Please comment on the scope of your review, particularly if there are any aspects which you do not feel qualified to comment on.

Rubric

The rubric is designed to standardise the structure of the all reviewers’ analysis and comments around six key aspects. Scores are provided alongside comments to aid readers in quickly differentiating work but they are not intended to be a substitute for the review. You must follow the guidance when choosing scores and ensure that comments and scores are coherent. Remember that very few articles would be expected to score strongly across all aspects; highly novel blue-sky research often has an unclear significance while work that diligently attempts to reproduce other’s work is very unlikely to be novel. All of these can be important contributions to a healthy scholarly and technical discourse.

Novelty & Originality

On a more subjective note, do you feel that the results presented are of immediate interest to many people in your own discipline, and/or to people from several disciplines? Replication study. When set, the novelty score is shown with context, or you could mark novelty as “not applicable.” Replications then aren’t punished. Overclaimed novelty. The paper claims “first” when prior work exists. This is useful even when the actual score is a respectable 3. Possible redundant or duplicate publication. Overlap with the authors’ own earlier work. This is an integrity issue, not a novelty score, so it should route differently. Unable to assess. The reviewer lacks the expertise to judge the literature. This is better than a forced middle score. Engagement with prior literature. Novelty asks the reviewer to find the closest prior work. Nothing scores whether the authors placed their work fairly: citing the relevant literature, benchmarking against the actual state of the art rather than weak comparisons, and avoiding excessive self-citation. This could be a flag (“important prior work not cited”) or a guidance bullet under conclusions, since unfair benchmarking is often what drives “best performing” claims.

Impact & Significance

• This score is an informed opinion, not a measurement. Significance is a forecast of how the work will be received and used, so reasonable reviewers can disagree. Write the justification as your own judgement (“In my view…”), and say where your expertise gives you a basis for it and where it doesn’t. This matters most for routes outside your own field, such as policy or industrial use.
• Consider all routes to impact. These include the technical field itself, technology and industrial application, policy and regulation, public understanding, and other disciplines. Score the strongest credible route, not the sum of all of them.
• Assume the results are correct. Validity is assessed under rigour.
• Keep it separate from novelty. Incremental work can be highly significant, and pioneering work can be niche.
• Justify with audience and mechanism. Name who benefits and how. For scores of 3 or below, say what would increase the impact.
• Treat 5 as rare. Most sound papers score 2–4.

Technical Validity

Appropriate use of statistics and treatment of uncertainties: all error bars should be defined in the corresponding figure legends; please comment if that is not the case. Please include in your report a specific comment on the appropriateness of any statistical tests, and the accuracy of the description of any error bars and probability values.

Guidance for reviewers:

• Scope is methods and analysis only. Assess whether the techniques were appropriate, correctly applied and correctly analysed. Whether the conclusions follow from the results is assessed separately under confidence in conclusions. For statistics, choosing and applying the right test is in scope here. Whether the result supports the paper’s claim is not.
• Be specific. For any score below 5, identify each issue by location (figure, table or section), say which results it affects, and say what would fix it.
• Judge against the norms of the field. Cite a relevant standard or guideline where one exists.
• Missing detail counts as unverified. If methods aren’t described well enough to judge, score no higher than 3 and say what’s missing.
• Integrity concerns are flagged, not scored. Suspected fabrication or data manipulation should be routed to the editor.

Confidence in Conclusions

Scope is the reasoning from results to conclusions. Take the reliability of the methods as assessed under technical validity. Where a conclusion rests on results you judged unreliable, say so here and refer to that assessment rather than repeating it.
Identify the central conclusion first. State in one sentence what you take the paper’s main claim to be, and say whether any issues you raise affect it.
Name each unsupported or overstated conclusion and say what evidence would support it or how it should be reworded.
Common overstatements to check for: “best performing” or “first” claims without a fair comparison, extrapolation beyond the conditions tested (for example, long-term durability inferred from short tests), causal claims from correlational data, and a single system presented as a general principle.
Extraordinary claims need extraordinary evidence. The more strongly a conclusion contradicts established knowledge, the stronger the evidence must be to score 4 or 5. Contradicting one or two earlier studies is normal and isn’t grounds for level 1. Level 1 is for claims that conflict with fundamental principles or a broad, consistent body of evidence without evidence to match.
This is expert judgement, not certainty. Say where your confidence rests on your own expertise and where it’s limited.

Reporting & Reproducibility

• Scope is whether the work could be repeated, not whether it’s correct. Technical validity asks whether there’s enough information to judge the methods. This axis asks whether there’s enough to redo them. A paper can pass the first and fail the second.
• Justified restrictions don’t count against the work. Examples are commercial confidentiality, personal data and safety. If the restriction is explained and a controlled access route exists, score as if the data were available. Unexplained withholding does count against it.
• Check that it’s really there. Confirm that links and identifiers resolve and that the deposits contain what the paper says they do.
• Computational work needs its own details: code version, software environment, parameters and settings, and random seeds where relevant.
• Apply field reporting guidelines where they exist, including clinical or animal-research checklists and field-specific reporting checklists. Name the one you applied.
• Be specific. List the missing items, such as parameters, datasets or scripts, so readers know exactly what would be needed to reproduce the work.

Proofing

Judge communication, not the science. Scientific problems belong on the other axes. Don’t penalise non-native phrasing where the meaning is clear. The test is clarity, not style.
Uncertainty is split between axes. Whether uncertainty was quantified is assessed under technical validity. Here, assess only whether it is displayed clearly, for example error bars that are shown and defined in the caption.
Be specific. List errors by location. If there are many, give representative examples and say that a full language edit is needed.
Conventions depend on the article type and field. A letter, a full paper and a review have different norms. Check nomenclature, units and any reporting checklist that applies.
A score of 1 affects the other axes. If presentation stopped you assessing part of the work, mark the affected criteria “unable to assess” rather than guessing.Does this manuscript reference previous literature appropriately? If not, what references should be included or excluded? Does the manuscript contain any language that is inappropriate or potentially libellous? Is the abstract clear and accessible? Are the abstract, introduction and conclusions appropriate?

Expected Conduct

Reviewers are expected to adhere to best practices ². In short reviewers should:

• Provide an honest and objective assessment on the work
• Report in line with the rubric with professional and constructive commentary
• Maintain the strictest confidentiality
• Declare real or perceived conflicts of interest
• Return reports within the agreed time frame
• Carry out and write the review themselves, delegating neither other people nor AI tools.