Peer review can improve a paper and expose important weaknesses before publication. It does not make the resulting article immune to error. To use publication status sensibly, distinguish review of a manuscript from independent reproduction of its measurements, analysis or conclusions.
Identify what kind of review occurred
ICMJE describes peer review as critical assessment of submitted manuscripts by experts and notes that journals vary in their processes, reviewers and manuscript types sent for review. Editorial decisions also involve suitability for the journal, beyond the assessment of scientific quality alone.ICMJE — Responsibilities in the Submission and Peer-Review Process (opens in a new tab)
A research article, editorial, letter and correction may pass through different processes within the same journal. Find the publication type and journal policy relevant to the item rather than assuming that every page carrying the journal’s name had identical scrutiny.
| Process | What to ask |
|---|---|
| Editorial assessment | Was the manuscript suitable for consideration? |
| Peer review | What questions did reviewers examine? |
| Data reanalysis | Was the reported computation independently checked? |
| New experiment | Was the finding tested with new observations? |
These activities can overlap, but the existence of one does not establish that all occurred. In particular, reading a methods section is not the same action as performing the experiment again.
Use review reports when they are available
An openly available review history can show which issues were raised and how the authors responded. It may reveal a clarified control, an additional analysis or a narrower conclusion in the final version.
In an original hypothetical example, a reviewer asks whether n represents independent cultures or repeated readings. The authors revise the caption and analysis. The useful evidence is the clarified replicate structure and its consequences, not merely the fact that a review exchange exists.
Check that the response corresponds to the final manuscript. A promise to add a control is not evidence that the control appears in the accepted paper, and an early criticism may have been resolved in a later version.
Do not assume every underlying record was audited
ICMJE notes that journal requirements for independent data analysis and data availability vary. Some processes include additional statistical or data checks, while others rely on the submitted manuscript and supporting materials.ICMJE — Responsibilities in the Submission and Peer-Review Process (opens in a new tab)
The specific record matters. If a journal states that a dataset was independently reanalysed, identify the scope of that reanalysis. It may check the reported calculations without authenticating every original measurement or resolving every design limitation.
Errors can also concern details that were not central to the original reviewers’ questions, such as a mislabeled supplementary file or a later-discovered material identity issue. Publication status does not make those details irrelevant.
Avoid the opposite inference that peer review is useless because it is imperfect. A constructive review can improve reporting and reasoning while leaving residual uncertainty that later scrutiny can address.
Place the reviewed article in a continuing evidence record
Before relying on a paper, check for corrections, expressions of concern, retractions or substantive follow-up. A later notice can change the interpretation of a result even when the original article was peer reviewed.
Then assess the actual design, material, readout, analysis and claim. A prestigious publication venue does not resolve a weak control, while a less familiar venue does not automatically make sound observations false.
An evidence summary can state that the article was published through the journal’s described review process, while separately describing any available data checks and replications. This gives each contribution its proper meaning.
For this Research Hub, source checking and editorial review by the authoring assistant are not independent scientific peer review. Keeping that distinction visible is part of the same standard we apply when describing other research.
Sources and further detail
- ICMJE — Responsibilities in the Submission and Peer-Review Process (opens in a new tab)
Peer-review definition, variable journal processes and data-analysis requirements read. The culture-versus-reading example is original; no claim about the review quality of a named journal or paper is made.
Sources checked 19 September 2026. Worked examples are illustrative unless a supplied report is explicitly identified. This article has not undergone independent scientific peer review.