Peer review

Why new claims are checked

A result reported by one scientist or research group does not immediately become accepted scientific knowledge. Methods can contain weaknesses, data can be interpreted incorrectly and a conclusion may go further than the evidence allows. Scientists can also make honest mistakes, and their background, experience or interests may affect their judgements. The scientific community therefore evaluates new claims critically.

Scientists may legitimately reach different conclusions from the same data. Disagreement is not necessarily evidence of dishonesty: the evidence may be incomplete or open to more than one interpretation. Peer review allows qualified scientists to examine the work and explain where its reasoning is strong or weak.

The peer review process

A researcher writes a paper explaining the question investigated, the method, results and conclusions, then submits it to a scientific journal. A journal editor selects other scientists with relevant expertise—the researcher’s peers—to evaluate it.

The reviewers consider questions such as:

• Is the method appropriate and described clearly?
• Were suitable controls used?
• Is the sample sufficiently large and representative?
• Have the results been interpreted correctly?
• Are the conclusions supported by the data?
• Is the work sufficiently valid, significant, original and clear?

Reviewers report their judgements and may identify errors, ask for clearer explanations or suggest further evidence that would strengthen the argument. They recommend whether the paper should be published, revised or rejected. The authors may then amend the paper and respond to the criticisms before the editor makes the publication decision.

What peer review achieves

Peer review acts as a quality check before a claim enters the scientific literature. It can prevent seriously flawed work from being published and can improve work that is nearly ready by strengthening its evidence, reasoning or presentation. Publication also makes the methods and findings available to the wider scientific community, allowing other researchers to scrutinise the work, use it and try to reproduce its findings.

This matters because scientists are usually sceptical of unexpected findings until an investigation has been repeated by the original researchers or reproduced independently by others. Agreement across further investigations contributes more convincing evidence than one paper alone.

What peer review cannot guarantee

Peer review is not a certificate that a claim is certainly true. Reviewers generally assess the written report and data; they are not normally expected to repeat all the experiments themselves. An unnoticed mistake, selective reporting or fabricated data may therefore escape detection.

Reviewers also exercise judgement, so different reviewers can reach different conclusions and bias may influence an evaluation. Using several suitably qualified reviewers can provide more than one expert opinion, but it does not remove every limitation.

A peer-reviewed claim should therefore be treated as having passed an important initial check, not as permanently proven. Its reliability becomes clearer as further evidence accumulates, other scientists attempt to reproduce the findings and the claim is considered alongside the wider body of evidence.

Recall the sequence

1. Researchers write up a new claim and submit the paper to a journal.
2. An editor sends it to scientists with relevant expertise.
3. Reviewers check the methods, evidence, interpretation and conclusions.
4. They identify weaknesses and recommend publication, revision or rejection.
5. Authors may revise the paper; the editor decides whether to publish it.
6. After publication, the wider scientific community can scrutinise and try to reproduce the findings.

Purpose

Peer review checks and improves research before publication. It helps filter out unsupported claims and gives other scientists access to work that has received expert evaluation.

Essential distinction

Peer review: experts evaluate the reported work.

Reproduction: other scientists independently carry out an investigation to see whether comparable findings can be obtained.

Reviewers do not usually repeat every experiment.

Limitation

Peer review raises confidence but does not prove that a claim is correct. Errors or bias can be missed, reviewers may disagree, and confidence depends on later scrutiny, reproduction and the wider body of evidence.