Peer review is the process by which scholarly work is evaluated by independent experts in the same field before it is accepted for publication. Its purpose is to assess whether a manuscript is methodologically sound, original and clearly reported, helping editors decide what to publish and helping authors improve their work. Peer review is a cornerstone of research integrity, though it is not infallible and has attracted serious reform efforts.
At its simplest, peer review answers a question on behalf of readers who cannot check every claim themselves: have qualified experts scrutinised this work and judged it credible enough to enter the scholarly record?
The peer review process step by step
While details vary by journal, the core sequence is broadly consistent.
- Submission and editorial triage. An editor checks the manuscript for scope, basic quality and adherence to journal policy, and may desk-reject before review.
- Reviewer selection. The editor invites independent experts, usually two or more, who have relevant subject knowledge and no disqualifying conflict of interest.
- Assessment. Reviewers evaluate the methods, analysis, originality and clarity, and write structured reports with a recommendation.
- Decision. The editor weighs the reports and decides: accept, request minor or major revisions, or reject.
- Revision and iteration. Authors respond to comments, and the manuscript may go through further rounds before a final decision.
The editor, not the reviewers, makes the final decision; reviewers advise. Authors preparing for this process can consult our guidance for authors on responding to reviewer reports constructively.
Models of peer review
Different journals manage the identities of authors and reviewers in different ways, with consequences for fairness and accountability.
| Model | Who knows whom | Main strength | Main concern |
|---|---|---|---|
| Single-blind | Reviewers know authors; authors do not know reviewers | Reviewers can comment candidly | Possible bias against authors or institutions |
| Double-blind | Neither side knows the other | Reduces identity-based bias | Authors can sometimes be identified from the work |
| Open | Identities are disclosed, reports may be published | Accountability and transparency | Reviewers may hesitate to be critical |
| Post-publication | Review continues after release | Ongoing scrutiny and correction | Less gatekeeping before publication |
The benefits and open questions of the transparent models are examined further in our article on open peer review models.
Ethics and COPE
The Committee on Publication Ethics (COPE) provides widely adopted guidance for editors, reviewers and publishers on handling ethical issues such as conflicts of interest, confidentiality, authorship disputes, plagiarism and research misconduct. Reviewers are generally expected to treat manuscripts as confidential, to declare competing interests, to review only within their competence, and to avoid using privileged information for personal advantage. These norms underpin trust in the system and connect peer review to the wider concerns of our standards dictionary.
Limitations of peer review
Peer review is valuable but imperfect. It can be slow, inconsistent between reviewers, and vulnerable to bias relating to gender, geography or institutional prestige. It is generally better at detecting flawed reasoning than deliberate fabrication, and reproducibility problems can pass through undetected. Recognising these limits is essential to using peer review responsibly rather than treating a published paper as automatically correct.
Reforms: registered reports and transparent review
Several reforms aim to address these weaknesses. The registered reports format reviews the research question and methodology before data are collected, granting in-principle acceptance on the strength of the design rather than the results. This reduces publication bias against negative findings and discourages questionable research practices. Transparent peer review publishes the review reports and author responses alongside the article, allowing readers to see how conclusions were scrutinised. These approaches reflect a broader movement, discussed across our responsible assessment coverage, towards judging research on its substance and process rather than on prestige alone.
Frequently asked questions
Who chooses the reviewers?
The handling editor selects reviewers, looking for relevant expertise and the absence of conflicts of interest. Authors may sometimes suggest or oppose particular reviewers, but the decision rests with the editor.
What is the difference between single-blind and double-blind review?
In single-blind review the reviewers know who the authors are but not vice versa. In double-blind review neither side knows the other’s identity, which is intended to reduce bias linked to author identity, institution or reputation.
What are registered reports?
Registered reports are a publishing format in which the study’s rationale and methods are peer reviewed before the research is carried out. If the design is sound the journal commits to publishing the results regardless of outcome, reducing publication bias.
Does peer review guarantee a paper is correct?
No. Peer review improves quality and filters out many weak submissions, but it cannot guarantee correctness, detect all misconduct, or ensure reproducibility. It is one safeguard among several in the scholarly record.
