Peer review is a process in which research is evaluated by people with relevant expertise before or, in some forms of publication, after it becomes part of the scientific literature. The term peer refers to researchers who possess knowledge sufficiently relevant to assess aspects of the work, such as its methods, analysis, interpretation, originality, and significance.
In traditional journal peer review, researchers submit a manuscript to a journal, an editor determines whether it is suitable for further consideration, and selected reviewers examine the work and provide comments or recommendations. Reviewers may identify methodological weaknesses, unsupported claims, unclear explanations, statistical problems, missing references, inappropriate conclusions, or issues that require clarification. Peer review is therefore a quality control and scholarly communication mechanism, not a process that proves a paper is correct.
The modern peer review system developed gradually alongside scientific journals and professional scientific communities. Early scientific communication relied heavily on correspondence, societies, demonstrations, books, and proceedings. As the volume of research increased and specialised disciplines developed, formal editorial and reviewing practices became more important. The precise form of peer review has never been uniform. Different journals, disciplines, conferences, funding bodies, and publishers use different procedures.
Some use single blind review, in which reviewers know the authors but authors do not know the reviewers. Double blind review attempts to conceal identities from both sides. Open peer review may disclose reviewer identities or publish reviewer reports. Post publication peer review allows researchers to comment on published work. These variations reflect different attempts to manage bias, accountability, confidentiality, and practical constraints.
The process usually begins when authors submit a manuscript describing their research. An editor or editorial team first assesses whether the submission fits the journal's scope and meets basic requirements. If it proceeds, the editor selects reviewers whose expertise is relevant to the subject and methodological approach. Reviewers assess aspects such as research design, statistical methods, evidence, interpretation, literature coverage, clarity, ethical considerations, and the relationship between conclusions and results.
They normally provide a recommendation such as acceptance, revision, or rejection, although the exact categories vary by journal. The editor makes the publication decision using the reviews and their own assessment. Authors may be asked to respond to reviewer comments and revise the manuscript. Several rounds can occur, especially for technically complex research.
Peer review can examine whether the evidence supports the conclusions rather than merely whether the topic is interesting. A reviewer may notice that a study uses an inappropriate control group, that a statistical model violates important assumptions, that a measurement technique lacks validation, or that an observed association is being incorrectly described as causation. Reviewers can also identify missing context, ambiguous terminology, unsupported extrapolation, inadequate documentation, or relevant previous work that should be considered.
Their role is not necessarily to agree with the authors. Critical evaluation is an essential part of the process because scientific publication is intended to create a record that other researchers can scrutinise and build upon. Reviewers therefore contribute expertise that editors cannot always provide themselves, particularly in highly specialised research areas.
Peer review has important limitations. Reviewers can make mistakes, misunderstand an argument, overlook an error, disagree about appropriate methods, or apply different standards to similar studies. The process can also be affected by conflicts of interest, disciplinary assumptions, prestige effects, confirmation bias, publication pressure, and reviewer availability. Highly novel research may be difficult to evaluate because established reviewers may not yet agree about the appropriate framework.
Reviewers normally have limited time and cannot reproduce every experiment or independently verify every dataset. Peer review also does not guarantee that fraudulent research will be detected, because deliberate misconduct can be difficult to identify from a manuscript alone. These limitations mean that peer reviewed status should be treated as evidence that a work has undergone scholarly scrutiny, not as proof that every statement in it is true.
Peer review differs from replication, editorial review, fact checking, and scientific consensus. Editorial review primarily concerns whether a submission fits a publication's scope and standards, while peer review involves substantive assessment by subject specialists. Replication involves conducting the research again, potentially independently, to determine whether the findings recur. Scientific consensus refers to the state of agreement among specialists based on the accumulated evidence and does not arise from one peer review decision.
Fact checking can assess whether particular claims are supported by available sources but may not involve the methodological scrutiny characteristic of specialist peer review. These processes can complement one another. A published study may be peer reviewed yet later fail to replicate, while a replicated result may contribute to a broader body of evidence that eventually changes the field's understanding.
Peer review also varies by research output. Journal articles commonly undergo manuscript review, while research proposals may undergo grant review, books may undergo academic review, and conference submissions may receive shorter evaluations. Some journals use registered reports in which important aspects of a study's research question and methodology are reviewed before data collection, reducing some forms of outcome dependent publication decisions. Data and code can also receive specialist review. Post publication commentary and corrections provide additional scrutiny after formal publication.
Digital publishing has made rapid public criticism possible, but informal discussion does not automatically have the same evidential status as systematic peer review. The scientific community increasingly treats peer review as one component of a broader system of quality assurance rather than as a single decisive checkpoint.
The importance of peer review ultimately lies in structured expert scrutiny within an ongoing scientific process. It helps filter out some methodological and interpretive problems, improves manuscripts through criticism, makes researchers justify their methods, and provides readers with information that the work has undergone a formal evaluation. At the same time, the reliability of a scientific conclusion depends on the entire subsequent evidence base, including replication, independent studies, data quality, theoretical coherence, and real world testing. Peer review can reject poor research and improve good research, but it cannot transform uncertain evidence into certainty. Its proper role is therefore best understood as a mechanism for disciplined scholarly criticism and communication within a larger system of self correction.
