Peer review

• 1. Introduction

  ›

  Peer review is the process by which a piece of work is judged not by its creator, but by someone equally qualified to have made it. That simple idea sits at the heart of how science, medicine, law, engineering, and education have organised themselves for centuries. Top journals today reject over ninety percent of what they receive. The reviewers who decide that fate are experts in the same field as the author, and they remain anonymous throughout. The author does not know who reads their work, and the reviewer cannot be pressured. How did a practice this powerful become standard across so many disciplines? And why, despite its influence, does it remain one of the most contested and inconsistent processes in professional life?

• 2. A Long History

  ›

  Ishaq ibn Ali al-Ruhawi, who lived from 854 to 931, described what may be the earliest recorded version of professional peer review in a work called the Ethics of the Physician. His system was precise. A visiting physician was required to make duplicate notes of a patient's condition on every visit. When the patient recovered or died, a local medical council of other physicians examined those notes to determine whether the standard of care had been met. The principle was already clear: performance should be measured against shared professional standards by people with the standing to judge.

  Henry Oldenburg, born in 1619 and died in 1677, is recognised as the father of modern scientific peer review. A German-born British philosopher, Oldenburg helped establish the practice of having experts evaluate work before publication in scientific circles. The process continued to develop over the following centuries, and the journal Nature made scholarly peer review standard practice in 1973. The term "peer review" itself did not enter common usage until the early 1970s. A monument to peer review has stood at the Higher School of Economics in Moscow since 2017.

• 3. Across Professions

  ›

  Clinical peer review in health care assesses a physician's handling of patient care and feeds directly into decisions about credentialing and privileging. The practice extends beyond medicine into physician peer review, nursing peer review, and dentistry peer review as separate streams. Accounting, law, aviation, engineering, and even forest fire management have all developed their own forms of peer evaluation.

  In engineering, technical peer review is a defined process carried out by small teams, typically limited to six or fewer people, who hold assigned roles and look specifically for defects. These reviews occur within development phases rather than at the end of a project. The European Union began using peer review to coordinate active labour market policy across member states in 1999, and in 2004 it extended the approach to social inclusion, sponsoring around eight peer review meetings per year. In those meetings, a host country opens a specific policy to examination by roughly half a dozen other countries and relevant European-level organisations, usually over two days that include site visits. The State of California went further than any other US state: in 1997, the governor signed Senate Bill 1320, known as the Sher bill, which mandated external scientific peer review for any regulatory decision by a California Environmental Protection Agency body. That requirement is written into California Health and Safety Code Section 57004.

• 4. In the Classroom

  ›

  Peer review entered education as a tool for reaching what Bloom's taxonomy describes as higher-order processes in both the affective and cognitive domains. In practice, this means groups of students reading and responding to each other's writing. Proponents argue the method helps students understand how their work lands with a diverse readership before a teacher grades it, builds professional habits, and can strengthen confidence on both sides of the exchange. Research has found that students tend to leave more positive comments than negative ones when reviewing classmates' work.

  Benjamin Keating, in an article titled "A Good Development Thing: A Longitudinal Analysis of Peer Review and Authority in Undergraduate Writing," followed two groups of students over time, one majoring in writing and one not, to compare how they valued peer feedback. The conclusion was specific: students in non-writing majors tended to discount mandatory peer review, while students majoring in writing gave classmates' comments considerably more weight. Elizabeth Ellis Miller, Cameron Mozafari, Justin Lohr, and Jessica Enoch identified three recurring structural problems in classroom peer review: insufficient training, limited student engagement, and time constraints that push feedback toward the surface rather than the substance of the writing. Pamela Bedore and Brian O'Sullivan reached a related conclusion after comparing peer review forms across two universities: the real goal is not better sentences but helping an author achieve their original writing vision, and most non-professional reviewers do not operate at that level.

• 5. The Bias Problem

  ›

  "Role duality" is one of the central tensions in peer review. A person who is simultaneously an evaluator and a person being evaluated has a stake in the outcome on both sides. Research has found that this parallel position leads evaluators to make strategic choices designed to improve their own chances of receiving a positive assessment.

  The editorial peer review process has also shown a persistent bias against studies that produce negative results, meaning studies that found something did not work. Richard Smith, writing in the Journal of the Royal Society of Medicine, framed the problem plainly: "Who wants to read something that doesn't work? That's boring." When journals systematically avoid publishing negative results, the information base that medicine and science build on becomes skewed. Magda Tigchelaar ran an experiment that divided students into three groups: one doing self-assessment, one doing peer review, and one doing neither. Across four writing projects, only the self-assessment group showed significant improvement. Her analysis attributed this to the fact that the author knows the revision goals most clearly and can apply a systematic approach, while peer feedback is often scattered and fails to meet the author's actual needs. Meanwhile, a survey conducted by Nature confirmed through interviews that artificial intelligence is being used, sometimes without disclosure, to assist with or carry out peer review. A small number of documented cases involve scholars inserting human-invisible prompts into their preprints specifically to influence any automated reviewing system that might process the paper.

• 6. Reform and Alternatives

  ›

  Stephanie Conner and Jennifer Gray argue that many peer review sessions fail because students who are reviewing do not feel confident enough in their own writing to offer useful guidance. Among the practical improvements they propose is segmenting the review into small groups where one student presents a paper, others take notes and analyse it, and the discussion then opens to the full class. A narrower technique suggested elsewhere involves a writer asking their reviewer exactly three focused questions about the paper, which gives the review a concrete scope and can help build trust between the two parties.

  Mimi Li studied an online peer review software tool used in a freshman writing class over one semester. Unlike traditional in-class methods, the software provided reviewers with a set of structured questions they could pose and allowed comments to be attached directly to selected passages of text. Students showed varying degrees of improvement in grades and writing quality, and they praised the technology. Researchers have also proposed alternatives to peer review in contexts beyond writing, including the use of funding-by-lottery as a substitute for peer review in science funding decisions. Instructional assistants have been recommended in place of student peers for graded school assignments on the grounds that they bring more consistent expertise and fewer competing biases to the task.