The linguistic characteristics of peer review reports are not influenced by research area, type of review or reviewer gender, which is evidence for the robustness of peer review.

Generative artificial intelligence, such as OpenAI’s ChatGPT, can be used to analyze scientific texts with specialized constructions, including peer review reports.

A survey of journals and editors in five areas of research - ecology, economics, medicine, physics and psychology - reveals a range of differences in their approach to peer review.

Early career researchers commonly peer review manuscripts on behalf of invited reviewers, often without receiving feedback or being named to the journal.

An analysis of peer review and funding outcomes of NIH research applications shows that funding disparities of topics preferred by African American Black investigators are not due to peer review preferences or biases.

A study of more than 75,000 grant proposals to the European Union indicates that the outcomes of the peer review process remain stable in response to changes in the way that peer review is organized.

eLife is changing its editorial process to emphasize public reviews and assessments of preprints by eliminating accept/reject decisions after peer review.

Gender-bias in peer reviewing might persist even when gender-equity is reached because both male and female editors operate with a same-gender preference whose characteristics differ by editor-gender.

Redaction, intended to obscure personal and institutional identity, reduced the gap between Black vs. White applicants in simulated review of real NIH scientific grant applications, thus providing qualified support for double-blinded models of peer review.

When deciding which submissions should be peer reviewed, eLife editors consider whether they will be able to find high-quality reviewers, and whether the reviews will be valuable to the scientific community.