Peer review perpetuates barriers for historically excluded groups

A meta-analysis of peer-review data from over 300,000 biological sciences manuscripts reveals worse review outcomes for authors from historically excluded groups, and limited data evaluating the effectiveness of interventions to address bias in peer review. Peer review is central to the scientific process and scientists' career advancement, but bias at various stages of the review process disadvantages some authors. Here we use peer review data from 312,740 biological sciences manuscripts across 31 studies to (1) examine evidence for differential peer review outcomes based on author demographics, (2) evaluate the efficacy of solutions to reduce bias and (3) describe the current landscape of peer review policies for 541 ecology and evolution journals. We found notably worse review outcomes (for example, lower overall acceptance rates) for authors whose institutional affiliations were in Asia, for authors whose country's primary language is not English and in countries with relatively low Human Development Indices. We found few data evaluating efficacy of interventions outside of reducing gender bias through double-blind review or diversifying reviewer/editorial boards. Despite evidence for review outcome gaps based on author demographics, few journals currently implement policies intended to mitigate bias (for example, 15.9% of journals practised double-blind review and 2.03% had reviewer guidelines that mentioned social justice issues). The lack of demographic equity signals an urgent need to better understand and implement evidence-based bias mitigation strategies.

Automated summary

A meta-analysis of peer-review data from over 300,000 biological sciences manuscripts indicates that authors from historically excluded groups experience worse review outcomes, and there is limited data on interventions to address bias in peer review. The study highlights the need for evidence-based strategies to mitigate bias and improve diversity in the peer review process, as well as the lack of implementation of such policies in current journals.