This article has been reviewed according to Science X's editorial process and policies. Editors have highlighted the following attributes while ensuring the content's credibility: An analysis of almost 40,000 research funding review reports found that female researchers wrote longer, more positive assessments than their male counterparts and addressed the formal evaluation criteria more comprehensively. New research led by University College Dublin in collaboration with the Swiss National Science Foundation (SNSF), published in PLOS One, found that the style and content of grant peer review reports varied by gender and academic discipline.
Using machine learning, researchers analyzed more than 1.3 million sentences from 39,280 peer review reports submitted to the Swiss National Science Foundation between 2016 and 2023. The study examined what reviewers discussed and the tone of their assessments. It found that female reviewers wrote longer reviews across all research areas examined and, after accounting for other factors, made fewer negative comments than male reviewers.
A similar gender difference appeared among applicants, with female applicants receiving fewer negative comments than male applicants. Reviews submitted by female researchers were also more closely aligned with the evaluation criteria. "This study with the Swiss National Science Foundation opened the 'black box' of grant peer review.
Over the past three years, we developed and validated privacy-protecting machine-learning classifiers that allow us to analyze peer review reports at scale," said lead author Associate Professor Stefan Müller of the UCD School of Politics and International Relations. "We could show how a typical grant peer review is written, and how the focus on evaluation criteria and sentiment differs depending on the discipline and the gender of both applicant and reviewer. We hope the patterns we reveal are a starting point for conversations about reforming peer review procedures." The research, conducted by Müller (UCD), Gabriel Okasa, Michaela Strinzel, Anne Jorstad, Katrin Milzow (all SNSF), and Matthias Egger (University of Zurich, Bern, and Bristol), also revealed substantial differences among academic disciplines.
Reviews in the social sciences and humanities (SSH) averaged 928 words, compared with 748 in life sciences (LS) and 716 in mathematics, informatics, natural sciences and technology (MINT). The content of the reviews differed, too. Reviews in MINT placed greater emphasis on applicants' track records and the feasibility of the research.
Reviews in SSH focused more on the relevance and originality of the proposed research than reviews in LS. Reviews in LS, in turn, placed greater emphasis on the suitability of the methods than reviews in either SSH or MINT. The tone of reviews also varied.
Extract — continue reading at the source.
