Sex bias in peer review and citation practices: Implications for research evaluation

• We analysed 2001–2003 research products submitted to the first Italian assessment exercise. • Female-authored papers receive lower peer-review scores and citations. • Peer review shows stronger gender bias than citation-based evaluation. • Sex disparities persist across citation windows despite long-term effects. • Findings inform policies promoting equity in research assessment. This study examines the effectiveness of peer review versus bibliometric analysis in reducing sex bias in research evaluations. Drawing on theoretical insights from psychological, sociological, and rational choice frameworks, as well as empirical data from Italy’s first national research assessment exercise (VTR 2001–2003), we conduct a comprehensive comparison between peer-review scores and citation-based metrics across various scientific fields. Our findings reveal substantial and consistent evidence of gender-based disparities disadvantaging female-authored publications. Through ordered logistic regression analyses, we demonstrate that while both evaluation methods exhibit sex bias, peer review systematically penalizes women more severely than citation-based metrics. This disparity remains evident across varying citation windows, despite a long-term Matthew effect that slightly increases sex disparities over extended citation periods. We conclude that bibliometric evaluations, though not entirely free from bias, represent a relatively fairer alternative for assessing research output, holding crucial implications for institutional policy and the pursuit of academic equity.