Examining the ‘gendered’ places and spaces of <scp>UK</scp> doctoral education using multilevel modelling

Abstract Research on gender in higher education frequently focuses on micro‐ or macro‐scale factors—power relations and working practices, or disciplinary norms and the educational ‘pipeline’—overlooking the meso‐scale of ‘place’ embodied in departments and institutions. This study bridges that gap by applying data science and multilevel modelling within a quantitative feminist geographical framework to analyse gendered PhD completion patterns in the UK from 1990 to 2020. Using the British Library's E‐Thesis Online Service (EThOS), we identify ‘department‐like units’ as pivotal, accounting for 10.7% of the variation in the likelihood of a PhD student being female—more than any other grouping. Overall, STEM fields remain male‐skewed, while the Arts and Humanities, and Social Sciences show a female skew. Institutional histories and geography also matter: in Scottish and Northern Irish universities students are more likely to be female compared with English universities, while the likelihood in Wales is lower. The use of statistical methods through a feminist lens offers a foundation for targeted interventions: our results suggest a greater focus on departments in equality initiatives like Athena SWAN and adjustments to funding policies to enhance diversity in PhD cohorts, particularly in male‐dominated disciplines. Future work should integrate intersectional approaches to deepen understanding of these dynamics, but these findings emphasise the value of place‐based analysis in addressing gender disparities and guiding policy.