["The Geographical Journal, Volume 192, Issue 3, September 2026. ", "\nShort Abstract\nThis study utilised local spatial regression techniques—geographically weighted regression (GWR) and multiscale geographically weighted regression (MGWR)—to examine divorce patterns among educated women in Kuwait, addressing a gap in literature that treated sociodemographic drivers as spatially uniform. The MGWR model outperformed traditional approaches by revealing that factors such as the number of sons, fertility rates and women in principal jobs operate on local urban scales rather than globally. These findings demonstrate the value of spatial modelling for informing targeted policy interventions aimed at reducing family dissolution.\n\nABSTRACT\nMarital instability is influenced by socioeconomic status, education and gender. In Arab nations, divorce among educated women has increased as cultural and socioeconomic barriers to marital dissolution have weakened. More specifically, in Gulf Cooperation Council countries, particularly Kuwait, patterns of divorce among women with higher education are changing; yet, notably, the spatial dimensions of these trends remain underexplored. Accordingly, this study utilises subnational datasets and employs a global ordinary least squares (OLS) model, in conjunction with geographically weighted regression (GWR) and multiscale geographically weighted regression (MGWR), to investigate the influence of sociodemographic factors on divorce rates among educated women in urban areas of Kuwait. Importantly, the MGWR model enhances the model fit compared to the GWR, thereby demonstrating that the predictors function at varying geographic scales. The findings indicate that women's employment in government, healthcare and STEM fields, as well as leadership positions, is positively associated with divorce among educated women, whereas, conversely, fertility and having sons are negatively correlated with divorce. Furthermore, these associations differed across neighbourhoods. Consequently, the results identify regions where female labour market participation and family structure are most strongly connected to divorce, thus highlighting the potential for localised interventions. Ultimately, this study exemplifies the efficacy of geospatial approaches in modelling social outcomes and advocates targeted strategies to mitigate family dissolution risks in Kuwait and similar urban areas.\n"]