Spatial patterns of divorce incidence in urban neighborhoods of Kuwait: mapping family dissolution

Divorce is a common global concern that affects familial harmony and stability. This study assessed the advantages of adopting spatial local models, including geographically weighted regression and geographically weighted Poisson regression, compared to the global ordinary least squares model for divorce incidence in Kuwait. This study utilized 2022 administrative data points from the Public Authority for Civil Information, offering neighborhood-level insights into divorced women and their sociodemographic characteristics. The data were incorporated into a geo-referenced spatial dataset by linking statistical records with urban neighborhood polygons within (GIS). This integration facilitated local regression modeling and mapping of divorce incidence and its determinants. The local GWPR model demonstrated an optimal fit to the data, as evidenced by a lower Akaike Information Criterion corrected (AICc) value of 5628 compared to the global model’s 14,298, and a high R2 value of 0.99. Additionally, the model generated residuals devoid of spatial autocorrelation, as indicated by the Moran’s I. The GWR and GWPR models effectively reduced residual clustering compared with OLS and more accurately represented regional variability in the associations between divorce and sociodemographic factors. Result reveals notable spatial disparities in these associations at local subnational administrative boundaries, with population density, female education, widowed women, number of children and percentage of housewives emerging as statistically significant predictors of divorce patterns. These findings contribute to a more nuanced understanding of the relationship between family dissolution and sociodemographic and cultural factors across local neighborhoods and may support geographically targeted programs to assist at-risk spouses and promote marital and familial stability.