Investigating access to support centers for Violence Against Women in Apulia: A Spatial analysis over multiple years

TL;DR

This paper models the spatial and temporal variability of violence against women in Apulia using Bayesian methods, revealing how support center access and socioeconomic factors influence reporting and incidence.

Contribution

It introduces a Bayesian spatio-temporal Poisson regression model to analyze gender-based violence and access to support centers across municipalities in a specific Italian region.

Findings

Access to support centers decreases with distance from residences.

Lower education levels correlate with under-reporting in disadvantaged areas.

Higher economic development associates with lower reported violence incidence.

Abstract

In this study, we address the challenge of modelling the spatial variability in violence against women across municipalities in a Southern Italian region by proposing a Bayesian spatio-temporal Poisson regression model. Using data on access to Local Anti-Violence Centers in the Apulia region from 2021 to 2024, we investigate the impact of municipality-level socioeconomic characteristics and local vulnerabilities on both the incidence and reporting of gender-based violence. To explicitly account for spatial dependence, we compare four spatial models within the Integrated Nested Laplace Approximation framework for Bayesian model estimation. We assess the relative fit of the competing models, discussing their prior assumptions, spatial confounding effects, and inferential implications. Our findings indicate that access to support services decreases with distance from the residential municipality, highlighting spatial constraints in reporting and the strategic importance of support center location. Furthermore, lower education levels appear to contribute to under-reporting in disadvantaged areas, while higher economic development may be associated with a lower incidence of reported violence. This study emphasises the critical role of spatial modelling in capturing reporting dynamics and informing policy interventions.