Mapping the geography of violence against children in Namibia: A Geographically Weighted Regression Approach

Abstract. This study investigated the geospatial patterns and determinants of violence against children and young people in Namibia using data from 5,191 individuals aged 13–24 years, interviewed across 79 constituencies in the 2019 Violence Against Children and Youth Survey. We employed Global Moran’s I and Local Indicators of Spatial Autocorrelation to identify global spatial autocorrelation and hotspots. Ordinary least squares regression was used to identify significant area-level risk factors, whereas Geographically Weighted Regression was used to model local variations. We found significant regional variations in the prevalence of past-year sexual(SV), physical (PV), and emotional violence (EV), with the highest rates observed in constituencies in central and northern Namibia. Global spatial autocorrelation was evident for SV and EV but not for PV, suggesting distinct spatial patterns for each form of violence in the study area.Our findings also revealed notable spatial variations in risk factors, with area-level factors showing strong influences in certain areas, while being less impactful or negligible in others. These findings emphasise the need for geographically targeted interventions and policies to address localised risk factors and reduce violence against children and young people in Namibia.