A modified spatial+ approach to remove bias in fixed effects estimates in multivariate spatial models for areal count data
Conference
65th ISI World Statistics Congress 2025
Format: IPS Abstract - WSC 2025
Session: IPS 809 - New Avenues in Disease Mapping
Wednesday 8 October 10:50 a.m. - 12:30 p.m. (Europe/Amsterdam)
Abstract
Spatial confounding can be defined as the impossibility to separate fix and random effects in spatial models. Difficulties are accentuated in a multivariate framework as covariates can affect the responses differently. Though several procedures have been proposed in the literature, no definite solution has been achieved yet. Here, we propose a modified spatial+ procedure, consisting on a partition of the covariate into two components describing large and short scale spatial dependence. As illustration, we jointly analyse two types of crimes against women in Uttar Pradesh, India.