Random parameters and spatial heterogeneity using rchoice package in R

Abstract

This document provides a brief introduction to models with spatial heterogeneity using a random parameter approach. Specifically, this paper shows how this modelling strategy can be used to capture and model spatial heterogeneity and locally varying coefficients for different latent structure. To show the main advantages of this modeling strategy, the Rchoice package (Sarrias 2016) in R is used. The examples will be focused on the ordered probit model with spatially varying coefficients using self-assessed health status as the dependent variable.