Bayesian Logistic Regression Model for Sub-Areas

Many population-based surveys have binary responses from a large number of individuals in each household within small areas. One example is the Nepal Living Standards Survey (NLSS II), which health status data (good versus poor) for individual sampled households (sub-areas) are available wards (small areas). To make an inference finite population proportion household, we use sub-area logistic regression model with reliable auxiliary information. The contribution this twofold. First, extend area-level to level model. Second, because there numerous sub-areas, standard Markov chain Monte Carlo (MCMC) methods find joint posterior density very time-consuming. Therefore, provide sampling-based method, integrated nested normal approximation (INNA), permits fast computation. Our main goal describe hierarchical Bayesian and show that computation much faster than exact MCMC method also reasonably accurate. performance our studied by using NLSS II data. can borrow strength both areas sub-areas obtain more efficient precise estimates. structure captures variation well.