A unit-level one-inflated beta model for small area prediction of seat-belt use rates

Abstract

We develop a unit-level one-inflated beta model for the purpose of small area estimation. Our specific interest is in estimation of seat-belt use rates for Iowa counties using data from the Iowa Seat-Belt Use Survey. As a result of small county sample sizes, small area estimation methods are needed. We propose frequentist and Bayesian implementations of a unit-level one-inflated beta model. We compare the Bayesian and frequentist predictors to simpler alternatives through simulation. We apply the proposed Bayesian and frequentist procedures to data from the Iowa Seat-Belt Use Survey.