Model-based small area estimates of overweight prevalence using sample selection adjustment.

Using a hierarchical model with an adjustment for sample selection, we estimate the overweight prevalence for adults, by states, using data from the Third National Health and Nutrition Examination Survey (NHANES III). A two-stage hierarchical model was selected to account for geographic variability of outcomes and to model possible overdispersion of estimates due to cluster sampling. We compare our model-based estimates with design-based estimates at the national level and obtain excellent agreement. We also provide a check of our model at the state level by comparing estimates with design-based and synthetic estimates.