SMALL AREA ESTIMATION UNDER A SPATIAL MODEL USING DATA FROM TWO SURVEYS

In this paper, we describe small area estimation (SAE) under a spatial dependent random effects model by combining data from two independent surveys. The spatial dependence is introduced using simultaneous autoregressive (SAR) structure in the random area effects part of the model. We use data from two independent surveys. The first survey, small in sample size, collects both variable of interest as well as auxiliary variables and the second survey, relatively larger in sample size, has some auxiliary variables common to the first survey. Our empirical results, based on simulation studies, show that proposed SAE method using the data from two surveys is efficient as compared to the one based on data from single survey. Use of spatial information further enhances the efficiency of the proposed estimator.