Bayesian Estimators for Small Area Models when Auxiliary Information is Measured with Error