Study of efficiency of stratified estimator of population mean under correlated measurement error

Abstract

Relaxing the general assumption that measurement errors are uncorrelated, this research work extends the works done on correlated measurement errors under simple random sampling scheme to stratified random sampling scheme and examines the consequence of correlated measurement error on separate regression estimator of population mean. The properties of the suggested estimator up to first order approximation were obtained. Theoretical as well as empirical comparison of the suggested estimator with the traditional unbiased stratified random sampling estimator when measurement errors are correlated was carried out. It was detected that the suggested estimator is a biased estimator and that correlated measurement errors inflate mean squared error of the suggested estimator but have no effect on the bias of the suggested estimator. The suggested estimator also recorded high loss in efficiency due presence of correlated measurement error. The paper concluded that the suggested estimator is more efficient than usual unbiased estimator.