Sequential data weighting procedures for combined ratio estimators ın complex sample surveys

Abstract

In sample surveys weighting is applied to data to increase the quality of estimates. Data weighting can be used for several purposes. Sample design weights can be used to adjust the differences in selection probabilities for non-self weighting sample designs. Sample design weights, adjusted for nonresponse and noncoverage through the sequential data weighting process. The unequal selection probability designs represented the complex sampling designs. Among many reasons of weighting, the most important reasons are weighting for unequal probability of selection, compensation for nonresponse, and post-stratification. Many highly efficient estimation methods in survey sampling require strong information about auxiliary variables, x. The most common estimation methods using auxiliary information in estimation stage are regression and ratio estimator. This paper proposes a sequential data weighting procedure for the estimators of combined ratio mean in complex sample surveys and general variance estimation for the population ratio mean. To illustrate the utility of the proposed estimator, Turkish Demographic and Health Survey 2003 real life data is used. It is shown that the use of auxiliary information on weights can considerably improve the efficiency of the estimates.