ROBUST REGRESSION TYPE ESTIMATORS FOR BODY MASS INDEX UNDER EXTREME RANKED SET AND QUARTILE RANKED SET SAMPLING

Abstract

. Robust regression-type estimators of population mean that use auxiliary variable information are proposed by considering robust methods under extreme ranked set sampling (ERSS) and quartile ranked set sampling (QRSS). We have used the data concerning body mass index (BMI) for 800 people in Turkey in 2014. The real data example is applied to see efficiency of the estimators in ERSS and QRSS designs and it is found that the proposed estimators are better in these designs than the classical ranked set sampling (RSS) design. In addition, mean square error (MSE) and percent relative efficiency (PRE) are used to compare the performance of the adapted and proposed estimators.