Improved exponential ratio type estimators in stratified sampling

This paper proposes two new exponential ratio type estimators for estimating finite population mean in stratified random m sampling (st rs). the proposed estimators perform better than the traditional combined ratio and combined difference estimators. they also perform better than the combined family of estimators given by koyuncu and kadilar (2010a),a general class of estimators by singh and vishwakarma (2010),and the exponential type estimator suggested by shabbir and gupta (2011). the expressions for the bias and mean squared errors (mses) of several estimators considered here,correct to first order of approximation,are presented. the performances of these estimators have been compared both theoretically and numerically. four data sets are used for comparisons. © 2013 pakistan journal of statistics.