On the Improvement of Multivariate Ratio Method of Estimation in Sample Surveys by Calibration Weightings

The most challenging limitation of the ratio estimation is that of deriving variance estimator that admits more than two auxiliary variables. This paper introduces a new calibration weights that prompt the formulation of a multivariate ratio estimator by the calibration tuning parameter subject to a pooled-calibration constraint. Analytical framework for deriving variance estimator that admits as many auxiliary variables as desired is developed. The efficiency gains of the proposed estimator vis-a-vis the Generalized Regression (GREG) Estimator are studied through simulation. Simulation results proved the dominance of the new proposals over existing ones.