Bootstrap Variance Estimation Technique under Dual Frame Ranked Set Sampling.

Abstract

Multiple frames are preferably used when a satisfactory sampling frame, covering the whole population in question, is unavailable or even if such a frame is available it may not be economically advantageous to use that frame for survey because of high cost of sampling per unit. In this paper, we dealt with the problem of variance estimation of the dual frame ranked set sample (DFRSS) estimator. We propose two rescaling Bootstrap variance estimation methods viz. strata based and cluster based, to obtain an unbiased estimator of the sampling variance of the proposed estimator. The comparison of performance of the proposed rescaled bootstrap methods with standard bootstrap methods were investigated through a simulation study. The simulation results show that the proposed methods are more stable and have lesser relative bias than the standard approaches. Among the two rescaling Bootstrap variance estimation methods, the strata based rescaling Bootstrap variance estimation approach is more powerful than its counterpart.