No abstract

This work is an extension of stratified sub-sampling in which both probability proportional to size (PPS) sampling and ranked set sampling (RSS) is introduced for estimating the population mean. This method first stratifies the population consisting of clusters or primary stage units. Then, clusters within each stratum are selected using PPS with replacement. The selection of units in each cluster is done by RSS without replacement in the second stage of sampling. Using real-life population based on the Census of India (2011) which relates to the total number of agricultural labourers in Imphal West district, Manipur, the population mean is estimated under the proposed estimator and its efficiency with respect to the conventional stratified two-stage cluster sampling is also studied.

A new estimator based on two-stage sampling using ranked set sampling in the second stage is constructed in this paper. We demonstrate a preliminary test regression estimator for two auxiliary variables using double sampling in the sampling scheme when there is partial information on one of them. Its bias and mean squared error are derived and the efficiency of the suggested estimator is also computed.