Abstract:
In this paper we have studied empirically the role of single or two auxiliary variables under which cluster sampling in conjunction with ratio method of estimation is more efficient (or less efficient) than using conventional estimator in simple random sampling (SRS), if intra class correlation of the elements within the cluster has positive correlation. We have considered some extreme populations which may help the survey practitioners to know the magnitude of gain or loss in efficiency of cluster sampling using single or two auxiliary variables in conjunction with ratio method of estimation. To have more insight, the clusters are formed for sample sizes 2, 3 and 4.
