Multistage Sampling

Abstract

In chapter 10, we have considered sampling procedures in which all the elements of the selected clusters are enumerated. It was seen that though cluster sampling is generally economical, but it is usually less efficient than sampling of same number of ultimate units directly from the population. This is because the former strategy restricts the spread of the sample over the population. It can, therefore, be logically expected that, for a given number of units in the sample, greater precision can be attained if (1) the units are distributed over a larger number of clusters, and (2) instead of completely enumerating all the units in each selected cluster, only a sample of units is observed. This logic gives rise to the following definition:

Definition 11.1 The procedure of sampling, which consists in first selecting the clusters and then randomly choosing a specified number of units from each selected cluster, is known as two-stage sampling.