Non-Overlapping, Overlapping, Post, and Adaptive Cluster Sampling

In survey sampling the basic assumption is that the population consists of a finite number of distinct and identifiable units. A group of such units is called a cluster. If, instead of randomly selecting a unit for sample, a group of units is selected as a single unit in the sample, it is called cluster sampling. If the entire area containing the population under study is divided into smaller segments, and if each unit of the population belongs to only one segment, the procedure is called area sampling or non-overlapping cluster sampling. If one or a few units appears in more than one segment or cluster, then such a procedure is called overlapping cluster sampling. The main purpose of cluster sampling is to divide the population into small groups with each group serving as a sample unit. Clusters are generally made up of neighbouring elements; therefore the elements within a cluster tend to be homogeneous. However at some stage in the research we become interested in heterogeneous clusters rather than homogeneous. More broadly, the concept of forming strata in the previous chapter was to form homogeneous groups, whereas in this chapter the concept of forming clusters will be to form groups of a heterogeneous nature. After dividing the population into clusters the sample of clusters can be selected with either equal or unequal probability. The concept of unequal probability may be based on the size of the cluster; that is, the larger the cluster, the larger the probability of its being selected in the sample. All the units in the selected cluster will be enumerated. As a simple rule the number of units in a cluster should be small and the number of clusters should be large.