ABSTRACT
Some populations are arranged hierarchically, and sampling may be conducted in a manner that mirrors that hierarchy. For example, in agricultural yield studies, exemplified often in Sukhatme & Sukhatme (1970), villages comprise a number of farms, and each farm has a number of fields planted to wheat, rice, or other crop of interest. It seems natural that a sample of villages is selected, followed by a sample of farms within each sampled village, followed further by a sample of fields within each sampled farm. It is straightforward to imagine a fourth level of hierarchy in this setting which could consist of a tessellation of each field into nonoverlapping cells or strips, such that the ultimate sampling unit would be one of these areal units. Sampling of a population in this hierarchical fashion is known as multistage sampling. When confined to two stages or levels of sampling it is known, appropriately, as two-stage sampling-the principal topic of this chapter. We defer to Sukhatme & Sukhatme (1970) for a more comprehensive treatment of the subject than is provided here.
