Analytical Surveys with Cluster Sampling

Summary This paper concerns the designing of certain analytical surveys (utilizing cluster sampling). The general concepts and problems peculiar to this type of study are discussed. The question of the relevant population for inference purposes is treated, and this is reflected in the model considered. Illustrations are given to clarify the situations where this type of model may be applicable. In general, the objective in planning such a survey is to select the sample so that maximum precision for the comparisons to be made is obtained for a specified total cost. Sample size allocations are carried out for several specifications of survey goals—these specifications being related to the type of comparisons to be made.