The Efficient Use of Supplementary Information in Standard Sampling Procedures

Abstract

SUMMARY Use is made of a well-known heteroscedastic super-population model to obtain comparisons of the efficiencies of six methods of sampling in common use. These are simple random sampling with unbiased estimators, simple random sampling with ratio estimators, and sampling with probability proportional to size with unbiased estimators; each with and without replacement. The resulting comparisons are readily interpretable, thus affording simple criteria for comparing the efficiencies of these widely used sample plans. They also provide heuristic explanations of their properties.