Empirical Study on Maximizing (Or Minimizing) the Number of Retentions in PPS Sampling Without Replacement: Two Units Per Stratum

摘要

Sampling strategies are determined to maximize (or minimize) the expected number of overlaps between two successive samples using the linear programming approach. The abilities to control overlaps between two successive samples are compared for five methods of sampling with two units selected, in terms of the maximum (or minimum) expected number of overlaps achieved. Six different types of artificial populations are considered to support the empirical study and to illustrate the differences among the five sampling procedures. (ERA citation 09:047227)