Conditional and restricted Pareto sampling: Two new methods for unequal probability sampling

摘要

Purpose:To introduce two new unequal probability sampling methods are introduced namely, conditional and restricted Pareto sampling that have high entropy.Summary:Unequal probability sampling is a widely applied sampling method which is also called 7rps sampling (ps stands for proportional to size). This paper focuses on improvements and extensions of Rosen's (Refs. 1 and 2) Pareto order sampling design. The Pareto design sampling schemes are its simple, of high entropy, and uses random numbers that can be considered as permanent. Permanent random numbers are popular nowadays since usage of them induces good change estimates. The sampling method when used for surveys is modified to conditional Pareto sampling for which the factual inclusion probabilities are closer to the desired ones. The Pareto sampling method is also extended to be able to handle cases where there are several restrictions on the sample. Restricted Pareto sampling advantages such as the entropy of the sampling design is high and permanent random numbers are used. (21 refs.)