On Pareto \(\pi\)ps sampling: reflections on unequal probability sampling strategies (Q2740071)

The authors consider Pareto \(\pi\)ps sampling and Poisson mixture sampling. These designs are compared with model-based simple random stratified sampling. From these comparisons, based on anticipated variances and a simulation study from a population that mirrors factual conditions, follows that the model-based stratified sampling approach still is a strong alternative in applications where the probability proportional-to-size sampling is considered to be appropriate.