Year: 2019 Vol.: 68 No.: 2
Authors: Daniel David M. Pamplona
Sampling with probability proportional to aggregate size (PPAS) is compared with traditional design-unbiased sampling methods under different simulated population scenarios in the estimation of the population total. The study considered both accuracy and precision of the estimates in the comparison. Heterogeneous populations were simulated by exploring varying behaviors of an auxiliary variable and its relationship with the target variable. Results show that the optimality of estimates using PPAS sampling improve as the association between the target variable and auxiliary variable strengthens. Furthermore, PPAS sampling estimates are more stable under large variability in the population.
Keywords: Probability Proportional to Aggregate Size Sampling, Nonparametric Bootstrap, Simple Random Sampling, Probability Proportional to Size Systematic Sampling