Efficient unequal probability resampling from finite populations

A resampling technique for probability-proportional-to size sampling designs is proposed. It essentially based on a special form of variable probability, without replacement applied directly to the sample data, yet according pseudo-population approach. From theoretical point view, it asymptotically correct: as both and population increase, under mild regularity conditions proposed design tends coincide with original which data were collected. computational methodology easy be implemented efficient, because neither requires actual construction nor any randomization ensure integer weights sizes. Empirical evidence simulation study 1 indicates that outperforms its two main competitors confidence interval various parameters including quantiles.