Compromise Allocation in Univariate Stratified Sampling

Abstract

To reduce the loss of efficiency in the Neyman allocation caused by using the estimators instead of the unknown strata standard deviations of population, we suggest a compromise allocation that the Neyman allocation using an estimator of the pooled standard deviation of combined strata and the proportional allocation are used together. It is shown that the compromise allocation makes the estimator more efficient than the proportional allocation and the Neyman allocation using the estimated strata standard deviations. Simulation study is carried out for the numerical comparison and the results are reported.

Keywords:

• Compromise allocation
• Neyman allocation
• Proportional allocation

Mathematics Subject Classification:

• Primary 62D05

Acknowledgments

This research was supported by Maeji research fund of 2006 (S. Na) and by KOSEF through the Statistical Research Center for Complex Systems at Seoul National University (J. Jeon).