Abstract

An estimator (a statistic) cannot be better than the parameter it estimates. Multi-stage sampling thrives because of its cost-effectiveness and not because its estimator is better. This study emphasis the need to develop higher-order estimators of multi-stage sampling and their use for large-scale sample survey. In this study, the researchers developed estimators of population total and variance of four-stage cluster sampling and illustrated their use with numerical example. Researchers should not sacrifice the development and application of higher-order estimators of multi-stage sampling that are robust statistical techniques on the altar of the complexity of design nature of large population. To reap the actual fruit of sampling that culminates in the high level of predication and estimation accuracy in sample survey, higher-order multi-stage estimators should be used in large-scale sample surveys.

Description