Estimating distribution functions with auxiliary information using poststratification

The estimation of a finite population distribution function is considered when auxiliary population information is available. A simple poststratification procedure is compared to a number of more sophisticated methods which have been proposed in the literature. Comparisons are made with respect to a number of criteria and via a simulation study based on two real populations. Whereas some more sophisticated procedures can display greater efficiency in certain circumstances, this can be compensated for by lack of robustness or by other practical disadvantages. Poststratification emerges as a simple and practical procedure offering some useful gains in efficiency.