Abstract:
|
Mashreghi, Haziza and Léger (2016), published in Statistics Surveys, presented a survey of all important existing bootstrap methods for survey data taken under stratified simple random sampling and unequal probability sampling designs. The methods were presented in a unified way by classifying them in three classes: pseudo-population, direct, and survey weights methods. Variance estimation and the construction of confidence intervals were addressed. In this paper, we present the results of a large simulation study to compare the various methods. In some cases, the methods involve a tuning parameter and the effect of its choice is studied. Most of the bootstrap methods have been derived in such a way that they reproduce the usual estimate of variance for the mean (or total) of a population. We evaluate the method on many other statistics, such as the correlation coefficient, the median and the Gini index. This talk should be of interest to any survey statistician using the bootstrap.
|