DMCA
Variance estimation for complex surveys in the presence of outliers (2006)
| Venue: | In Proceedings of the Section on Survey Research Methods |
| Citations: | 1 - 0 self |
BibTeX
@INPROCEEDINGS{Hulliger06varianceestimation,
author = {Beat Hulliger and Ralf Münnich},
title = {Variance estimation for complex surveys in the presence of outliers},
booktitle = {In Proceedings of the Section on Survey Research Methods},
year = {2006},
pages = {3153--3156}
}
Share
 |
 |
 |
 |
||
OpenURL
Abstract
Quantitative variables in surveys often have a markedly skew distribution and, in addition, contain outliers. Robust estima-tors, which may be used in this situation, generally are bi-ased. In addition linearized variance estimators tend to un-derestimate the true variance considerably. Alternatives are Bootstrap variance estimators or estimators based on multi-ple imputation. A simulation study with data from the Swiss Household Budget Survey shows the effects of outliers on esti-mators and their variance estimators. Three poverty measures and proposals for the estimation of their variance are included in the simulation study.
