|
Activity Number:
|
177
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 7, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #304922 |
|
Title:
|
Variance Estimation for Complex Surveys in the Presence of Outliers
|
|
Author(s):
|
Ralf T. Münnich*+ and Beat Hulliger
|
|
Companies:
|
University of Trier and Swiss Federal Statistical Office
|
|
Address:
|
Economics and Social Statistics Department, Trier, 54286, Germany
|
|
Keywords:
|
variance estimation ; outlier ; imputation ; robust estimation
|
|
Abstract:
|
Statistical production often faces the difficulty of finding adequate methods for detecting and treating outliers. This generally is solved by either applying robust estimation methods or identifying outliers and replacing them or adjusting their weight. While, for robust estimators, approximate variance estimations exist for other methods, often one has to resort to resampling methods. Based on the EU projects EUREDIT and DACSEIS, the accuracy and sensitivity of selected variance estimation strategies with respect to outliers will be investigated in a Monte Carlo simulation. The study will be conducted on data from the Swiss Household and Budget Survey and the Swiss Environment Protection Expenditure Survey.
|
