|
Activity Number:
|
82
|
|
Type:
|
Invited
|
|
Date/Time:
|
Monday, August 4, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #300020 |
|
Title:
|
Robust Small-Area Estimation Under Unit-Level Models
|
|
Author(s):
|
J. N. K. Rao*+
|
|
Companies:
|
Carleton University
|
|
Address:
|
School of Mathematics and Statistics, Ottawa, ON, K2G 4H8, Canada
|
|
Keywords:
|
unit level models ; resistant methods ; outliers ; parametric bootstrap ; mean squared error ; random effects
|
|
Abstract:
|
Empirical best linear unbiased prediction (EBLUP) estimators of small area means have been obtained under unit level nested error regression models. But EBLUP estimators can be highly influenced by the presence of outliers in the data. We propose a resistant method for small area estimation which is useful for downweighting any influential observations in the data when estimating small area means. A parametric bootstrap method is used to estimate the mean squared error (MSE). A simulation study is conducted to study the efficiency of the proposed robust estimators relative to EBLUP estimators and the relative bias of the bootstrap MSE estimators in the presence of outliers. The proposed robust method is also applied to some real data reported in the published literature.
|
