Activity Number:
|
5
|
Type:
|
Invited
|
Date/Time:
|
Sunday, August 11, 2002 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Section on Bayesian Stat. Sciences*
|
Abstract - #300774 |
Title:
|
Small Area Estimation Using Area-Level and Unit-Level Models
|
Author(s):
|
J. N. Rao*+ and Yong You
|
Affiliation(s):
|
Carleton University and Statistics Canada
|
Address:
|
, Ottawa, Ontario, K1S-5B6, Canada
|
Keywords:
|
Benchmarking ; census undercoverage ; Gibbs sampling ; unmatched model
|
Abstract:
|
Both unit-level and area-level models are widely used for small-area estimation. Area-level unmatched sampling and linking models are studied using a hierarchical Bayes (HB) approach. The proposed method is applied to Canadian census undercoverage estimation at the provincial level. The sampling model is based on the survey estimate of undercount, while the linking model is a loglinear model for undercoverage rate. An empirical study is also conducted to compare inferences under unmatched models with those obtained under the customary matched models. A basic unit level model is also studied, and design-consistent, model-based estimators are developed using survey weights and a pseudo-EBLUP or pseudo-HB approach. The proposed estimators are shown to have a nice self-benchmarking property. A real data set is used to compare the proposed method to a purely model-based EBLUP or HB method.
|