|
Activity Number:
|
186
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 4, 2008 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #302079 |
|
Title:
|
Spatial Modeling and Prediction of County-Level Employment Growth Data
|
|
Author(s):
|
Nadarajasundaram Ganesh*+
|
|
Companies:
|
National Opinion Research Center
|
|
Address:
|
4350 East-West Highway, Suite 800, Bethesda, MD, 20814 ,
|
|
Keywords:
|
Fay-Herriot model ; Infill asymptotics ; Increasing domain asymptotics ; Empirical best linear unbiased predictor ; Spatial Statistics ; Small Area Estimation
|
|
Abstract:
|
For correlated sample survey estimates, a linear model with covariance matrix in which small areas are grouped into clusters by a similarity measure based on spatial locations is proposed. In the context of correlated data, a novel asymptotic framework, a hybrid of infill asymptotics and increasing domain asymptotics is introduced. The hybrid asymptotic framework assumes that the number of clusters and the number of small areas in each cluster grows with sample size. Under the previously mentioned asymptotic framework, the proposed parameter estimators are sqrt(k) consistent, where k is the number of clusters. The proposed model is implemented for county-level civilian employment growth data.
|
