|
Activity Number:
|
133
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #307153 |
|
Title:
|
Small-Area Estimation Using Nonparametric Regression
|
|
Author(s):
|
Tathagata Bandyopadhyay*+
|
|
Companies:
|
Indian Institute of Management Ahmedabad
|
|
Address:
|
Wing14 Vastrapur, Ahmedabad, 380015, India
|
|
Keywords:
|
small area estimate ; nonparametric regression
|
|
Abstract:
|
Small-area estimates are being used increasingly by policymakers for making important decisions. However, small sample sizes---often available for small areas---make the estimates unreliable. An often used approach for producing reliable small-area estimates is to employ a mixed linear model. It borrows strength from survey data collected in other small areas for getting an estimate of a small area. The success of this model-based approach depends largely on the choice of the model and, of course, the availability of good auxiliary data. We address the question of choosing a correct regression model. More specifically, we consider finding small-area estimates using a nonparametric model.
|
