|
Activity Number:
|
479
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #308472 |
|
Title:
|
An Evaluation of Basic Calibration Estimators in Presence of Nonresponse
|
|
Author(s):
|
Yves Thibaudeau*+ and Jun Shao and Jeri Mulrow
|
|
Companies:
|
U.S. Census Bureau and University of Wisconsin-Madison and National Science Foundation
|
|
Address:
|
1037 17th St S, Arlington, VA, 22202,
|
|
Keywords:
|
Calibration ; nonresponse ; Variance Estimation
|
|
Abstract:
|
Sarndal et al. (1992) discuss calibration estimators for complex surveys. These estimators rely on core survey information carried by covariates to reduce mean squared error. Sarndal and Lundstrom (2005) discuss using calibration estimators in presence of nonresponse. The Survey of Industrial Research and Development sponsored by the National Science Foundation exhibits several patterns of nonresponse involving its core items. But, some covariates are less often missing. We can take advantage of this situation and calibrate core items to reported information when available. Our ultimate objective is to evaluate jackknife variance estimation (Shao and Tu 1995) in presence of nonresponse for this type of calibration estimator.
|
