Activity Number:
|
613
|
Type:
|
Contributed
|
Date/Time:
|
Wednesday, August 3, 2016 : 2:00 PM to 3:50 PM
|
Sponsor:
|
Survey Research Methods Section
|
Abstract #319848
|
View Presentation
|
Title:
|
A Calibrated Bayesian Method for Propensity Score Estimation
|
Author(s):
|
Hejian Sang* and Jae-kwang Kim
|
Companies:
|
Iowa State University and Iowa State University
|
Keywords:
|
Nonresponse ;
propensity score ;
Bayesian approach ;
ignorable ;
nonignorable
|
Abstract:
|
Decreasing response rates in sample surveys is a serious problem and nonresponse weight adjustment using propensity score method is commonly adopted to compensate for unit nonresponse. However, the inference after nonresponse weight adjustment is complicated because the effect of estimating propensity score method needs to be incorporated. In this paper, we develop a new Bayesian approach to handling unit nonresponse under a parametric model for the response probabilities. The proposed Bayesian method is calibrated to frequentist inference. The computation does not involve Taylor linearization. Our proposed method is applicable to both ignorable and nonignorable response mechanism. Some asymptotic properties are established and the results from two simulation studies are presented.
|
Authors who are presenting talks have a * after their name.