|
Activity Number:
|
66
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 6, 2006 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #305475 |
|
Title:
|
Estimation of Low Incidence Rates under Selection Bias
|
|
Author(s):
|
Bin Wang*+ and Jiayang Sun
|
|
Companies:
|
University of South Alabama and Case Western Reserve University
|
|
Address:
|
9713 Concord Place, Mobile, AL, 36695,
|
|
Keywords:
|
missing data ; semi-parametric ; biased sampling
|
|
Abstract:
|
This paper is motivated by a study of the cancer risks of Vietnamese Americans along the central Gulf Coast. In the study, researchers encountered two difficulties: selection bias and low incidence rate. This may cause under-estimate in estimation by a standard logistic regression model. In this paper, the authors considered using a semiparametric method to evaluate the risk of rare events from biased data with nonignorable missing values. A generalized additive model is used and a modified iterative reweighted least square estimator is developed to correct the selection bias and account for missing values. The new method will be compared with existing methods, and simulation will be performed to illustrate the performance of the new estimators.
|
