|
Activity Number:
|
275
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Nonparametric Statistics
|
| Abstract - #305146 |
|
Title:
|
Estimating Equations in Biased Sampling Problems
|
|
Author(s):
|
Bin Zhang*+ and Jianguo Sun and Jing Qin
|
|
Companies:
|
University of Missouri and University of Missouri and National Institute of Allergy and Infectious Diseases
|
|
Address:
|
510 High Street Apt. 316, Columbia, MO, 65201,
|
|
Keywords:
|
Biased data ; Bivariate data ; Empirical likelihood ; Estimating equation ; Likelihood ratio test
|
|
Abstract:
|
An outcome-dependent sampling (ODS) scheme is a retrospective sampling scheme where one observes the exposure/covariates with a probability that depends on the outcome variable. It is widely used in epidemiologic observation studies because of the efficiency of the method. In 2002, Zhou et. al. considered a semiparametric empirical likelihood inference procedure for bivariate data in which the underlying distribution of covariances is left unknown. It requires the conditional distribution of the two variables. In this project, we do not require such condition and leave the joint distribution unspecified. We proved the asymptotic distribution of the likelihood ration test by assuming the information of the distribution function is given by the estimating equations. Simulation studies show the approach is appropriate for practical use and an application is given to illustrate the approach.
|
