|
Activity Number:
|
521
|
|
Type:
|
Invited
|
|
Date/Time:
|
Thursday, August 10, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
WNAR
|
| Abstract - #305307 |
|
Title:
|
Consistent Model Selection and Goodness-of-Fit Test for Marginal Regression Analysis of Longitudinal Data
|
|
Author(s):
|
Lan Wang*+ and Annie Qu
|
|
Companies:
|
University of Minnesota and Oregon State University
|
|
Address:
|
385 Ford Hall, Minneapolis, MN, 55347,
|
|
Keywords:
|
GEE ; marginal model ; lack-of-fit ; model selection
|
|
Abstract:
|
An important approach to analyze longitudinal data is to model the marginal expectation and treat the correlation as a nuisance, such as the method of GEE. The validity of this approach depends on the correct specification of the marginal regression model. We proposes a BIC type procedure to select a marginal regression model from a group of candidate models. The procedure uses the quadratic inference function (QIF) recently proposed by Qu, Lindsay and Li (2000) and does not need to specify the full likelihood. We establish that with probability approaching to one the proposed procedure selects the most parsimonious correct model. Furthermore, a data-driven smooth test is constructed to check the adequacy of a proposed marginal model. Simulation results and data analysis are provided.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2006 program |
