Paper Info
Title
Selection of Working Correlation Structure in Weighted Generalized Estimating Equation Method for Incomplete Longitudinal Data.
Abstract
The weighted generalized estimating equation (WGEE), an extension of the generalized estimating equation (GEE) method, is a method for analyzing incomplete longitudinal data. An inappropriate specification of the working correlation structure results in the loss of efficiency of the GEE estimation. In this study, we evaluated the efficiency of WGEE estimation for incomplete longitudinal data when the working correlation structure was misspecified. As a result, we found that the efficiency of the WGEE estimation was lower when an improper working correlation structure was selected, similar to the case of the GEE method. Furthermore, we modified the criterion proposed by Gosho etal. (2011) for selecting a working correlation structure, such that the GEE and WGEE methods can be applied to incomplete longitudinal data, and we investigated the performance of the modified criterion. The results revealed that when the modified criterion was adopted, the proportion that the true correlation structure was selected was likely higher than that in the case of adopting other competing approaches.
Year
DOI
Venue
2014
10.1080/03610918.2012.698772
COMMUNICATIONS IN STATISTICS-SIMULATION AND COMPUTATION
Keywords
DocType
Volume
Criterion,Efficiency,Incomplete longitudinal data,Weighted generalized estimating equation,Working correlation structure
Journal
43
Issue
ISSN
Citations 
1
0361-0918
0
PageRank 
References 
Authors
0.34
0
3
Name
Order
Citations
PageRank
Masahiko Gosho100.34
Chikuma Hamada221.17
Isao Yoshimura35521.02