|
Activity Number:
|
116
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Epidemiology
|
| Abstract - #304231 |
|
Title:
|
Empirical Likelihood Methods in Efficient Design and Inference of Randomized Clinical Trials
|
|
Author(s):
|
Xiaoru Wu*+ and Zhiliang Ying
|
|
Companies:
|
Columbia University and Columbia University
|
|
Address:
|
Room 1021, SSW, 1255 Amsterdam Ave., New York, NY, 10027,
|
|
Keywords:
|
randomized clinical trial ; empirical likelihood with infinitely many constraints
|
|
Abstract:
|
In many randomized clinical trials, the primary goal is to estimate the marginal treatment effect. Many attempts have been made to improve the efficiency of inferences when the auxiliary baseline covariate information is also available. This article proposes an empirical likelihood approach to estimate the marginal effect consistently and efficiently while having the computational superiority over the existing techniques. Certain rate at which the number of estimating functions is growing with the increasing sample size is provided, which guarantees the maximum empirical likelihood estimate is consistent and efficient under some regularity conditions. Epidemiological applications demonstrate the performance of the method.
|
