|
Activity Number:
|
70
- Utilizing High-Dimensional and Complex Data in Personalized Medicine
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 30, 2017 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Mental Health Statistics Section
|
|
Abstract #324976
|
View Presentation
|
|
Title:
|
Empirical Likelihood Estimator of Causal Effect under Subclassification
|
|
Author(s):
|
Junvie Pailden*
|
|
Companies:
|
Southern Illinois University Edwardsville
|
|
Keywords:
|
propensity scores ;
subclassification ;
causal inference ;
empirical likelihood
|
|
Abstract:
|
Subclassification or blocking on the estimated propensity score is an effective design stage method to achieve covariate balance in observational studies. The simple blocking estimator for the causal effect is obtained by first computing the block-specific average effects and averaging over the blocks. We use the empirical likelihood approach to construct an alternative estimator of the block-specific average effects by assigning unequal sampling weights on the control and treated units. These sampling weights are obtained by maximizing the nonparametric likelihood under equal covariate moment constraints on the control and treated samples. Performance improvement is shown through simulation results and application example.
|
Authors who are presenting talks have a * after their name.
