Activity Number:
|
154
|
Type:
|
Contributed
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biometrics Section
|
Abstract - #308510 |
Title:
|
Development and Application of a Multivariate Mantel-Haenszel Mean Scores Test
|
Author(s):
|
Rebecca McNeil*+ and Robert Woolson and Carol Wagner
|
Companies:
|
Medical University of South Carolina and Medical University of South Carolina and Medical University of South Carolina
|
Address:
|
Department of Biometry, Charleston, SC, 29425,
|
Keywords:
|
Mantel-Haenszel ; multivariate ; stratified ; clustered ; mean scores test ; clinical trial
|
Abstract:
|
Several methods are appropriate for data from clinical trials with stratified randomization: these include linear models and the Mantel-Haenszel mean scores test. It has been demonstrated that the mean scores test has superior power and size to the linear model alternatives in univariate data. Thus, we provide the theoretical foundations for a convenient multivariate extension to this highly useful test. In a series of Monte Carlo simulations of the comparative behavior of the Mantel-Haenszel and two linear models approaches in multivariate data, we found that the Mantel-Haenszel extension performed as well or better than the linear models in a wide range of data structures. We expect that this randomization model approach to multivariate or clustered stratified data will be a useful addition to the statistician's toolbox.
|