|
Activity Number:
|
154
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #310275 |
|
Title:
|
Type I and Type II Error Properties for White's Robust Covariance Matrix Estimator in Longitudinal Designs with Several Repeats
|
|
Author(s):
|
Keith Williams*+ and Deborah W. Biddle
|
|
Companies:
|
University of Arkansas for Medical Sciences and University of Arkansas for Medical Sciences
|
|
Address:
|
4301 W Markham Street, Little Rock, AR, 72205,
|
|
Keywords:
|
Longitudinal ; Robust Covariance Estimation ; Repeated Measures ; White's Sandwich Estimator
|
|
Abstract:
|
SAS Proc Mixed and Genmod offer an empirical option to estimate the variance-covariance structure for repeated measure designs. The empirical option uses an estimation approach that is referred to as 'White's sandwich estimator,' and is regarded to possess robustness and statistical accuracy. Its use is generally reserved for settings in which the researcher does not have solid knowledge of a particular covariance structure that is appropriate for a data set at hand. This study investigates the Type I and II error performance of White's estimator when compared to competing approaches in various repeated measure design scenarios.
|
