|
Activity Number:
|
24
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #309370 |
|
Title:
|
Giving Treatment to Controls: When Is It a Good Idea?
|
|
Author(s):
|
Kimberly Walters*+
|
|
Companies:
|
The Ohio State University
|
|
Address:
|
Cockins Hall, Columbus, OH, 43210,
|
|
Keywords:
|
linear mixed effects ; longitudinal methodology ; simulation studies ; treatment estimation
|
|
Abstract:
|
The power of the likelihood ratio test to detect treatment effect was compared for a full linear mixed effects model and a proposed reduction. Simulation studies were conducted under conditions of equal or unequal slopes in the treatment and control groups following intervention. Initial parameter values were generated by fitting the full model to data from a study in which all participants received treatment, either upon entry or following a control period. The full model does not assume equal post-intervention slopes while the reduced model does, therein combining information from both groups to determine the effect of intervention. We anticipate the reduced model will offer more power to detect treatment effect and an improved estimator of that effect when the post-intervention slopes are not too dissimilar.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
