|
Activity Number:
|
164
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 9, 2004 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #300422 |
|
Title:
|
A Novelty Test for Testing Treatment Differences under Unparallel Studies
|
|
Author(s):
|
Yi-Ting Hwang*+ and Chien-Hua Wu and Yu-Sheng Wang
|
|
Companies:
|
National Taipei University and Chung Yuan Christian University and National Taipei University
|
|
Address:
|
67, Sec. 3, Ming-Sheng E. Rd., Taipei, International, , Taiwan
|
|
Keywords:
|
analysis of covariance ; unparallel studies ; F test ; Johnson-Neyman technique
|
|
Abstract:
|
The Analysis of Covariance (ANCOVA) has been widely implemented in the various areas to assess the effect of the treatments on the variable of interest controlling for the effects of prognostic variables. Although ANCOVA is a major tool in the statistical analysis, it requires many assumptions such as normality assumption of the error distribution and homogeneity of variance within groups. Furthermore, the coefficient on the prognostic variable is usually assumed to be the same accross treatment groups since the derivation is simple. There are only limited studies that can handle the unparallel problem. However, it requires the mean values of the two prognostic variables are equal. A new test adopting the idea from regression estimators is proposed for testing treatment differences under unparallel studies. The proposed method preserves descent power and Type I error under parallel or unparallel studies, whereas the regular F test is only good for the parallel studies.
|
- 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 2004 program |
