|
Activity Number:
|
504
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 10, 2006 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Bayesian Statistical Science
|
| Abstract - #305395 |
|
Title:
|
Bayes Approach to Dependent Multiple Comparisons
|
|
Author(s):
|
Lemuel Moye*+
|
|
Companies:
|
The University of Texas Health Science Center at Houston
|
|
Address:
|
School of Public Health, RAS Building E815, Houston, TX, 77025,
|
|
Keywords:
|
Bayes ; multiple comparisons ; type I error level ; dependency
|
|
Abstract:
|
The statistician working in the setting of hypothesis testing has several useful tools. The approach to dependent hypothesis testing that models the conditional probability of the type I error of one hypothesis test in a set, given knowledge of a type I error on another hypothesis test in the same set, is a useful approach. This procedure permits investigators who wish to conserve alpha-level errors in a dependency setting to retain control of the type I error levels for the hypotheses they wish to test. However, this latter approach is hampered by requiring investigators to select the dependency parameter D (0< =D< =1). A probability distribution for D is provided. A closed-form solution for the type I error levels for dependent statistical hypothesis tests is available. An example of the procedure is offered.
|
