|
Activity Number:
|
281
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Consulting
|
| Abstract - #310303 |
|
Title:
|
Small Sample Properties of Information-Based Monitoring
|
|
Author(s):
|
David Sawrie*+ and Christopher S. Coffey
|
|
Companies:
|
The University of Alabama at Birmingham and The University of Alabama at Birmingham
|
|
Address:
|
Department of Biostatistics, 128 Heather Lane, Pelham, AL, 34124,
|
|
Keywords:
|
|
|
Abstract:
|
In the case of normally distributed outcome measures, sample size depends upon the variance, a nuisance parameter. If variance were known, we could easily plan to collect the exact number of observations required to perform a two sample test at the desired level of power. Unfortunately, true variance is unknown. Recent literature suggests the use of information based monitoring to update an initial variance estimate based upon accumulating study data. Specifically, the literature proposes the integration of information based monitoring into a group sequential framework within which data are already monitored periodically for early indications of efficacy. The approach successfully assures sufficient power and maintains approximately nominal alpha levels for large samples. This paper investigates the effect of this method upon power and type I error when small sample sizes are considered.
|
- 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 |
