|
Activity Number:
|
285
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #307315 |
|
Title:
|
Sample-Size Calculations in Logistic Regression: Commnets on Hsieh, Bloch, and Larsen's Method
|
|
Author(s):
|
Mohammed K. Alam*+ and Marepalli Rao and Rupa Mitra
|
|
Companies:
|
University of Cincinnati and University of Cincinnati and Minnesota State University
|
|
Address:
|
CTR for Genome Information, Cincinnati, OH, 45267,
|
|
Keywords:
|
|
|
Abstract:
|
Hsieh, Bloch, and Larsen (1998) have proposed a novel method of calculating sample size in the environment of Logistic regression. Suppose Y is a binary random variable with single covariate X, which has a standard normal distribution, with gamma1 the slope in the model. Let MU1 and MU2 be the expected values of X given Y=1 and Y=0 respectively. It can be shown that gamma1 is zero iff the conditional distributions are standard normal. Thus to test the null hypothesis that gamma1 is zero is equivalent to testing MU1 is equal to MU2. Hsieh, Bloch, and Larsen used two sample methodology to calculate sample size for given level and power for testing the null hypothesis that gamma1 is zero against the alternative that gamma1 is a specified number. In their calculations, they made a number of assumptions. In this presentation, the focus is on the assumptions they have made.
|
- 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 2006 program |
