|
Activity Number:
|
254
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2003 : 12:00 PM to 1:50 PM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #301855 |
|
Title:
|
On Tests for Proportion
|
|
Author(s):
|
Wei-Min Huang*+
|
|
Companies:
|
Lehigh University
|
|
Address:
|
14 E Packer Avenue, Bethlehem, PA, 18015-3175,
|
|
Keywords:
|
Bernoulli trials ; tests for proportion ; normal approximation ; probability of Type I error and power ; bootstrap
|
|
Abstract:
|
Standard tests for proportion based on Bernoulli trials are examined. There are two common approximate tests for proportion in the standard probability and statistics textbooks. Both use normal approximation to determine the rejection regions. The only difference is the standard deviation used in standardizing the sample proportion. Of the two most popular tests, one uses the standard deviation of the sample proportion under the null hypothesis as the normalizing standard deviation. The second option is to use the standard error of the sample proportion as the normalizing standard deviation. This article examines the exact probability of Type I error and the exact power of these standard tests. We also propose a modified test which has better control on the approximate probability of Type I error. Some approximations using bootstrap are also included in this study. In fact, an exact test should probably be used, since it is becoming so easy to obtain the binomial probability using a scientific calculator or a computer.
|
- 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 2003 program |
