|
Activity Number:
|
409
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Thursday, August 15, 2002 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Computing*
|
| Abstract - #301501 |
|
Title:
|
A Monte Carlo Study of Two Normal Based Homogeneity of Variance Tests for Unbalanced Designs
|
|
Author(s):
|
Peter Wludyka*+ and Anastasia Bergen and Xiaoni Liu
|
|
Affiliation(s):
|
University of North Florida and University of North Florida and University of North Florida
|
|
Address:
|
, , , ,
|
|
Keywords:
|
robustness ; ANOMV ; Bartlett ; power
|
|
Abstract:
|
Few parametric test tests for homogeneity of variance (HOV) are available for analyzing data arising from unbalanced experiments. The power and Type I error stability of two tests that can be used with data arising from independent samples from normal populations will be compared: Bartlett's test and the Analysis of Means for Variances (ANOMV). The ANOMV has the advantage that a decision chart that allows for judging practical, as well as statistical, significance can be produced. Conclusions regarding the nature of distributions (i.e., those that differ from the normal with respect to skewness and kurtosis) for which these tests can safely be used will be provided. Tables useful for answering sample size questions will be provided.
|
- 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 2002 program |
