|
Activity Number:
|
169
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistical Education
|
| Abstract - #304172 |
|
Title:
|
Hartley F-max Statistic for Unequal Sample Sizes
|
|
Author(s):
|
William Warde*+ and Guohui Yang
|
|
Companies:
|
Oklahoma State University and Oklahoma State University
|
|
Address:
|
Departmetn of Statistics, 301 MSCS, Stillwater, OK, 74078,
|
|
Keywords:
|
F-max ; Hartley ; Homogeneity of Variance
|
|
Abstract:
|
Hartley's F-max test is a simple and convenient test used to test for homogeneity of variances before conducting an ANOVA. The original tables give the 1% and 5% critical values for equal sample sizes. Beckman and Tietjen (1973) extended these table to the 10% and 25% critical values for equal sample sizes. We extend these further to the 50%, 75%, 90%, 95% and 99% critical values and also examine methods to deal with unequal sample sizes. A Monte Carlo routine was programmed to obtain critical values for unequal sample sizes and the use of several functions of the sample sizes was studied to try and determine quick and easy approximate critical values.
|
