|
Activity Number:
|
113
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 3, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistical Computing
|
| Abstract - #305290 |
|
Title:
|
A Generalized F-Test
|
|
Author(s):
|
George R. Terrell*+
|
|
Companies:
|
Virginia Polytechnic Institute and State University
|
|
Address:
|
Department of Statistics, Blacksburg, VA, 24061,
|
|
Keywords:
|
Regression ; ANOVA ; Invariance ; Duality
|
|
Abstract:
|
F-tests are an elegant and flexible device for testing hypotheses in normal linear models. However, when we use more general fitting methods, robust for example, we often find ourselves forced to use asymptotic tests of dubious utility. We here use statistics from recently developed co-residual methods of fitting linear methods to accomplish the same ends for finite samples. This approach generalizes an invariance principle due to Jensen and Good. We illustrate our tests for robust regression problems.
|
- 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 2009 program |
