|
Activity Number:
|
481
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
General Methodology
|
| Abstract - #307383 |
|
Title:
|
An Examination of the Utility of Bonferroni Adjustments for Tests of Regression Coefficients
|
|
Author(s):
|
Daniel Mundfrom*+ and Jamis Perrett and Jay Schaffer and Adam Piccone
|
|
Companies:
|
University of Northern Colorado and University of Northern Colorado and University of Northern Colorado
|
|
Address:
|
501 20th Street, Greeley, CO, 80639,
|
|
Keywords:
|
multiple linear regression ; simulation ; Bonferroni adjustments
|
|
Abstract:
|
A common research question for which a MLR analysis would be used involves the determination of which, if any, of the predictor variables display a unique ability to explain variation in the response. Often a test of the adequacy of the overall model is followed with separate tests of each predictor to determine which variables are needed in the model to explain the variation in the response. It is not uncommon for each of these tests to be conducted at the nominal 5% level. This research investigated whether conducting each test with a Bonferroni-type adjusted alpha-level is useful for controlling the overall alpha-level closer to the nominal level. In each case, the Bonferroni-adjusted tests more consistently maintained the overall alpha-level below the nominal level. As the number of regressors increased, the utility of making the adjustment to the overall alpha-level also increased.
|
