|
Activity Number:
|
447
|
|
Type:
|
Invited
|
|
Date/Time:
|
Wednesday, August 9, 2006 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Section on Quality and Productivity
|
| Abstract - #304900 |
|
Title:
|
A Bayesian Approach to the Analysis of Split-Plot Experiments
|
|
Author(s):
|
Peter Goos*+ and Steven G. Gilmour
|
|
Companies:
|
Universiteit Antwerpen and Queen Mary, University of London
|
|
Address:
|
Faculty of Applied Economics, Antwerpen, 2000, Belgium
|
|
Keywords:
|
generalized least squares ; variance component estimation ; degrees of freedom
|
|
Abstract:
|
Industrial experiments often involve hard-to-change variables not reset for every run of the experiment. The resulting experimental designs are of the split-plot type. A proper classical statistical analysis requires the use of generalized least squares procedures and, hence, the estimation of the variance components in the statistical model. In most split-plot designs, the hard-to-change variables are reset a small number of times, such that estimation of the variance components corresponding to the whole plots stratum of the experiment is either impossible or inefficient. As a result, these variance components often are estimated to be zero and the generalized least squares inferences collapse to ordinary least squares ones. This problem can be avoided by incorporating prior beliefs regarding the magnitudes of the variance components in a Bayesian data analysis.
|
