|
Title
|
Room
|
|
Analyzing Fractional Factorial Experiments in the Presence of Dispersion
|
M-International Salon B
|
|
Date / Time
|
Sponsor
|
Type
|
|
08/04/2001
8:15 AM
-
4:15 PM
|
ASA, Section on Quality & Productivity*
|
Other
|
|
Organizer:
|
n/a
|
|
Chair:
|
n/a
|
|
Discussant:
|
|
|
CE Presenter
|
Dennis Lin
Dennis Lin
Richard Mcgrath
Richard Mcgrath
|
Description
Fractional factorial designs are commonly used in industry to determine what factors have a substantial effect on the mean response. Quite often, procedures used to identify these factors having location effects assume homogeneity of variance, i.e. no dispersion effects. This course will show that not only are dispersion effects of interest themselves, but that they impact location effect inference causing complications. Participants will learn how unidentified dispersion effects impact location effect identification, and how unidentified location effects impact dispersion effect identification. It will be shown how to remove the confounding between dispersion and location effects through partial replication of the experiment or through a newly developed test. State-of-art techniques will be discussed and employed and participants will receive software to perform analyses. A basic knowledge of experimental design (but not necessarily fractional factorials) will be assumed.
|
|
