|
Activity Number:
|
282
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #308525 |
|
Title:
|
Complex Experimental Design and Simple Data Analysis
|
|
Author(s):
|
Joseph G. Pigeon*+
|
|
Companies:
|
Villanova University
|
|
Address:
|
Department of Mathematical Sciences, Villanova, PA, 19085,
|
|
Keywords:
|
experimental design ; fractional facorial ; strip plot
|
|
Abstract:
|
Very often in industrial factorial experiments, not all factors have the same error variance associated with them. This generally arises because it may be more costly or simply not feasible to completely randomize the order of the experimental runs. The result of this restricted randomization is that multiple error measures are introduced into the experiment and the resulting analysis needs to account for these multiple error measures. Experimental designs that combine fractional factorials may have a multiple error structure that is even more complicated than the usual split plot experiment. These designs are sometimes called strip plot or multiway split unit designs. This poster uses a pharmaceutical example to emphasize the need for experimenters to recognize these types of experiments and suggests the use of simple data analysis methods for analyzing these types of experiments.
|
