|
Activity Number:
|
337
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 11, 2004 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Physical and Engineering Sciences
|
| Abstract - #301220 |
|
Title:
|
Average Correlations in Projections Designs
|
|
Author(s):
|
Cheryl Dingus*+ and Angela Dean
|
|
Companies:
|
Ohio State University and Ohio State University
|
|
Address:
|
404 Cockins Hall, Columbus, OH, 43210-1247,
|
|
Keywords:
|
orthogonal arrays ; projection properties ; generalized wordlength pattern ; isomorphism ; main effect plan
|
|
Abstract:
|
Orthogonal arrays are used as screening designs to identify active main effects, after which the properties of the subdesign for estimating these effects and possibly their interactions becomes important. The subdesign consisting of the columns of the orthogonal array corresponding to the active factors is known as a projection design. Two projection designs are said to be isomorphic or combinatorially equivalent if one can be obtained from the other by a sequence of row and column permutations and permutations of symbols in each column. Nonisomorphic projection designs are often evaluated by the generalized wordlength pattern. This can distinguish many, but not all, nonisomorphic projections. An alternative criterion is presented. This criterion is based on the average squared correlations of complete sets of orthonormal contrasts and it will be shown that the criterion is independent of the contrasts selected. Examples will demonstrate that the criterion is better able to rank order and distinguish projections from three-level orthogonal arrays than the generalized wordlength pattern. (Joint work with H. Evangeleros and C. Koukouvinos.)
|
