Step 5 -- Choose an Optimal Experiment Design.

This is a budget driven choice. A good design sets the upper and lower limits for each factor as far apart a possible to allow for more possibilities while holding everything else fixed. A scrambled run order is also important to ensure that extraneous factors do not contaminate the response values.

A QUADRATIC DESIGN:

Uniform Shell

Interaction plus Star

Optimal(Hardin-Sloane)

AN INTERACTION DESIGN:

All the corners of a cube

A fraction of the corners

Optimal(Hardin-Sloane)

A MAIN EFFECTS DESIGN:

Plackett-Burman Design or its reflection.

Plackett-Burman Design and its reflection

Optimal(Hardin-Sloane)

 STRATEGY for Windows can generate the types of designs above. For custom Hardin-Sloane designs, call or fax The Experiment Strategies Foundation.

Previous Step

Next Step

The Nine Step Strategic Method