Model Limits

Stat-Ease will automatically fit up to sixth-order polynomials to model experimental results. The program also allows the user to define unique polynomial terms. Ordinary polynomials will be used for response surface and factorial designs. Mixture designs use Scheffe models. Scheffe models have the intercept built into the coefficients.

Response Surface Designs: Stat-Ease will fit higher order models that have fewer than 400 terms. This implies the following constraints:

4 factors – up to sixth order model

5 factors – up to fifth order model

6-7 factors – up to quartic model

8-10 factors – up to cubic model

Mixture Designs: Stat-Ease will fit up to a quartic Scheffé model that has fewer than 400 terms as follows:

3-8 components – up to quartic model

9 components – up to special quartic model

10-12 components – up to cubic model

13 components – up to special cubic model

14-24 components – up to quadratic model

Mixture/Process Crossed Designs: Stat-Ease will not build crossed designs larger than cubic by cubic and any design is limited to 400 coefficients.