Model Graphs

Stat-Ease software provides various graphs to help interpret the model selected. These are generated by clicking on the Model Graph button after fitting a model.

For Factorial designs, the primary graphs of interest are the one factor, interaction, and cube. The one factor graph should is used to look at main effects for factors that are not part of an interaction. The interaction graph can be used to display any two-factor interactions. The cube plot is useful to viewing the relationship among three factors.

For Response Surface, and Mixture designs the primary graphs will be the contour and 3D surface.

Combined Designs have variations on the 3D and Contour plots. They can be viewed as just the mixture (up to 3 components at a time) just the process (two-factors at a time), or a Mix x Process graph showing two mixture components and a single factor. When the design has two independent mixtures there is also the Mix – Mix graph showing two components from each mixture.

The best practice when looking at model graphs is to focus on the terms with significant effects. Use the graphs that best help explain the behavior of the process.