If a response has data, clicking on the corresponding node in the navigation tool starts the analysis.
The first step in the analysis is choosing the correct Transformation. Leave the transformation choice set to “None” until the analysis and/or diagnostics indicate another choice should be made.
For factorial designs, including custom designs with all categoric factors, terms are picked by looking at their effects relative to other possible effects and, if available, estimates of pure error.
For response surface, mixture and custom designs with numeric factors the Fit Summary provides suggestions for where to start the analysis and forwards the starting point on the Model tab, where either manual or computer added selection is used to choose the model to be tested.
The ANOVA is where the statistical tests, and descriptive statistics are displayed for the effects chosen for the model.
Diagnostics and Influence plots provide a way to check the assumptions about the residual distributions to validate the ANOVA results.
Model Graphs provide the best method for interpreting the models. By default, the displayed factors are limited to the ranges used when building the design to discourage extrapolating the model.