Easy DOE uses JMP modeling platforms to analyze the designs. The model used depends on the type of design and the mode.
For the Guided Mode, the best model is fit according to the following rules.
• Fit Stepwise is used with the BIC criterion to select terms, and the parameter estimates and confidence intervals are calculated using the Standard Least Squares personality of the Fit Model platform. This is analogous to using Fit Stepwise to select terms and then clicking the Run Model button.
• If terms are removed from the model, Fit Least Squares is used to recalculate the parameter estimates.
For the Flexible Mode, the models are fit according to the following rules:
• In JMP with no Hard or Very Hard to change factors, Generalized Regression with forward selection and AICc Validation is used.
• In JMP Pro with no Hard or Very Hard to change factors, Generalized Regression with the Best Subset Estimation Method with AICc Validation is used.
• In JMP Pro, Fit Mixed is used to fit models for designs with one or more Hard or Very Hard to change factors.
• In JMP, Fit Least Squares with the REML method is used to fit models for designs with one or more Hard or Very Hard to change factors.