As an entry-level user of JMP but with a bit of statistical knowledge. I encountered with a problem when trying to use the stepwise fit platform to develope a regression model for my DoE dataset.
I read the JMP help book and understand how combined rule applied to p-value based stepwise regression (a joint F-test and another partial F-test of interaction effect when added precedent main effects). However, I am a bit confused when transfer such understanding to AICc or BIC criteria. I noticed that my AICc-combined model and AICc-no rule model are different.
Therefore my question is: when I select AICc and combine rules together. How does the stepwise regression work? It still do a joint addition (say an interaction effect and 2 associated main effects) and calculate AICc? Besides, also calculate another AICc of adding only interaction effect. Then how the comparison of AICc is carried out?
Hope my description is clear and feel free to let me know if further clarification is needed.
Best
Yun
