Thank you very much for your help sir. When I applied what you said, the analysis part was successful, thank you very much again. But I have 2 more questions. The f-value and p-value values appear as '*' when doing anova. You suggested that we extract the smallest effect parameter on different subjects. Is it okay for me to do it this way too? My other question is, sir, why do we need to follow a path like this? I really appreciate your help. If you have a study on Taguchi, I would like to benefit and cite it.
Taguchi *error*
- Thread starter sefaaa
- Start date
Yes, this is the correct way to deal with this issue. The reason this occurs is that when you use every available column on a Taguchi orthogonal array, you end up with a saturated model with no remaining degrees of freedom to calculate the F-value (the p-value is calculated from the F-value).
Saturated models are usually avoided by performing replicates of the design. This is recommended when using classical designs. However, Taguchi encouraged simplicity as much as possible and advocated using repeats instead of replicates. He provided two options: 1) leave an unused column to avoid a saturated model, or 2) remove the smallest effect term from the model to free up additional degrees of freedom.
Note: I learned these methods from Shin Taguchi, Genichi Taguchi's son and have Genichi Taguchi's books.
Saturated models are usually avoided by performing replicates of the design. This is recommended when using classical designs. However, Taguchi encouraged simplicity as much as possible and advocated using repeats instead of replicates. He provided two options: 1) leave an unused column to avoid a saturated model, or 2) remove the smallest effect term from the model to free up additional degrees of freedom.
Note: I learned these methods from Shin Taguchi, Genichi Taguchi's son and have Genichi Taguchi's books.
thank you very much sir. I am grateful to you.