Hi all,
I have a customer who ran a 28 run RSM with 5 continuous factors. Since he took the work from someone who left, we assume it includes a center point or additional run (since JMP suggests 27 runs).
Now they want to extend the analysis by adding a factor, which they hadn't considered before. Here is a great description of how to accomplish this, but several questions came up.
1. When the box "Group new runs into separate block" is checked, the "new factor" is added as a model effect (main effect; alongside the blocking factor). If it is not checked, the factor is not added by default.
2. If the initial design is RSM, should the new factor also be added as RSM (via RSM button)?
Attached is the table with the initial design and the new factor added.
Sidenote: They want to spend max. 8 extra runs for this augmentation.
Appreciate any help!
Best regards from Heidelberg,
Florian
Thanks @Phil_Kay , it does help. I'll talk to Support to see whether this (1.) is a bug.
Best
Florian
Hi Florian,
1. Seems like a bug to me. I can't think why you would not want to add the new factor to the model. It should not matter whether you have a blocking factor.
2. Given that the original design was for RSM, it would make sense to make the model RSM for the augmentation too. However, that means there are 7 new parameters to estimate: ME and X^2 for new factor, and 2-way interactions between new factor and the other 5 factors. That might be a challenge with just 8 more runs. As always, it will depend on the science/engineering objective for the experiment.
I hope this helps,
Phil
