How should I treat a factor with range or value that is dependent on the level of another factor in a DOE?

[#1]

Hello,

I am designing a DOE with 4 factors. My problem is that the range for 2 factors are dependent on the level of 1 of the factors. For example:

Factor 1 - low (1), medium (2), high (3)

If Factor 1 is at low setting, then Factor 2 settings are low (1), medium (2), high (3).

If Factor 1 is at high setting, then Factor 2 settings are low (3), medium (4), high (6).

The range or values for low/medium/high settings for Factor 2 changes depending on the low, medium and high settings of Factor 1.

Should I use coded values (-1,0,1) when inputing the values for each level? Or should I treat Factor 1 and 2 as categorial factors and run the experiment at their corresponding settings? I am stuck on how I should approach this problem. Thanks!

[#2]

Hi @CuriousExplorer ,

This sounds like it could be a good case for the Custom Design platform and use the Disallowed Combinations to set up your experiment with the limitations you've described in your post.

In regards to whether to set your factors to numeric or continuous - you need to think about your end goal with the factors as to whether or not you're looking for more detail on the interactions and how you may control them. One thing to consider could be rolling Factor 1 and 2 together - as you've said, Factor 2 can only be run at certain conditions of Factor 1 - so do they need to be separate?

Thanks,

Ben

[#3]

I interpret this as a nested situation.  Factor 2 is nested in factor 1.  You will not be able to estimate interactions between factor 1 & 2.

The model is Y = F1 + F2[F1]