Definitive Screening Design with asymmetrical Setpoint

[#1]

Helly everyone,

I want to design an experiment as a definitive screening design. However, for one factor I don't want to include 0 but -0.5 as a setpoint level. This represents the current standard. We cannot screen symmetrical around the current setpoint value because of technical obstructions. Is there a way to still use a definitive screening design? And if so, how would I do this in JMP?

I'm looking forward to answers!

Best,

Anja

[#2]

I don't understand your question, but here are my thoughts:

1. the levels in a definitive screening design are coded -1, 0, 1.  These do not represent actual levels, so you can use whatever actual levels you want.  I don't know what you mean by "setpoint level"?

2. Regarding your factor that currently runs at -0.5 and can't be tested <-0.5, you can use the -0.5 as the -1 setting and choose 2 other levels equally spaced for the 0 and 1 coding.  The three levels in a definitive screening design are so the quadratic effect can be estimated.

[#3]

I'm sorry for the confusion. I was already speaking in terms of a coded factor. Let me refrase the question:
Is it possible to design a DSD with -1, -0.5, 1 instead of -1, 0, 1?

[#4 · Accepted solution]

You can make a DSD and then recode the 0 levels to -0.5 after you make a data table. This result is no longer a DSD. It will not have the properties of a DSD or support the Fit Definitive Screeing platform and analysis.

You might instead use a Custom Design and choose Discrete Numeric factors instead of Continuous factors. You can set the 3 levels as you want. In this case, you must leave all the additional power terms for such factors in the model or the algorithm will remove levels. You can add terms, thoigh.