mixed (2 & 3 level) fractional factorials

[jcwesthoff]

I have 3 factors at 2 levels and 1 factor at 3 levels. So, the full factorial woudl be 24 runs. However, D.C. Montomery in his book Design and Analysis of Experiments (6th Ed) states that a resolution V fractional factorial wiht 16 runs can be designed in this situation. The context of his statement is in a section where he is describing how to use a 2^3, 8 run full factorial design and conbine columns B & C into a new 3 level factor so that you end up with 1 factor at two levels and 1 factor at 3 levels. He gives a general description of the ANOVA, but no example with numbers. I'm trying to understand this better and extend the concept to a 2^(5-1) fractional factorial to apply to my situation. Is DOE++ capable of analyzing this type of design?

[David]

DOE++ can analyze the design once the design matrix and data is entered into DOE++ folio. The design matrix can be generated in Excel and then use “Add From File” to import it into DOE++.

I hope this helps.