How can judge design by evaluate design platform

Question

Our customer have some question when they use the evaluate design platform.
There are serval indices I can get from evaluate design platform, and I wanna know if any index that must with great perforance and we can say the design is good.
Somthing like Power in power analysis must all over 0.7, or the maximun Prediction Variance must lower than 1, or the D efficiency of Design Diagnostics must over 50,etc.
customer want a standard process to decide the design can acceptable or not, and want to know the good design must reach which index with which criteria.
Thank you for any suggestion.

Phil_Kay · Accepted Answer

Hi Eric,

There are not really any rules for a minimum power or maximum prediction variance.

I tell customers that these measures should be used for comparing one design against another.

I have posted some things on LinkedIn about power and other ways to evaluate designs.

The best resource for understanding what each of these mean is the book, "Optimal Design of Experiments: A Case Study Approach" by Peter Goos and our own Bradley Jones.

I recommend that new users start with the Color Map on Correlations as it is very visual, useful, and relatively easy to understand.

I also tell customers that many people have been extremely successful with DOE without ever looking at any of the design diagnostics.

I hope that helps,

Phil