topic Re: Analysis of split plot DOE results in Discussions

Analysis of split plot DOE results

frankderuyck — Thu, 08 Jun 2023 21:12:10 GMT

In analysis of a (split) split plot REML as well as Generalized regression is used, which anaysis method is preferred?

Re: Analysis of split plot DOE results

Phil_Kay — Wed, 05 Oct 2022 10:11:35 GMT

REML is recommended as you should include a random effect for whole plot to capture the variance between whole plots. I recommend the book, Optimal Design Of Experiments: A Case Study Approach, for a more detailed explanation of this and some good examples of split plot and split-split plot experiments.

Re: Analysis of split plot DOE results

statman — Wed, 05 Oct 2022 15:42:00 GMT

There are a number of good books on split-plots. I will add my suggestion for the following papers:

Box, G.E.P., Stephen Jones (1992), “Split-plot designs for robust product experimentation”, Journal of Applied Statistics, Vol. 19, No. 1

Jones, Bradley, Christopher J. Nachtsheim (2009) “Split-Plot Designs: What, Why, and How”, Journal of Quality Technology, Vol. 41, No. 4, pp. 340-361

Re: Analysis of split plot DOE results

frankderuyck — Wed, 05 Oct 2022 19:12:09 GMT

I have this great & very helpful) book; I read on p. 236 related to the spit plot wind tunnel experiment that Brad & Peter are usign generalized least squares for the analysis. I have tried this but got an error message claiming that generalized regression analysis does not accept (whole plot) random effects?

Re: Analysis of split plot DOE results

Phil_Kay — Mon, 10 Oct 2022 17:48:28 GMT

Hi @frankderuyck,

"Generalized least squares analysis" and the Generalized Regression platform in JMP Pro are not the same thing. The generalized least squares analysis in the Goos & Jones split plot case study would be carried out in JMP in Fit Model using Standard Least Squares personality and REML method. You can find more detail about generalized least squares analysis on page 243 of the book.

I hope that helps,

Phil