cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
FabioRoselet
Level I

8-run Plackett-Burman with JMP14

Hi everyone,

How can I perfrom an 8-run Plackett-Burman design with JMP14? Browsing the online documentation about Plackett-Burman suggests to include 4 additional runs and construct a 12-run Plackett-Burman design instead. However, a 12-run is just too much as I only have 4 factors... so a PB-8 would be enough for me. 

Unfortunately, after choosing Classical > Screening Design and setting the responses (1) and factors (4) the list of designs do not offer a PB-8 option, only fractional factorials with some 2-factors interactions.

Is it possible to construct a PB-8 with JMP14?

Thanks!

 

3 REPLIES 3
msharp
Super User (Alumni)

Re: 8-run Plackett-Burman with JMP14

It doesn't look like JMP likes Plackett-Burman for 1 response 4 factors.  The default example in their documentation is a 1 resposne 5 factors and this works as expected in both JMP 13/14.  Removing one of the factors removes PB designs from the options in both JMP versions.  Whether or not this is intended would be a question for JMP.

 

https://www.jmp.com/support/help/14/plackett-burman-design.shtml

Re: 8-run Plackett-Burman with JMP14

I believe that the Plackett-Burman fractional factorial design is not shown because the regular fractional factorial in 8 runs (1) is the same, (2) has equivalent resolution, or (3) has better resolution.

Re: 8-run Plackett-Burman with JMP14

Four factors does not seem like a screening experiment to me. It is a situation in which the screening principles are not likely to apply.

 

Eight runs does not have much power unless the effects are many times the standard deviation of the response.

 

You can't afford twelve runs and accept eight runs, so it seems that you know that main effects, for which regular and irregular fractional factorial designs are intended, are sufficient to model the response.

 

Why not use custom design?