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
ataylor
Level III

Evaluate DoE not generated by JMP

I have a DoE that was given to me as a CSV.  I'd like to see the usual diagnostics (correlation plot and power) that one gets when the design is generated by JMP for this DoE. Is that possible?  I put the parameters into JMP and generated a DoE using JMP, but it is not quite the same and I'd like to compare the diagnostics for the two designs.  

 

Thanks!

Amelia

1 ACCEPTED SOLUTION

Accepted Solutions
KarenC
Super User (Alumni)

Re: Evaluate DoE not generated by JMP

 

Hello,

Not surprsing that you woudl get a slightly different design as how the designs are genearted may use a random starting point (if they are not a standard text-book type design). However, you can still evaluate the design regardless of where it was generated.  See:

 

https://www.jmp.com/support/help/14/overview-of-the-evaluate-design-platform.shtml


Quick Instructions:

In the DOE Menu:  DOE>Design Diagnositcs>Evaluate Design

You will have to do a little more work than with a design created in JMP in that you will have to tell JMP what columns are what (factors and responses) and then specify the model.  Once you do that you should be back in familiar teritory.   

View solution in original post

6 REPLIES 6
txnelson
Super User

Re: Evaluate DoE not generated by JMP

I am assuming that what is causing an issue for you is that the Fit Model or Model script that is embedded into the results from the "Make Table" button in the DOE, is not generating all of the results you expected.  Different DOE designs will generate different default analysis specifics.  However, additional displays and analyses are available from the "red triangles" available.  I suggest that you look at making sure the DOE model you are defining from the CSV is exactly what is being called for, and once that is determined, take a look at the options available to you under the "red triangles" to find the options and displays you are looking for.  

Look for documentation on DOE and Fitting Linear Models in the documents

     Help==>Books==>Design of Experiments Guide

and

     Help==>Books==>Fitting Linear Models

Jim
ataylor
Level III

Re: Evaluate DoE not generated by JMP

Hi.  I am asking a different question.  I have a csv with a designed experiment that was sent to me by someone else.  Not the results of the experiment, just the design.  I'd like to get the same sort of design analysis dialog I get when I design within JMP for this design that was sent to me.  When I make my own designs in JMP, I know how to do all this, including using the "triangles" from the matrix generated.  This is different because the design sent to me was not done using JMP, can I still use JMP to get design diagnostics?

txnelson
Super User

Re: Evaluate DoE not generated by JMP

If the Raw Data is contained in the CSV file, and you can precisely create the DOE in JMP from the description of the DOE design in the CSV, you should be able to replicate the analysis in JMP.

Not knowing exactly what the CSV file contains leaves one a bit in the dark on possible issues as to why you are not seeing what you want. 

Jim
ataylor
Level III

Re: Evaluate DoE not generated by JMP

Jim,  What is interesting is that when I take the values from the DoE sent to me (as suggested) and use JMP to get a DoE, what I get is a bit different from what my colleague sent me.  More detail:  looking at the csv sent to me, it looks like 2 continuous variables, 1 categorical 2-level and 1 categorical 3-level.  Taking the recommended 24 conditions from JMP (which matches the number of conditions I was sent), the fundamental difference is that JMP suggests 9 conditions that have the first value for the 3-level categorical variable, 7 for the second value and 8 for the third value and the design sent to me has exactly 8 conditions for each of the 3-level categorical levels.  I was hoping to compare the diagnostics of these two DoE's.  I suspect they are very similar, but want to be sure.

 

It sounds like, from our discussion that JMP can only give diagnostics for designs created using JMP directly.  

KarenC
Super User (Alumni)

Re: Evaluate DoE not generated by JMP

 

Hello,

Not surprsing that you woudl get a slightly different design as how the designs are genearted may use a random starting point (if they are not a standard text-book type design). However, you can still evaluate the design regardless of where it was generated.  See:

 

https://www.jmp.com/support/help/14/overview-of-the-evaluate-design-platform.shtml


Quick Instructions:

In the DOE Menu:  DOE>Design Diagnositcs>Evaluate Design

You will have to do a little more work than with a design created in JMP in that you will have to tell JMP what columns are what (factors and responses) and then specify the model.  Once you do that you should be back in familiar teritory.   

ataylor
Level III

Re: Evaluate DoE not generated by JMP

Perfect.  Thanks.  That was what I was missing is the design diagnostics selection under DOE.  Thank you!