BookmarkSubscribeSubscribe to RSS Feed

Re: Why same DOE give me different modelling results?

JMPBoy

New Contributor

Joined:

Aug 9, 2018

I used response surface to design and finish my experiments. It gave me a very good model. But once I copy all these data together with original design into a new data table to run this analysis again. It gave a different results. I chose exactly same modelling method. Why???? Do I have to save this design as its original jmp to do analysis?

1 ACCEPTED SOLUTION

Accepted Solutions
gzmorgan0

Community Trekker

Joined:

Jul 25, 2016

Solution

Check that the data types and modeling types are the same in both the original design and combined table.

 

I agree with Cameron's assessment.

4 REPLIES
cwillden

Community Trekker

Joined:

May 1, 2017

There likely is some difference between between the tables you are not accounting for.  It would be really difficult to diagnose the problem without seeing both tables.  Can you post them?

-- Cameron Willden
Highlighted
JMPBoy

New Contributor

Joined:

Aug 9, 2018

Thank you cwillden. This was solved, but another problem between initial DOE jmp file and new copied file. When I evaluate these two models, it gave me different power analysis results. Do you think it's normal?
cwillden

Community Trekker

Joined:

May 1, 2017

That should absolutely not happen if everything is identical.  If you lost some column properties along the way, that could explain the difference.  Pay particular attention to the coding column property.

-- Cameron Willden
gzmorgan0

Community Trekker

Joined:

Jul 25, 2016

Solution

Check that the data types and modeling types are the same in both the original design and combined table.

 

I agree with Cameron's assessment.