01-19-2015 04:40 PM
Can someone please provide direction on how to compare two different slopes from two different regression curves for equivalency? I have data sets with one response and the effect being time. I have another effect, manufacturer that I want to include to see if there's any difference in the two different regression curves by manufacuturer. I believe I want to perform an Analysis of Covarience, but can't locate where it is (Using Jmp Ver 8.0.2). Hope someone can help me.
01-19-2015 05:00 PM
Just use the Fit Model platform.
Include the 2 effects, and also the interaction. If the interaction (time*manufacturer) is significant, then there is difference in the slope.
01-20-2015 09:28 AM
Thanks for the reply. So as to be certain, I would add the response (call it activity) as my Y variable and the products age (in months) and the manufacture (a nominal charatcter) as construct model effects? Or do I have to make the manufactures a numeric/nominal value?
01-20-2015 10:29 AM
If manufacturer is numeric, be careful that its modelling type is nominal.
If manufacturer is character it will be automatically nominal.
01-20-2015 10:58 AM
thanks so much for your help.
Here's what I then get (Hope it's readable). I don't see the interaction you previously mentioned - (time*manufacturer) given.
Am I doing something incorrectly?
01-20-2015 11:08 AM
I don't have JMP right now (and I am using JMP 11), but somewhere in the dialog where you inclue model effects, you can also include the interaction.
In JMP 11 you need to highlight the effects (one effect on the right, one on the left pane), and then click "cross".