Why the confidence interval from conditional prediction profiler has an irregular shape?

[#1]

Hi all,

I would like to understand why the conditional prediction profiler from a mixed model has an irregular shape.

Here I attached a dataset with 1 responses, and you will see the irregular shape of the CIs in conditional prediction profiler from a mixed model. 

[#2]

There is various information on profilers built into JMP (https://www.jmp.com/support/help/en/16.0/#page/jmp/profiler.shtml#)

Your model is not very good.  Have you analyzed the residuals?  There is evidence of non-constant variance and possibly multiple distribution of errors.  If you try removing some the effects from your model, you can watch the changes to the Rsquare-Rsquare adjusted delta and residual plots and subsequently the confidence curves.

[#3]

Hi @YangHong ,

  I would also add that it is not that common to have a two-factor interaction be more important than a main effect. It certainly can happen, but if the DOE is designed to look at the right factors, the main effects will almost always show more importance than two-factor interactions, which is related the the concepts of effect hierarchy and effect heredity. You can read more about it here.

DS