Missing P Values in Effect Summary OF Fit Model

[#1]

Hi, I am using JMP-14.

I was using Fit Model and somehow my Effect Summary Dialogue is missing Data, no P-Values for the data.

Is there a reason why this would happen?

Thanks in advance 

[#2]

Hi, it is difficult to figure out what may have gone wrong without more details but the first thing that comes to mind is the large number of effect in your model. I wonder if your data coverage may not be sufficient to explore all these factors.

Would it be possible to share an example of the data you are working on? Also, are there key factors that would be worth exploring first? Finally, have you considered the Screening platform to priotitize the factors in your model.

Best,

TS

[#3]

If your model is saturated (all DF's are accounted for in the model), there are no DF's to estimate the error term and therefore there are no tests of significance.

[#4]

Hi I have a similar problem and I don't think my model is saturated because there are 28 experiment and 6 variables. The two missing p values correspond to the two categorical variables. Why?

[#5]

What does the Analysis of Variance, Parameter Estimates, and Effect Tests tables show?

[#6]

[#7]

So this indicates that you do not have adequate data to estimate your model. Fitting a model requires that you have enough data, but also you need to have the data collected in the proper locations. Think of it this way: I want to estimate a straight line relationship between Y and X. I collect 100 data points, but X is always 10. Lots of data, but I cannot estimate the line.

In your situation, I believe you also have a Singularity Details report. If you open that up you will see that you have a relationship (or confounding) between two of your input variables. It looks like it is between Screen Plate and SDS. JMP cannot estimate both of those effects because there is not sufficient data to do so. Make a plot of these two input variables. What does it look like? I think that might show you the issue.

[#8]

Thanks, Dan. This is very helpful.

[#9]

Hi, could you explain what specifically about the Analysis of Variance, Parameter Estimates, and Effect Tests tables stuck out to you regarding the lack of data for the model? I would appreciate any input or papers on this topic. Thank you.

[#10]

The items that told me the data could not support the model:

Singularity Details report (although it is not shown, I know it is there)

From the Parameter Estimates table:  The words Biased and Zeroed. We don't want to be biased! That should tell you to not trust the results -- they are biased! And Zeroed means that JMP "deleted" those terms from the model in order to try and give you some results. However, JMP really does not know the system so there is no way for JMP to know which terms should have been removed. A mathematical choice is being made based solely on the order of the terms you specified in the model.

From the Effect Tests table: LostDFs which leads to no statistical testing for those effects.