topic Re: Fit Model with >1 factor results in NonEstimable Std Error? in Discussions

Fit Model with >1 factor results in NonEstimable Std Error?

bayesfactor1 — Fri, 09 Jun 2023 00:31:31 GMT

When I run a Fit Model with one factor, I get an output that includes Std Error estimates from which I could calculate confidence limits. When I run a Fit Model with more than 1 factor, I get an output with Std Error blank and a column that says "NonEstimable". Why is that? I'd like to know confidence limits on the parameter estimates!

Re: Fit Model with >1 factor results in NonEstimable Std Error?

Thierry_S — Mon, 05 Apr 2021 05:11:41 GMT

Hi,
It is a bit challenging to give you a precise answer based on the limited information you provided but I wonder if your Samples variable is only represented in specific Batches, as shown below:
Batch 1 - Sample A
Batch 1 - Sample B
Batch 1 - Sample C
Batch 2 - Sample D
Batch 2 - Sample E
...
If it is the case, there is not information for JMP to test the effect across your variables. You may want to explore the functionality of Nesting Sample within Batch: Sample[Batch]?
Best,
TS

Re: Fit Model with >1 factor results in NonEstimable Std Error?

Mark_Bailey — Mon, 05 Apr 2021 16:22:41 GMT

The Singularity Details notice at the beginning of the report suggests that you have confounded effects with your current specification of the model.