I love JMP but the way it runs statistical tests is a little bit not crystal clear. For instance, I have data organized in three columns A (numeric), B (Ordinal), C (numeric). I want to test the effects of both A and B on C. So I do Analyze -> Fit Model, and I put A and B on the "model" and C on the "Y". JMP then reports the effect test which is probably an F-test. My question is: is there any way I can run a non-parametric test in this specific case?
Fit Model doesn't exactly have non-parametric tests. Once you've fit a model using least squares estimation, you've already done something parametric, and anything else you do from there is a post-hoc operation on a parametric model. So, a non-parametric method wouldn't really make sense.
One caveat: If you have JMP Pro, you could bootstrap the parameter estimates via a right-click. That is one non-parametric thing you could do.