I'm not sure what you mean by "Mediation test". But in the Fit Y by X platform, from the report hot spot, the Nonparametric dropdown offers a median test.
Here's the link to the JMP online documentation for more details:
I would also be curious to know if anyone has attempted to address this in JMP.
my guess is xiaoyaj0 is looking for a mediation test under this definition: Mediation (statistics) - Wikipedia, the free encyclopedia
the most advanced technique at the moment is perhaps the one offered by Andrew Hayes in this link: The PROCESS macro for SPSS and SAS - PROCESS macro for SPSS and SAS
Interesting approach. While there isn't a what I'll call direct path to this general method within JMP...I could envision someone replicating the process to some degree using JMP Pro. I could envision using many different modeling approaches in JMP Pro such as the Bootstrap Forest or Generalized Regression for fitting multiple models with and without the mediator and direct terms. Then using the Model Comparison platform for evaluating model performance.
Having just recently taught statistics in the Psychology Department, I became aware of their frequent use of the terms moderation and mediation. Moderation appears to be just another term for two-factor interactions. However, mediation is quite different. It is defined as a situation when the relationship between a predictor variable and outcome variable can be explained by their relationship to a third variable (the mediator)..I was able to "salvage" a little bit of my ego by using SPSS, as was suggested previously.
I wonder if JMP might add these two terms into the JMP vocabulary and possibly increase its usage by those in the psychology and maybe even medical fields?
Process Macro seems to be the solution. Can I use it with JMP?
A more general question: Can I use SAS code with JMP? How does the conversion look like? I never used SAS, yet I kind of know that SAS is capable of things JMP cannot do.
It's definitely not a mediation test, but you might try Neural Networks. This paper out of Vanderbilt (link) discusses some of the similarities and differences between the techniques. I'm not going to say its a perfect fit, but it may help answer the question. Particularly with the control over the Neural Net provided in JMP Pro.