Generalized linear multilevel modeling

[michaelyoung]

I've been desperately wanting to see generalized multilevel modeling for years.  The absence of this ability in JMP is what drives me to R on a regular basis.  Whether this is accomplished via the mixed models platform or simply by allowing random effects in a generalized linear model, I'd be thrilled to see this feature finally appear. 

[Phil_Kay]

I have to agree that this seems like an obvious gap. When I was doing multilevel modelling as part of my Stats MSc I had to go to R for the logistic regression part of the coursework and wasted many hours of my life that I will never get back.

[Madsen]

I do think if JMP wants to attract more users it has to incorporate the ability to enter random effects (i.e. to avoid pseudoreplication and also several other features associated with the icorporation of random effects) in both GLMM's and in logistic models.  I simply can not understand that an expensive software like JMP14, developed in 2018 (?), does not allow the use of these basic functions!  In order to perform such analyses I have to use SPSS.  I have already uninstalled the trail version of JMP 14 that you sent me and I am now going to purchase the latest version of SPSS, that allows the use of these essiential components in statistical analyses and consequently cancel my JMP account. 

Thomas Madsen

[jszarka]

Yes, I come across many situations where I'd like to incorporate the solution in JMP but end up in SAS with proc genmod or proc glimmix.

[GM]

Having the ability to run mixed models with categorical or discrete outcomes in JMP would make my work more seamless. 

[mia_stephens]

Fitting GLMM's within Fit Model is under development in JMP Pro 17, due out in October, 2022.