[Crash]: Bayesian Generalized Linear Mixed Models

[fplatz]

System Info

-------- Application Info --------
JASP Version: JASP 0.95 (Apple Silicon)
Build Branch: development
Build Date: Jul 24 2025 12:32:20 (Netherlands)
Last Commit: c4045c3cd65cb374145aac47e69d59c7f90b67bb

-------- Basic Info --------
Operating System: macOS Sequoia (15.6)
Product Version: 15.6
Kernel Type: darwin
Kernel Version: 24.6.0
Architecture: arm64
Install Path: /Applications/JASP.app/Contents/MacOS
Platfotm Name: cocoa
System Local: en_DE

When doing a Bayesian Generalized Linear Mixed Model with a dichotomous response variable, several 6 independent variables (one interval, 5 ordinal) as fixed effects and two random factors using a binomial family and a Logit link function I got the following error message:

Please, report the following error message at JASP GitHub https://github.com/jasp-stats/jasp-issues: Error in rstan::summary(stanfit, probs = probs, digits = 10)$summary :
$ operator is invalid for atomic vectors

JASP Module

Regression, Mixed Models

Is there a specific analysis that caused the crash?

Bayesian Generalized Linear Mixed Models

Log files

JASP 2025-08-17 19_17_38 Desktop.log
JASP 2025-08-17 19_17_38 Engine 0.log

Steps to Reproduce

Bayesian Generalized Linear Mixed Models

If you would like, I can send you the JASP-File privately.

[FBartos]

Hi @fplatz ,

could you please send me the JASP file to @FBartos ?
I will see what I can do!

Cheers,
Frantisek

[fplatz]

Thanks for your offer, @FBartos! I wanted to send you the requested file, but unfortunately do not know how to proceed here in order to send the file privately.

Cheers
Friedrich

[FBartos]

sorty, I wanted to add my email address f.bartos96@gmail.com :)

[tomtomme]

@fplatz
any news for us on this one?

[fplatz]

Hi! Thanks for the reminder. Meanwhile we changed the model. It seems that the deactivation of the „correlation“ option solved the problem. Maybe there were too much parameters to be estimated. Also, what I could observe is, that problems could be reduced when no labels were used but instead only the numbers. When we removed the labels, jasp was running more stable. However, working with labels („dislike“ instead of „2“) is something that should be possible, because the numbers behind the labels were attached. Tables with labels are quite nicer for presentation and interpretation than working solely with values (of categorical variables, for instance), which the authors have to recode into meaningful labels. But all in all, it is fascinating to see, how jasp is growing and getting better from day to day! Best wishes, Friedrich

…

Am 26.08.2025 um 17:58 schrieb Thomas Langkamp ***@***.***>: tomtomme left a comment (jasp-stats/jasp-issues#3610) <#3610 (comment)> @fplatz <https://github.com/fplatz> any news for us on this one? — Reply to this email directly, view it on GitHub <#3610 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADT3DYLT33GZOANIHJVKO7T3PR7Y5AVCNFSM6AAAAACEDEWO2GVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTEMRUG44TMMZYHE>. You are receiving this because you were mentioned.