Bug in Gamlj multinomial regression

Discuss the jamovi platform, possible improvements, etc.
Post Reply
User avatar
Arnaud Mortier
Posts: 32
Joined: Tue Apr 27, 2021 11:39 am

Bug in Gamlj multinomial regression

Post by Arnaud Mortier »

Hi,

I'm running Gamlj 2.6.3 with jamovi 2.3.13.0.

My DV is categorical and I have 1 factor and 1 covariate. In the simple effects section, using the covariate as moderator, the output table is not consistent: the row heads repeat as you can see in the picture below (we have the combination [ FB - F ; Mean-1-SD ; 8-6 ] twice and we never get to see [ FB - F ; Mean ; 8-6 ] for example.

Thanks!
Arnaud
Screenshot from 2022-07-01 11-40-04.png
Screenshot from 2022-07-01 11-40-04.png (163.32 KiB) Viewed 1590 times
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: Bug in Gamlj multinomial regression

Post by mcfanda@gmail.com »

.
User avatar
Arnaud Mortier
Posts: 32
Joined: Tue Apr 27, 2021 11:39 am

Re: Bug in Gamlj multinomial regression

Post by Arnaud Mortier »

Hi,
Sorry I don't understand, is this a post-it note for yourself or again some kind of bug?
Thanks!
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: Bug in Gamlj multinomial regression

Post by mcfanda@gmail.com »

Sorry, the message went blank. Yes, the multinomial simple effects table gets the labels wrong with three levels of the IV. A new version 2.6.6 is coming out (tomorrow or so) that fixes the issue. The results in the current version were correct, but of course, it is not clear what the coefficients correspond to.
Thanks for spotting it
Please check if the bug has been solved using your example
User avatar
Arnaud Mortier
Posts: 32
Joined: Tue Apr 27, 2021 11:39 am

Re: Bug in Gamlj multinomial regression

Post by Arnaud Mortier »

Hi,
Thanks for your answer.
For firewall-related reasons I can only sideload packages, and it seems that the linux version of Gamlj available from the library online is still 2.6.3.
All the best,
Arnaud
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: Bug in Gamlj multinomial regression

Post by mcfanda@gmail.com »

Try this https://github.com/mcfanda/binaries/tree/master/gamlj
User avatar
Arnaud Mortier
Posts: 32
Joined: Tue Apr 27, 2021 11:39 am

Re: Bug in Gamlj multinomial regression

Post by Arnaud Mortier »

Works perfectly! Thank you so much!
Post Reply