GAMLj: GLM logistic regression plots

General help and assistance with jamovi. Bug reports can be made at our issues page: https://github.com/jamovi/jamovi/issues . (If you're unsure feel free to discuss it here)
Post Reply
TomerK
Posts: 2
Joined: Sun Oct 18, 2020 6:51 pm

GAMLj: GLM logistic regression plots

Post by TomerK »

Hi all,

Let me first state that I'm new to Jamovi, so I hope my questions aren't too blunt...
I'm running a logistic regression analysis through the GAMLj package, and the 'plots' tab options are unclear to me.
The only information I could find online was the following:

https://gamlj.github.io/gzlm.html

My questions regarding these plots:
1. What are these plots representing? Are they main effect plots? What is represented by the curve?

2. How come there are negative values to the x axis (and the consequently the curve) when all the values of my independent variables are greater then 0?

I've tried looking for the source syntax in R to try and figure it out but couldn't find access to that.
3. Is there a way to access the actual syntax for the GAMLj package (specifically these plots)? The syntax mode in Jamovi doesn't seem to show the entire code.

All my independent variables are continuous and I'm looking for an informative way to plot the univariates of the model.
I have no factors and no interactions.

Thank you very much in advance,

Tomer
User avatar
mcfanda@gmail.com
Posts: 457
Joined: Thu Mar 23, 2017 9:24 pm

Re: GAMLj: GLM logistic regression plots

Post by mcfanda@gmail.com »

Hi,
here are some info regarding your questions.

1. gamlj generalized linear models plots (by default) produce the plots of the predicted values of the model, transformed back into the response metric. Thus, for the logistic and probit models plots represent predicted probabilities, for poisson model they represent frequencies.

What effects are represented depends on how the user sets up the plots. One can ask to plot main effects or interactions, up to 3-way interaction. Plots can be for continuous or categorical independent variables.

2. GAMLj, by defaults, centers the continuous independent variables to their means, so you get negative values for values smaller than the variable means. You can change that in the "Covariates Scaling" pannel, setting your variables to "none"

3. Yes, https://github.com/gamlj/gamlj
cheers
mc
TomerK
Posts: 2
Joined: Sun Oct 18, 2020 6:51 pm

Re: GAMLj: GLM logistic regression plots

Post by TomerK »

Hi mc,

Got it.

Thank you very much for your reply.

Best wishes,

Tomer
Post Reply