r2.glmmTMB "Warning: get_variance ignores effects of dispersoin model."

[gimickwu]

Hello,

I'm using the package performance (love it), and came across an issue. It seems the r2() function can't deal with glmmTMB models that include a dispersion model (dispformula). Would there be a "quick" way to make this possible?

repro:

library(glmmTMB)
library(performance)
data(mtcars)
mod <- glmmTMB(mpg ~ cyl, dispformula = ~ cyl, data = mtcars)
performance::r2(mod) 

This gives the message "Warning: get_variance ignores effects of dispersion model. Random effect variances not available. Returned R2 does not account for random effects." (The random effect part is ok; I didn't put one, but the part is the issue and it happens also whene there is a random effect).

Thank you in advance for your input! Mick

[bwiernik]

The problem is that that there isn't one R2 for a model with a dispersion model. There's one for every possible value of sigma.

We should perhaps consider treating these models more like generalized models like binomial or Poisson models