How to exponentiate model coefficients with brms models

[pgseye]

I have started dipping my toes into Bayesian models and as I use tbl_regression for presenting all of my (frequentist) model results, naturally tried to do the same with the brm model. It works to present the coefficient but is there a way to exponentiate them if required (e.g. a logistic model)

If I try this I get:
! exponentiate = TRUE is not valid for this type of model and was ignored.

[larmarange]

Are you referring to brm package or brms?

Could you provide a reproductive example?

[pgseye]

Thanks - I think I have managed to put together a small example and supporting code:

small_dat <- structure(list(id = c(1L, 2L, 3L, 4L, 5L, 6L, 7L, 8L, 9L, 10L, 
                                   11L, 12L, 12L, 13L, 14L, 15L, 16L, 17L, 18L, 18L, 19L, 20L, 21L, 
                                   22L, 23L, 24L, 25L, 26L, 27L, 27L, 27L, 27L, 28L, 29L, 30L, 31L, 
                                   32L, 33L, 33L, 34L, 35L, 36L, 36L, 37L, 38L, 39L, 39L, 40L, 41L, 
                                   42L), x = structure(c(2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
                                                         2L, 1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 
                                                         2L, 2L, 2L, 2L, 2L, 2L, 1L, 2L, 2L, 2L, 2L, 1L, 2L, 2L, 2L, 2L, 
                                                         2L, 2L, 2L, 2L, 1L, 2L, 2L, 2L), levels = c("low", "high"), class = "factor"), 
                            y = c(1, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1, 1, 1, 0, 
                                  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 
                                  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0)), row.names = c(NA, 
                                                                                            -50L), class = c("tbl_df", "tbl", "data.frame"))


library(lme4)
library(brms)
library(gtsummary)
mod <- glmer(y ~ x + (1|id), data = small_dat, family = "binomial")
summary(mod)
tbl_regression(mod, exp = T) # this exponentiates

mod2 <- brm(y ~ x + (1|id), data = small_dat, family = "bernoulli")
summary(mod2)
tbl_regression(mod2, exp = T) # this doesn't exponentiate

[larmarange]

First, the correct syntax is tbl_regression(mod2, exponentiate = TRUE).

Second, broom.mixed fails to tidy mod2. There is therefore a bug to be reported in this package: https://github.com/bbolker/broom.mixed/issues

> rlang::last_trace()
<error/dplyr:::mutate_error>
Error in `mutate()`:
ℹ In argument: `exp`.
Caused by error:
! `exp` must be a vector, not a primitive function.
---
Backtrace:
     ▆
  1. ├─broom.mixed::tidy(mod2, exponentiate = TRUE)
  2. ├─broom.mixed:::tidy.brmsfit(mod2, exponentiate = TRUE)
  3. │ └─out %>% mutate(across(contains(vv)), exp) %>% ...
  4. ├─dplyr::mutate(., across(std.error, ~. * estimate))
  5. ├─dplyr::mutate(., across(contains(vv)), exp)
  6. ├─dplyr:::mutate.data.frame(., across(contains(vv)), exp)
  7. │ └─dplyr:::mutate_cols(.data, dplyr_quosures(...), by)
  8. │   ├─base::withCallingHandlers(...)
  9. │   └─dplyr:::mutate_col(dots[[i]], data, mask, new_columns)
 10. │     └─mask$eval_all_mutate(quo)
 11. │       └─dplyr (local) eval()
 12. └─dplyr:::dplyr_internal_error("dplyr:::mutate_not_vector", `<named list>`)
Run rlang::last_trace(drop = FALSE) to see 5 hidden frames.

[ddsjoberg]

Thank you @pgseye and @larmarange !

@pgseye please re-open this issue of there is anything more to attend to after the fix in broom.mixed is live.