Brms conditional_effects show exact value

[ananas]

Hi,

I’m currently using conditional_effects(fit, “condition”) that shows this graph for a ordinal regression model:

I want to show the exact numbers of mean and sd in this graph. How should I do that?

Thank you!

[joels]

You can modify the plot generated by conditional_effects or create your own plot. I show both methods below with a reproducible example.

library(tidyverse)
theme_set(theme_bw())
library(brms)

m = brm(Petal.Width ~ Species, data=iris, cores=4, 
        file="tmp", backend="cmdstanr")

ce = conditional_effects(m)

conditional_effects actually returns a list containing a data frame for each predictor variable in the model . These data frames are used to generate the conditional effects plots. You can get the exact values from these data frames.

# Look at the structure of the conditinal_effects() object
str(ce)

List of 1
 $ Species:'data.frame':	3 obs. of  8 variables:
  ..$ Species    : Factor w/ 3 levels "setosa","versicolor",..: 1 2 3
  ..$ Petal.Width: num [1:3] 1.2 1.2 1.2
  ..$ cond__     : Factor w/ 1 level "1": 1 1 1
  ..$ effect1__  : Factor w/ 3 levels "setosa","versicolor",..: 1 2 3
  ..$ estimate__ : num [1:3] 0.247 1.326 2.026
  ..$ se__       : num [1:3] 0.0297 0.0277 0.0285
  ..$ lower__    : num [1:3] 0.189 1.268 1.967
  ..$ upper__    : num [1:3] 0.304 1.383 2.085
  ..- attr(*, "effects")= chr "Species"
  ..- attr(*, "response")= chr "Petal.Width"
  ..- attr(*, "surface")= logi FALSE
  ..- attr(*, "categorical")= logi FALSE
  ..- attr(*, "ordinal")= logi FALSE
  ..- attr(*, "points")='data.frame':	150 obs. of  4 variables:
  .. ..$ Species  : Factor w/ 3 levels "setosa","versicolor",..: 1 1 1 1 1 1 1 1 1 1 ...
  .. ..$ resp__   : num [1:150] 0.2 0.2 0.2 0.2 0.2 0.4 0.3 0.2 0.2 0.1 ...
  .. ..$ cond__   : Factor w/ 1 level "1": 1 1 1 1 1 1 1 1 1 1 ...
  .. ..$ effect1__: Factor w/ 3 levels "setosa","versicolor",..: 1 1 1 1 1 1 1 1 1 1 ...
 - attr(*, "class")= chr "brms_conditional_effects"

# Look at the data frame conditional_effects created for Species
ce$Species

     Species Petal.Width cond__  effect1__ estimate__       se__   lower__   upper__
1     setosa    1.199333      1     setosa  0.2473169 0.02974676 0.1886225 0.3040592
2 versicolor    1.199333      1 versicolor  1.3256387 0.02766471 1.2675946 1.3831360
3  virginica    1.199333      1  virginica  2.0259349 0.02848083 1.9668068 2.0846869

# Modify conditional effects plot
#  Use cat_args to remove point markers
plot(ce, ask=FALSE, cat_args=list(colour=NA))[[1]] + 
  geom_label(aes(label=round(estimate__, 2)), size=3) +
  geom_text(aes(label=sprintf("SE: %0.3f", se__)), nudge_x=0.35, 
            hjust=1, colour="red", size=3)

# Create your own plot
ce$Species %>% 
  ggplot(aes(Species, estimate__)) +
  geom_errorbar(aes(ymin=lower__, ymax=upper__), width=0.3) +
  geom_label(aes(label=round(estimate__, 2)), size=3) +
  geom_text(aes(label=sprintf("SE: %0.3f", se__)), nudge_x=0.35, 
            hjust=1, colour="red", size=3) +
  labs(y="Petal Width")