Is there a way to customise panel titles in an effects plot?

Question

I have created the plot below to visualize interactions but I was wondering if I can customize the parts of the plot that say education = College, education = Graduate, education = High School.

Code:

# Generate example data
set.seed(123)
education <- factor(rep(c("High School", "College", "Graduate"), each = 30))
study_hours <- rep(c(10, 20, 30), each = 10, times = 3)
exam_scores <- rnorm(90, mean = 70, sd = 10)

# Create dataframe
example_df <- data.frame(education, study_hours, exam_scores)
# Fit the model with interaction term
library(glmmTMB)
models_example <- glmmTMB(exam_scores ~ education * study_hours, data = example_df)

# Generate the effects plot for the interaction term
library(effects)
effects_plot_example <- plot(allEffects(models_example, data = example_df))

# Display the plot
print(effects_plot_example)

Examples of what I tried:

1. I attempted to rename the variable to Educational Level before plotting but the space gave me this error

Error in parse(text = paste("~", paste(extras, collapse = "+"))) : 
  <text>:1:28: unexpected symbol

2. I tried several arguments in the plot/allEffects function like panel.title but I can't find which could modify this part of the plot, I am not even sure how this part of the plot is called.

Answer 1

If you use effects:::plot.eff() directly, it will produce a plot.eff object for which you can modify the strip function which is used to create the labels you want to change.

library(lattice)
# save the plot object
ef.plot <- effects:::plot.eff(allEffects(models_example, data = example_df)[[1]])

# change the function which draws strips at the top of the panels to a custom one
ef.plot$strip <- strip.custom(
  var.name = 'Educational Level', 
  factor.levels = c('College', 'Graduate', 'High School'),
  strip.names = c(T, T),
  strip.levels = c(T, T),
  sep = ': ')

# draw the plot with the new strip function
ef.plot