stacked regression model tables using modelsummary and kableextra

stacked.R

library(tidyverse)
library(modelsummary)
library(kableExtra)

model_am1 = mtcars %>% 
  filter(am == 1) %>% 
  lm(mpg ~ wt, data = .)

model_am0 = mtcars %>% 
  filter(am == 0) %>% 
  lm(mpg ~ wt, data = .)

model_am1_hp = mtcars %>% 
  filter(am == 1) %>% 
  lm(mpg ~ wt + hp, data = .)

model_am0_hp = mtcars %>% 
  filter(am == 0) %>% 
  lm(mpg ~ wt + hp, data = .)

models <- list(
  list(model_am1, model_am0),
  list(model_am1_hp, model_am0_hp)
)

modelsummary(
  models,
  shape = "rcollapse",
  gof_map = "nobs",
  coef_map = "wt",
  output = "data.frame"
) %>% 
  rename(row_names = 1) %>% 
  mutate(row_names = if_else(row_number() == 1, "No controls", row_names)) %>%
  mutate(row_names = if_else(row_number() == 3, "Additional controls", row_names)) %>% 
  kbl(booktabs = T, format = "latex", col.names = NULL) %>% 
  kable_classic(full_width = FALSE) %>% 
  row_spec(c(2, 4), extra_latex_after = "\\\\[-0.3cm]") %>% 
  add_header_above(c(" " = 1, "Automatic transmission" = 1, "Manual transmission" = 1))