Tidyverse approach to fitting multiple models

multiple_models_tidy.R

library(tidyverse)
library(palmerpenguins) # for `penguins` data set

# fit a model for each variable (bill length, bill depth, flipper length)
penguin_models <- penguins %>%
  pivot_longer(c(bill_length_mm, bill_depth_mm, flipper_length_mm),
               names_to = "outcome_name",
               values_to = "outcome") %>%
  group_by(outcome_name) %>%
  summarise(model = list(lm(outcome ~ species, data = cur_data())))

# regression coefficients
library(broom)
penguin_models %>%
  rowwise(outcome_name) %>%
  summarise(tidy(model, conf.int = TRUE))

# estimated means
library(emmeans)
penguin_models %>%
  rowwise(outcome_name) %>%
  summarise(emmeans(model, "species") %>%
              as_tibble())

# pairwise comparisons
penguin_models %>%
  rowwise(outcome_name) %>%
  summarise(emmeans(model, "species") %>%
              pairs(adjust = "none", infer = TRUE) %>%
              as_tibble())

minor correction: cur_group() should be cur_data()