CODEWITHSUNDEEP
rregressiontidyverseinteraction
melbez
melbez

Reputation: 1000

Running several regressions with interaction terms simultaneously in R

I have a dataframe (df) that looks like the following:

yourbehavior  condition  tryreduce  comfortable  vegvalidity
1             2          4          3            3
2             5          7          4            5
3             7          1          2            4

I would like to figure out a way to run the following regressions efficiently.

lm(yourbehavior ~ condition + tryreduce + (condition*tryreduce), data = df)
lm(yourbehavior ~ condition + comfortable + (condition*comfortable), data = df)
lm(yourbehavior ~ condition + vegvalidity + (condition*vegvalidity), data = df)

Does anyone know how I can perform these regression analyses simultaneously? I have several more in my actual dataset than I am including in this example.

I've read that using a for-loop is typically not the most efficient way to conduct these types of analyses in R. If possible, I'd prefer to use the tidyverse since that's what I understand best.

Upvotes: 3

Views: 519

Answers (1)

eipi10
eipi10

Reputation: 93761

You could do something like this, which will create a list where each element is one of the regression models and the name of each list element will be the variable that is interacted with condition:

library(tidyverse)

interaction.vars = c("tryreduce", "comfortable", "vegvalidity")

form = paste("yourbehavior ~ condition *", interaction.vars)

models = form %>% 
  set_names(interaction.vars) %>% 
  map(~lm(.x, data=df))

If you want to iterate over every column except the first two you could do:

interaction.vars = names(df)[!names(df) %in% c("yourbehavior","condition")]

Note that in a model formula a*b is equivalent to a + b + a*b, so you don't need to repeat each column name.

for loops aren't necessarily bad, and the map function above is essentially iterating in the same way as a for loop. However, many functions in R are vectorized which is much faster and more succinct than looping.

Upvotes: 4

Related Questions