Is there an easy way to include all possible two-way interactions in a model in R?
Given this model:
lm(a~b+c+d)
What syntax would be used so that the model would include b, c, d, bc, bd, and cd as explanatory variables, were bc is the interaction term of main effects b and c.
You can write the following:
lm(a ~ (b + c + d)^2)
This creates all combinations of two-way interactions between b, c, and d
For example:
lm(mpg ~ (cyl+disp+hp)^2, data = mtcars)
gives:
Call:
lm(formula = mpg ~ (cyl + disp + hp)^2, data = mtcars)
Coefficients:
(Intercept) cyl disp hp cyl:disp cyl:hp disp:hp
5.601e+01 -4.427e+00 -1.184e-01 -1.142e-01 1.439e-02 1.556e-02 -8.567e-05