Dummy Variable

Also known as · Dummy Variables

A 0/1 regressor marking membership of a category. In yi=β0+β1xi+β2 is.Ai+uiy_i = \beta_0 + \beta_1 x_i + \beta_2\,is.A_i + u_i the dummy keeps the slope the same and shifts the intercept by β2\beta_2, so β2\beta_2 is the A-versus-not-A difference holding xx fixed.

When to use

A variable with KK categories needs K−1K-1 dummies, with one left out as the Base Category. Test the whole categorical variable with an F-test on all its dummies, and a single binary category with a t-test.

See Lecture 10.

Appears in