Degrees of freedom (regression)

In regression output...

C. Total DF = 1 + 50
C. Total DF = 51

C. Total DF = 52 - 1
C. Total DF = 51

Okay... but actually explain it to me now.

df (Total) = N - 1

N is the number of observations.

You may notice this is the same, simple degrees of freedom formula we used to use in t-tests!

df (Model) = p

p is the number of predictor variables in the model.

The degrees of freedom in the model is simply however many predictor variables are present!

df (Error) = N - p - 1

N is the number of observations.
p is the number of predictor variables in the model.

I like to think of these as the "leftover" degrees of freedom after accounting for the model.

Activate AutoScroll