Degrees of Freedom

The number of observations that are still free to vary once the parameters have been estimated; in OLS with KK regressors and an intercept it is N−K−1N-K-1.

When to use

It sets which t-distribution to use for critical values and confidence intervals, and the variance estimate divides by it. In F-tests, qq is the difference in degrees of freedom of the restricted and unrestricted models.

See Lec 5.

Appears in

No references yet.