Adjusted R-squared

Also known as · Adjusted R squared

R2R^2 with a penalty for the number of regressors, Rˉ2=1−(1−R2)N−1N−K−1\bar R^2 = 1 - (1-R^2)\frac{N-1}{N-K-1}. It rises only if a new variable improves fit by more than a random regressor would, so it falls when you add a useless variable.

When to use

Use it to choose between models of different size. It still cannot compare models with different outcome variables or models with no intercept, and under Multicollinearity it barely changes when one of two collinear variables is dropped.

See Lecture 8.

Appears in