Is there a difference between the $R^2$ statistic of a linear regression and pearson correlation?

Question

Is there any difference between $R^2$ that we get from linear regression and pearson correlation?

$\sqrt{}$ is your friend :) – epsilone Nov 06 '16 at 23:38 — epsilone, Nov 06 '16 at 23:38

score 2 · Answer 1 · answered Nov 06 '16 at 23:48

A simple answer: in a simple regression model (with a single independent variable) where the independent is interval (not nominal) - Pearsons correlation$^2$ equals $R^2$. But in a multiple regression model it does not. Pearson correlation cannot control for confounders, while a regression model can, thus the explained variance of the dependent variable is a combination of the multiple independent variables' effect.

See: coefficient of determination

Is there a difference between the $R^2$ statistic of a linear regression and pearson correlation?

1 Answers1