Quick Scoop

R-squared, often written as R2R^2R2, is a statistic that tells you how much of the variation in a dependent variable is explained by a regression model. A value closer to 1 means the model fits the data better, while a value closer to 0 means it explains little of the variation.

What it means

  • R2=0R^2=0R2=0: the model explains none of the outcome’s variation.
  • R2=1R^2=1R2=1: the model explains all of the outcome’s variation.
  • In between: it shows the proportion of variance explained by the model.

Simple example

If a regression model has R2=0.72R^2=0.72R2=0.72, that means about 72% of the variation in the outcome is explained by the predictors in the model.

Important note

A high R2R^2R2 does not automatically mean the model is good, and a low R2R^2R2 is not always bad; it depends on the field, the data, and the purpose of the model.

TL;DR

R2R^2R2 is a goodness-of-fit measure for regression. It answers: “How much of the outcome can my model explain?”