Overview
R-squared shows how much of the variation in your data a regression model explains, from 0 to 1. Adjusted R-squared corrects for the number of predictors.
Enter the sum of squared errors (SSE), the total sum of squares (SST), the sample size and the number of predictors.
Worked examples
| Inputs | Result |
|---|---|
| SSE (Sum Squared Error)120SST (Total Sum Squares)480Sample Size (n)60Number of Predictors3Residual for a Point-1.8 | R-squared: 0.75Adjusted R-squared: 0.74 |
Show 2 more examples
| SSE (Sum Squared Error)50SST (Total Sum Squares)500Sample Size (n)30Number of Predictors1Residual for a Point0 | R-squared: 0.9Adjusted R-squared: 0.9 |
| SSE (Sum Squared Error)300SST (Total Sum Squares)400Sample Size (n)100Number of Predictors5Residual for a Point0 | R-squared: 0.25Adjusted R-squared: 0.21 |
R-squared by SST
With the calculator's starting values, the r-squared is 0.75. The table shows how it changes with the SST.
| SST (Total Sum Squares) | R-squared |
|---|---|
| 300 | 0.6 |
| 400 | 0.7 |
| 480starting value | 0.75 |
| 500 | 0.76 |
| 600 | 0.8 |
| 700 | 0.83 |
Other inputs: SSE: 120 Sum Squared Error · Sample Size: 60 n · Number of Predictors: 3 · Residual for a Point: -1.8. Change them in the calculator above for your own numbers.
How it works
R² is 1 − SSE ÷ SST. Adding predictors never lowers R², so adjusted R² penalizes them: 1 − (1 − R²) × (n − 1) ÷ (n − predictors − 1). Compare models with adjusted R² when they have different numbers of predictors.
The formula
| SSE | Sum of squared residuals |
| SST | Sum of squared differences from the mean |
Practical tips
- SSE 120 and SST 480 give R² = 0.75, and 0.74 adjusted with 60 observations and 3 predictors.
- A high R² doesn't prove the model is right. Check the residuals.
- R² can't tell you whether a predictor causes anything.
Frequently asked questions
What's a good R²?
It depends on the field. Physical sciences often see 0.9+, social sciences much lower.
Can R² be negative?
For a model worse than predicting the mean, yes, with some calculations. Here, SSE above SST gives a negative value.
Why use adjusted R²?
It only rises if a new predictor improves the model more than chance would.