On this page
What this lesson covers
Regression goodness of fit measures how much dependent-variable variation the fitted line explains and how widely observations scatter around it.
The lesson decomposes SST into SSR and SSE, calculates R-squared and constructs an ANOVA F-test. It distinguishes mean square error from the standard error of the estimate.
How do you calculate R-squared, F and the standard error of the estimate?
Measure explained variation with SSR/SST, compare mean squares with the F-statistic and take the square root of MSE for the standard error of the estimate.
R-squared = explained sum of squares (SSR) / total sum of squares (SST)
- Fitted value
- Sample mean of Y
F = [SSR / k] / [SSE / (n − k − 1)]
- Number of independent variables
- Sample size
Standard error of the estimate = sqrt(MSE) = sqrt[SSE / (n − 2)]
- Standard error of the estimate
For example, for a simple regression with an intercept and six companies, SST = 316, SSR = 252 and SSE = 64. R-squared = 252 / 316 ≈ 0.797; with k = 1, MSE = 64 / (6 − 2) = 16, F = 252 / 16 = 15.75 and the standard error of the estimate is sqrt(16) = 4.
Common mistakes
- Treating the smallest possible SSE as proof of a good fit overstates what OLS guarantees.
- Using R-squared alone to test whether the slope differs from zero treats a descriptive measure as a statistical test.
- Reporting mean square error as the standard error of the estimate omits the required square root.
Try a practice question
Choose an answer, then check it to read the explanation for every option.
Continue studying in Quizara
In Quizara the lesson comes with more practice questions on the topic, explanations for every option and an AI tutor for your follow-up questions.
Written by the Quizara Editorial Team under our editorial standards.