Standard error of the estimate
The standard error of the estimate measures how far actual data points typically fall from a fitted regression line, computed from the error sum of squares and used to build prediction intervals and test the slope's significance.
What it is
See it move
Loading infographic...
Five months of machine hours and maintenance cost produce residuals of €4, −€3, €0, −€7 and €6, squaring and summing to an error sum of squares of 110. With n − 2 = 3 degrees of freedom, the standard error of the estimate is √(110 ÷ 3) = €6.06 — the typical distance between actual cost and the fitted line.
The formula
LaTeX
Variables
- Error sum of squares (squared units of y)
- Number of observations
Estimates the typical distance between actual values and the fitted regression line, using n − 2 degrees of freedom for a simple regression.