Residual Calculator
Observed minus predicted, with standardised residuals.
Enter your values
ⓘ Live results
Results
- Residuals
- 0.5, -0.8, 0.8, -1.1, 0.6
- Residual sum of squares
- 3.1
- Standard error of the estimate
- 1.0165
- Largest residual
- -1.1
- Mean residual
- 0
- Standardised residuals
- 0.49, -0.79, 0.79, -1.08, 0.59
Quick answer
A residual is the observed value minus the predicted value; positive residuals sit above the model line and negative ones below.
residual = observed − predicted
How to use the Residual Calculator
- 1Enter your observed data points.
- 2Enter the value your model predicts for each one, in the same order.
- 3Scan for standardised residuals beyond ±2 — those points deserve a second look.
Worked example
Using the values the calculator loads with:
Inputs
- Observed values12, 15, 19, 22, 27
- Predicted values11.5, 15.8, 18.2, 23.1, 26.4
Results
- Residuals0.5, -0.8, 0.8, -1.1, 0.6
- Residual sum of squares3.1
- Standard error of the estimate1.0165
- Largest residual-1.1
- Mean residual0
- Standardised residuals0.49, -0.79, 0.79, -1.08, 0.59
Frequently asked questions
What should a residual plot look like?
A shapeless cloud around zero. Curves, funnels or trends mean the model is missing something.
Should the mean residual be zero?
For a fitted least-squares line, yes, essentially. A clearly non-zero mean means the predictions are biased.
About this calculator
- Built and reviewed by James Whitfield · How we build and test calculators
- Report an error
Results are estimates for general information. For medical, legal, structural or financial decisions, confirm with a qualified professional.