Key takeaways
Percent error is the gap between a measured and an accepted value, as a share of the accepted value. It tells you how good an estimate is, independent of units.
- Percent error = |measured − accepted| ÷ |accepted| × 100.
- Use the accepted value, not the measured one, as the denominator.
- Near zero, percent error explodes; use absolute error instead.
The formula
Percent error = |measured − accepted| ÷ |accepted| × 100.
A scale reads 98 g for a 100 g reference mass: |98 − 100| ÷ 100 × 100 = 2 %.
Signed or absolute?
Scientists usually drop the sign and report 2 %. Engineers sometimes keep it, so −2 % means the reading is low. State which you use.
When it misleads
If the accepted value is close to zero, a small absolute miss becomes a huge percentage. A thermometer that reads 1 °C when the truth is 0.5 °C is "100 % off" yet only half a degree out. Report absolute error in those cases.
Related questions
- Can percent error be negative?
