How this calculation works
Subtract the reference from the measurement, then scale the difference by the reference magnitude. The absolute result measures size. The signed result retains the direction of the measured-minus-reference difference.
Absolute percent error = |measured − reference| ÷ |reference| × 100.A worked example
A measured 97 compared with a reference of 100 has a difference of −3 and an absolute percent error of 3%.
Good to know
Both inputs must use the same unit. Percentage error is undefined when the reference is zero.
How the inputs affect the result
Report the measured and reference values alongside the percentage when the denominator is close to zero. A small raw difference can produce a large percentage in that case.
The signed result changes direction while the absolute error remains 3%.
Where the calculation applies
Both inputs must use the same unit. Percentage error is undefined when the reference is zero.
For a related calculation, see the Percentage change calculator.
A few questions, answered
Which number is the denominator?
Use the known or accepted reference. If the two values have equal status, a symmetric percent-difference formula answers a different question.
What does the “Measurement above the reference” setting illustrate?
The signed result changes direction while the absolute error remains 3%.