An estimate of systematic error, obtained by averaging repeated measurements of a reference standard and subtracting the standard's assigned value. A torque analyzer averaging 50.6 N·m on a 50.0 N·m reference has a bias of +0.6 N·m. Labs either correct for bias or, when it is small and left uncorrected, carry it explicitly in the uncertainty budget so it cannot hide.
Check standard shows a bias of plus 3 ppm since the move, apply the correction until we recalibrate.
Never fold a known, significant bias into the uncertainty just to avoid applying a correction; assessors spot it fast.
Also heard as
- measurement bias
Do not mix it up with
- Offset: Offset usually means the reading at zero input; bias is the average deviation from a reference at any point.
Related terms
Systematic error
Error component that pushes every repeat reading the same way, or follows a pattern you can predict and correct.
Trueness
Where a method's long-run average result lands relative to a reference value; small bias means good trueness.
Correction
The amount added to a raw reading to cancel a known, repeatable error, equal to that error with its sign flipped.
Check standard
A stable item measured regularly to confirm a measurement process still behaves as it did when characterized.