Bias (Systematic Deviation)
Difference between the mean of repeated measurements and the known reference value.
Also known as: systematic deviation, systematic error, measurement bias
The bias of a measurement system is the systematic deviation of its mean from the true (reference) value. It is the counterpart to random dispersion: bias stays no matter how many times you measure, and only calibration or a design change can remove it.
A non-zero bias violates the trueness of the system. In MSA Type 1, bias enters the Cgk index: Cg captures precision only, Cgk additionally accounts for how much tolerance the bias consumes.
- Bias ≈ 0 and high Cg → true and precise (capable)
- Large bias, high Cg → precise but shifted — recalibrate
- Small bias, low Cg → correct on average but too spread out
- Large bias and low Cg → neither true nor precise (not capable)
See also
Used in
In the Algorithm Lab
Sources
- DIN ISO 5725-1: Genauigkeit (Richtigkeit und Präzision) von Messverfahren und Messergebnissen — Teil 1
- VDA Band 5 — Prüfprozesseignung, 3. Auflage
- AIAG — Measurement Systems Analysis Reference Manual, 4th Edition (MSA-4)