Bias

Bias, also called systematic error or offset, is the consistent deviation of measurements from the true value. Unlike random error, which causes readings to scatter unpredictably, bias shifts all readings in the same direction by approximately the same amount. If a scale consistently reads 0.5 grams high, it has a bias of +0.5 grams.

Bias is determined by comparing the average of multiple measurements against a known reference value. The difference is the bias. In calibration, bias is one of the primary characteristics evaluated: the as-found readings are compared to the reference standard values, and any consistent offset is identified. If the bias exceeds the allowed tolerance, the instrument may be adjusted (zeroed, offset-corrected, or realigned) to reduce the bias to an acceptable level.

For calibration management, understanding bias is important because it is correctable. Once quantified, bias can be eliminated through adjustment or compensated through correction factors applied to measurements. However, uncorrected bias directly degrades accuracy and can lead to systematic quality problems in manufacturing. Tracking bias over time (through as-found data from successive calibrations) reveals instrument drift and helps optimize calibration intervals. A steadily growing bias may indicate wear, aging components, or environmental damage.