Bias in Simple Terms
A gauge can give very consistent readings and still be wrong. If a micrometer always reads 0.02 mm too high, its repeatability is good but its bias is 0.02 mm. Bias is about accuracy, while Gauge R&R is about precision. Both are part of MSA.
How a Bias Study Works
- Take a reference part or master whose value is known, usually from a calibrated lab measurement.
- One appraiser measures it at least 10 times under normal conditions.
- Calculate the average and subtract the reference value. That difference is the bias.
- Test whether the bias is statistically different from zero, as described in the AIAG MSA manual.
Bias is different from linearity, which looks at whether bias changes across the gauge range, and from stability, which looks at whether it drifts over time.
Example
A height gauge used on a pump housing is checked against a 50.000 mm gauge block. Fifteen readings average 50.012 mm, so the bias is plus 0.012 mm. With a tolerance of plus or minus 0.05 mm, this is a large share of the tolerance, so the gauge is pulled for calibration and the probe is found to be worn.
Common Causes of Bias
- Wear on the gauge or probe.
- A master that is damaged or out of calibration.
- Wrong zero setting or calibration.
- Temperature differences between the part and the gauge.
- Appraisers using a different method from the one in the instruction.
OperaHive calibration management software tracks calibration due dates and certificates and helps investigate gauges that fail.