InsightFab
Knowledge Base/Gauge Bias and Linearity: Is Your Gauge Accurate Across the Entire Measurement Range?
MSA7 min read

Gauge Bias and Linearity: Is Your Gauge Accurate Across the Entire Measurement Range?

Gage R&R only measures gauge repeatability and reproducibility but not Bias and Linearity. These two issues can cause a gauge to be accurate in some measurement ranges but systematically high or low in others.

Scenario

Your caliper passed Gage R&R. However, you noticed that when measuring small parts (approx. 10mm), the readings are accurate, but when measuring large parts (approx. 50mm), they are systematically 0.05mm too high. This is not an issue that Gage R&R can detect.

Bias

Bias = Measurement Average - True Value (Reference Value)

A gauge's systematic error – regardless of who measures or how many times it's measured – will consistently be either too high or too low.

Sources:

  • Gauge not calibrated
  • Gauge design flaws
  • Improper usage (zero point misset)

Determination: Measure a standard part 10-15 times, calculate the average, and perform a t-test to determine if the bias is significantly different from 0.

Correction Methods: Calibration (adjusting gauge zero point or gain)

Linearity

Linearity = Does the bias remain consistent across the entire measurement range

The degree to which a gauge's bias changes at different measurement values.

Example:

  • Measure 10mm: Bias = +0.01mm
  • Measure 30mm: Bias = +0.05mm
  • Measure 50mm: Bias = +0.09mm

Bias increases as the measured value increases; this indicates poor linearity.

Sources:

  • Uneven gauge wear
  • Mechanical structural issues in gauge design
  • Usage outside the gauge's design range

How to Evaluate Linearity

Steps:

  1. Select 5 standard parts evenly distributed across the measurement range
  2. Measure each standard part 10-15 times
  3. Calculate the bias for each standard part
  4. Plot a scatter diagram with "Measured Value" on the X-axis and "Bias" on the Y-axis, then perform regression.
  5. If the slope of the regression line is significantly non-zero, it indicates a linearity problem.

Acceptance Criteria:

  • Slope p-value > 0.05: Linearity is acceptable
  • Slope p-value < 0.05: Linearity problem exists, needs correction or gauge replacement

Complete MSA Framework for Bias and Linearity

MSA StudyQuestion Addressed
Gage R&RHow is the gauge's repeatability and reproducibility?
BiasIs the gauge systematically high or low?
LinearityIs the bias consistent across the entire measurement range?
StabilityDoes the gauge performance change over time?

Complete measurement system validation requires all four studies.

Golden Quote

"A gauge passing Gage R&R is just the entry ticket; Bias and Linearity are what tell you if the gauge is trustworthy within your operating range."

Want to try it yourself?

Every tool mentioned in this article is available on InsightFab — just upload a CSV to analyze.

Go to Tools