Quality and Inspection

Gauge R and R: Measurement System Analysis in Manufacturing

Gauge R&R measures how much of your measurement variation comes from the gauge and appraisers versus the actual parts. Here is how to run a study, interpret results, and act on them.

Gauge R and R quantifies how much measurement variation comes from the measurement system instead of the part. Repeatability is the variation when the same appraiser measures the same part multiple times with the same gauge. Reproducibility is the variation when different appraisers measure the same part. Total Gauge R and R = sqrt(Repeatability^2 + Reproducibility^2), and %GRR = (Total Gauge R and R / Total Process Variation) x 100%. By AIAG convention, below 10% is generally acceptable, 10% to 30% may be acceptable depending on risk, and above 30% is usually unacceptable.

A solid Gauge R and R study uses at least 2 appraisers, 3 measurements per appraiser, and 10 parts that span the real process range. That means a minimum of 60 measurements before analysis. The parts need to represent actual variation, not a stack of easy nominal parts that make the gauge look better than it is. Study data can be analyzed with the average and range method or with ANOVA, and ANOVA is usually better when you want to separate interaction effects. Gauge resolution should also be checked, and a good rule is resolution at least 10 times finer than the tolerance.

The biggest mistake is running the study on too few parts or with parts that do not cover the full range. Another common error is blaming the gauge when the %GRR is high only because the process itself is very tight and part-to-part variation is tiny. This process masking effect is why you should report the number of distinct categories, not just %GRR. You usually want ndc of 5 or more if the system will be used for process control. Worn gauge contacts, weak fixturing, and poor appraiser technique are also common causes of inflated GRR.

Use the result to decide whether the measurement system is good enough for acceptance, SPC, or capability studies. If reproducibility is the biggest contributor, focus on training, method clarity, and fixturing. If repeatability dominates, the gauge itself likely needs calibration, repair, or replacement. That makes Gauge R and R a practical decision tool, not just a quality document. There is no point chasing a 1.67 Cpk target if the gauge is adding 35% of the observed variation.

Gauge results should feed directly into process improvement planning. A poor measurement system creates false alarms, hides real shifts, and wastes time sorting good parts from bad. Related checks such as calibration history, bias, linearity, and stability studies help explain why GRR is high on a critical dimension. Review GRR any time you change a fixture, gauge type, or measurement method. The fastest way to waste a Six Sigma project is to improve a process using data from a weak gauge.

Published 2026-05-28.