Gage R&R measures how much of your observed variation comes from the measurement system itself—not the process. In the Measure phase of DMAIC, this distinction is critical. If your gauge is adding significant noise, every decision you make downstream is built on unreliable data. This article walks you through each step of a proper Gage R&R study, from study design to result interpretation and corrective action.
You will find step-by-step instructions for setting up a crossed Gage R&R, the formulas behind both the Average and Range method and ANOVA Gage R&R, accepted thresholds under the 10/30 rule, and clear guidance on what actions to take when a measurement system fails. Whether you are conducting your first study or reviewing someone else's results, this guide gives you the practical detail you need.
Key Takeaways
- Gage R&R checks if your measurement system is reliable.
- It separates equipment error from operator error.
- A crossed study usually uses 10 parts, 3 operators, and 2–3 trials.
- Under 10% GRR is acceptable; over 30% needs correction.
- Fix failed studies before using data for Six Sigma decisions.
What Gage R&R Actually Measures in MSA Six Sigma
Gage repeatability and reproducibility decomposes total measurement variation into its root sources. Repeatability refers to variation from the equipment itself—same operator, same part, repeated measurements. Reproducibility captures variation between operators measuring the same part with the same gauge.
Together, these two components form the measurement system variation, or GRR. The goal is to compare that GRR to total observed variation, which includes both part-to-part variation and measurement error. When the measurement system contributes too large a share, you cannot trust your process data.
According to ASQ, GR&R is a core measurement system analysis tool used widely in quality and Six Sigma programs to evaluate whether a gauge produces accurate, repeatable, and reproducible results. This matters most during the Measure phase of DMAIC. Before you analyze any process data, you need to confirm the data is worth analyzing. A failed measurement system does not just affect one project—it can corrupt baseline data, capability studies, and control charts across an entire production line.
Step-by-Step Guide to Setting Up a Crossed Gage R&R Study
A crossed Gage R&R is the most common design for non‑destructive manufacturing and quality applications. In this setup, every operator measures every part, so all combinations of parts and operators are included. Here is how to set it up correctly before collecting a single data point.
Step 1: Define the Study Parameters — Parts, Operators, and Replicates
Select 10 parts that represent the full range of process variation, not just good parts. Choose 2 to 3 operators who regularly use the gauge in production. Plan for 2 to 3 replicates per operator per part, giving you enough data to estimate variance components reliably.
A common design is 10 parts × 3 operators × 2 replicates, yielding 60 total measurements. This structure provides enough data to estimate variation between parts, operators, and replicates with reasonable power in most applications.
Step 2: Randomize the Measurement Order
Operators should not know which part they measured previously, and parts should be presented in random order. Randomization prevents operators from adjusting their technique based on prior results, which would bias repeatability estimates.
Step 3: Keep Operators Blind to Each Other's Results
Each operator should complete all their measurements independently. Sharing results mid-study inflates reproducibility artificially and produces an overly optimistic %GRR.
Step 4: Record Measurements to Sufficient Resolution
The gauge must have enough resolution to detect differences between parts. A common rule is that the gauge discrimination should be at least one-tenth of the process tolerance or the expected part-to-part variation range.
Step 5: Choose Your Analysis Method Before You Start
Decide upfront whether you will use the Average and Range method or ANOVA Gage R&R. ANOVA is preferred because it separates out the operator-by-part interaction, which the range method cannot detect. Air Academy Associates covers both methods in their Measurement System Analysis course, giving practitioners the skills to apply the right approach for each study type.
Gage R&R Formulas: Average and Range Method vs. ANOVA Gage R&R
Both methods estimate the same variance components but differ in depth and precision. The Average and Range method is faster and easier to compute by hand. ANOVA Gage R&R uses a random-effects statistical model and produces more accurate estimates, especially when operator-by-part interaction is present.
Average and Range Method — Core Formulas
The range method calculates repeatability from the average range of repeated measurements within each operator. Key calculations include:
- Equipment Variation (EV): EV = R-bar x K1, where R-bar is the average range across all operators and K1 is a constant based on the number of replicates.
- Appraiser Variation (AV): AV = square root of [(X-diff x K2)^2 – (EV^2 / (n x r))], where X-diff is the range of operator averages, K2 depends on operator count, n is parts, and r is replicates.
- GRR: GRR = square root of (EV^2 + AV^2).
- %GRR: %GRR = (GRR / TV) x 100, where TV is total variation.
ANOVA Gage R&R — Random-Effects Model
ANOVA Gage R&R fits a random-effects model with three variance components: parts, operators, and the operator-by-part interaction. According to Wikipedia's explanation of ANOVA gauge R&R, this approach provides a statistically rigorous decomposition of total variance.
- Total Variation (TV): TV^2 = sigma^2_parts + sigma^2_operators + sigma^2_interaction + sigma^2_repeatability.
- %Contribution: Each component's variance divided by total variance, expressed as a percentage.
- %GRR: (sigma^2_GRR / sigma^2_Total) x 100, where sigma^2_GRR = sigma^2_repeatability + sigma^2_reproducibility.
The operator-by-part interaction term is particularly important. If certain operators measure specific parts differently from others, that interaction inflates your reproducibility estimate and points to a training or technique issue. The Advanced Measurement System Analysis course at Air Academy Associates goes deep into ANOVA-based MSA, including interaction effects, expanded studies, and nested designs for more complex measurement scenarios.
How to Interpret %GRR Results Against the 10/30 Rule
Once the study is complete, the primary metric is %GRR—the percentage of total variation attributable to the measurement system. The 10/30 rule is the industry-standard framework for evaluating that number. Here is what each band means in practice.
| %GRR Result | Interpretation | Action Required |
|---|---|---|
| Under 10% | Acceptable measurement system | Proceed with data collection and analysis |
| 10% to 30% | Conditional — may be acceptable depending on application | Evaluate cost, risk, and application context before proceeding |
| Over 30% | Unacceptable measurement system | Do not proceed — investigate and correct before using data |
These thresholds are widely used guidelines, not strict standards, so practitioners should also weigh decision risk, cost, and the criticality of the characteristic being measured.
GRR is compared against total variability to define the capability of the measurement system. A result under 10% means the gauge contributes a small enough share of variation that your data is trustworthy for decision-making.
The 10 to 30% conditional band requires judgment. In some low-risk applications, a 20% GRR may be acceptable if the cost of improving the gauge is high and the measurement is not used for critical accept/reject decisions. For safety-critical or highly capable processes, that same 20% could be disqualifying.
You might be wondering what to check beyond %GRR alone. Two additional metrics provide important context:
- Number of Distinct Categories (ndc): Should be 5 or greater. This tells you how many groups the gauge can reliably distinguish within the part variation range.
- P/T Ratio: Precision-to-Tolerance ratio, calculated as 5.15 x GRR divided by the specification tolerance. A P/T ratio at or below 0.10 is the standard acceptance guideline per Purdue University lecture notes.
What to Do When the Gage R&R Study Fails
A %GRR over 30% is not just a number to report—it is a signal that requires a structured response. Before any process improvement work continues, the measurement system must be addressed. Proceeding with bad measurement data leads to false conclusions, wasted resources, and missed defects.
According to 1Factory's practical guide to Gage R&R, variance components from the study point directly to the source of the problem. Use those components to guide your corrective action.
When Repeatability Is the Dominant Source
High repeatability variation points to the equipment or measurement method. Check for worn gauge components, inconsistent clamping or fixture setup, environmental factors like vibration or temperature, and inadequate gauge resolution relative to the measurement range.
When Reproducibility Is the Dominant Source
High reproducibility variation points to operator technique or training. Review the measurement procedure for ambiguity, standardize how parts are held and aligned, and provide retraining on consistent measurement technique across all operators.
When the Operator-by-Part Interaction Is Significant
This pattern means certain operators struggle with specific part types—often due to part geometry, surface finish, or feature accessibility. Targeted coaching and clearer measurement instructions for those specific part characteristics typically resolve this.
When the Gauge Itself Is the Problem
If equipment variation dominates even after operator training and procedure improvements, the gauge may need recalibration, repair, or replacement. In some cases, the measurement technology is simply not capable enough for the tolerance being controlled.
Rerun the Study After Corrective Action
Always repeat the Gage R&R study after making changes. A corrective action that looks reasonable on paper may not fully resolve the variation in practice. Confirm the improvement with data before moving forward in the DMAIC Measure phase.
Deepen Your MSA Skills With These Air Academy Associates Courses
Understanding the theory behind Gage R&R is one thing—executing a study correctly under real production conditions takes applied practice. Air Academy Associates offers targeted courses that build exactly that capability, from foundational MSA concepts through advanced variance component analysis.
Measurement System Analysis
This course covers the full scope of MSA tools used in Six Sigma and quality engineering, with direct application to Gage R&R study design and execution.
- Covers repeatability, reproducibility, bias, linearity, and stability studies.
- Teaches both the Average and Range method and ANOVA-based analysis.
- Applies the 10/30 rule and P/T ratio interpretation in real study scenarios.
- Designed for practitioners who need to run or review MSA studies on the job.
Advanced Measurement System Analysis
For practitioners ready to go beyond standard crossed Gage R&R, this course addresses complex study designs and deeper statistical analysis of measurement error.
- Explores nested, expanded, and attribute Gage R&R study structures.
- Examines operator-by-part interaction effects and their practical implications.
- Addresses measurement systems for destructive testing and non-replicable scenarios.
- Builds confidence in interpreting variance component outputs from statistical software.
Graphical and Measurement Tools Short Course
This short course connects measurement system analysis to the broader set of graphical tools used throughout DMAIC, giving practitioners a more complete analytical toolkit.
- Covers control charts, histograms, scatter plots, and measurement tools in one focused module.
- Shows how MSA results feed into graphical analysis and process monitoring decisions.
- Ideal for Green Belts and Black Belts who want to sharpen specific analytical skills quickly.
Process Capability Short Course
Process capability analysis depends entirely on clean measurement data—which is exactly why Gage R&R must come first. This short course teaches you how to calculate and interpret Cp, Cpk, and Pp/Ppk correctly once your measurement system is confirmed.
- Explains how measurement error inflates or masks true process capability estimates.
- Covers short-term vs. long-term capability indices and when each applies.
- Directly applicable to the Measure and Analyze phases of DMAIC projects.
Wrapping Up the Gage R&R Study Process
A Gage R&R study is not a formality—it is the foundation that makes every downstream DMAIC decision credible. Running it correctly, interpreting the results against the 10/30 rule, and acting on failures before moving forward protects the integrity of your entire improvement project. Air Academy Associates has equipped more than 250,000 professionals with the measurement system analysis skills needed to do exactly that—through practical, instructor-led training built on over 30 years of real-world Six Sigma application. If your team is ready to build that capability, explore our MSA and process improvement courses at Air Academy Associates or contact us directly at 1-800-748-1277.
Air Academy Associates offers expert Lean Six Sigma certification and training to sharpen your measurement system analysis skills. Our Master Black Belt instructors deliver real-world Gage R&R application you can use immediately. Get started today and build lasting process improvement capability.
FAQs
What Is Gage R&R and Why Is It Used?
Gage R&R (Gage Repeatability and Reproducibility) is a measurement system analysis method used to quantify how much of the observed variation in data comes from the measurement process itself—equipment (repeatability) and people (reproducibility). Six Sigma practitioners use it to confirm the measurement system is reliable before making decisions about process capability, improvement, or control—an emphasis we build into our Lean Six Sigma and DOE training through practical, real-world examples.
How Do You Calculate Gage R&R (GRR)?
GRR is typically calculated by having multiple operators measure multiple parts multiple times, then separating variation into repeatability (equipment), reproducibility (operator), and part-to-part variation using either the ANOVA method (most common today) or the Average & Range method. The GRR result is reported as a standard deviation (or variance) and often as a percent of total variation or tolerance—an approach we teach step-by-step so teams can interpret results correctly and act on them.
What Is an Acceptable Gage R&R Percentage?
A common guideline is: <10% is generally acceptable, 10–30% may be acceptable depending on risk and application, and >30% is typically unacceptable and needs improvement. The best threshold depends on whether you're comparing GRR to total process variation or to tolerance, and on the decision risk—something our instructors stress when coaching teams to avoid "one-size-fits-all" conclusions.
What Is the Difference Between Gage R&R and MSA?
MSA (Measurement System Analysis) is the broader discipline of evaluating measurement quality (e.g., bias, linearity, stability, attribute agreement), while Gage R&R is a specific MSA study focused on repeatability and reproducibility for variable measurement systems. In practice, GRR is often one key component of a complete MSA plan.
How Many Parts and Operators Are Needed for a Gage R&R Study?
A common, effective design is 10 parts × 3 operators × 2–3 trials (60–90 measurements total), with parts selected to represent the full expected process range. Depending on constraints and the type of measurement, you may adjust the design, but ensuring representative parts and consistent measurement procedures is critical—an area where our consulting and certification programs help teams avoid costly study design mistakes.
Related Articles:
