Using Minitab’s Crossed Gage R&R to Quantify Measurement System Variability

Problem: Is your measurement system adding too much noise?

Before you trust capability indices or control limits, you need to know how much of the observed variation comes from the measurement process itself. If repeatability and reproducibility (R&R) are high, any conclusions about the product may be misleading.

Takeaway

Running a crossed Gage R&R study in Minitab gives you %Study Var for repeatability and reproducibility. If this value is below 10 % the measurement system is excellent; 10‑30 % is borderline acceptable; above 30 % signals a need for improvement.

1. Prepare your data

Minitab accepts either a stacked format (one column for measurements, plus columns for Operator, Part, and Replicate) or an unstacked format (each operator’s measurements in separate columns). For a simple crossed design you need the same number of replicates for every operator‑part combination.

Example layout (stacked)

C1   C2   C3   C4
Measurement Operator Part Replicate
12.4      A        1      1
12.6      A        1      2
11.9      A        2      1
...

Enter the data into a worksheet, then save it (e.g., MyGageRR.MTW).

2. Run the crossed Gage R&R

1. Choose Stat → Quality Tools → Gage Study → Gage R&R Study (Crossed).
2. In the dialog:
   • If your data are stacked, select Each measurement in its own column and map Measurement, Operator, Part, and Replicate columns.
   • If unstacked, choose Each operator’s measurements in separate columns and select the appropriate columns.
3. Click Options to specify the number of replicates per cell (Minitab will highlight any imbalance).
4. Press OK to run the analysis.

Permissions: You need read access to the worksheet and the ability to create new output (no admin rights required).

Risk: If the design is unbalanced or contains outliers, Minitab will show a warning but will not correct the estimates; the results may be biased.

3. Interpret the output

The session window shows three key percentages for each source:

• %Contribution – proportion of total variance.
• %Study Var – proportion of the study variation (6 × σ). This is the metric most engineers use.
• %Tolerance – proportion of the specified tolerance range.

Guidelines (commonly cited in AIAG MSA):

• <10 % – measurement system is excellent.
• 10‑30 % – may be acceptable depending on the application; consider improvement if you are pushing process limits.
• >30 % – measurement system needs improvement.

Worked example: Three operators (A, B, C) measure five parts (1‑5) twice each. After running the analysis, Minitab reports:

• Repeatability %Study Var = 8 %
• Reproducibility %Study Var = 10 %
• Combined %Study Var = 18 %

The combined value of 18 % falls in the borderline acceptable range. You might decide to reduce operator‑related variation (e.g., through training) before using the data for capability analysis.

4. Limitations and practical checks

• Assumptions: Normally distributed measurement errors and a balanced design. Minitab will flag missing cells but does not automatically estimate them.
• Stability: Gage R&R does not assess drift over time. For long‑term monitoring, run a control chart on measurement standards.
• Verification: Open the built‑in sample worksheet GageRR.MTW (File → Open Worksheet) and repeat the same steps. Compare the generated %Study Var with the values shown in Minitab’s help documentation for that sample.
• Residual check: After the analysis, request the residual plots (Graph → Gage R&R Residuals). Look for patterns; random scatter supports the normality assumption.

Actionable next step

Collect a balanced set of measurements (same replicates per operator‑part), run the crossed Gage R&R as described, and examine the combined %Study Var. If it exceeds 30 %, plan a root‑cause investigation (e.g., operator training, fixture repeatability, or instrument calibration) before proceeding with further process capability studies.