Diagnosing Unacceptable Gage R&R in Minitab

Recognizable Condition

When the Gage R&R output shows %Study Var (or %Contribution) for total Gage R&R above 30 %, repeatability above 10 %, or reproducibility above 30 %, the measurement system is flagged as potentially inadequate for process control or capability analysis.

Cause/Diagnostic Table

Observed Pattern	Likely Cause
High appraiser %Contribution	Appraiser inconsistency (training, technique)
Low part %Contribution relative to gage	Parts do not span process spread or gage lacks discrimination
Significant operator*part interaction (p‑value < 0.05)	Operators measure parts differently; possible bias or fixturing issue
Non‑random residuals or trends in residual plots	Instrument drift, calibration error, or environmental influence

Ordered Checks

1. Verify that the data were collected in the prescribed random order with the correct number of operators, parts, and trials (e.g., 3 operators, 10 parts, 2 trials). In Minitab, check the worksheet layout or the data collection plan used.
2. Confirm that the gage was calibrated and zero‑balanced before each session. Look for calibration logs or perform a quick check with gauge blocks.
3. Examine the residual plots (Menu: Stat > Regression > Regression > Fit Regression Model or directly from the Gage R&R output: Graphs > Residual Plots) for normality (approx. straight line on normal probability plot) and independence (no patterns vs. fitted values).
4. Review the ANOVA table in the Gage R&R report for %Contribution of each source (Repeatability, Reproducibility, Part‑to‑Part, Interaction). Note which percentages exceed the acceptance limits.
5. Inspect the operator*part interaction plot (Stat > Quality Tools > Gage Study > Gage R&R Study (Crossed) then Graphs and select Interaction Plot). Look for non‑parallel lines indicating significant interaction.
6. If interaction is present or %Study Var remains high, repeat the study with an increased number of trials (e.g., from 2 to 3) to obtain a more stable estimate.

Fixes Tied to Findings

• High appraiser variation: Conduct additional training, create a standardized measurement work‑instruction, and retest.
• Low part variation relative to gage: Select parts that better represent the process spread (e.g., include extremes) or use a gage with higher resolution.
• Calibration/drift evidence: Schedule regular maintenance, verify with gauge blocks before each session, and log results.
• Significant interaction: Retrain operators to follow identical procedures, review fixturing, and ensure consistent part presentation.
• Overall %Study Var still >30 % after actions: Consider redesigning the gage or fixturing, or acquire a gage with higher discrimination.

Escalation Criteria

If corrective actions do not reduce total Gage R&R %Study Var below the applicable threshold (commonly 30 % or a stricter customer limit), escalate to:

• Measurement system redesign or acquisition of a higher‑resolution gage.
• Involvement of a Six Sigma Black Belt or quality engineer for a deeper MSA review.
• Consult the AIAG MSA Manual (4th ed.) for industry‑specific limits.

Verification

Reproduce the analysis using Minitab’s built‑in sample data: Help > Sample Data > Measurement Systems Analysis > Gage R&R. Follow the ordered checks above; the output should show low %Study Var (≈7 %) and no significant interaction, confirming that the diagnostic steps work as intended.