Repeating measurement-system analysis

Learning objectives

• Explain the purpose of repeating measurement-system analysis in a garment-factory improvement project.
• Apply the described method to a representative shop-floor situation.
• Recognize the common mistakes and how to avoid them.

Concept

Repeat MSA confirms that new ranges, methods, gauges, inspectors, automation, and critical X controls remain trustworthy after improvement.

Reassess after gauge, fixture, software, method, tolerance, product range, appraiser, environment, or calibration change.

Compare with original acceptance criteria; do not weaken thresholds to pass.

Variables, units, and assumptions

%GRR (%), P/T ratio (%), bias (units), stability (drift over time).

Representative parts spanning tolerance, trained appraisers, blinded runs.

Free-spreadsheet workflow

Free worksheet: 2-3 appraisers × 5-10 parts × 2-3 trials; ANOVA or range method; compare %GRR vs 10-30% acceptance.

Interpretation and limitations

%GRR ≤ 10% acceptable; 10-30% conditional; >30% unusable.

Attribute MSA (kappa/agreement) needed for pass/fail; small studies have wide confidence.

Garment-factory example

A new digital measurement fixture changes operator interaction, so repeatability, reproducibility, bias, and stability are rechecked.

Method

1. Review the change.
2. Select representative parts/appraisers.
3. Blind and randomize.
4. Analyze with a free MSA worksheet.
5. Correct system.
6. Document approval and surveillance.

Common mistakes

• Assuming a prior MSA covers a redesigned method.
• Testing only easy parts near the middle of the tolerance.