Definition

Measurement system analysis evaluates whether a measurement process provides data suitable for its intended decisions. It considers the combined effects of the measuring equipment, method, people, parts, and operating conditions rather than the instrument alone.

In this glossary, the full term is used to avoid confusion with MSA, which denotes Master Service Agreement in the commercial-contract entry. The two concepts are unrelated and require separate records.

Start with the decision the measurement supports

A system used to sort conforming from nonconforming parts may face different requirements from one used to detect small process changes. Define the intended use before choosing a study.

VPP recommends asking what decision would change if the measured value were slightly different. That question helps expose whether the measurement is adequate for the consequence.

Consider the relevant product range, operators, environment, and measurement procedure. A study performed under unusually favorable conditions may not represent routine production.

Distinguish calibration from system performance

Calibration addresses the instrument’s relationship to reference standards under specified conditions. It is necessary where applicable, but it does not fully describe the variation introduced by actual measurement practice.

For example, fixture placement, operator technique, part condition, and environmental changes can affect results. A calibrated instrument can still be used in a measurement process that is unsuitable for a particular decision.

Measurement-system analysis and calibration should therefore be connected, not treated as substitutes. The appropriate study depends on the data type, application, and required method.

Evaluate the relevant sources of variation

For variable measurements, studies may examine repeatability, reproducibility, bias, stability, and linearity as appropriate. Attribute-based decisions require methods suited to classifications such as pass/fail or defect categories.

  • Repeatability: Can the same method reproduce results under the relevant repeated conditions?
  • Reproducibility: How do results change across operators or other specified conditions?
  • Bias: Is there a systematic difference from an accepted reference?
  • Stability: Does performance remain consistent over time?
  • Range: Is the system suitable across the measurement range being used?

Use the applicable technical guidance to design and interpret the study. A single acceptance threshold should not be copied across unrelated applications without justification.

A hypothetical supplier disagreement

Assume a buyer and supplier disagree about whether a component dimension meets specification. Each uses a calibrated instrument, but their reported results differ near the acceptance boundary.

Before concluding that the supplier’s process is incapable, the teams compare methods, fixtures, part conditioning, operator practice, and reference checks. They then design an appropriate measurement study.

The investigation may identify a measurement-system issue, a product issue, or both. The operating decision should follow the evidence rather than the assumption that one calibrated device must be correct in every context.

Containment and customer protection should follow the applicable quality process while the issue is investigated. The study is not permission to release uncertain material.

Connect the result to process decisions

An unsuitable measurement system can distort capability indices and control-chart signals. Address material measurement limitations before drawing strong conclusions about the underlying process.

Maintain the approved method after the study. Changes to equipment, fixtures, software, operators, or measurement conditions may create a need for reassessment.

Retain study design, data, interpretation, and approval. A summary slide stating “measurement passed” is less useful than evidence showing what was evaluated and for which purpose.

Frequently asked questions

  • Does calibration replace measurement-system analysis? No. They address related but different questions.
  • Is gage R&R the whole discipline? No. It is one type of analysis within a broader assessment of measurement suitability.
  • Should every study use the same limits? No. Use applicable standards, customer requirements, and the intended decision.

Validate the evidence before validating the supplier

VPP helps procurement and operating teams make quality evidence part of supplier readiness. Technical measurement studies should be designed and interpreted by qualified quality specialists.