MSA – Measurement System Analysis for Assessing the Suitability of Measuring Systems

Measurement system analysis (MSA) is a collection of statistical procedures for assessing whether a measuring system is suitable for inspecting a specific characteristic. It breaks down the observed measurement variation into its causes – in particular into components caused by the measuring equipment itself and by the appraiser – and compares them with the variation of the inspected characteristic. The aim is a statement as to whether a measuring system delivers measured values that can be trusted for the intended decision (e.g. good part/bad part, process capability assessment).

What Is the Objective of Measurement System Analysis?

MSA is intended to determine whether the variation of a measuring system is sufficiently small in relation to the variation of the characteristic to be inspected to make reliable decisions on the basis of the measured values. To this end, the total observed variation is broken down mathematically into a product-related component (part variation) and a measuring-system-related component. MSA makes a statement exclusively about the measuring system examined under the conditions examined; it replaces neither the evaluation of the inspection process as a whole nor a statement about the cause of an identified lack of suitability.

Which Variation Components Does MSA Distinguish?

For variable (continuously measurable) characteristics, MSA breaks down the measuring system variation into several components:

• Equipment variation (EV) – repeatability: variation that arises when one appraiser repeatedly measures the same characteristic on the same part with the same measuring equipment.
• Appraiser variation (AV) – reproducibility: variation that arises when different appraisers measure the same characteristic on the same parts with the same measuring equipment.
• Interaction (IA): the share of variation that results from the interaction of appraiser and part, e.g. when individual appraisers systematically obtain deviating values for certain parts.

From EV, AV, and – depending on the calculation method – IA, the combined measuring system variation gage repeatability and reproducibility (GRR) is formed, which together with the part variation (PV) results in the total variation of the inspection process. Whether the interaction IA is reported separately depends on the evaluation method chosen: in the analysis of variance (ANOVA), IA is calculated as a separate variation component; in the average and range method, it is mathematically included in AV and is not reported separately.

Which Procedures Does MSA Comprise for Variable Characteristics?

In industrial practice, particularly in the German automotive supply chain, a division into several procedures has become established for variable measuring systems, based on the methods described in the AIAG reference manual:

• Type 1 study: examination of the capability of the measuring equipment alone, without appraiser influence. A single appraiser repeatedly measures a reference part or standard; the indices Cg and Cgk are evaluated.
• Type 2 study: examination of the entire measurement process including appraiser influence, using several appraisers, several parts, and repeated measurements. The evaluation is carried out using the average and range method or an analysis of variance (ANOVA); the central result is the index %GRR.
• Type 3 study: a special case of the type 2 study for measuring systems without or with negligible operator influence, for example automated or integrated measuring systems. The appraiser component of the test setup is omitted accordingly.

The AIAG reference manual itself does not structure these methods as a rigid numbering “type 1 to 3”, but describes them as a family of methods for repeatability, reproducibility, bias, linearity, and stability. The numbering type 1–3 is the designation widely used in operational practice, based on the AIAG concept.

Which Indices Does MSA Provide for Variable Measuring Systems?

In addition to EV, AV, IA, and %GRR, MSA uses the following indices:

• Cg/Cgk: indices from the type 1 study. Cg relates the spread of the repeated measurements to a specified proportion of the tolerance; Cgk additionally takes into account the systematic deviation (bias) from the reference value. In operational practice, a minimum value of Cg and Cgk ≥ 1.33 is frequently used as a guide; the specific limit is defined company- or customer-specifically and is not uniformly standardized.
• Bias: systematic measurement deviation – the difference between the mean of repeated measurements and the reference value assumed to be correct.
• Linearity: behavior of the bias over the entire measuring range of the measuring system; insufficient linearity means that the bias differs depending on the size of the characteristic.
• Stability: behavior of the bias over time, determined by repeated measurements of the same reference value at different points in time. In practice, the evaluation is frequently carried out using a control chart with which bias or variation is tracked over a longer period.
• %GRR: ratio of the combined measuring system variation (GRR) to the total variation or to the tolerance of the characteristic. In technical practice, based on the guide values of the AIAG reference manual, the following usually applies: %GRR ≤ 10% as capable, 10–30% as conditionally capable (depending on the application and criticality of the characteristic), and > 30% as not capable. VDA Volume 5 uses partly different indices and a risk-based approach to measurement uncertainty to evaluate the capability of measurement and inspection processes; the specific limit values to be applied depend on the reference source agreed between customer and supplier.
• ndc (number of distinct categories): the number of categories into which the measuring system can still resolve the part variation distinguishably. An ndc of at least 5 is considered a guide value; the value is always rounded down.

How Does MSA Work for Attribute Characteristics?

If a characteristic cannot be measured on a continuous scale but can only be assessed comparatively – for example in color, surface, or go/no-go inspections – attribute MSA is used. Several appraisers repeatedly evaluate the same parts, whose actual condition is usually known in advance through a reference procedure. The evaluation covers, among other things, the agreement of the appraisers with themselves (repeatability), the agreement of the appraisers with each other (reproducibility), and the agreement with the reference value, using cross-tabulations, the kappa coefficient, or hypothesis tests, among other methods. Attribute MSA procedures therefore make a different kind of statement than variable procedures: they evaluate the consistency and hit rate of decisions, not the variation of a numerical value.

How Is MSA Distinguished from Calibration?

Calibration determines and documents the relationship between the reference values of a measurement standard and the indications of a measuring instrument under specified conditions, including the associated measurement uncertainty. It thus describes the behavior of the measuring equipment in relation to a standard at a specific point in time. MSA goes further: it examines the complete measurement process including repeatability, reproducibility, appraiser influence, and – depending on the procedure – linearity and stability over time. A calibrated measuring instrument is therefore not automatically a measuring system suitable for a specific inspection task; conversely, a meaningful MSA usually requires a traceably calibrated reference standard to which the measured values of the system under examination are oriented.

Nor does an MSA replace ongoing inspection equipment monitoring. This ensures organizationally that measuring equipment is calibrated on time, managed, and withdrawn if necessary; MSA supplements this with a statistical suitability assessment of the entire measurement process for a specific characteristic and replaces neither calibration nor inspection equipment monitoring.

What Role Does MSA Play for Process Capability?

Process capability indices such as Cp, Cpk, or Ppk are based on the statistical evaluation of measured values from the manufacturing process. This evaluation is only reliable if the observed variation is predominantly attributable to the process itself and not to the measuring system. An MSA with insufficient %GRR or ndc indicates that a relevant part of the observed variation is caused by the measuring system; in this case, process capability or SPC evaluations based on these measured values cannot be interpreted reliably, regardless of how capable the underlying manufacturing process actually is. In the automotive environment, evidence of a completed MSA is typically part of the initial sample inspection within the PPAP process (element 10). An MSA is usually repeated when the measuring equipment, the inspection personnel, the inspection method, or the requirements for the measured characteristic change.

How Is MSA Anchored in Standards in Quality Management and the Automotive Industry?

ISO 9001 requires suitable measures to ensure valid and reliable measurement results as part of the resources for monitoring and measurement, but does not name MSA as a separate, mandatory procedure. In the automotive industry, on the other hand, measurement system analysis is anchored as one of the so-called Core Tools via IATF 16949: the standard requires statistical studies to evaluate the variation of the measuring and inspection equipment specified in the control plan, provided that numerical results from measurements are used for the characteristics concerned. The methodological reference sources are primarily the AIAG reference manual “Measurement Systems Analysis” and – particularly in the German and European automotive environment – VDA Volume 5, which evaluates the capability of measurement and inspection processes using a risk-based approach to measurement uncertainty and thus differs methodologically from the classic GRR study of the AIAG. Which of the two reference sources is to be applied is usually defined customer-specifically or within the quality management system.

FAQ – Frequently Asked Questions About Measurement System Analysis (MSA)

What is meant by MSA in quality management?

MSA (measurement system analysis) refers to a collection of statistical procedures used to assess whether a measuring system is suitable for inspecting a specific characteristic by breaking down the observed variation into its causes.

What is the difference between MSA and calibration?

Calibration determines and documents the relationship between the indications of a measuring instrument and a reference standard at a specific point in time. MSA additionally examines the entire measurement process, including repeatability, reproducibility, and appraiser influence.

Which variation components are distinguished in MSA?

For variable characteristics, MSA distinguishes repeatability (equipment variation, EV), reproducibility (appraiser variation, AV), and the interaction between appraiser and part (interaction, IA).

What does %GRR mean and which value is considered acceptable?

%GRR relates the combined measuring system variation to the total variation or tolerance. Commonly used guide values in practice are ≤ 10% as capable, 10–30% as conditionally capable, and over 30% as not capable; the binding limit values are usually defined by the quality management system or the customer.

What is the ndc value and why is it relevant?

The ndc (number of distinct categories) indicates how many distinguishable categories a measuring system can resolve within the part variation. An ndc of at least 5 is considered a guide value for a measuring system with sufficient resolution.

Are the acceptance criteria for %GRR identical in AIAG and VDA?

No. AIAG MSA and VDA Volume 5 use partly different indices and acceptance limits to evaluate the suitability of measuring systems; which set of rules with which limit values is to be applied depends on customer- or company-specific requirements.

When is the MSA type 3 study used?

The type 3 study is applied when a measuring system has no or only negligible operator influence, for example automated or integrated measuring systems; the appraiser component of the test setup is omitted accordingly.

How does MSA work for attribute characteristics?

For characteristics that cannot be measured continuously, several appraisers repeatedly evaluate the same parts; the agreement with themselves, the agreement with each other, and the agreement with a reference value are evaluated, among other things using cross-tabulations or the kappa coefficient.

Is an MSA mandatory under ISO 9001?

No. ISO 9001 generally requires suitable resources for monitoring and measurement, but does not prescribe MSA as a separate procedure. In the automotive industry, MSA is anchored as a Core Tool via IATF 16949.

Standards and Norms

• AIAG reference manual “Measurement Systems Analysis” (MSA), 4th edition – aiag.org (description/scope freely accessible, full text for a fee). Together with APQP, PPAP, FMEA, and SPC, the MSA manual is one of the five so-called AIAG Core Tools of the automotive industry.
• VDA Volume 5 “Capability of Measurement Processes”, 3rd edition 2021 – webshop.vda.de (description freely accessible, full text for a fee)
• JCGM 200:2012 – International Vocabulary of Metrology (VIM) – bipm.org (full text freely accessible)
• ISO 9001:2015 – Quality management systems – Requirements – iso.org (scope only freely accessible)
• IATF 16949:2016 including Sanctioned Interpretations and FAQs – iatfglobaloversight.org (Sanctioned Interpretations/FAQs freely accessible, standard text for a fee)
• VDI/VDE/DGQ/DKD 2618 guideline series (inspection equipment monitoring) – vdi.de (title/scope only freely accessible)

Further reading: Quality management, Quality assurance, Statistical process control, ISO 9001, Measurement system analysis, Measurement uncertainty (Wikipedia).

Further Internal Content

• Calibration as a prerequisite for meaningful MSA results
• Process capability and the suitability of the underlying measuring system
• Inspection equipment management in the CAQ system

AI-assisted · Editing and publication: Christian Sedlag · Last updated: 28 October 2026