We are a data-driven civilization. Every day, we constantly use data in business on a wide range of aspects. A well-thought-out and implemented Measurement System Analysis (MSA) may help lay a substantial basis for any data-driven decision-making process.
For those who are in quality management, it is important to understand the basics of measurement system analysis as it is a key element in six sigma.
What is Measurement System Analysis (MSA)
A measurement systems analysis is an extensive examination of a process that usually comprises a carefully constructed experiment to discover the origins of variance in the measurement process. MSA analysis is a mathematical and experimental means to estimate the level of variance present in a measurement procedure. Measurement System Analysis is used to validate a measuring system's accuracy, precision, and stability before being employed.
The following factors are considered in a measuring systems analysis:
- Choosing the right measurement and strategy.
- Examining the measurement instrument.
- Examining processes and personnel.
- Considering any measurement interactions.
- Calculating individual measurement device and measurement system measurement uncertainty.
Why Measurement System Analysis is Necessary?
A suitable Measurement System Analysis process can ensure that the data being collected is correct and that the data collection method is appropriate for the process. Appropriate, trustworthy data may help you avoid wasting time, labor, and scrap in a manufacturing process. Faulty components can be accepted, and perfect parts might be rejected because of an inadequate measuring method, resulting in unsatisfied customers and excessive scrap.
Fundamentals of Measurement Systems Analysis
- Determine how many assessors, how many sample parts, and how many repeat readings there will be. Larger numbers of pieces and repeat measurements produce more confident answers, but the benefits must be weighed against the time, expense, and disturbance required.
- Use assessors who are knowledgeable about the methods and who make the measurements regularly.
- Determine that all assessors adhere to a standard, documented measuring technique.
- To depict the complete process spread, choose the sample components. This is a really important aspect. Measurement error may be exaggerated if the process spread is not properly reflected.
- As described in the Requirements section, ensure that the measuring instrument has enough discrimination/resolution.
- The assessors should not know the number assigned to each part or any prior measurement value since the parts should be numbered and measurements were done in random order. A third party should keep a table with the dimensions, appraiser, trial number, and number for each part.
Related Blogs: Six Sigma Tools
Types of Measurement System Analysis
MSA types are generally grouped based on the nature of the data being measured — whether it's continuous, categorical, destructive, or automated.
1. Variable MSA
Used for continuous, numeric data such as length, weight, or temperature. This type commonly relies on Gage R&R studies to assess how accurate and consistent the measurements are.
2. Attribute MSA
Applies to binary or categorical data — think pass/fail or yes/no type outcomes. Here, Attribute Agreement Analysis is the standard method used to evaluate how consistent decisions are across inspectors.
3. Destructive Testing MSA
Used in situations where the part being measured is damaged or destroyed in the process, such as crash testing. Since repeated measurements on the same item aren't possible in these cases, this type of MSA requires a different approach than standard repeatability testing.
4. Automated System MSA
Focuses on checking the performance of automated measurement systems — sensors, machines, or other automated tools — ensuring their readings stay consistent and function as intended without human variability in the loop.
Attribute Agreement Analysis
Attribute Agreement Analysis (AAA) is a statistical technique used to evaluate how consistently different appraisers — or a single appraiser over time — classify categorical (attribute) data, such as Pass/Fail, Good/Bad, or Approved/Denied decisions. It serves as a measurement system analysis technique for discrete data, ensuring that inspectors or evaluators categorize items correctly against a known standard and against each other.
What It Actually Measures
AAA evaluates agreement across three distinct dimensions:
Inter-Appraiser Agreement – whether different appraisers reach the same conclusion when evaluating the same item. High agreement here shows the measurement system produces consistent results regardless of who performs the evaluation, while disagreement often points to gaps in training or unclear standards.
Intra-Appraiser Agreement (Repeatability) – whether the same appraiser reaches the same decision when evaluating the identical item a second time.
Agreement with a Known Standard – how accurately appraisers' decisions match a correct or reference result. These reference values are typically established through expert evaluation, consensus among master appraisers, or more precise measurement methods.
Identify Measurement System Errors
The data might either be accurate or contain system faults when doing an MSA analysis. These measuring system flaws are identified by Measurement System Analysis and are characterized by precision and accuracy.
- Precision: Precision refers to how near the sample data points are to one another.
- Accuracy: The accuracy of the sample data points to the goal value.
System Errors Classifications
Precision is divided into two categories: repeatability and reproducibility. To assess the combined measurement of repeatability and reproducibility, a gauge R&R research is employed.
Accuracy can also be divided into three categories: linearity, stability, and bias. To establish if measuring instruments are taking correct measurements, gauge linearity and bias analysis are performed.
PRECISION
- Repeatability: The capacity to obtain the same results each time the same operator performs the same measurement. This gives you information about the equipment's variability.
- Reproducibility: Reproducibility is the efficiency of one operator to produce the same outcomes as another. This information is useful in determining operator variability.
ACCURACY
- Linearity is a term that describes how a measurement device's accuracy changes across its operational range. Is there a difference in precision when measuring a component that weighs 5 lbs. vs 30 lbs.?
- Stability refers to the constancy with which the study is conducted throughout time. Is accuracy varying because of the operator employing different strategies to collect the sample today than one month ago?
- Bias is a term that describes the disparities between a sample data set's average and the actual value. For example, if a thermometer reads 72 degrees outside but the temperature is 70 degrees, the thermometer has a +2 degree bias since it reads higher than the actual temperature.
The Procedure of MSA: Gage R&R Study
Gage Repeatability and reproducibility (Gage R & R) may assess the number of uncertainty in a measuring system for gauges or devices that gather varied continuous data. To begin a Gage R & R, choose the gage to be assessed.
Then take the following actions:
- At least 10 random samples of components made during a standard production run should be taken.
- Choose three operators who execute the inspection regularly.
- Measure the sample pieces and record the results for each operator.
- Repeat the measuring method three times with the same parts for each operator.
- Calculate the average (mean) readings for each of the operators and the range of the trial averages.
- Calculate the difference between the averages of each operator, the average range, and the measurement range for each sample component utilized in the research.
- Determine the level of equipment variation by calculating repeatability.
- Calculate repeatability to determine how much variance the operators introduce.
- Calculate the percentages of variance in the components and the overall variation.
How to Interpret Gage R&R Results
Gage R&R results are typically expressed as a percentage of total process variation, and the standard AIAG benchmark is straightforward:
- Below 10% – The measurement system is acceptable; most of the variation comes from the parts themselves, not the measurement system.
- 10–30% – The system may be accepted, but there should be a plan to review and improve it, depending on the application's criticality, measurement cost, and process capability.
- Above 30% – The system needs improvement, since appraisers and equipment are contributing to more than 30% of the total variation.
Gage R&R Acceptance Criteria
A Gage R&R study is judged against two AIAG MSA 4th Edition benchmarks: %GRR and the number of distinct categories (NDC). Both must be met — passing one while failing the other still means the measurement system doesn't qualify.
%GRR | Classification | What It Means |
Below 10% | Acceptable | The measurement system is contributing minimal noise and is accepted without condition |
10% to 30% | Conditional | May be acceptable depending on the characteristic's criticality, process capability, and cost of improving the gauge — requires documented evaluation |
Above 30% | Not Acceptable | The system should be rejected or improved before use, unless customer engineering sign-off is obtained |
Number of Distinct Categories (NDC)
The number of distinct categories represents the number of non-overlapping confidence intervals that span the range of product variation — essentially, the number of groups within your process data that a measurement system can actually discern.
The Formula
NDC = 1.41 × (PV / GRR), where PV is the part variation from the study and GRR is the calculated Gage R&R result. The resulting value is truncated to a whole number — for example, a calculated value of 15.8 gets truncated to 15, not rounded up, since the system isn't quite capable of distinguishing 16 categories.
What Is the Significance of Measurement System Analysis?
Faulty measuring approaches might allow low-quality units to sneak through the gaps while rejecting high-quality parts. Manufacturers may assure that their measuring equipment and operation comply with their precision and accuracy standards by performing an MSA analysis.
As previously stated, the Measurement System Analysis is an important aspect of six sigma. It offers a firm foundation for any claims made during the system analysis. The data gathered could not be trusted without the presence of a trustworthy measuring system. This would make it difficult to accept or reject any hypothesis presented during the defined phase of the six sigma.
Read More: Lean Principles
Conclusion-
Now that you have a good idea of what is Measurement System Analysis and why is it considered such an important thing in the six sigma methodology. If you're keen to learn more about the six sigma methodology, we provide a full-fledged six sigma green belt certification that will teach you everything in and out about quality management and will polish you into an expert.
FAQs
1. When should an MSA study be conducted?
MSA should be conducted when new measuring equipment or software is introduced, when changes are made to existing measurement systems, when new operators or inspectors begin using the equipment, if a quality issue is suspected to be caused by measurement error, or when there's a need to validate data accuracy before a process improvement project like Six Sigma or Lean.
2. What are the acceptance criteria for a Gage R&R study?
A Gage R&R study is judged against two thresholds: %GRR below 10% is acceptable, 10–30% is conditional (requiring documented evaluation of the characteristic's criticality and cost of improvement), and above 30% is not acceptable. Alongside %GRR, the number of distinct categories (NDC) must be at least 5 — both criteria need to be met together.
3. How is the percentage of Gage R&R variation interpreted?
%GRR shows how much of the total observed variation comes from the measurement system itself rather than actual differences between parts. A low %GRR means the gauge and operators are adding minimal noise, so most of the variation genuinely reflects real part-to-part differences — which is what you want. A high %GRR means the measurement system is masking real process variation, making it harder to tell good parts from bad ones.
4. What is the difference between MSA and calibration?
Calibration checks whether a single instrument reads accurately against a known reference standard — it corrects bias in the tool itself. MSA is broader: it evaluates the entire measurement process, including the instrument, the operator, the environment, and the procedure, to determine whether the overall system produces consistent and reliable results. A gauge can be perfectly calibrated and still fail an MSA study if operators use it inconsistently.
5. How do you select parts for a Gage R&R study?
Parts should be chosen to represent the full range of variation expected in actual production — typically 10 or more parts, including a mix of sizes or types. Selecting parts that are too similar to each other is one of the most common causes of a misleadingly high %GRR or low NDC, since it narrows the part-to-part variation the study depends on.
6. What should you do if a Gage R&R study fails?
First, check whether the study design itself was the problem — parts pulled from a single narrow batch will inflate %GRR artificially. If the design was sound, look at whether reproducibility (operator variation) or repeatability (equipment variation) is the bigger contributor: operator issues call for retraining and standardized procedures, while equipment issues may require gauge repair, recalibration, or a higher-resolution instrument.
7. Can MSA be performed for attribute measurement systems?
Yes — this is called Attribute MSA, and it uses Attribute Agreement Analysis rather than Gage R&R. It's designed for categorical or binary data, such as pass/fail or go/no-go decisions, and evaluates how consistently different appraisers (or the same appraiser over time) classify the same items.
8. What is the difference between variable MSA and attribute MSA?
Variable MSA applies to continuous, numeric data (like length, weight, or temperature) and typically uses Gage R&R studies. Attribute MSA applies to categorical or binary data (like pass/fail) and uses Attribute Agreement Analysis instead, since there's no continuous scale to measure repeatability and reproducibility against.
9. How often should a measurement system be evaluated?
There's no single fixed interval — MSA is typically re-evaluated during routine equipment calibration cycles and quality audits, whenever new equipment or operators are introduced, or when measurement results start looking inconsistent despite the process itself being in control. Many organizations tie MSA reviews to their broader calibration schedule so both stay aligned.
10. What is the role of MSA in improving manufacturing process capability?
Process capability metrics like Cpk and Ppk are only meaningful if the underlying measurement data is trustworthy. MSA verifies that the data feeding into capability studies isn't distorted by measurement error, so any variation identified truly reflects the process itself — not noise from the gauge or the operator. Without a validated measurement system, capability analysis and control charts can lead teams to reject good parts or accept bad ones.
