A dimensional report can look precise while the measurement system adds enough variation to change the decision. On an injection molded part, the result may depend on the instrument, fixture, datum setup, operator, software, temperature, part condition and the feature being measured. Measurement System Analysis (MSA) gives buyers a structured way to ask whether the reported number is fit for its intended decision. It does not make a part conforming, replace drawing interpretation or prove process capability. It tests the evidence used to make those judgments.
For a sourcing or quality team, the practical question is not simply “Did the supplier perform gauge R&R?” It is “Can this measurement method distinguish an acceptable part from a nonconforming one at the feature and tolerance that matter?” Before comparing supplier reports, align the drawing revision, datums, measurement method, sample condition and decision rule. InjeForm’s [request-a-quote](/request-a-quote) workflow can help buyers put those evidence requirements into the RFQ.
Do not approve a dimensional result from a gauge study that does not represent the actual feature, operators, fixture, environment and decision being made.
01
1. Define the dimensional decision before requesting an MSA study
Start with the decision the measurement will support: tool correction, first-article approval, process release, incoming acceptance, sorting, capability analysis or a supplier corrective action. Identify the characteristic, nominal, tolerance, datum scheme, functional consequence and decision boundary. A broad MSA claim cannot answer a feature-specific question if the study used a different size range, orientation or setup.
Separate product variation from measurement variation in the request. Ask the supplier to identify whether the study concerns a digital height gauge, CMM, vision system, pin gauge, comparator, functional fixture or attribute inspection. The method should reflect how the part is actually checked in production. InjeForm’s [quality](/quality) and [CMM inspection](/injection-molding-cmm-inspection-gdt-buyer-guide) resources can help frame the evidence request, while the released drawing and quality agreement control acceptance.
| Buyer decision | MSA question | Evidence to retain |
|---|---|---|
| Tool or process correction | Can the system detect a meaningful shift at the CTQ? | Study design, raw readings, feature setup and action threshold. |
| First-article approval | Is the reported result dependable for this part and datum scheme? | Method, calibration status, report revision and sample identity. |
| Incoming acceptance | Can different inspectors or sites reach comparable decisions? | Operator plan, reproducibility result and boundary samples. |
| Capability analysis | Is measurement variation small enough for the intended capability interpretation? | MSA result linked to the same process and characteristic. |
| Supplier dispute | Could method or setup variation explain conflicting results? | Cross-check plan, retained parts and agreed referee method. |

02
2. Read repeatability and reproducibility as different risks
Repeatability is the variation when the same operator measures the same characteristic with the same method and equipment. It can expose instrument resolution, fixture instability, probing technique, software settings or part positioning problems. Reproducibility concerns differences between operators or other defined conditions in the measurement process. NIST’s gauge R&R guidance treats these as separate sources to characterize, alongside stability, resolution and bias considerations [1].
A high repeatability component suggests the method itself is noisy even when one person repeats it. A high reproducibility component suggests the work instruction, datum setup, fixturing or operator training is not sufficiently controlled. Do not treat a combined GR&R result as a universal pass or fail. Ask how the variance is split, whether the operators represent the intended production users and whether the parts span the actual process range rather than being convenient, visually similar samples.
For a molded feature, the study should also state part conditioning and handling. Flexible walls, snap features, residual stress, temperature and clamping force can change the measured result. A CMM can be highly repeatable for a rigid datum setup yet unsuitable if the fixture distorts a thin wall. The measurement method and restraint condition belong in the evidence package.
03
3. Check study design, bias and stability—not only the headline GR&R
Request the study design before accepting its conclusion. It should identify the number and selection of parts, operators, trials, equipment, feature, unit, environment, datum setup and randomization or measurement order. AIAG describes MSA as a decision-oriented discipline covering study planning, GR&R, attribute studies and interpretation rather than a single spreadsheet percentage [2]. A study with too little part spread can make the measurement system look large relative to the observed process, while a narrow tolerance can make the same method consequential for release.
Ask how calibration, resolution, bias, linearity and stability were considered. NIST notes that a gauge study can support uncertainty assessment only when it represents the working environment and that changes over time can make a short-term standard deviation misleading [3]. For injection molding, repeat the check after a significant tool, fixture, software, operator or inspection-location change. Consider a reference artifact or retained sample to detect drift between production lots.
If the supplier reports a favorable GR&R but the measurement method has a systematic offset, the result may still misclassify parts. Ask for comparison to a suitable reference, the calibration trace or verification record, and any correction applied. Avoid asking the molder to promise a universal acceptance percentage; the product risk, tolerance, measurement technology and customer standard should determine the decision rule.

04
4. Convert MSA findings into a release and supplier-control decision
Use the study to choose an action, not to decorate a quality file. If the method is adequate, record the approved equipment, fixture, work instruction, operators, environmental conditions, sampling frequency and change triggers. If it is marginal, improve the method before using the data for capability or acceptance. Possible actions include a better fixture, clearer datum simulation, higher-resolution equipment, operator training, environmental control, a defined conditioning period or an agreed referee method.
When a system is not adequate for a critical feature, do not automatically reject every part or waive the requirement. First determine whether the result can support a limited decision, such as screening well inside the specification, while a more capable method is implemented. Any temporary arrangement should state its scope, expiry, responsible approver and disposition of borderline parts. The buyer owns the product decision; the supplier should provide truthful evidence and disclose method limitations.
Put the control into the purchase and quality records. Specify the drawing revision, CTQ list, measurement method, sample size, report format, calibration expectations, retained data, change notification and dispute-resolution method. Link the MSA requirement to the [injection molding quality control](/injection-molding-quality-control-inspection-plan) plan and the [FAI approval](/injection-molding-fai-inspection-buyer-approval-guide) gate instead of treating it as an isolated certificate.
| MSA outcome | Buyer response | Procurement evidence |
|---|---|---|
| Adequate for intended use | Approve the method for the named feature and decision. | Signed study, method sheet, calibration and revision-controlled report. |
| Repeatability concern | Improve equipment, resolution, fixture or technique; re-study. | Root cause, corrective action and new study date. |
| Reproducibility concern | Standardize setup, work instruction and operator training; re-study. | Operator results, training record and controlled method. |
| Bias, drift or environment concern | Control conditions, verify against a reference and define monitoring. | Verification records, stability checks and escalation rule. |
| Not adequate for release | Use a more capable method or hold the decision; do not overstate certainty. | Deviation approval, boundary samples and replacement plan. |
05
5. Compare supplier dimensional evidence on a common MSA basis
When comparing two suppliers, normalize more than the reported result. Ask whether both measured the same part revision, cavity, resin lot, conditioning state, datum scheme, feature location and tolerance. Check whether one supplier used a production operator and the other used a metrology specialist, or whether one measured a free-state part and the other used a constrained fixture. A different method may be valid, but it is not automatically comparable.
Build an evidence scorecard that records the study scope, variance components, calibration and stability status, raw or summarized data, report date, open actions and the person who approved the conclusion. For multi-cavity tools, identify cavity and part position; for a family of CTQs, do not assume one feature’s study covers every feature. Ask the supplier to flag characteristics that need a separate method or functional test.
A useful RFQ asks for the proposed measurement approach before production, then makes final approval conditional on representative samples and agreed evidence. InjeForm’s [DFM and tool design](/injection-mold-design-dfm) discussion can help connect datum strategy, shrinkage and feature access to inspection planning. For a quote or audit, request the method and decision boundary—not just the phrase “MSA compliant.”

Practical takeaway
Before you move
to the next review.
- Name the CTQ, tolerance, datums, functional risk and release decision before asking for MSA evidence.
- Require the study to identify representative parts, operators, equipment, fixture, environment, trials and measurement order.
- Review repeatability, reproducibility, resolution, bias, stability and uncertainty context instead of relying on one GR&R percentage.
- Define the action for an inadequate system, including method improvement, re-study, temporary containment and approval authority.
- Compare supplier reports on the same revision, cavity, resin condition, datum setup, feature and decision rule, then retain the evidence.
FAQ
Common questions
What is MSA for injection molded parts?
MSA evaluates whether the measurement process is suitable for the decision being made. It examines sources such as repeatability, reproducibility, resolution, bias and stability for a defined feature and method.
Is gauge R&R the same as MSA?
No. Gauge R&R is one important MSA study focused on repeatability and reproducibility. MSA can also consider bias, linearity, stability, resolution, attribute agreement and uncertainty-related factors.
What does high repeatability mean in a molded-part inspection?
It means repeated measurements under the same defined conditions vary more than desired. Investigate instrument resolution, fixture or datum stability, probing technique, part handling and software before using the method for release.
Can a supplier pass GR&R and still report the wrong dimension?
Yes. A study can show acceptable short-term precision while a systematic bias, unsuitable datum setup, drift, conditioning effect or nonrepresentative study design remains. Ask how those risks were checked for the actual feature.
Should buyers require one universal GR&R threshold?
Not as a substitute for engineering judgment. The appropriate decision depends on tolerance, product risk, measurement technology, customer requirements and intended use. Set the rule in the quality agreement and explain how borderline results are handled.
References
Source material
- [1] U.S. National Institute of Standards and Technology, “Gauge R & R studies.”
- [2] Automotive Industry Action Group, “Measurement Systems Analysis (MSA) with Applications.”
- [3] U.S. National Institute of Standards and Technology, “Quantifying uncertainties from a gauge study.”
- [4] ISO, “ISO 9001 — Quality management systems.”
- [5] ISO, “ISO 20457:2018 Plastics moulded parts — Tolerances and acceptance conditions.”
This guide is general information for early procurement and engineering discussion. It does not replace product-specific DFM, qualification, validation, regulatory assessment or legal advice.