Gauge R&R tests how consistent your measurement process is. Calibration proves an instrument is accurate and traceable. Learn how they differ and how to use both.
Gauge R&R and calibration answer two different questions, and neither can stand in for the other. Calibration shows whether an instrument reads true against a traceable reference. Gauge R&R shows whether your whole measurement process, including operators, fixtures and method, gives consistent results. A quality system needs both, and the sensible order is to calibrate first so that the study is not run on a biased instrument. For the calibration side, we at PCS Precision provide NATA calibration on site or in our workshop.
Calibration compares an instrument with a reference standard and corrects or documents its bias. According to the American Society for Quality, it establishes accuracy and traceability. In practice you get a certificate that says what the instrument read against a known value, at which points, and with what uncertainty.
Traceability is the part auditors look at first. NATA defines it as a property of a measurement result that links it to a reference through an unbroken, documented chain of calibrations, each adding to the uncertainty. Its General Accreditation Criteria on metrological traceability say facilities must establish that traceability, and that where required it can come from a suitable national metrology institute service or an appropriately accredited calibration laboratory, with supporting evidence.
What calibration does not tell you is how the instrument behaves in your hands. A scale can be dead on at 500 g on the bench of an accredited laboratory and still produce scattered readings on a vibrating production floor with three operators loading it differently.
A gauge repeatability and reproducibility study quantifies the variation that the measurement system adds to your data. The ASQ guide defines the two terms this way:
The result is compared with the total variation you see in the process. Minitab explains that %Study Var sets measurement-system variation against total variation, and that ideally most of the observed variation should come from the parts, not the gauge.
This is why Gauge R&R belongs to measurement system analysis (MSA) rather than to the calibration schedule. It judges a system, not a single device.
| Question | Calibration | Gauge R&R |
|---|---|---|
| What it asks | Does the instrument read true against a reference? | Is the measurement process consistent? |
| What it addresses | Accuracy, bias, traceability | Precision: repeatability and reproducibility |
| Scope | The instrument | Instrument, operators, fixture, method, environment |
| Typical output | Certificate with results and uncertainty | Percentage of variation, number of distinct categories |
| Who usually does it | An accredited laboratory or technician | Your quality team, using production parts |
| When | At purchase, after repair, periodically, when accuracy is in doubt | New gauges or methods, process changes, suspect results |
The split between the two is summarised well in the ASQ material: calibration compares an instrument with a reference to establish accuracy and traceability, while Gauge R&R looks at the system as used. Consequently, neither replaces the other.
The ASQ guide notes that calibration does not establish that operators, fixturing, method or the measurement process are consistent enough. A calibrated instrument can therefore still show excessive Gauge R&R.
Common causes on the shop floor:
The reverse also holds. A process can be tightly repeatable and reproducible while being systematically biased, because nobody has compared the instrument with a reference. Ten operators who all read 0.3 g high will produce a beautiful Gauge R&R result and a bad product. That is why Gauge R&R cannot replace calibration.
Take a checkweigher line filling 500 g packs with a tolerance of plus or minus 5 g.
Now reverse it. The study looks excellent, but a later audit finds the verification scale was never calibrated after a repair. Every result is precise and none can be trusted. The first case needs a process fix. The second needs calibration, then a repeat of the study.
Run the calibration first. A Gauge R&R study on an instrument with unknown bias tells you about its consistency but nothing about whether it is right, and a failed study may simply send you chasing a fault that a calibration or repair would have found immediately.
A practical sequence for a quality manager:
A typical crossed study uses at least 10 parts from a regular production run, three regular operators and three measurements per operator, as set out in Quality-One’s MSA guidance. The study design matters as much as the arithmetic. Rolls-Royce’s Measurement System Analysis How-to Guide recommends collecting measurements in random order and keeping operators from seeing earlier results. Randomising spreads time-related and environmental effects, and blind measurement avoids bias.
Points that most often go wrong:
There are two common calculation methods, average and range, and ANOVA. ANOVA is generally preferred when you want to see operator-by-part interaction, and most statistical software offers it.
The familiar automotive guideline treats measurement-system variation under 10% of total variation as generally acceptable, 10% to 30% as possibly acceptable depending on the application, and over 30% as unacceptable. Minitab, citing AIAG guidance, describes under 10% as acceptable. The number of distinct categories (ndc) is commonly recommended to be five or more.
Treat these as guidelines. Your customer may set its own criteria, and the right threshold depends on the tolerance, the cost of a wrong decision and the purpose of the measurement. A 20% result on a critical aerospace dimension is a different problem from 20% on a non-critical packaging weight.
The split between repeatability and reproducibility tells you where to look, as described in Quality-One’s guidance:
After any improvement, repeat the study to confirm it worked. If the fix involved repair or adjustment, calibrate again before you repeat it. Our note on calibration versus adjustment explains why the as-found and as-left records both matter at audit.
Australian rules speak firmly about calibration and traceability, and say little about Gauge R&R thresholds. Where you work determines what applies.
If your work falls under trade measurement, see our trade measurement page, since verification of trading instruments is a separate legal process from either calibration or Gauge R&R.
For customers in regulated sectors, a calibration certificate is the document that auditors ask for. A Gauge R&R report supports your measurement-system claims, but it sits on top of valid calibration, not in place of it. If you are unsure whether a certificate is traceable or accredited, our guide to verifying traceable versus accredited calibration lists the documents to check, and our article on calibration uncertainty shows how to read the uncertainty on a certificate.
Most of the measurement problems we are asked about sit between the two ideas. A quality manager has a Gauge R&R result they do not like, or a certificate they are not sure how to read, and the two are treated as rivals. They are not. The certificate tells you the instrument can be trusted to read true. The study tells you the process can be trusted to read the same way twice.
Our position is simple. Start with a calibration that holds up at audit: accredited, traceable, covering the range you actually use. We are NATA accredited and ISO 9001 certified, and we calibrate on site or in our workshop, across multiple equipment brands regardless of where the equipment was bought. Then run your study on that foundation, so that what it shows is the method, the fixture and the people, not a hidden offset.
Calibration is also the cheapest way to remove a variable. When a study fails, knowing the instrument is already verified saves days of guessing. That is why we encourage proactive calibration schedules instead of waiting for a failed audit or an unplanned stoppage.
Kaz
If your next Gauge R&R study is coming up, the calibration behind it should be current and traceable first. We offer NATA calibration and traceable calibration for precision instruments, together with preventative maintenance, repairs, trade verification and hire equipment, and we back this with 24/7 emergency support. If a study points to an instrument problem, the same team can repair or replace it, and you can browse balances, scales and measuring equipment in our range. Tell us the instruments, the ranges and the tolerances you work to, and we will tailor the service to your business.
No. Calibration compares an instrument with a reference standard to establish accuracy and traceability. Gauge R&R measures how consistent the whole measurement process is across repeated measurements and different operators. A quality system uses both.
Yes, it is the sensible order. A study on an instrument with unknown bias can show good consistency while every reading is wrong. Confirming a current, traceable calibration first means that any problem the study finds points to the method, fixture or operators.
Under 10% of total variation is generally treated as acceptable under AIAG guidance, 10% to 30% may be acceptable depending on the application, and over 30% is generally unacceptable. An ndc of five or more is commonly recommended. Your customer or your own risk assessment may require stricter limits.
It depends on the instrument, its use and its history. CASA’s guidance for aviation maintenance and production suggests an initial maximum of 12 months where no manufacturer interval exists, then review. Recalibrate after any repair, and whenever you doubt the accuracy.
Look at the split between repeatability and reproducibility. High repeatability points to the gauge, its condition or its resolution. High reproducibility points to operator training, the method or a fixture. Make the change, then repeat the study.
