Key takeaways
Short answer: Repeatability and reproducibility are the two components of measurement variation, and a Gage R&R study separates them. Repeatability is how much the same operator, using the same gauge, varies when measuring the same part repeatedly — the within-operator, equipment side. Reproducibility is how much different operators (or instruments) vary when measuring the same part — the between-operator side. One points at the gauge and method, the other at people and technique. For why a consistent reading is not enough, see precision vs accuracy.
Repeatability is the variation you get when one operator measures the same characteristic on the same part, with the same gauge, several times. Ideally the readings would be identical; in reality they scatter a little, and that scatter is repeatability — sometimes called equipment variation or within-operator variation, because the operator and method are held constant and only the gauge and the act of measuring vary. Good repeatability means the measurement system gives nearly the same answer when nothing has changed. Poor repeatability — a single operator getting noticeably different numbers on identical remeasurements — points at the gauge itself or the measuring method: wear, play, an unstable fixture, or an ambiguous technique.
Reproducibility is the variation you get when different operators (or different instruments, or different labs) measure the same characteristic on the same part. It isolates the between-operator component: when everything except the person changes, how much does the answer move? Reproducibility captures the human and procedural differences — how each operator seats the part, reads the scale, interprets the procedure, applies pressure. Good reproducibility means different people get the same result; poor reproducibility means the answer depends on who is holding the gauge, which is a training, technique, or procedure problem rather than a hardware one. It is the difference between operators, not the difference within one.
The two are assessed together in a Gage R&R (repeatability and reproducibility) study, a standard tool of measurement-system analysis. Several operators each measure several parts several times, and the analysis partitions the total measurement variation into its repeatability and reproducibility components — and compares that combined measurement variation against the actual part-to-part variation or the tolerance. The output tells you how much of the variation you see in your data is real product variation versus noise from the measurement system itself. If the measurement system eats up too much of the tolerance, your quality data is partly fiction, and no amount of process control will fix a process you cannot measure trustworthily.
Three operators each measure ten parts twice with the same bore gauge. The analysis shows repeatability is excellent — each operator's two readings on a given part agree closely — but reproducibility is poor: operator A reads consistently larger than operators B and C. The gauge is fine (good repeatability); the problem is technique (poor reproducibility) — operator A seats the gauge differently. The fix is not a new gauge but training and a clearer procedure to standardise how all three seat and read it. Had it been the reverse — each operator disagreeing with their own remeasurements — the gauge or method would be the culprit. Separating the two components pointed straight at the real cause.
The diagnosis drives the action. Poor repeatability — an operator disagreeing with their own remeasurements — points to the equipment or method: replace or service the gauge, stabilise the fixture, remove ambiguity from the technique, or pick an instrument with adequate resolution for the tolerance. Poor reproducibility — operators disagreeing with each other — points to people and procedure: standardise the method, train to it, and remove the room for interpretation that lets different operators do it differently. Spending on a new gauge when the problem is reproducibility wastes money and changes nothing, and retraining operators when the gauge itself is worn will not help either. Identify the component first.
Repeatability and reproducibility sit underneath the quality factor of OEE, just like precision and accuracy. The quality factor counts good versus defective units, and that count is only as trustworthy as the measurement system making the call. If your Gage R&R is poor, the system is partly sorting parts at random — passing some defectives and scrapping some good units — so the OEE quality number, and the yield and scrap behind it, become unreliable. Trustworthy measurement is the unglamorous prerequisite for trustworthy OEE; without it you are optimising against noise.
Fabrico consumes the good and reject counts your measurement system produces, so the integrity of that system flows straight into the reliability of its OEE. By trending quality results over time, it can help surface the symptoms of a measurement problem — reject rates that shift with operator or shift rather than with the process — prompting a Gage R&R before you chase a process ghost. Reliable measurement upstream and honest OEE downstream go hand in hand. Book a demo to see how trustworthy quality data drives trustworthy OEE.
Repeatability is the variation when the same operator measures the same part with the same gauge repeatedly (within-operator, equipment variation). Reproducibility is the variation when different operators or devices measure the same part (between-operator variation). Together they form measurement-system variation.
A Gage R&R (repeatability and reproducibility) study is a measurement-system analysis in which several operators measure several parts multiple times. It partitions the measurement variation into repeatability and reproducibility and compares it against the tolerance or part variation.
Poor repeatability — one operator disagreeing with their own remeasurements — points to the gauge or method: wear, play, an unstable fixture, inadequate resolution, or an ambiguous technique. It is an equipment or method problem, not a people problem.
Poor reproducibility — different operators disagreeing with each other — points to people and procedure: differences in how each seats the part, reads the scale, or interprets the method. The fix is standardising the procedure and training to it.
The quality factor of OEE counts good versus defective units, and that count depends on the measurement system. Poor repeatability or reproducibility means the system partly sorts parts at random, making the OEE quality number and the yield and scrap behind it unreliable.
Programați o întâlnire individuală cu experții noștri sau înscrieți-vă direct în planul nostru gratuit.
Nu este nevoie de card de credit!
Login
Blog
