Direct comparison
Repeatability vs Reproducibility in Precision
Repeatability and reproducibility are sets of conditions, not methods. What to vary, why sR is never smaller than sr, and how to use the limits r and R.
Written and maintained by CASRAI Editorial Board
Last updated
Ask CASRAI · included with Regulatory Radar
Ask about Repeatability vs Reproducibility in Precision
Ask CASRAI answers research-administration questions and cites the passages behind every claim — and says so when the corpus does not cover something, instead of guessing. It comes with a Regulatory Radar subscription at $29 a month, alongside the daily digest of regulatory changes and the dashboard of what changed.
150 questions a day, on this site, over the API, or inside your own tools through the CASRAI MCP server.
Everything CASRAI publishes — this page, the dictionary, the guides and the news — stays free to read, with no account and no card.
How do Repeatability, Intermediate precision, Reproducibility compare side by side?
The table below compares Repeatability, Intermediate precision, Reproducibility across 9 procurement-relevant dimensions, from what is held constant through when regulators expect it.
Side-by-side comparison
| Dimension | Repeatability | Intermediate precision | Reproducibility |
|---|---|---|---|
| What is held constant | Same procedure, same operator, same equipment under the same conditions, same location, short interval of time (ISO 5725-1:2023, 3.14). | Same measurement procedure and same location; the laboratory and the method stay fixed (VIM 2.22). | The test item and, under ISO 5725, the measurement method. Nothing else about the setting is fixed. |
| What is deliberately varied | Nothing except the act of measuring. Replicates must still run the whole procedure, including sample preparation. | Days, analysts, instruments, calibrations, calibrators, environmental conditions — over an extended period, inside one lab. | Laboratory, operator and equipment together. Under VIM the measuring systems may even use different measurement procedures. |
| How the study is normally run | One analyst, one instrument, one short session. Eurachem suggests 6–15 replicates per material; ICH Q2(R2) asks for a minimum of 9 determinations (3 concentrations x 3 replicates) or 6 at 100% of test concentration. | A nested or run-structured design analysed by ANOVA — for example CLSI EP05 single-site: 20 days, 2 runs per day, 2 replicates per run. | An inter-laboratory (collaborative) trial in which many laboratories analyse the same materials using the same written method. |
| Statistic reported | Repeatability standard deviation sr (or RSDr); repeatability limit r. | Intermediate precision standard deviation sI, combining within-run and between-run components. | Reproducibility standard deviation sR (or RSDR); reproducibility limit R. |
| Expected relative size | The smallest of the three. It is the floor of the method’s random variation. | Always at least sr, because it adds a between-run variance component on top of the repeatability variance. | The largest. AOAC reports that within-laboratory RSDs are typically one-half to two-thirds of among-laboratory RSDs. |
| What the number entitles you to claim | How closely one analyst can agree with themselves in one sitting. Says nothing about tomorrow or about another site. | How closely your laboratory can agree with itself across routine day-to-day operation. This is the number that predicts your own QC chart. | How closely independent laboratories running the published method will agree — the basis of a standard method’s stated precision. |
| Where the definition comes from | ISO 5725-1:2023 3.13–3.17; VIM 2.20–2.21; ICH Q2(R2) glossary (‘intra-assay precision’). | VIM 2.22–2.23; ICH Q2(R2) section 3.3.2.2. ISO 5725-1 itself does not define it as a headline term; it is developed in ISO 5725-3. | ISO 5725-1:2023 3.18–3.21; VIM 2.24–2.25; ICH Q2(R2) section 3.3.2.3. |
| Most common mis-report | Injecting the same prepared vial six times and calling it repeatability. Instrument precision is not method repeatability — the replicates were not independent. | Reporting it as ‘reproducibility’ because two analysts were involved. Two analysts in one building is still intra-laboratory. | Claiming it from a single-site study, or importing an automotive gauge R&R result, where ‘reproducibility’ means operator-to-operator variation on one gauge in one room. |
| When regulators expect it | Always, for any quantitative analytical procedure being validated. | Whenever the procedure will be used routinely; ICH Q2(R2) says the extent depends on the intended conditions of use and encourages design of experiments. | ICH Q2(R2): usually not required for a regulatory submission, but expected for standardisation — pharmacopoeial inclusion, or a procedure run at multiple sites. |
Common questions
Common questions about Repeatability vs Intermediate precision vs Reproducibility
Can reproducibility ever be smaller than repeatability?
+
No, not as a property of the method. Reproducibility variance is the repeatability variance plus a between-laboratory component, so sR is bounded below by sr. If your arithmetic says otherwise, the between-laboratory component was estimated with too few degrees of freedom, an outlying laboratory was removed asymmetrically, or the two numbers came from different concentration levels.
Is intermediate precision just a weaker form of reproducibility?
+
It is a different condition, not a weaker one. Intermediate precision keeps the location and the measurement procedure fixed and varies time, analysts and instruments. Reproducibility changes the laboratory itself. Eurachem notes that intermediate precision is sometimes called ‘within-laboratory reproducibility’, which is exactly why the mis-labelling is so common.
What is the difference between sr and the repeatability limit r?
+
sr is a standard deviation describing the spread of single results. r is a critical difference: the value the absolute difference between two results obtained under repeatability conditions should not exceed, with 95% probability. Eurachem gives r = t x sqrt(2) x sr, approximated as 2.8 sr when the degrees of freedom are large. Using r as a tolerance for one result rather than for a pair is a routine error.
Does a gauge R&R study give me ISO 5725 reproducibility?
+
No. In measurement systems analysis the R&R decomposition splits gauge variation into equipment variation (repeatability) and appraiser variation (reproducibility) — all inside one facility. That maps onto intermediate precision, not onto ISO 5725 reproducibility, which requires different laboratories. Acceptance bands for %GRR are set by the AIAG MSA reference manual and by customer requirements rather than by ISO 5725.
How many replicates does a precision study need?
+
Enough for a usable number of degrees of freedom, which matters more than the raw count. Eurachem suggests typically 6–15 replicates per material and notes that 8 groups of 2 replicates yields 8 and 7 degrees of freedom for the within- and between-run estimates respectively. ICH Q2(R2) sets a floor of 9 determinations for repeatability, or 6 at 100% of the test concentration.
Do I need one precision figure or several?
+
Several. Eurachem is explicit that precision generally depends on analyte concentration and should be determined at a number of levels across the range of interest, including any regulatory limit and the ends of the measuring interval. Where precision scales with concentration, report it as a relative standard deviation.
