Relative Standard Deviation Calculator
Relative standard deviation turns a pile of repeat measurements into one number that says how precise they are. Paste your data into the calculator below and it reports the mean, standard deviation, variance, the RSD as a percentage, and a plain-language precision rating, with an optional second data set for a head-to-head precision comparison.
Quick Answer
The ten measurements pre-filled in the tool (12.5, 13.2, 11.8, 12.9, 13.1, 12.3, 12.7, 12.0, 13.0, 12.6) have mean 12.61, sample standard deviation 0.4677, and RSD 3.71%, which rates as very good precision. Dataset B (12.5, 13.9, 11.2, 14.0, 12.1) scores 9.43%, so the tool declares Dataset A more precise by 5.72 percentage points.
Key takeaways
- RSD is the standard deviation as a percentage of the mean: one unit-free number that lets you compare precision across data sets with different means and units.
- Sample versus population changes the denominator (n-1 versus n), so it changes the answer. Pick the one that matches your data.
- The precision bands are general comparison guidelines, not laboratory acceptance criteria. Regulated labs validate their own stricter limits, for example bioanalytical method validation commonly requires precision within 15% RSD.
- RSD breaks down when the mean is near zero, because a tiny denominator inflates the percentage. Compare the plain standard deviation instead in that case.
3.71%
RSD of the worked example, rated very good precision
12.61
Mean of the ten measurements
0.4677
Sample standard deviation (divides by n-1)
- Quick Answer
- How to use the relative standard deviation calculator
- The formula: RSD is the standard deviation as a percentage of the mean
- Worked example: 10 measurements checked against the tool
- Compare two data sets head to head
- How to read the precision rating
- When RSD misleads
- What competing RSD calculators miss
- Frequently asked questions
- Add This Calculator to Your Website
- References
How to use the relative standard deviation calculator
Paste your measurements, pick the right calculation type, and click Calculate. Results stay hidden until you do, and Reset restores the worked example.
- Paste your data values into the first box, separated by commas, spaces, or new lines. You need at least two numbers; anything that is not a number stops the calculation with an error.
- Add Dataset B if you want a comparison. Paste a second set of measurements and the tool reports its mean, standard deviation, and RSD beside Dataset A, then declares which set is more precise and by how many percentage points. Leave it blank to skip this.
- Choose sample or population. Sample (the default) divides the squared deviations by n-1 and is the right choice when your data is a sample from a larger population, which is the usual case in experiments and quality control. Population divides by n, only when your data set is the entire population.
- Set the decimal places (2 to 6) to control how the reported figures are rounded.
- Click Calculate. You get the RSD as the headline figure, the count, mean, standard deviation, variance, a precision assessment, and the comparison verdict when Dataset B is present. The Copy and Share buttons export the full result as text.
The formula: RSD is the standard deviation as a percentage of the mean
RSD = (standard deviation ÷ |mean|) × 100%. The absolute value of the mean keeps RSD positive when data averages below zero, and the standard deviation itself is never negative, so RSD is always zero or positive.
Sample SD = √[Σ(x − x̄)² ÷ (n − 1)]
Population SD = √[Σ(x − x̄)² ÷ n]
RSD and the coefficient of variation (CV) are the same quantity. Textbooks tend to say CV; laboratories tend to say RSD. This tool reports the single number and labels it RSD, so there is no confusion from showing two identical figures. If you want the CV framing with more on interpretation, see the coefficient of variation calculator, which treats the identical ratio from the textbook side. The definition matches the authoritative one: NIST's statistical reference defines relative standard deviation with the absolute value of the mean in the denominator, and GraphPad's statistics guide defines the coefficient of variation as the standard deviation divided by the mean, using the sample standard deviation by default. One more check on your side: a cousin of this idea is relative error, which also normalises a quantity against a reference value as a percentage, but against a true value rather than a mean.
Worked example: 10 measurements checked against the tool
Data: 12.5, 13.2, 11.8, 12.9, 13.1, 12.3, 12.7, 12.0, 13.0, 12.6 (n = 10). Every number below is the calculator's own output for the pre-filled inputs.
- Mean = 126.1 ÷ 10 = 12.61
- Sum of squared deviations = 1.969
- Sample variance = 1.969 ÷ 9 = 0.2188
- Sample SD = √0.2188 = 0.4677
- RSD = (0.4677 ÷ 12.61) × 100 = 3.71%, rated very good precision
Switch the calculation type to population and the same data gives 3.52%, because dividing by n instead of n-1 shrinks the standard deviation. That difference is exactly why the choice matters: two analysts can report different RSDs from identical data if one silently uses the population formula. On the worked example the tool's own readout confirms both figures, so pick sample when your data is a sample.
Compare two data sets head to head
RSD exists so you can compare precision across data sets with different means, and the calculator does the comparison for you. Paste a second set into Dataset B and the tool reports its statistics beside Dataset A's, then states the verdict as a difference in percentage points.
| Data set | n | Mean | SD (sample) | RSD |
|---|---|---|---|---|
| A: 12.5, 13.2, 11.8, 12.9, 13.1, 12.3, 12.7, 12.0, 13.0, 12.6 | 10 | 12.61 | 0.47 | 3.71% |
| B: 12.5, 13.9, 11.2, 14.0, 12.1 | 5 | 12.74 | 1.20 | 9.43% |
The verdict: Dataset A is more precise: RSD 3.71% versus 9.43%, a difference of 5.72 percentage points. The bars below show both RSDs against the precision bands, so the gap reads at a glance.
How to read the precision rating
The tool rates your RSD with a general rule of thumb for comparing data sets. These are comparison guidelines, not laboratory acceptance criteria: regulated labs validate their own RSD limits per method, often far below 10%.
| RSD | Rating |
|---|---|
| Under 10% | Very good precision |
| 10–20% | Good precision |
| 20–30% | Acceptable precision |
| Above 30% | Not acceptable – high relative variability |
A low RSD means your measurements cluster tightly around the mean; a high RSD means they scatter widely relative to it. Context is everything: in bioanalytical method validation, precision is typically required within 15% RSD (20% at the lower limit of quantification), so an RSD of 9.43% would be excellent there and merely fine here.
When RSD misleads
RSD breaks down when the mean is near zero, because dividing by a tiny number explodes the percentage. Two data sets with identical absolute spread can show wildly different RSDs if their means differ, so never compare RSDs across data sets whose means are close to zero or differ in sign. In those cases, compare the plain standard deviation instead. And when you do want a statement about the mean itself rather than the scatter around it, feed the same standard deviation into the confidence interval calculator, which turns it into a range that is likely to contain the true mean. The tool blocks a zero mean outright and warns you when the mean is merely close to zero.
Two more honest limits. First, RSD only makes sense for ratio-scale data with a meaningful zero, such as mass or concentration; it is meaningless for values on an interval scale like temperature in degrees Celsius, or for logarithms such as pH, where zero is arbitrary. Second, RSD measures precision (repeatability), not accuracy: a tightly clustered set of measurements can all be wrong together if the instrument is miscalibrated. Precision and accuracy answer different questions, and RSD answers only the precision one.
What competing RSD calculators miss
The top results for this keyword each compute one data set's RSD and stop there:
- MiniWebTool RSD calculator: paste values and get %RSD, CV, mean and SD with steps, sample or population, and honest notes that RSD does not establish data quality and that a mean near zero makes the ratio unstable. Gap: one data set only, no comparison, no precision verdict.
- OnlineMiniTools coefficient of variation calculator: formulas, interactive visualisation, sample and population modes, decimal precision from 2 to 15. Gap: the comparison case (the actual reason RSD exists) is left to the user and a second run.
- Testbook RSD calculator: calculator plus worked examples, but its worked example divides by n (population) without telling the reader the choice was made, and it never offers the alternative.
- TheTechEdvocate and UMA Technology guides: clear step-by-step arithmetic, but the same hidden-choice problem: one uses population (divides by n), the other sample (n-1), and neither flags that the reader's software may use the other one.
- Google Sheets guides (STDEV.S / AVERAGE): practical for spreadsheet users. Gap: a spreadsheet recipe, not a calculator, and no comparison or rating.
- The gap this page fills: the sample-or-population choice is explicit and changes the result live, the precision rating reads the number for you, and the two-data-set comparison with the percentage-point verdict does the job RSD was invented for: comparing precision across data sets. None of the rivals gives the verdict line.
Frequently asked questions
Relative standard deviation (RSD) expresses the standard deviation as a percentage of the mean: RSD = (SD ÷ |mean|) × 100%. It is a unit-free measure of how much data varies relative to its average, which is why it can compare the precision of data sets with different means and units.
Yes. Coefficient of variation (CV) and relative standard deviation (RSD) are two names for the same quantity: the standard deviation divided by the mean, times 100%. Textbooks tend to say CV; laboratories tend to say RSD. This tool reports the single number and labels it RSD.
Use the sample version (divide by n-1) when your data is a sample from a larger population, which is the usual case in experiments and quality control. Use the population version (divide by n) only when your data set is the entire population of interest. The choice changes the RSD, so it should always be stated.
As a general comparison guideline, below 10% is very good, 10–20% is good, 20–30% is acceptable, and above 30% is high. Regulated laboratories set their own stricter method-specific limits, so these bands are for comparing data sets, not for passing lab validation. Bioanalytical method validation, for example, commonly requires precision within 15% RSD.
The formula divides by the mean, and division by zero is undefined. Even a mean merely close to zero makes RSD unstable and misleading, since tiny changes in the mean swing the percentage wildly. The tool blocks a zero mean outright and warns when the mean is close to zero.
No. The formula uses the absolute value of the mean, and standard deviation is never negative, so RSD is always zero or positive. A negative mean does not produce a negative RSD.
Paste a second set of measurements into Dataset B and click Calculate. The tool reports Dataset B's mean, standard deviation, and RSD next to Dataset A's, then declares which set is more precise and by how many percentage points. Leave Dataset B blank to skip the comparison.
Add This Calculator to Your Website
Embed this calculator on your own site with the code below. Pick a size to preview it live, then copy the code.
Powered by Calculator Zilo
References
- National Institute of Standards and Technology, DATAPLOT Reference Manual. RELATIVE STANDARD DEVIATION Statistics LET Subcommand. Accessed October 23, 2026.
- National Institute of Standards and Technology, DATAPLOT Reference Manual. COEFFICIENT OF VARIATION CONFIDENCE LIMIT. Accessed October 23, 2026.
- GraphPad. Prism 10 Statistics Guide: Interpreting results, Coefficient of Variation. Accessed October 23, 2026.
- U.S. Food and Drug Administration. M10 Bioanalytical Method Validation and Study Sample Analysis, Guidance Document. Accessed October 23, 2026.
- U.S. Food and Drug Administration. Bioanalytical Method Validation Guidance for Industry (PDF). Accessed October 23, 2026.
- AAPS Journal. Why is Biomarker Assay Validation Different from that of Pharmacokinetic Assays?. Accessed October 23, 2026.
