๐ Coefficient of Variation Calculator (CV / RSD)
By Shihab Mia ยท Updated 2026-07-01
CV = (standard deviation / mean) x 100. Sample CV uses the n-1 SD; population CV uses the n SD.
The coefficient of variation (CV), also called relative standard deviation (RSD), is the standard deviation divided by the mean, multiplied by 100 to give a percentage. It measures how spread out your data is relative to its average. Paste a list of numbers below and this calculator returns the count, sum, mean, both the sample and population standard deviation, and the CV as a percentage. Because the CV is unitless, it lets you compare the variability of two data sets even when they have different means or different units.
What is the Coefficient of Variation Calculator?
The coefficient of variation answers a question a plain standard deviation cannot: how big is the spread compared with the size of the values themselves? It is defined as CV = (standard deviation / mean) x 100, expressed as a percentage. A salary data set with a standard deviation of $5,000 sounds volatile, but if the mean salary is $200,000 the CV is only 2.5%, which is very stable. The same $5,000 spread around a mean of $20,000 gives a CV of 25%, far more variable. By dividing the spread by the mean, the CV strips away the scale and leaves pure relative variability.
Because it is a ratio of two quantities in the same units, the CV itself has no units. That is what makes it useful for comparison. You can compare the consistency of a manufacturing process measured in millimetres against one measured in grams, or compare the volatility of a stock priced at $30 against one priced at $3,000, on a single common scale. A lower CV means the data clusters tightly around its mean (more consistent), while a higher CV means the values are more scattered relative to the average.
There is one important condition: the CV is only meaningful for data measured on a ratio scale with a true, meaningful zero and a positive mean. Temperature in Celsius, for example, is an interval scale (its zero is arbitrary), so a CV computed on Celsius readings is misleading. The CV also becomes unstable or undefined when the mean is near or equal to zero, because you are dividing by a tiny or zero denominator. This calculator reports the CV using both the sample standard deviation (dividing by n-1) and the population standard deviation (dividing by n), so you can pick the one that matches your situation: use the sample version when your numbers are a sample drawn from a larger group, and the population version when they are the entire group.
The CV shows up under different names across disciplines, but the arithmetic is identical. Analytical chemists and lab technicians call it the relative standard deviation (RSD) and use it to judge assay precision, where a low RSD across replicate measurements signals a repeatable method. In finance it is the risk-to-reward ratio: divide an investment's standard deviation (its risk) by its expected return (its reward), and the lower the CV, the more return you earn per unit of risk. Agronomists use it to gauge how uniform a field trial was, and quality engineers use it to confirm a process stays within tolerance. Same formula, different labels.
Interpretation is always context-dependent, so there is no single threshold that means good everywhere. A CV under 5% is often demanded in pharmaceutical and clinical lab work, while agricultural field experiments routinely accept CVs of 20% to 30% because outdoor conditions vary. A CV above 100% is entirely possible when the standard deviation exceeds the mean, which is common for skewed, heavy-tailed data such as incomes or insurance claims. Rather than memorising a universal cutoff, compare your CV against the accepted norm for your own field, and treat the ranges in the tables below as a rough starting point, not a rule.
One practical caution: the CV summarises spread with a single number, so it hides the shape of the distribution. Two data sets can share the same CV while one is symmetric and the other heavily skewed. For that reason, pair the CV with a quick look at the raw values or a histogram before drawing conclusions, especially when the mean is small or the data contains outliers that can inflate both the standard deviation and the ratio.
When to use it
- Comparing the variability of two data sets that have different means or different units, where a raw standard deviation would be misleading.
- Quality control in labs and manufacturing, where a low CV (often under 2% to 5%) signals a precise, repeatable process.
- Comparing the risk-per-unit-return of investments, where a lower CV means less volatility for each unit of expected return.
- Checking the reliability of repeated measurements or assay replicates by reporting the relative standard deviation (RSD).
- Assessing the uniformity of an agricultural field trial or experiment, where a lower CV indicates more consistent plot-to-plot results.
- Reporting the precision of an analytical method or instrument during method validation and calibration.
How to use the Coefficient of Variation Calculator
- Paste or type your numbers into the box, separated by commas, spaces, or new lines.
- Read off the count, sum, and mean that the calculator computes from your data.
- Note the sample standard deviation (n-1) and population standard deviation (n).
- Use the sample CV if your numbers are a sample, or the population CV if they are the whole group. The CV is shown as a percentage.
Formula & method
Worked examples
Find the population and sample CV of the data set 2, 4, 4, 4, 5, 5, 7, 9.
- Count n = 8 and sum = 40, so mean = 40 / 8 = 5
- Squared deviations from the mean: 9, 1, 1, 1, 0, 0, 4, 16, which sum to 32
- Population variance = 32 / 8 = 4, so population SD = sqrt(4) = 2
- Population CV = (2 / 5) x 100 = 40%
- Sample variance = 32 / (8 - 1) = 4.5714, so sample SD = sqrt(4.5714) = 2.1380
- Sample CV = (2.1380 / 5) x 100 = 42.76%
Result: Mean 5, Population CV 40%, Sample CV about 42.76%
Find the population and sample CV of the data set 10, 12, 14, 16, 18.
- Count n = 5 and sum = 70, so mean = 70 / 5 = 14
- Squared deviations from the mean: 16, 4, 0, 4, 16, which sum to 40
- Population variance = 40 / 5 = 8, so population SD = sqrt(8) = 2.8284
- Population CV = (2.8284 / 14) x 100 = 20.20%
- Sample variance = 40 / (5 - 1) = 10, so sample SD = sqrt(10) = 3.1623
- Sample CV = (3.1623 / 14) x 100 = 22.59%
Result: Mean 14, Population CV about 20.20%, Sample CV about 22.59%
Finance: compare two investments to see which carries less risk per unit of return. Investment A has an expected return of 10% and a standard deviation of 5%. Investment B has an expected return of 12% and a standard deviation of 8%.
- For Investment A, CV = standard deviation / mean return = 5 / 10 = 0.50
- For Investment B, CV = 8 / 12 = 0.667
- The lower CV wins because it means less volatility for each unit of expected return
- Investment A carries about 0.50 units of risk per unit of return, versus about 0.67 for Investment B
Result: Investment A has the lower CV (0.50 vs 0.67), so it offers the better risk-to-reward trade-off despite the slightly lower expected return.
Interpreting the coefficient of variation (rough, context-dependent guide)
| CV range | Relative variability | Typical reading |
|---|---|---|
| Under 10% | Low | Data is tightly clustered around the mean (very consistent). |
| 10% to 20% | Moderate | Reasonable spread; common in many real-world measurements. |
| 20% to 30% | High | Noticeably scattered relative to the average. |
| Over 30% | Very high | Highly variable; the mean may not represent the data well. |
Typical acceptable CV by field (general norms, always defer to your own standards)
| Field or use | Commonly accepted CV | Why |
|---|---|---|
| Clinical and pharmaceutical labs | Under 5% (often under 2%) | High precision is required; a CV of 10% or more suggests poor method performance. |
| Manufacturing and quality control | Under 10% | A low CV confirms a stable, repeatable process within tolerance. |
| Analytical chemistry (RSD) | Under 5% for good technique | Replicate measurements should agree closely to trust the assay. |
| Agricultural field experiments | Up to 20% to 30% | Outdoor plot-to-plot variation is naturally high; pot experiments are stricter. |
| Finance and investing | No fixed cutoff | Market returns are inherently variable; use the CV to rank risk per return, not to pass or fail. |
Standard deviation versus coefficient of variation
| Property | Standard deviation | Coefficient of variation |
|---|---|---|
| Units | Same as the data | Unitless (a percentage) |
| Measures | Absolute spread | Spread relative to the mean |
| Compare across data sets | Only if same units and scale | Yes, even across units or scales |
| Needs a meaningful zero | No | Yes (ratio scale, positive mean) |
Common mistakes to avoid
- Using the CV on interval-scale data. The CV is only valid on a ratio scale with a true zero, such as length, mass, or count. Applying it to temperature in Celsius or Fahrenheit, or to calendar years, gives a meaningless number because the zero point is arbitrary.
- Computing the CV when the mean is near zero. Dividing by a mean close to zero makes the CV blow up to huge or unstable values, and a mean of exactly zero leaves it undefined. For data centred near zero, report the standard deviation instead.
- Mixing up sample and population standard deviation. Sample SD divides by n-1 and population SD divides by n, so they give slightly different CVs. Use the sample version for a sample of a larger group and the population version only when your data is the entire group.
- Forgetting the percentage multiplier. CV is usually reported as a percentage, so multiply the SD-over-mean ratio by 100. A ratio of 0.4 is a CV of 40%, not 0.4%. Reporting the raw ratio confuses readers expecting a percent.
- Applying the CV to data with negative values. When a data set mixes positive and negative numbers, the mean can be small or misleading and the CV loses its meaning. The CV assumes all values are positive and on a ratio scale.
- Treating one CV threshold as universal. A CV that is excellent in agriculture (say 20%) would fail in a clinical lab that demands under 5%. Always compare your CV against the accepted norm for your specific field rather than a single fixed cutoff.
Glossary
- Coefficient of variation (CV)
- The standard deviation divided by the mean, times 100. A unitless measure of relative variability, expressed as a percentage.
- Relative standard deviation (RSD)
- Another name for the coefficient of variation, used widely in chemistry and lab work to describe assay precision.
- Standard deviation
- A measure of how far the values typically fall from the mean, in the same units as the data.
- Variance
- The average of the squared deviations from the mean. The standard deviation is its square root.
- Mean
- The arithmetic average: the sum of the values divided by how many there are.
- Ratio scale
- A measurement scale with a true, meaningful zero (length, mass, count), where ratios between values make sense.
- Interval scale
- A scale with equal spacing but an arbitrary zero, such as Celsius temperature, where the CV is not meaningful.
- Sample vs population
- A sample is a subset drawn from a larger group; the population is the entire group. The sample SD divides by n-1, the population SD by n.
Frequently asked questions
What is the coefficient of variation?
The coefficient of variation (CV) is the standard deviation divided by the mean, multiplied by 100 to give a percentage. It measures how large the spread of a data set is relative to its average, so it is often called relative variability or relative standard deviation.
How do I calculate the coefficient of variation?
Find the mean of your data, find the standard deviation, then compute CV = (standard deviation / mean) x 100. This calculator does it automatically and shows the result using both the sample standard deviation (divide by n-1) and the population standard deviation (divide by n).
What is the formula for the coefficient of variation?
The formula is CV = (standard deviation / mean) x 100. Use the sample standard deviation s = sqrt(sum of (x - mean)^2 / (n - 1)) when your data is a sample, or the population standard deviation sigma = sqrt(sum of (x - mean)^2 / n) when it is the whole group.
Is the coefficient of variation the same as relative standard deviation?
Yes. Relative standard deviation (RSD) is just another name for the coefficient of variation. The two terms are interchangeable, with RSD being more common in laboratory and analytical chemistry contexts.
What is a good coefficient of variation?
It depends on the field. In quality control and clinical labs a CV under about 2% to 5% is considered very precise. More broadly, under 10% is low variability, 10% to 30% is moderate to high, and over 30% means the values are very scattered relative to the mean. Agricultural field trials often accept 20% to 30%. Always judge it against norms for your own data.
Why use the CV instead of the standard deviation?
The standard deviation is in the units of your data, so you cannot fairly compare a spread in millimetres with one in grams, or compare two data sets with very different means. The CV is unitless, so it puts relative variability on one common scale and makes such comparisons valid.
Can the coefficient of variation be greater than 100%?
Yes. A CV above 100% simply means the standard deviation is larger than the mean, which happens with highly skewed or heavy-tailed data such as incomes, insurance claims, or rare-event counts. It is a signal of very high relative variability, not an error.
Should I use the sample or population CV?
Use the sample CV (standard deviation with n-1) when your numbers are a sample drawn from a larger population, which is the most common case. Use the population CV (standard deviation with n) only when your data covers the entire group you care about. The sample CV is slightly larger.
What happens when the mean is zero?
The coefficient of variation is undefined when the mean is zero, because the formula divides by the mean. When the mean is merely close to zero, the CV becomes very large and unstable. In those cases report the standard deviation instead, or use a measure suited to data centred near zero.
How is the coefficient of variation used in finance?
In finance the CV is the risk-to-reward ratio: divide an investment standard deviation (its risk) by its expected return (its reward). A lower CV means you earn more return per unit of risk, so when comparing two investments the one with the lower CV offers the better risk-adjusted trade-off.