Coefficient of Variation Calculator

CV = (σ / μ) × 100%

Calculate the coefficient of variation from raw data or summary statistics. Compare variability across datasets with different units — the only metric that makes heights in centimetres and weights in kilograms directly comparable.

CV = (σ / μ) × 100 RSD = s / x̄ × 100 Raw Data or Summary Stats Dataset Comparison Sample & Population

Share this tool

Calculating…

Enter your data above and click Calculate to compute the coefficient of variation.

What Is the Coefficient of Variation?

The coefficient of variation (CV) — also called the relative standard deviation (RSD) — is a standardised measure of statistical dispersion. It expresses the standard deviation as a percentage of the mean, making it possible to compare variability across datasets that use completely different units or have very different means.

A standard deviation of 10 kg tells you very little on its own. Is that a lot of variation? It depends entirely on the mean. If the mean weight is 70 kg, a 10 kg spread is substantial. If the mean were 1,000 kg (a dataset of elephants), 10 kg would be negligible. The CV solves this by always expressing dispersion relative to the centre of the data.

CV is widely used in research methods courses, analytical chemistry (lab precision), finance (comparing investment risk), biology (population studies), and manufacturing quality control. Whenever you need to compare "how spread out" two or more datasets are — regardless of scale or units — CV is the right tool.

Coefficient of Variation Formula

Population CV

CV = (σ / μ) × 100

σ = population standard deviation
μ = population mean
Use when you have data for the entire population

Sample CV

CV = (s / x̄) × 100

s = sample standard deviation (Bessel's correction)
x̄ = sample mean
Use when your data is a sample from a larger population

Sample standard deviation uses n−1 in the denominator (Bessel's correction) to produce an unbiased estimate of the population variance. Population standard deviation uses n. In practice, most statistical analysis uses the sample version because you rarely have complete population data.

The result is always expressed as a percentage. A CV of 15% means the standard deviation is 15% of the mean. CV has no units — a CV of 15% is equally interpretable whether your data is in metres, dollars, or seconds.

Worked example — Heights dataset

Data: 165, 170, 172, 168, 175, 160, 178 (n = 7)

Mean x̄ = (165+170+172+168+175+160+178) / 7 = 169.71

Sample std dev s = 6.02

CV = (6.02 / 169.71) × 100 = 3.55%

How to Interpret CV Values

There is no single universally agreed threshold for "good" or "bad" CV — it depends on the field and the precision needed. The following guidelines are widely used in research methods and quality-control contexts as starting points.

CV Range	Label	Typical Meaning
< 10%	Low	Very consistent data; high precision instrument or homogeneous group
10% – 20%	Moderate	Acceptable variability in most biological and social science studies
20% – 30%	High	Notable spread; worth investigating the source of variability
≥ 30%	Very High	Large dispersion; data may be heterogeneous or contain outliers

Domain context matters: In analytical chemistry, a CV above 5% may be unacceptable. In financial markets, a CV of 50% might be normal for a volatile asset. In epidemiology, CV thresholds vary with sample size and outcome measurement. Always interpret your CV in the context of your field's norms.

CV vs. Standard Deviation — When to Use Each

The standard deviation and CV both measure spread, but they answer different questions. Understanding which to use prevents common analytical mistakes.

Use Standard Deviation when…

You are comparing datasets in the same units with similar means
You need an absolute measure of spread (e.g. "scores varied by ±8 points")
Computing confidence intervals or running t-tests
Your data include zero or negative values (CV breaks down)

Use Coefficient of Variation when…

Comparing datasets with different units (e.g. heights vs. weights)
Comparing datasets with very different means (e.g. children vs. adults)
Assessing lab or instrument precision relative to the measurement scale
Comparing investment risk across assets at different price levels

Important limitation

CV is undefined when the mean is zero, and unreliable when data contain both positive and negative values (the mean can be near zero even if spread is large). CV is also problematic for data measured on interval scales without a true zero (e.g. Celsius temperature) — use Fahrenheit or Kelvin data with caution.

Visual Guide: Why CV Makes Datasets Comparable

Standard deviation alone cannot compare datasets with different units. CV solves this by normalising against the mean.

Heights

Mean (μ) 172 cm

Std Dev (σ) 10 cm

Can we compare "10 cm" to "14 kg"?

No — different units, not comparable

CV = (10 / 172) × 100

5.8%

Low variability

vs CV removes
units

Weights

Mean (μ) 70 kg

Std Dev (σ) 14 kg

Can we compare "14 kg" to "10 cm"?

No — different units, not comparable

CV = (14 / 70) × 100

20.0%

High variability

CV makes the comparison possible

Despite different units, CV reveals that weight (20%) is far more variable than height (5.8%) relative to their respective means. This comparison is impossible with raw standard deviations alone.

CV comparison

Heights (cm) 5.8%

Weights (kg) 20.0%

Coefficient of Variation in Different Fields

Finance & Investing

CV compares the risk-to-reward ratio of investments at different price levels. A stock with mean return 8% and std dev 4% has CV = 50%. A bond with mean 3% and std dev 0.5% has CV = 17%. The bond is more consistent relative to its return, regardless of the absolute dollar values.

Analytical Chemistry & Lab Science

CV (often called RSD in chemistry) measures instrument precision. A typical acceptance threshold is CV < 2% for high-precision assays and CV < 5% for routine analytical methods. Comparing precision across instruments with different measurement scales requires CV, not raw std dev.

Biology & Medicine

Used to compare variability in biological measurements across species, age groups, or treatment conditions. CV is essential when comparing, say, enzyme activity between mice (mean 50 units) and humans (mean 2,000 units) — a direct std dev comparison would be meaningless.

Manufacturing & Quality Control

Process capability and Six Sigma methodologies use CV to assess consistency across production lines operating at different scales. A line producing 1,000 kg batches and one producing 10 kg batches can be compared on consistency using CV even though their standard deviations will naturally differ.

Coefficient of Variation FAQ

What is the coefficient of variation formula?

The coefficient of variation formula is CV = (σ / μ) × 100 for population data, or CV = (s / x̄) × 100 for sample data, where σ is population standard deviation, μ is population mean, s is sample standard deviation (using n−1), and x̄ is sample mean. The result is expressed as a percentage — a unit-free measure of relative dispersion.

What is a good coefficient of variation?

It depends heavily on your field. As a general guide: CV < 10% is considered low (very consistent); 10–20% is moderate; 20–30% is high; and ≥ 30% is very high. In analytical chemistry, CV must often be below 2–5%. In social science surveys, CV of 30–40% may be acceptable. In financial analysis, a CV of 50%+ is common for equities. Always compare your CV to the accepted norms in your specific discipline.

What is the difference between CV and standard deviation?

Standard deviation is an absolute measure — it carries the same units as the data (e.g. 5 kg, 3 seconds). CV is a relative measure — it is unit-free and expressed as a percentage. Use standard deviation when you need the actual spread in context (e.g. "temperatures vary by ±3°C"). Use CV when you need to compare datasets with different units or different magnitudes — the standard deviation alone can be misleading when means differ greatly.

When should I use population CV vs. sample CV?

Use sample CV (with Bessel's correction, n−1 denominator) when your data is a random sample from a larger population — which is almost always the case in research. Use population CV (n denominator) only when you have measured every member of the population with certainty. The difference is small for large samples (n > 30) but can be significant for very small samples.

Can the coefficient of variation be greater than 100%?

Yes. A CV greater than 100% simply means the standard deviation is larger than the mean. This can happen with highly skewed distributions, data containing a mix of very small and very large values, or data from heavy-tailed distributions such as income data or insurance claims. There is no mathematical cap on CV — values of 200–500% are observed in practice for highly variable phenomena.

Why is CV undefined when the mean is zero?

CV involves dividing the standard deviation by the mean. When the mean equals zero, this results in division by zero, which is mathematically undefined. Even when the mean is close to zero, CV can produce arbitrarily large and misleading values. This is why CV should not be used for data that can span positive and negative values (where the mean might be zero or near-zero) — such as financial returns centred around zero, or temperature data in Celsius where 0°C is not a true absence of temperature.

What is relative standard deviation (RSD)?

Relative standard deviation (RSD) is simply another name for the coefficient of variation, used especially in analytical chemistry and laboratory science. The formulas are identical: RSD = (s / x̄) × 100%. The term "RSD" is preferred in ISO/IUPAC standards for laboratory method validation, while "CV" is more common in statistics, social science, and medicine. This calculator computes both — they are the same number.

Related Calculators

Standard Deviation Calculator

Sample & population std dev from raw data

Variance Calculator

Population and sample variance with Bessel's correction

Mean, Median & Mode Calculator

Central tendency statistics from any dataset

Z-Score Calculator

Convert raw scores to z-scores and percentile ranks

Frequency Distribution Calculator

Build a frequency table and histogram from raw data

Confidence Interval Calculator

Margin of error, CI bounds, and critical values

External Resources

Wikipedia — Coefficient of variation: definitions, properties, and use cases NIST/SEMATECH e-Handbook — Relative Standard Deviation in analytical measurement