Coefficient of Variation Calculator
Standard deviation tells you how spread out a data set is, but it's expressed in the same units as the data itself, which makes it awkward to use when comparing variability between two data sets that use different units or vastly different scales — comparing the standard deviation of a stock priced in dollars against one priced in a different currency, for instance, or comparing variability in height (measured in centimeters) against variability in weight (measured in kilograms), doesn't give a meaningful apples-to-apples comparison on its own. The coefficient of variation solves this by expressing standard deviation as a percentage of the mean, producing a unitless number that can be compared directly across data sets regardless of their original scale or units. This calculator takes a data set's mean and standard deviation (or the raw data itself) and returns the coefficient of variation, letting you compare relative variability meaningfully across otherwise incomparable data.
How the Coefficient of Variation formula works
The coefficient of variation formula:
CV = (Standard deviation / Mean) × 100
Expressed as a percentage, CV represents the standard deviation as a proportion of the mean — a CV of 20% means the data's spread is, on average, about a fifth of the size of the mean value itself, regardless of what units or scale the original data used.
Step-by-step calculation
- Calculate (or enter) the data set's mean and standard deviation.
- Divide the standard deviation by the mean.
- Multiply by 100 to express the result as a percentage.
Worked example
Data set A has a mean of 50 and standard deviation of 5: CV = (5/50) × 100 = 10%. Data set B has a mean of 500 and standard deviation of 40: CV = (40/500) × 100 = 8%. Even though B's standard deviation is much larger in absolute terms, its coefficient of variation is actually lower — meaning B is relatively less variable once scale is accounted for.
When to use coefficient of variation instead of raw standard deviation
The clearest use case for coefficient of variation is comparing variability between two or more data sets that don't share the same units or the same scale of magnitude. A raw standard deviation of 5 sounds small next to a standard deviation of 500, but that comparison is meaningless without knowing the respective means — a standard deviation of 5 around a mean of 10 represents enormous relative variability (CV of 50%), while a standard deviation of 500 around a mean of 50,000 represents comparatively little (CV of 1%). Financial analysts commonly use CV to compare the relative risk (volatility) of different investments regardless of their price scale, since a $5 stock and a $500 stock can have wildly different standard deviations in dollar terms purely because of their price level, not because one is genuinely more volatile than the other in a meaningful risk sense.
CV shows up regularly in scientific and laboratory contexts too, particularly for assessing measurement precision and repeatability. A lab technique that consistently produces results with a CV under roughly 5% is often considered highly precise, while a CV above 15-20% might flag a measurement process worth investigating for sources of inconsistency — thresholds vary by field and application, but the underlying logic is the same: CV gives a scale-independent way to judge whether variability is 'a lot' or 'a little' relative to the values actually being measured.
One important limitation is worth flagging clearly: coefficient of variation becomes unstable and can be misleading when the mean is close to zero, since dividing by a very small number produces an inflated or even meaningless CV even for genuinely modest variability. CV is also generally reserved for data measured on a ratio scale (where zero represents a true absence of the quantity, and negative values either don't exist or don't make sense) — for data that can meaningfully take negative values or is measured on an interval scale without a true zero point, like temperature in Celsius, CV doesn't have a sensible interpretation and shouldn't be used as a comparison metric.
Frequently asked questions
Why is CV expressed as a percentage instead of a raw ratio?
Expressing it as a percentage (multiplying by 100) is simply a convention that makes the number more intuitively readable — a CV of 15% is easier to interpret at a glance than a ratio of 0.15, though both represent the same underlying relationship between standard deviation and mean.
Can coefficient of variation be used to compare data sets with negative values?
It's generally not meaningful for data that can be negative or is measured without a true zero point (like temperature in Celsius), since the ratio of standard deviation to mean loses its intuitive interpretation in those cases. CV works best for data on a ratio scale where zero represents true absence of the quantity, like weight, price, or count data.
What counts as a 'high' or 'low' coefficient of variation?
There's no universal threshold — it depends heavily on the field and what's being measured. In some scientific contexts, a CV under 5-10% might be considered low variability, while in fields like finance or social science data, considerably higher CVs are common and not necessarily a red flag. Compare against typical values for your specific type of data.