Coefficient of Variation — Definition, Formula & Example
The coefficient of variation is a measure of how spread out data are compared with their average. It is usually written as a percentage, so a CV of 10% means the typical distance from the mean is about one-tenth of the mean.
For a data set with mean (population) or (sample) and standard deviation or , the coefficient of variation is the ratio or . It is a unitless quantity, often multiplied by . It is defined only when the mean is nonzero. It is meaningful only for ratio-scale data, meaning data measured on a scale with a true zero, such as weight or length. Temperature in degrees Fahrenheit is not ratio-scale.
Key Formula
Where:
- = Population or sample standard deviation, in the same units as the data
- = Population or sample mean, which must be nonzero
- = Coefficient of variation, a unitless ratio usually written as a percent
Worked Example
Problem: A vet records the weights of puppies and of adult dogs of one breed. The puppies have a mean of 20 pounds and a standard deviation of 4 pounds. The adult dogs have a mean of 60 pounds and a standard deviation of 6 pounds. Which group has more variability relative to its average weight?
Step 1: Find the CV for the puppies by dividing the standard deviation by the mean.
Step 2: Find the CV for the adult dogs in the same way.
Step 3: Compare the two values. The adults have the larger standard deviation (6 pounds versus 4 pounds). Relative to their mean, however, the puppies vary twice as much.
Answer: The puppies have more relative variability, with a CV of 20% compared with 10% for the adult dogs.
Why It Matters
In chemistry and biology labs, the precision of repeated measurements is often reported as percent relative standard deviation (%RSD), which is the CV. In finance, investors use CV to compare the risk per dollar of expected return across investments with very different average returns.
Common Mistakes
Mistake: Comparing standard deviations directly when two data sets have very different means or different units.
Correction: The adult dogs' standard deviation of 6 pounds looks bigger than the puppies' 4 pounds, but relative to their means (60 versus 20 pounds) it is smaller: 10% versus 20%. Divide by the mean first, so you compare unitless CVs.
