Coefficient of Variation (CV)

The coefficient of variation (CV) is a statistical measure that represents the ratio of the standard deviation to the mean. It is a highly useful statistic for comparing the degree of variation from one data series to another, even if the means are drastically different from one another.

While standard deviation measures the absolute dispersion of data, it falls short when you need to compare datasets with different units or vastly different scales. By dividing the standard deviation by the mean, the coefficient of variation normalizes the data. It is usually expressed as a percentage, making it an intuitive gauge of relative volatility or risk.

When to Use CV Instead of Standard Deviation

• Comparing Different Scales: If you are comparing the price volatility of a penny stock valued at $2 against a blue-chip stock valued at $500, a $1 price swing is catastrophic for the former but meaningless for the latter. Standard deviation cannot accurately compare these two, but CV normalizes the risk.
• Comparing Different Units: If you want to compare the variability in the weight of a local population (measured in kilograms) against their height (measured in centimeters), standard deviation is useless because the units don’t match. CV is unitless, allowing for direct comparison.

A lower coefficient of variation indicates less volatility and a tighter grouping of data around the mean. A higher CV indicates greater dispersion and higher volatility.

The Coefficient of Variation Formulas

Depending on whether you are analyzing a complete population or just a sample subset, the formula shifts slightly to account for the different standard deviation calculations.

Population Formula

Use this when you have data representing the entire group you want to draw conclusions about.

• σ = Population Standard Deviation
• μ = Population Mean

Sample Formula

Use this when your data is a fraction of a larger population. This is the most common application in real-world research.

• s = Sample Standard Deviation
• x̄ = Sample Mean

How to Calculate the Coefficient of Variation

Let’s manually calculate the sample CV using a straightforward dataset: 10, 34, 23, 54, 9.

Step 1: Find the Mean (x̄) Add all the numbers together and divide by the total count (n = 5). (10 + 34 + 23 + 54 + 9) / 5 = 130 / 5 = 26

Step 2: Calculate the Standard Deviation (s) Find the squared differences from the mean for each number, sum them up, divide by n – 1 (which is 4 in this case), and find the square root. For this dataset, the sample standard deviation is approximately 18.77.

Step 3: Divide the Standard Deviation by the Mean 18.77 / 26 = 0.722

Step 4: Convert to a Percentage 0.722 × 100 = 72.2%

The relative variability of this specific dataset is 72.2%. You can verify this result using the interactive calculator above.

Real-World Applications

• Investing and Finance: Analysts use the “risk-to-reward ratio” to evaluate portfolios. By looking at the CV of historical returns, investors can identify which assets offer the best return for the lowest relative risk.
• Manufacturing and Quality Control: Factories measure product consistency. If a machine cutting 100mm steel pipes has a CV of 0.5%, it is highly precise. A rising CV indicates machine wear and calibration issues.
• Medical Research: Epidemiologists and lab technicians use CV to evaluate the precision of assay tests. A low CV ensures that blood test results are reproducible and reliable.

How to Find CV in Excel or Google Sheets

There is no dedicated “CV” function built into spreadsheet software, but you can easily calculate it by combining two existing functions.

To find the sample coefficient of variation for data located in cells A1 through A10, type the following formula into a blank cell:

=(STDEV.S(A1:A10) / AVERAGE(A1:A10)) * 100

Note: If you are calculating the population CV, replace STDEV.S with STDEV.P.

Important Limitations to Keep in Mind

1. When the mean is close to zero: Because the formula requires dividing by the mean, a mean that hovers around zero will cause the CV to spike to infinity, producing a highly misleading percentage that does not accurately reflect the actual data spread.
2. With interval data: CV is only valid for ratio data (data that has a true, meaningful zero point, like weight, height, or salary). It should never be used with interval scales like temperature (Celsius or Fahrenheit), where the “zero” is arbitrary.