Variance and Coefficient of Variation Calculator
Calculate sample or population variance, standard deviation, and the signed coefficient of variation from a numeric list. The calculator also reports the mean and count, marks CV undefined when the mean is zero, and warns when a negative mean makes relative comparisons questionable.
By Dr. Hassan Khames EliwaLast updated
Result
- Count
- 4
- Mean
- 4
- Standard deviation
- 1.632993
- Coefficient of variationCV is meaningful for ratio-scale data with a meaningful zero.
- 40.824829%
How it works
Choose the denominator deliberately
Variance measures squared distance from the arithmetic mean. For n observations, first calculate mean = sum of values / n. Next calculate S = sum of (value − mean)². Population variance is S / n; sample variance is S / (n − 1). Standard deviation is the square root of the selected variance, and CV = 100 × standard deviation / mean.
Use population mode when the list is the complete group you intend to describe, such as every completed batch in a particular shift. Use sample mode when those batches represent a wider production process. Dividing by n − 1 accounts for estimating the unknown population mean from the same sample. It does not correct a biased sampling plan or prove the process is stable.
Keep the measurement scale in view
If observations are seconds, standard deviation is in seconds and variance is in seconds squared. CV is a percentage because its numerator and denominator share units. That cancellation only supports sensible comparisons when zero means an absence of the measured quantity. Celsius temperatures, indexed scores, and many accounting balances fail that condition even when a spreadsheet produces a number.
The implementation accepts up to 10,000 numbers, each within ±1e100. Separate observations with commas, whitespace, or semicolons; do not use thousands separators. One observation is sufficient for population variance, but sample variance needs at least two. All calculations use unrounded inputs; displayed decimals are a presentation choice, not an instruction to round intermediate deviations.
Worked example
Four inspection times
Suppose four inspection times are 2, 4, 4, and 6 minutes. Their sum is 16 minutes and their mean is 4 minutes. The squared deviations are 4, 0, 0, and 4 minutes squared, totaling 8 minutes squared.
In sample mode the count is 4, variance is 2.666667 minutes squared, standard deviation is 1.632993 minutes, and CV is 40.824829%. In population mode the same observations give variance 2 minutes squared, standard deviation 1.414214 minutes, and CV 35.355339%. Neither answer contradicts the other: the denominator answers a different statistical question.
For a zero-mean list such as −1 and 1, variance still exists but CV is undefined. For −6, −4, −4, and −2, sample CV is −40.824829%; the warning matters more than the sign because a negative relative spread is rarely a useful business comparison.
Expert tips
Common mistakes
- Do not enter 1,000 to mean one thousand: the list parser reads that as two observations. Enter 1000 instead, and keep decimal points unambiguous.
- Do not square the standard deviation after rounding it and expect the printed variance to match exactly. The calculator retains more precision internally.
- Do not interpret a small CV as evidence of accuracy. Measurements can cluster tightly around a systematically wrong value; spread and calibration describe separate problems.
Frequently asked questions
Does CV use an absolute mean?
No. This calculator reports the signed expression, standard deviation divided by mean times 100. A negative mean triggers a caution rather than silently changing the definition.
What happens when every value is equal?
Variance and standard deviation are zero. CV is zero if the common value is nonzero, but remains undefined when all observations are zero.
Can I paste negative observations?
Yes. Variance and standard deviation support negative values. The interpretability of CV depends on the measurement scale and mean, not merely whether entry is allowed.
Why are sample results larger?
For the same nonconstant list, division by n minus one produces a larger variance than division by n. The difference decreases as the count increases.
Can a CV exceed 100 percent?
Yes. That means the standard deviation exceeds the mean when the mean is positive. It is not a probability and has no universal upper bound.
Where can I learn the underlying spread calculation?
See the standard deviation explanation for interpretation and the average calculator to inspect the center separately.
Related calculators
Related guides
MathWeighted averages: where simple averages misleadChoose weights that match the question, combine rates from their denominators, and avoid misleading averages of prices, speeds and percentages.16 November 2026 · 8 min read
MathStandard deviation without the jargonThe average tells you the centre. Standard deviation helps describe the spread, provided you know what the data represents.14 November 2026 · 7 min read
MathPercentages people get wrong: change, difference and pointsThe denominator determines the answer. Learn to distinguish percentage change, symmetric difference and percentage-point movements.14 November 2026 · 7 min readAbout the author
Dr. Hassan Khames Eliwa holds a PhD in Construction & Built Environment and has more than 20 years of international project controls experience. He is Director of Eliwa Project Controls Limited in Auckland and founder of PMMilestone.org and PMMilestone.com.
Buy me a coffee