Standard Deviation vs Variance
Standard deviation and variance measure the same spread, but variance is squared and standard deviation is in the original units. This guide explains the difference, shows why variances add and standard deviations do not, and gives worked numbers for when to report each one.
Variance and standard deviation are the same spread seen two ways. Variance is the average squared distance from the mean. Standard deviation is the square root of that β the same spread, back in the units you started with. Everything on this page follows from that one fact.
What Is Variance?
Variance is the mean squared deviation from the mean: take each value, subtract the mean, square the result, and average those squares. For a sample, divide by n β 1; for a population, divide by n. The squaring is what makes variance what it is β it stops positive and negative deviations from cancelling, and it punishes large deviations far more than small ones.
Because the deviations are squared, variance is measured in squared units: if your data are heights in centimetres, the variance is in square centimetres. That is the price of squaring, and it is why variance is awkward to report on its own.
Work it out step by step in the how to calculate standard deviation guide, or let the standard deviation calculator do the arithmetic. For data already grouped into a frequency table, the standard deviation from frequency table guide works through the same calculation.
What Is Standard Deviation?
Standard deviation is the square root of the variance. Taking the square root undoes the squaring and brings the spread back into the original units: centimetres, not square centimetres. That is why standard deviation is the number people actually report β it sits on the same scale as the data, so “the mean is 170 cm with a standard deviation of 8 cm” is immediately readable.
Variance and standard deviation carry exactly the same information. One is the squared version of the other. The choice between them is a choice about units and about what the maths needs next, not about which is “right”.
The Units Tell the Story
| Quantity | Units (data in cm) | What it says |
|---|---|---|
| Variance | cmΒ² (square centimetres) | The spread, squared |
| Standard deviation | cm (centimetres) | The spread, in the original units |
| Mean | cm (centimetres) | The centre, in the original units |
If a number’s units do not match the data, it is a variance. If they do, it is a standard deviation (or a mean). That is the quickest way to tell them apart in a paper or a report.
Misconception 1: “Standard Deviation Is Always Smaller Than Variance”
This sounds right, because taking a square root usually makes a number smaller. It is false whenever the variance is below 1. Take a variance of 0.25: its square root is 0.5000 β the standard deviation is BIGGER than the variance. The flip happens at 1: a variance of 4 has a standard deviation of 2, which is smaller. Below 1 the square root is larger than the number; above 1 it is smaller. So the rule “SD is smaller than variance” only holds for variances above 1, and it says nothing about the data β only about the arithmetic of square roots.
Misconception 2: “Standard Deviations Add” β They Do Not
Take two independent measurements. X has variance 4 (standard deviation 2) and Y has variance 9 (standard deviation 3). Add them: the variance of X + Y is 4 + 9 = 13, so the standard deviation of X + Y is β13 = 3.6056.
The naive answer is to add the standard deviations: 2 + 3 = 5. That is wrong. Variances add; standard deviations do not. The square root of a sum is not the sum of the square roots β β13 is 3.6056, not 5. Whenever you combine independent sources of spread, add the variances first and take the square root last.
The rule: Var(X + Y) = Var(X) + Var(Y) for independent X and Y. So SD(X + Y) = β(Var(X) + Var(Y)) β never SD(X) + SD(Y).
This is the same reason covariance matters: the identity Cov = r Γ SDx Γ SDy ties the two spreads together, and the covariance calculator works out the covariance alongside the correlation for any paired data.
When to Report Each One
Report the standard deviation when the reader needs the spread in the original units β which is most of the time. A mean with its standard deviation tells a reader how far a typical value sits from the centre, on the scale of the data itself.
Report the variance when the maths needs it. Variances add (standard deviations do not), so any calculation that combines independent sources of spread β ANOVA, combining measurement errors, theoretical work β runs on variances, and the standard deviation is taken only at the end.
Population or sample? The population standard deviation Ο divides by n; the sample standard deviation s divides by n β 1. The full Ο-vs-s distinction is in the population vs sample standard deviation guide.
Do not confuse this comparison with standard deviation vs standard error: the standard error is the standard deviation of a statistic (how far a sample mean typically sits from the population mean), not a measure of the data’s spread. That comparison is covered in the standard error vs standard deviation guide.
Frequently Asked Questions
Why do we square the deviations?
Because deviations from the mean cancel if you just average them β positive and negative deviations sum to zero. Squaring makes every deviation positive, so they add up instead of cancelling, and it punishes large deviations far more than small ones. The standard deviation then takes the square root to bring the result back into the original units.
When should I report variance instead of SD?
When the maths needs it. Variances add, so any calculation that combines independent sources of spread β ANOVA, combining measurement errors, theoretical work β runs on variances. Report the standard deviation when the reader needs the spread in the original units, which is most of the time.
Do variances always add?
Only for independent variables. For independent X and Y, Var(X + Y) = Var(X) + Var(Y). If X and Y are correlated, the covariance enters: Var(X + Y) = Var(X) + Var(Y) + 2Β·Cov(X, Y). That is why the covariance matters whenever variables move together.
Check Tools
Work out the numbers on this page with the site’s free calculators:
Standard Deviation Calculator
Mean, variance and standard deviation for any dataset, step by step.
Try the standard deviation calculator β
Statistics Calculator
Summarise any dataset: mean, median, mode, standard deviation and more.
Try the statistics calculator β
Covariance Calculator
Sample and population covariance for paired data, with the correlation and a scatter plot.
Try the covariance calculator β
Explore All Statistical Tools
See every calculator on the site in one place.
Check Articles
Go deeper with these guides:
Population vs Sample Standard Deviation
The Ο-vs-s distinction: when to divide by n and when to divide by n β 1.
Standard Error vs Standard Deviation
The standard error is the spread of a statistic, not of the data β see how they differ.