Skip to content
OneKitly

The Variance of a Height Is Measured in Square Centimetres

Published 11/13/2026 · 3 min read · Everyday calculators

Lena Hoffmann

Lena Hoffmann — Science & education writer at OneKitly

Mathematics · Physics

Checked against 2 sources

View profile →
In short

Variance is the average of the squared distances from the mean, so its units are the units of the data squared. For that six-number set the mean is 18 and the variance is 182, which if the numbers were centimetres would be 182 square centimetres — an area, describing a spread of lengths. Taking the square root returns 13.49 centimetres, which is the standard deviation and the number anybody can picture. Statistics keeps both because the awkward one is the one that behaves: variances of independent quantities add, and standard deviations do not.

Advertisement

The set 4, 8, 15, 16, 23, 42 has a variance of 182 and a standard deviation of 13.49. One of those numbers has a meaning you can point at.

Why squaring and not absolute value

Both cancel the signs, and absolute values are easier to explain, so the choice needs a reason. It is that squares have a derivative everywhere and absolute values do not, which makes every optimisation built on variance solvable in closed form. Least squares regression, the Kalman filter and the whole apparatus of analysis of variance exist because the quantity being minimised is smooth. The cost of that convenience is an unit nobody can visualise, and the profession pays it every day without much complaint.

The property that pays for the awkwardness

Add two independent measurements and their variances add: a component with variance 4 stacked on one with variance 9 gives a stack with variance 13, and a standard deviation of 3.61 rather than the 5 someone might expect from adding 2 and 3. That is why tolerance stacking in manufacturing is done in variances, why portfolio risk is quoted in variances before being square-rooted for presentation, and why an error budget for an instrument is a column of variances with a single square root at the bottom. Nothing else has that property, which is what keeps the unreadable unit alive.

Variance
Two ways to report the same spread
DataVarianceStandard deviation
2, 4, 6, 8, 10103.16
4, 8, 15, 16, 23, 4218213.49
Each value doubled72826.98

Worked with our own calculator

Variance calculator

Given

Numbers (comma-separated)
4, 8, 15

Result

Sample variance
31
Mean
9

These figures are produced by the calculator below, not typed in by hand — they are recomputed whenever the tool changes.

Run it on your own figures →
Advertisement

Frequently asked questions

What happens to variance if I double every value?
It quadruples, from 182 to 728, while the standard deviation merely doubles from 13.49 to 26.98. Scaling behaviour is the quickest way to remember which quantity is which: anything that changes with the square of a scale factor is living in squared units. Shifting all the values instead — adding 100 to each — leaves both unchanged, since spread does not care where the data sits, only how far apart the points are.
Should a report ever quote variance directly?
Only to an audience that is going to combine it with another variance. For everyone else the standard deviation is the honest number, because it can be compared with the mean and with the data on the same axis. Quoting 182 to a room expecting centimetres invites a silent misreading, and adding the words 'square centimetres' invites the correct question of what a squared centimetre of height is supposed to be. The good habit is to compute in variances and publish in standard deviations.

Articles you may find interesting

All guides →

Related tools

Sources

Spotted a mistake in this article?