s = √(Σ(xᵢ − x̄)² / (n − 1))
The square root of the sample variance. It is the standard deviation you work out
when the values are a sample from a larger group, and it is what the calculator’s s button and the
spreadsheet’s STDEV.S give. The formula is a definition.
It differs from the standard deviation σ only
in the denominator: n − 1 instead of n. So s is a little larger than σ for the same list, by the factor
√(n / (n − 1)), unless all the values are equal.
Worked examples
Eight values. 2, 4, 4, 4, 5, 5, 7, 9 has a sum of squares of 32. The sample variance is
32 / 7 ≈ 4.571, and s = √(32 / 7) ≈ 2.138. The population σ for the same list is 2.
The difference as a factor. √(8/7) ≈ 1.069, so s is just under 7% larger than σ. With a hundred
values the factor shrinks to about 1.005.
Common mistakes
- Reporting s when the question is about the whole population. If the values are the whole group, σ
is what is asked for.
- Rounding the variance before taking the root. √4.57 ≈ 2.138, but √4.6 ≈ 2.145. Work with the full
value and round at the end.
- Reading the wrong button on the calculator. s (sometimes σₙ₋₁) divides by n − 1, σ (sometimes σₙ)
by n.
Common questions
Which one does the calculator give?
Usually both. The button marked s, Sx or σₙ₋₁ gives this one. The one marked σ, σx or σₙ gives the
population version. In a spreadsheet, STDEV.S gives this one and STDEV.P gives σ.
Is the difference from σ large?
For small samples it shows: for eight values s is just under 7% larger. For a hundred values the
difference is about half a percent and seldom shows in a rounded answer.
Why n − 1?
Because the mean is worked out from the same sample. The sum of squares around it therefore comes out
slightly too small on average, and n − 1 in the denominator makes up for that. The page on the
sample variance explains it further.