Standard Deviation Calculator

Paste up to 10,000 numbers: mean, median, mode, variance, standard deviation for sample and population, plus MAD, CV%, and a histogram.

Your data set never leaves your device: every statistic and the chart are computed locally in your browser.

Results

Which SD? Use the sample values (÷ n−1) when your numbers are a draw from a larger group; use the population values (÷ n) when they are the entire group. These are descriptive estimates, not professional statistical advice.

Frequency histogram (Sturges bins)

How It Works

Standard deviation is a five-step story: find the mean; measure each value's deviation from it; square those deviations so negatives stop cancelling positives; average the squares; take the square root to return to the original units. The Formula card above walks those steps with your own numbers substituted in, so the symbols on the page are the arithmetic you just ran.

Bessel's correction in plain words
Sample deviations are measured from the sample mean, which was itself chosen to minimize those squared distances — so the sum of squares is systematically too small as an estimate of the population's spread. Dividing by n−1 instead of n pays back exactly that undercharge (unbiased variance), which is why the tool always prints both: the n version describes your data, the n−1 version estimates the world behind it.
The units never change
Variance squares the units — meters become square-meters, seconds become square-seconds — which is convenient algebra but unreadable data. Taking the square root restores the original unit, so SD is always quoted in the same units as the input: a 3.2 m height dataset has an SD in metres. That is also why CV% (SD ÷ mean × 100) exists: it is the unit-free ratio you use to compare spread across datasets measured in different units.
Histogram binning
The chart uses Sturges' rule, k = ⌈log₂ n + 1⌉ bins of equal width across the range — a cheap heuristic that behaves well from about 10 to a few thousand points and never strands you with one bar or five hundred. Exact bin edges and counts are listed under the chart so you can verify every bar without hovering anything.

Frequently Asked Questions

Sample or population standard deviation — which one do I use?

Ask whether your numbers are the whole story or a sample of a bigger one. Test scores for every student in your class: population SD, divide by n. Reaction times from 30 volunteers meant to speak for all adults: sample SD, divide by n-1. When in doubt, use the sample version — it is what statistics courses, scientific papers, and most spreadsheets' STDEV function default to, and it only differs by a small factor once n grows.

Is the standard deviation robust to outliers?

No — SD is one of the least robust common statistics because deviations are squared before averaging. A single 100 in the dataset 2, 3, 2, 4, 100 pulls the mean up and dwarfs every other deviation: SD jumps to about 42 while the median absolute deviation stays at 1. That is why this page also reports MAD; when SD and MAD disagree wildly, an outlier, not the data, is driving the number.

Is standard deviation the same as standard error?

No, and mixing them up is a classic slip. SD describes how spread out your data points are. Standard error describes how uncertain your sample mean is as an estimate: SE = SD ÷ √n, so it shrinks as you collect more data while SD stays put. This tool gives you SD; for confidence intervals you need SE, which you can compute straight from the numbers shown here.

Why divide by n−1 instead of n? (Bessel's correction)

Because the deviations are measured from the sample mean, not the true population mean, and the sample mean has already been fitted to sit as close to your data as possible. That fit absorbs one degree of freedom, systematically shrinking the sum of squared deviations; multiplying by n/(n−1) undoes exactly that shrinkage, making the sample variance an unbiased estimate of the population variance.

Is anything uploaded to a server?

No. The stats, the histogram, and the downloadable summary are all computed in your browser with plain JavaScript and a canvas. Paste up to 10,000 numbers if you like — they never leave your device, and no statistical estimates here are a substitute for professional advice.