SuperDevTools

Standard deviation calculator

How numbers are read: the point is the decimal mark (12.5). Commas, spaces, semicolons and line breaks all separate values, so 1,500 is read as two values, 1 and 500. Write thousands without a separator: 1500.

σ — s —

Population (÷ n)Sample (÷ n − 1)
Variance——
Standard deviation——
Coefficient of variation——

Population: your values are the whole group you care about (every score in the class). Sample: they are only part of a larger group (30 parts pulled from a production run). Dividing by n − 1 corrects a sample's tendency to underestimate the spread.

Count n
—
Sum
—
Mean
—
Median
—
Mode
—
Minimum
—
Maximum
—
Range
—
First quartile Q1
—
Third quartile Q3
—
Interquartile range (Q3 − Q1)
—
Standard error of the mean (s ÷ √n)
—

Outliers (1.5 × IQR rule)

Histogram and box plot

Standard deviation step by step

Your last data set stays saved on this device, in this browser.

Standard deviation, variance and descriptive statistics in one place

The standard deviation tells you how spread out a set of numbers is around its mean. This calculator shows the population (σ) and sample (s) versions side by side, with the variance and other descriptive statistics (median, mode, quartiles, coefficient of variation, standard error). It flags outliers, draws a box plot and lays out the calculation line by line, handy for stats homework, an AP Statistics problem or a quality-control sheet.

How to use it

  1. In the "List of values" tab, type or paste your numbers. A column copied from a spreadsheet works as is.
  2. If your data is already grouped, switch to "Values and frequencies": one "Value" and its "Frequency" per row, and "Add a row" for more.
  3. Read σ and s at the top, then the "Population (÷ n)" and "Sample (÷ n − 1)" table.
  4. Pick the "Quartile method": "Linear interpolation (spreadsheets)" or "Nearest rank".
  5. Open "Standard deviation step by step" to see the table of deviations from the mean. With more than 30 values, the table starts collapsed.

How numbers are read: the point is the decimal mark, and commas, spaces, semicolons and line breaks all separate values. So 1,500 counts as two values, 1 and 500: type 1500 instead.

The formulas

  • Mean: x̄ = Σx ÷ n
  • Population variance: σ² = Σ(x − x̄)² ÷ n, standard deviation σ = √σ²
  • Sample variance: s² = Σ(x − x̄)² ÷ (n − 1), standard deviation s = √s²
  • Coefficient of variation: standard deviation ÷ mean, as a percentage
  • Standard error of the mean: s ÷ √n

With frequencies, each squared deviation is multiplied by its frequency f, and n is the total of the frequencies. Two passes keep it accurate with large numbers: 1000000001, 1000000002, 1000000003 gives σ = 0.8165 and s = 1.

Worked example: 2, 4, 4, 4, 5, 5, 7, 9

xx − 5(x − 5)²
2−39
4 (three times)−11 + 1 + 1
5 (twice)00 + 0
724
9416

The sum is 40, so the mean is 40 ÷ 8 = 5. The squared deviations add up to 9 + 3 + 0 + 4 + 16 = 32. Population variance: 32 ÷ 8 = 4, so σ = 2. Sample variance: 32 ÷ 7 = 4.5714, so s = 2.1381. The median is 4.5, the mode is 4 (three times) and the standard error is 0.7559.

Quartiles and outliers

With linear interpolation, Q1 = 4 and Q3 = 5.5, giving an IQR of 1.5. The 1.5 × IQR rule puts the fences at 4 − 2.25 = 1.75 and 5.5 + 2.25 = 7.75, so 9 is flagged as an outlier and shows up as a red dot on the box plot. With the nearest rank method, Q1 = 4 and Q3 = 5, the fences move to 2.5 and 6.5, and 2, 7 and 9 are all flagged. On small data sets the method matters.

A second example: seven commute times in minutes, 22, 25, 25, 28, 30, 31, 45. The mean is 29.43, σ = 6.99 and s = 7.55. The fences are 16.75 and 38.75, so the 45-minute day stands out as an outlier.

Frequently asked questions

Population or sample: which standard deviation should I use?

Use σ when you have every value you care about, such as the scores of the whole class. Use s when your values are a sample from a larger group, such as 50 customers surveyed out of thousands.

Which spreadsheet function gives the same result?

Spreadsheets have two standard deviation functions: the population one, whose name usually ends in P, and the sample one, which ends in S or has no suffix. The "inclusive" quartile function matches this tool's linear interpolation quartiles.

Why divide by n − 1?

A sample's mean is computed from those same values, so they sit a little closer to it than to the true mean. Dividing by n − 1 instead of n removes that bias from the variance.

What does a coefficient of variation of 40% mean?

The standard deviation is 40% of the mean, which is a lot of spread. It means nothing when the mean is zero or when values are both positive and negative.