Two students both score an average of 75. One is rock-steady: 74, 75, 76. The other is chaos: 60, 75, 90. Same mean, completely different story. The number that tells these stories apart is standard deviation — arguably the most useful statistic in everyday life.

A bell curve chart with data points and a magnifying glass

What standard deviation measures

Standard deviation (SD, or σ) measures spread: how far, on average, values sit from the mean. Small SD = data huddles near the average. Large SD = data scatters far.

The recipe:

1. Compute the mean.
2. Find each value's deviation from the mean, and square it.
3. Average the squared deviations (divide by n for a population, n−1 for a sample — the sample version is what calculators usually want).
4. Take the square root to return to original units.

Why square at all? Because deviations above and below the mean cancel out — (+15) and (−15) sum to zero, hiding the chaos. Squaring makes both positive before averaging, and the final square root brings the answer back into real units (points, dollars, centimeters).

The 75-average students, computed

Steady student: 74, 75, 76. Mean 75, deviations ±1 and 0 → SD = 1.

Chaotic student: 60, 75, 90. Mean 75, deviations −15, 0, +15 → SD = 12.2.

Same mean, SD twelve times larger. Which student would you bet on for the next exam? That's the whole point — the mean without the SD is half a story.

The 68–95–99.7 rule

For data that's roughly bell-shaped (and an enormous amount of natural data is), the SD carves up the population:

• 68% of values fall within 1 SD of the mean
• 95% within 2 SDs
• 99.7% within 3 SDs

This is why "within two standard deviations" is the everyday definition of normal. IQ has mean 100, SD 15 — so 95% of people score between 70 and 130. A manufacturing process, a farm's crop yields, your commute time: the same rule maps them all.

Where you meet SD in real life

• Finance — investment "risk" is literally the SD of returns. Same average return, bigger SD = bigger stomach required.
• Medicine — test results arrive with reference ranges defined by SDs of the healthy population.
• Quality control — a factory doesn't just want bolts averaging 10 mm; it wants SD tiny, or nothing fits.
• Sports & weather — consistency ratings, forecast confidence, all spread.

Watch the companions

SD pairs naturally with two friends:

• Variance — SD squared. Same information, squared units, loved by mathematicians.
• Mean, median & mode — if these three disagree, your data is skewed, and SD alone can mislead. Check all of them: our mean, median, mode calculator handles that in one pass.

A note on outliers

One wild value inflates SD dramatically, because squaring amplifies extremes. Income data is the classic case — one billionaire in the sample ruins every average. When data has heavy outliers, the median with a range often describes reality better than mean with SD.

Compute the spread of your own numbers: