What This Statistics Calculator Does
Paste or type any list of numbers — test scores, sales figures, lab measurements — and get every core descriptive statistic at once:
- Center: mean, median, mode
- Spread: range, variance, standard deviation
- Structure: count, sum, min/max, quartiles
Together, the center and spread numbers answer the two questions any dataset gets asked first: where is it, and how scattered is it?
Mean, Median, Mode: Which "Average" to Trust
They agree on symmetric data and split on skewed data — and the split is informative:
| Data: incomes | 30, 35, 38, 42, 45, 47, 250 (k) |
|---|---|
| Mean | 69.6k — dragged up by the outlier |
| Median | 42k — the typical member |
| Mode | none meaningful here |
Rule of thumb: median for skewed data (incomes, house prices), mean for symmetric data and any downstream math, mode for categories and discrete peaks. When mean ≫ median, suspect right skew or outliers — that comparison alone is a built-in data check.
Standard Deviation, Demystified
Standard deviation (σ) is the typical distance of points from the mean:
Worked example — data: 4, 7, 8, 9, 12
| Mean | 40 ÷ 5 = 8 |
| Squared deviations | 16, 1, 0, 1, 16 → Σ = 34 |
| Population σ | √(34/5) = √6.8 ≈ 2.61 |
| Sample s | √(34/4) = √8.5 ≈ 2.92 |
Why n−1 for samples: a sample's own mean sits artificially close to its points, biasing spread low; dividing by n−1 (Bessel's correction) compensates. Use sample when your data is a subset standing in for something larger — which is most real situations.
Reading σ: The Empirical Rule
For roughly bell-shaped data:
| Within | Share of data |
|---|---|
| ±1σ of the mean | ~68% |
| ±2σ | ~95% |
| ±3σ | ~99.7% |
Example: exam scores mean 74, σ = 8 → about 95% of students scored 58–90, and a 98 is a genuine outlier (3σ). This rule converts an abstract σ into concrete expectations — and flags data points worth double-checking.
Practical Notes for Clean Results
- Outliers: don't delete them silently — investigate. A typo (μ vs 10× entry) is fixable; a real extreme value is information.
- Compare spreads fairly with the coefficient of variation (σ ÷ mean) when datasets have different scales.
- Mode quirks: data can be bimodal (two peaks) — often a sign two different groups are mixed in one dataset.
- Small samples wobble: statistics from n < 10 are sketches, not portraits.
Related tools: probability questions live in the Random Number Generator's territory; score-to-grade conversions in the Grade Calculator.
The Bottom Line
Five numbers — mean, median, σ, min, max — tell you most of what any dataset has to say: where it lives, how it spreads, and whether something odd is hiding in it. Paste your numbers above, check mean-vs-median for skew, apply the 68/95/99.7 lens, and you're reading data the way analysts do.