Standard deviation is one of the most important numbers in all of statistics — and one of the most misunderstood. It shows up in exam scores, financial risk analysis, quality control, scientific research, and sports analytics. Yet many students can calculate it without really knowing what it means.
This guide changes that. You'll learn not just how to calculate standard deviation, but what it tells you about your data, when to use population versus sample formulas, and how to interpret results using the famous 68-95-99.7 rule. A full worked example with a deviation table is included, along with a free embedded calculator.
Standard deviation measures how spread out your data is around the mean (average). Think of it as the "typical distance" each data point sits from the center.
Two classes both average 70 on a test. In Class A, everyone scored between 68 and 72. In Class B, scores ranged from 40 to 100. Both classes have the same mean, but very different standard deviations. Class A's small SD shows consistency; Class B's large SD shows high variability.
Standard deviation answers: "If I pick a random data point, how far from the mean can I expect it to be?" A SD of 5 on a test means most students scored within 5 points of the class average. A SD of 20 means scores are all over the place.
Variance and standard deviation measure the same thing — spread — but in different units:
| Property | Variance | Standard Deviation |
|---|---|---|
| Formula | Average of squared deviations | Square root of variance |
| Units | Squared (e.g., cm², points²) | Same as data (e.g., cm, points) |
| Symbol | σ² (population) or s² (sample) | σ (population) or s (sample) |
| Interpretation | Harder to interpret directly | Easier — same scale as data |
| Use case | Mathematical calculations, ANOVA | Everyday reporting, charts, comparisons |
Both measure spread; standard deviation is simply variance made interpretable by taking the square root.
There are two versions of the formula depending on whether your data is a complete population or a sample drawn from a larger group.
The key difference is the denominator: N for population, N−1 for sample. Using N−1 for sample data is called Bessel's correction, and it corrects for the fact that a sample tends to underestimate the true population spread.
Both formulas follow the same five steps. Here they are, clearly laid out:
Add all data points together and divide by the count (N). This is your center value, denoted μ (population) or x̄ (sample).
For each data point xᵢ, subtract the mean: (xᵢ − mean). Some will be positive (above mean), some negative (below mean).
Square each result from Step 2: (xᵢ − mean)². Squaring eliminates negative signs and amplifies larger deviations.
Sum all squared deviations, then divide by N (population) or N−1 (sample). This gives you the variance.
Take the square root of the variance. The result is the standard deviation — back in the same units as your original data.
Let's calculate the sample standard deviation for five students' test scores: 72, 85, 90, 68, 95.
| Score (xᵢ) | Deviation (xᵢ − 82) | Squared Deviation (xᵢ − 82)² |
|---|---|---|
| 72 | 72 − 82 = −10 | (−10)² = 100 |
| 85 | 85 − 82 = +3 | (+3)² = 9 |
| 90 | 90 − 82 = +8 | (+8)² = 64 |
| 68 | 68 − 82 = −14 | (−14)² = 196 |
| 95 | 95 − 82 = +13 | (+13)² = 169 |
| Sum | −10+3+8−14+13 = 0 | Σ = 538 |
This means the typical test score deviates about 11.6 points from the class mean of 82. Knowing this, a score of 93 (1 SD above mean) would stand out clearly as above average.
In a normal distribution (bell curve), standard deviation tells you exactly how data is distributed. This is called the empirical rule or the 68-95-99.7 rule:
In a normal distribution: 68% of data falls within ±1σ · 95% within ±2σ · 99.7% within ±3σ
With a mean of 82 and SD of 11.6:
| Scenario | Formula to Use | Why |
|---|---|---|
| Test scores of an entire class (you have all data) | Population SD (σ) | You have every data point |
| Polling 1,000 people to estimate national opinion | Sample SD (s) | Your data is a subset of a larger group |
| Analyzing last year's company sales (complete data) | Population SD (σ) | Complete historical dataset |
| Quality control sample of 50 items from a production line | Sample SD (s) | Sample represents a larger production run |
| Height measurements of 30 students to estimate all students | Sample SD (s) | Inferring about a larger population |
When in doubt, use sample SD (N−1). Most real-world data analysis involves samples, not complete populations.
Enter your data points below, separated by commas. Choose population or sample standard deviation. The calculator shows mean, variance, and SD instantly.
Enter numbers separated by commas (e.g. 72, 85, 90, 68, 95)
For more statistical calculations including mean, median, mode, weighted average, and geometric mean, visit our full Math & Statistics Calculator.
Standard deviation is everywhere in professional and academic work:
In finance, standard deviation measures volatility and risk. A stock with high SD swings dramatically in price; one with low SD is stable. When comparing two investments with the same average return, the one with lower SD is the safer choice. The Sharpe ratio — a key metric in portfolio management — is literally return divided by standard deviation.
Standardized tests (SAT, GRE, IQ tests) are designed with a specific mean and standard deviation in mind. IQ is calibrated to a mean of 100 and SD of 15, so a score of 115 is exactly 1 standard deviation above average. Teachers use SD to spot unusually high or low scores that warrant attention.
In manufacturing, the "Six Sigma" methodology is built on standard deviation. "Six sigma quality" means defects occur at less than 3.4 per million opportunities — because 6 standard deviations from the mean covers 99.99966% of outcomes. Companies like Motorola and GE built their quality programs on this concept.
Medical researchers report mean ± standard deviation for clinical measurements (e.g., blood pressure, drug dosage response). When a treatment's effect is more than 2 SDs from the control group's mean, the result is typically considered statistically significant.
Sports analysts use SD to measure consistency. A basketball player who scores 22 points per game with a SD of 3 is more reliable than one who averages 24 with a SD of 12. Teams seeking playoff consistency often prefer the lower-SD player despite a lower average.
Standard deviation measures how spread out data points are from the mean. A low SD means data clusters near the average; a high SD means data is widely scattered. It is the square root of variance, expressed in the same units as the data.
Population SD (σ) divides by N and is used when you have all data points. Sample SD (s) divides by N−1 (Bessel's correction) and is used when your data is a sample from a larger group. Most real-world analyses use sample SD.
There is no universal "good" SD — it depends entirely on context. Relative to the mean, a low coefficient of variation (SD/mean × 100) indicates high consistency. For exam scores, an SD of 10–15 on a 100-point scale is typical. For investments, lower SD means lower volatility and risk.
Variance is the average of squared deviations from the mean. Standard deviation is simply the square root of variance. Variance is in squared units (difficult to interpret), while SD is in the original units, making it much more intuitive for reporting and comparison.
In a normal distribution, 68% of data falls within 1 SD of the mean, 95% within 2 SDs, and 99.7% within 3 SDs. This empirical rule allows you to immediately interpret where any data point stands relative to the distribution.
Standard deviation is the language of variability. Once you understand it, you can interpret data reports, assess risk, evaluate test results, and build stronger statistical arguments. The five steps — mean, deviation, square, average, root — are simple enough to apply by hand, but our free calculator handles any dataset instantly.
Use the 68-95-99.7 rule to interpret your results, choose the right formula (sample vs population) for your context, and remember: a smaller SD means more consistency, a larger SD means more spread. That's the core insight that makes standard deviation so powerful.