Standard Deviation

What It Measures

A mean tells you where the center of a dataset sits. Standard deviation adds the missing second half of the picture: how far typical observations tend to sit from that center. Two datasets can share the same average but feel very different once their spread is considered.

This is why standard deviation appears so often in descriptive statistics, z-scores, normal distributions, confidence work, process monitoring, and repeated measurement analysis.

Population vs Sample

Use the population formula when you have every value in the full group you care about. Use the sample formula when your data are only a subset and you want to estimate how spread out the larger population may be.

How To Compute It

1. Find the mean of the dataset.
2. Subtract the mean from each value to get deviations.
3. Square each deviation so negative and positive gaps do not cancel out.
4. Add the squared deviations.
5. Divide by `N` for a population or `n - 1` for a sample.
6. Take the square root.

Squaring makes large deviations matter more. That is useful when you want a spread measure that reacts strongly to unusual observations.

Interpreting the Result

Standard deviation is always expressed in the same units as the original data. If heights are measured in centimeters, the standard deviation is also in centimeters. That makes interpretation easier than a variance value, which is expressed in squared units.

That pattern is called the empirical rule. It is useful for quick checks, but only when the data are reasonably close to a bell-shaped distribution.

A Complete Worked Example

Six students take a quiz and score 70, 74, 78, 82, 86, 90. Working through the full calculation:

1. Mean: (70 + 74 + 78 + 82 + 86 + 90) ÷ 6 = 480 ÷ 6 = 80.
2. Deviations: −10, −6, −2, +2, +6, +10. Note they sum to zero — they always do, which is why we square them next.
3. Squared deviations: 100, 36, 4, 4, 36, 100, totalling 280.
4. Variance: treating the class as the entire population, 280 ÷ 6 ≈ 46.67. Treating it as a sample of a larger group, 280 ÷ 5 = 56.
5. Standard deviation: √46.67 ≈ 6.83 (population) or √56 ≈ 7.48 (sample).

Either way, a teacher can summarize the class as "scores average 80 and typically vary by about 7 points." If these data were roughly bell-shaped, the empirical rule would put most scores between about 73 and 87 — which matches the actual list well.

Standard Deviation vs Other Spread Measures

Standard deviation is not the only way to describe spread, and choosing the right measure depends on the data:

• Range (max − min) is the simplest but depends entirely on the two most extreme values, so one typo or outlier dominates it.
• Interquartile range (IQR) describes the middle 50% of the data and ignores the tails completely, making it the robust choice for skewed data such as incomes or response times.
• Standard deviation uses every observation and connects directly to z-scores, confidence intervals, and the normal model — but it is sensitive to outliers precisely because it uses everything.

A practical habit: report mean with standard deviation for roughly symmetric data, and median with IQR when the distribution is skewed. When the two pairs tell different stories, that disagreement is itself useful information about the shape of your data.

Where It Helps

• Comparing the consistency of two groups with similar averages.
• Checking whether repeated measurements are stable.
• Creating z-scores and interpreting percentile position.
• Summarizing spread alongside the mean in reports and experiments.
• Flagging unusual observations for closer inspection.

Common Mistakes

• Using the sample formula when you actually have the full population.
• Interpreting standard deviation without checking the data shape.
• Comparing spread across datasets with very different scales without also considering the mean.
• Relying on one summary number when the data may contain strong skew or outliers.

If a dataset is highly skewed or full of extreme values, consider pairing standard deviation with medians, quartiles, or robust spread measures.

Related Reading

• Normal Distribution
• Z-Score Calculator
• Descriptive Statistics Calculator

References

• Freedman, Pisani, and Purves. Statistics.
• Moore, McCabe, and Craig. Introduction to the Practice of Statistics.
• Triola. Elementary Statistics.