Standard Deviation

Introduction

Averages like the mean, median, and mode tell you what a typical value in a data set looks like, but they say nothing about how consistent or variable the data actually is. Two classes can score the exact same mean on a test while one class has everyone clustered tightly around that score and the other has results scattered all over the place. Standard deviation is the tool statisticians use to measure exactly that: how spread out a set of numbers is around its mean.

Why Range and IQR Are Not Always Enough

You may already know that range (highest minus lowest) and interquartile range (IQR) both describe spread. They are useful, but they only use a couple of values from the data set — the extremes, or the middle 50%. Standard deviation instead uses every single value, comparing each one to the mean, which makes it a much more precise and widely used measure of spread in science, finance, and research.

Key Terms

• Deviation: the distance of a single value from the mean (value minus mean)
• Variance: the mean of the squared deviations — squaring removes negative signs so deviations do not cancel out
• Standard deviation: the square root of the variance, which brings the measure back into the same units as the original data

The Formula

For a full data set (a population) of n values with mean x-bar:

Variance = sum of (each value minus the mean) squared, divided by n

Standard deviation = square root of the variance

Squaring the deviations is essential — without it, positive and negative deviations would cancel out and always sum to zero. Taking the square root at the end undoes the squaring, so the final answer is measured in the same units as the original data (e.g. marks, centimetres, dollars).

Step-by-Step Method

1. Calculate the mean of the data set.
2. Subtract the mean from each value to find its deviation.
3. Square each deviation.
4. Add up all the squared deviations.
5. Divide by the number of values (n) to get the variance.
6. Take the square root of the variance to get the standard deviation.

Worked Example

Data set: 6, 8, 5, 9, 7

Step 1 — Mean: (6 + 8 + 5 + 9 + 7) ÷ 5 = 35 ÷ 5 = 7

Step 2 — Deviations: -1, 1, -2, 2, 0

Step 3 — Squared deviations: 1, 1, 4, 4, 0

Step 4 — Sum of squares: 1 + 1 + 4 + 4 + 0 = 10

Step 5 — Variance: 10 ÷ 5 = 2

Step 6 — Standard deviation: square root of 2 ≈ 1.41

So, on average, each score in this data set is about 1.41 marks away from the mean of 7.

Comparing Two Data Sets

Standard deviation is most powerful when comparing two sets with the same or similar mean.

Class A scores: 5, 6, 7, 8, 9 — mean = 7, standard deviation ≈ 1.41 Class B scores: 6, 7, 7, 7, 8 — mean = 7, standard deviation ≈ 0.63

Both classes averaged exactly 7, but Class B's much smaller standard deviation shows its scores were far more tightly clustered around that average, while Class A's results were more spread out. A teacher looking only at the mean would miss this important difference in consistency.

Common Mistakes

• Forgetting to square the deviations — without squaring, the deviations always sum to zero and tell you nothing.
• Forgetting to take the square root at the end — stopping at the variance gives an answer in the wrong (squared) units.
• Dividing by the wrong number — always divide by the total number of values, n, when working with a full population.
• Confusing standard deviation with range — range only looks at the two extreme values; standard deviation accounts for every value in the set.
• Assuming a bigger data set automatically has a bigger standard deviation — spread depends on how varied the values are, not how many there are.

Tips and Tricks

• Standard deviation can never be negative — if you calculate a negative value, check your working for an arithmetic error.
• A standard deviation of 0 means every value in the data set is identical.
• When comparing two data sets, always compare an average (mean or median) together with a spread measure (standard deviation, range, or IQR) — either alone tells only half the story.
• Calculators with statistical modes can compute standard deviation directly, which is worth practising for exam efficiency once you understand the manual method.