Measures of Dispersion

When we collect data, knowing the average value is only part of the story. Two datasets can share the exact same mean yet look completely different — one tightly clustered, the other wildly scattered. Measures of dispersion (also called measures of spread) quantify how far apart data values are from each other and from the centre of the distribution. They answer the question: 'How consistent or variable is this data?' In statistics, the most commonly used measures of dispersion are the range, the interquartile range (IQR), the variance, and the standard deviation. Each captures spread in a slightly different way and is suited to different contexts. For example, the standard deviation is ideal when data are roughly symmetric, while the IQR is preferred when outliers are present or the distribution is skewed. Understanding spread is critical in real-world decision-making. A pharmaceutical company testing two drugs may find both reduce blood pressure by the same average amount, but one drug might be far more consistent (lower spread), making it safer and more predictable. From finance to engineering to sports analytics, dispersion measures are indispensable tools for interpreting data responsibly.

Key takeaways
  • Range
  • Interquartile Range (IQR)
  • Variance
  • Standard Deviation
  • Choosing the Right Measure

Range

The range is the simplest measure of spread. It is calculated as Range = Maximum value − Minimum value. For the dataset {3, 7, 7, 12, 19}, the range = 19 − 3 = 16. While easy to compute, the range is heavily influenced by outliers and ignores all intermediate values.

Interquartile Range (IQR)

The IQR measures the spread of the middle 50% of data. First, sort the data and find Q1 (the median of the lower half) and Q3 (the median of the upper half). Then IQR = Q3 − Q1. The IQR is robust to outliers because it ignores the extreme 25% at each end of the distribution. It is the spread measure used in box-and-whisker plots.

Variance

Variance measures the average squared deviation of each data point from the mean. For a population of N values, σ² = Σ(xᵢ − μ)² / N. For a sample of n values, s² = Σ(xᵢ − x̄)² / (n − 1). We divide by (n − 1) instead of n in the sample case to correct for bias — this is called Bessel's correction. Squaring the deviations ensures negative and positive deviations do not cancel, and also gives extra weight to values far from the mean.

Standard Deviation

The standard deviation is the square root of the variance: σ = √(σ²) for a population, or s = √(s²) for a sample. Because we take the square root, the standard deviation is expressed in the same units as the original data, making it far more interpretable than variance. For instance, if data are in metres, the standard deviation is also in metres. A small standard deviation means data points are concentrated near the mean; a large standard deviation means they are spread out.

Choosing the Right Measure

Use the range for a quick, rough indication of spread. Use the IQR when data contain outliers or are skewed — pair it with the median as your measure of centre. Use the standard deviation (and variance) when data are roughly symmetric and bell-shaped — pair it with the mean. In standardised testing, finance, and science, the standard deviation is the dominant measure of spread.

The Key Equation

s=∑i=1n(xi−xˉ)2n−1s = \sqrt{\frac{\sum_{i=1}^{n}(x_i - \bar{x})^2}{n-1}}
  • s — Sample standard deviation (same as data): The measure of spread around the sample mean, in the original units of the data.
  • x_i — Individual data value (same as data): Each observation in the dataset, indexed by i from 1 to n.
  • \bar{x} — Sample mean (same as data): The arithmetic average of all n data values.
  • n — Sample size (dimensionless): The total number of observations in the sample.
  • n-1 — Degrees of freedom (dimensionless): Bessel's correction: dividing by n−1 rather than n gives an unbiased estimate of the population variance.

Worked Example

Problem. Eight students scored the following marks out of 50 on a test: 32, 35, 38, 40, 42, 44, 45, 48. Calculate (a) the range, (b) the IQR, and (c) the sample standard deviation.

Range=16,IQR=8,s=2007≈5.35\text{Range} = 16, \quad \text{IQR} = 8, \quad s = \sqrt{\tfrac{200}{7}} \approx 5.35

Answer. Range = 16, IQR = 8, Sample Standard Deviation ≈ 5.35 marks.

Figure
Diagram
Measures of Spread on a Number Line3248Range = 16Q1 = 36.5Q3 = 44.5IQR = 83248x̄ = 40.5s ≈ 5.35Data points with mean and one standard deviation marked3248
Origins · 1 minKarl Pearson and how this was discovered
KP
Karl Pearson
1857–1936 · British

Karl Pearson was a British mathematician and statistician who founded the world's first university statistics department at University College London. He formalised and popularised the concept of the standard deviation, introducing the term in 1894. He also developed the Pearson correlation coefficient and the chi-squared test, laying the systematic foundations of modern mathematical statistics. Pearson was so dedicated to statistics that he compiled and published extensive mathematical tables by hand long before computers existed, enabling scientists worldwide to perform statistical calculations for decades.

  • Measures of Central Tendency — Measures of spread are always interpreted alongside measures of centre (mean, median, mode). Together they give a complete summary of a dataset's distribution.
  • Box-and-Whisker Plots — Box plots are a graphical representation that directly displays the IQR, median, and range, making them the standard visual tool for comparing spread across datasets.
  • The Normal Distribution — The standard deviation defines the shape of the normal distribution through the 68-95-99.7 rule, connecting measures of dispersion to probability theory.
  • Outliers and Their Effects — Outliers dramatically inflate the range and standard deviation while leaving the IQR relatively unchanged, illustrating why the choice of dispersion measure matters.