Mean Deviation

When analysing a data set, knowing the average value is only half the story. Two classes might both average 70% on a test, yet one class could have scores clustered tightly between 65% and 75% while the other ranges wildly from 30% to 100%. Measures of spread quantify this variability, and the mean deviation is one of the most intuitive of these measures. It answers a simple question: on average, how far is each data value from the centre of the data set? The mean deviation (also called the mean absolute deviation, or MAD) is calculated by finding the absolute difference between each data value and a chosen centre point — usually the mean or the median — and then averaging those absolute differences. The use of absolute values is crucial: without it, positive and negative deviations would cancel out and always sum to zero, giving no useful information about spread. Unlike the range, which depends only on the two extreme values, the mean deviation uses every data point in its calculation, making it a more complete descriptor of how spread out a data set is. It is expressed in the same units as the original data, which makes it easy to interpret. While the standard deviation is more commonly used in advanced statistics, the mean deviation is an excellent introduction to the concept of measuring variability and builds important intuition for understanding why spread matters.

Key takeaways
  • Find the centre point
  • Calculate each absolute deviation
  • Sum the absolute deviations
  • Divide by the number of data values
  • Interpret the result

Find the centre point

Decide whether to measure deviations from the mean or the median. The mean is more common, but the median is preferred when the data contains outliers. Calculate whichever you choose. For a data set {x₁, x₂, …, xₙ}, the mean is x̄ = (Σxᵢ)/n.

Calculate each absolute deviation

For every data value xᵢ, subtract the centre point (mean or median) and take the absolute value: |xᵢ − x̄|. Absolute value means you ignore the sign — a deviation of −5 and a deviation of +5 are both recorded as 5. This prevents negative and positive deviations from cancelling each other out.

Sum the absolute deviations

Add all the absolute deviations together: Σ|xᵢ − x̄|. This total tells you the combined distance of all data points from the centre, but it grows larger simply because there are more data points, so we need to average it.

Divide by the number of data values

Divide the sum of absolute deviations by n (the number of data values) to get the average absolute deviation. A small mean deviation indicates that values are clustered near the centre; a large mean deviation indicates they are widely spread. The result is in the same units as the original data.

Interpret the result

Compare mean deviations across different data sets to judge relative spread. If data set A has a mean deviation of 2.4 marks and data set B has a mean deviation of 8.1 marks, data set B is considerably more spread out. Always state the units when reporting the mean deviation (e.g., '3.2 kg', '5.1 points').

The Key Equation

MD=1n∑i=1n∣xi−xˉ∣\text{MD} = \frac{1}{n}\sum_{i=1}^{n}|x_i - \bar{x}|
  • \text{MD} — Mean Deviation (same as data): The average absolute distance of each data value from the mean.
  • n — Number of data values (dimensionless): The total count of values in the data set.
  • x_i — Individual data value (same as data): The i-th observation in the data set, where i runs from 1 to n.
  • \bar{x} — Mean (same as data): The arithmetic average of all data values, equal to (Σxᵢ)/n.
  • |x_i - \bar{x}| — Absolute deviation (same as data): The non-negative distance between data value xᵢ and the mean x̄.

Worked Example

Problem. Seven students scored the following marks out of 20 on a quiz: 12, 15, 9, 18, 11, 14, 17. Calculate the mean deviation from the mean.

MD=18.297≈2.61 marks\text{MD} = \frac{18.29}{7} \approx 2.61 \text{ marks}

Answer. Mean Deviation ≈ 2.61 marks

Figure
Diagram
Mean Deviation – Absolute Deviations from the Mean020x̄ = 13.719111214151718|9−13.71|=4.71|15−13.71|=1.29|18−13.71|=4.29Each coloured segment shows an absolute deviation from the mean (red dashed line).
Origins · 1 minFrancis Galton and how this was discovered
FG
Francis Galton
1822–1911 · British

Francis Galton was a Victorian polymath who pioneered the statistical study of human variation. He developed early measures of spread and correlation while studying how traits vary within populations. His work on quantifying deviation from the average laid groundwork for descriptive statistics, including absolute deviation measures. He introduced the concept of regression to the mean and was the first to systematically apply statistical ideas to the natural sciences. Galton was so obsessed with measuring and counting everything that he once devised a 'beauty map' of Britain by secretly recording how attractive he found people in each town he visited — ranking data points wherever he went!

  • Range — The range is the simplest measure of spread, using only the maximum and minimum values. Understanding range first provides context for why the mean deviation, which uses all data points, gives a more complete picture.
  • Standard Deviation — The standard deviation is the most widely used measure of spread. It is closely related to mean deviation but squares the deviations instead of taking absolute values, making it mathematically more tractable for advanced statistics.
  • Arithmetic Mean — The mean is the reference point most commonly used when computing mean deviation. A solid understanding of how to calculate and interpret the mean is a prerequisite for this concept.
  • Median — Mean deviation can also be computed from the median, which is preferred when outliers are present. The mean deviation from the median is actually the smallest possible mean absolute deviation for any choice of centre point.
  • Interquartile Range (IQR) — Like the mean deviation, the IQR is a robust measure of spread that avoids being heavily influenced by extreme values. Comparing IQR and mean deviation helps students choose the right tool for different data distributions.