The arithmetic mean serves as the primary measure of central location in classical statistical theory.
Mathematically, it represents the balance point of a distribution where the sum of signed deviations equals zero.
If each observation is visualised as an identical weight placed on a beam, the mean marks the exact fulcrum.
Calculating the sample mean involves summing all n observations and dividing that total sum by n.
In mathematical notation, it is denoted by placing a bar over the variable name, such as x-bar.
For example, six delivery times of ten, eleven, twelve, thirteen, fourteen, and thirty minutes yield a mean of fifteen minutes.
Notice that five of the six deliveries finished faster than the mean of fifteen minutes.
This occurs because the single thirty-minute delay exerts considerable leverage on the arithmetic total.
In statistics, the property of being vulnerable to a single extreme observation is called having a breakdown point of zero.
The mean also minimises the sum of squared errors between itself and all data points in the sample.
This least-squares property makes the mean central to regression analysis, analysis of variance, and linear modelling.
Beyond the arithmetic mean, specialised averages exist for specific data types, including geometric and harmonic means.
Geometric means multiply values and take the nth root, which suits compounding investment returns and growth ratios.
Harmonic means average rates and ratios, such as calculating the true average speed across equal travel distances.
A common mistake is reporting the arithmetic mean for highly skewed metrics like wealth, where the median gives a truer picture.