Rule of thumb · MathematicsNº 117 / 167

The mean is not the middle when data is skewed

A handful of extreme values can drag the average well away from where most of the data actually sits — the median, not the mean, is the more honest "typical value" for skewed distributions like income or home prices.

Why it works

Averages are pulled toward outliers because every value contributes proportionally to the sum — the median only cares about rank order, which is exactly why it resists a few extreme values that would distort the mean.

When it fails

The median is not automatically the honest one either. It ignores magnitude entirely, so it hides exactly the tail that matters in insurance losses or system latency — where the 99th percentile is the number with consequences and the median is the one that reassures.

Do it exactly

Estimate with the rule, then check it against the calculator that models it properly.

Open Statistics & Probability →

More in Mathematics

← PreviousA long run of heads is not evidence of anything Next →A graph is just the equation, drawn

Should you use the mean or the median?

A handful of extreme values can drag the average well away from where most of the data actually sits — the median, not the mean, is the more honest "typical value" for skewed distributions like income or home prices. Averages are pulled toward outliers because every value contributes proportionally to the sum — the median only cares about rank order, which is exactly why it resists a few extreme values that would distort the mean.

Browse all 167 rules of thumb →