Mean, Median and Mode: Which Average Should You Use?
Three numbers, three different questions
The word average casually covers three different calculations. The mean answers: if the total were shared out equally, what would each one get? The median answers: what is the value in the middle when everything is lined up? The mode answers: what value shows up most often?
All three are honest calculations, but they tell different stories — and choosing the wrong one is the single easiest way to mislead people with true numbers. A headline claiming to know what the average worker earns might be quoting a mean dragged upward by a few enormous paychecks, or a median sitting comfortably below them. One small dataset makes the difference impossible to ignore.
The three rules
The mean is the familiar one: add every value, divide by how many there are. Every value influences it, which is both its power and its weakness. Geometrically it is the balance point of the data — the value where the list would tip level on a seesaw.
The median is positional: sort the values from smallest to largest and take the one in the middle. With an even count, take the mean of the two middle values — for 4, 8, 15, 16 the middle two are 8 and 15, so the median is (8 + 15) ÷ 2 = 11.5. Extreme values can sit at the ends without moving the middle at all.
The mode is simply the most frequent value. A dataset can have no mode (nothing repeats), one mode, or several — there is no requirement that an answer exists.
mean = sum of all values ÷ number of valuesOne dataset, three answers
Five monthly salaries at a small company, in dollars: 2500, 2800, 3000, 3200 and 15000. Four employees and the owner, who pays himself 15000. Run all three averages over the same five numbers and watch them disagree.
The mean
Add the salaries: 2500 + 2800 + 3000 + 3200 + 15000 = 26500. Divide by the count: 26500 ÷ 5 = 5300.
mean = 5300
The median
Sort the values: 2500, 2800, 3000, 3200, 15000. With five values, the median is the third one — two values sit below it and two above it.
median = 3000
The mode
Look for a value that appears more than once. Nothing repeats, so this dataset has no mode. If two people earned 3000, the mode would be 3000.
no mode (no value repeats)
Why the mean lies here and the median does not
The mean of 5300 is higher than what 4 of the 5 people earn — only the owner's salary sits above it. One extreme value, called an outlier, dragged the mean 2300 dollars above the median, because every value counts toward the mean, no matter how large. That is exactly how the math is supposed to behave; it just answers a different question than the one people usually ask.
To feel how much leverage a single outlier has, drop the 15000 and recompute: the mean of the remaining four salaries is (2500 + 2800 + 3000 + 3200) ÷ 4 = 11500 ÷ 4 = 2875. One value moved the average by more than 2400 dollars.
The median ignores size and cares only about position. Change the top salary from 15000 to 150000 and the median does not move: still 3000. That stability is why the median is the honest number for salaries, house prices and anything else with a long tail of big values.
The mean is not wrong, though — it is the right tool when the total matters. The payroll is 5300 × 5 = 26500, and the mean is precisely that total split five ways. Mean for totals, median for typical values.
When to use which average
The choice depends on the data's shape and on the question being asked. Money and prices usually have outliers at the top, so the median wins there; totals and shares demand the mean; categories and popularity belong to the mode.
| Data you are summarizing | Best average | Why |
|---|---|---|
| Salaries and incomes | Median | a few very high earners drag the mean up |
| House prices in a city | Median | a handful of mansions skew the mean |
| Splitting a shared total | Mean | every value must count in the sum |
| Test scores, no outliers | Mean | uses every score equally and is stable |
| Most popular shoe size | Mode | you want the most frequent value |
| Survey answers, favorite color | Mode | categories are not quantities |
The classic misuse: the company average salary
A company with the five salaries above can honestly announce an average salary of 5300 — and give every reader a false picture of a normal paycheck. The trick runs in reverse too: a developer selling homes near a wealthy suburb may quote the median of a poorer neighboring area. The fix is not more math, it is a sharper question: when money is involved, ask for both numbers, and always ask which average was used. As a rule of thumb, a summary that involves money tends to flatter whoever benefits from the bigger figure, and the median describes the person actually in the middle.
One mechanical detail trips people up as well: the median requires sorted data, and even counts need one extra step. Drop the 2500 salary and the list becomes 2800, 3000, 3200, 15000 — four values, so the median is the mean of the two middle ones: (3000 + 3200) ÷ 2 = 3100.