Distinction · Statistics
Mean, Median and Mode: Which Average to Use
Three numbers all called the average, answering three different questions. Choosing wrongly is how statistics ends up misleading.
Three averages, three questions
Mean, median and mode are all called averages, but they measure different things. The mean balances the values, the median splits them into two halves, and the mode is whichever value occurs most often. On symmetric data they land close together; on skewed data they separate, and that separation is itself information.
| Measure | What it answers | Weak when |
|---|---|---|
| Mean | where is the balance point? | there are extreme outliers |
| Median | what is the middle value? | you need every value to count |
| Mode | what happens most often? | values rarely repeat |
Calculating each one
For the data set 4, 8, 9, 11 the three measures come out as follows.
- (4 + 8 + 9 + 11) ÷ 4 = 8the mean: sum divided by how many
- 4, 8, 9, 11sort the data before finding the median
- (8 + 9) ÷ 2 = 8.5the median: mean of the two middle values
- no modeno value repeats
Two details matter here. The median requires the data to be sorted first — taking the middle of an unsorted list gives a number with no meaning. And when there is an even count, the median is the mean of the two central values, not one of them.
When the mean misleads
The mean uses every value, which is usually a strength and occasionally a serious weakness. One extreme value drags it a long way.
Take five salaries: 30, 32, 35, 38 and 400 thousand. The mean is 107 thousand, which describes nobody in the room — four people earn far less and one earns far more. The median is 35 thousand, which describes the typical case accurately.
When the mode is the only sensible choice
The mode is the only average that works on data that is not numeric. The most common shoe size, the most frequent response on a survey, the best-selling colour — none of these can be summed or ordered meaningfully, but they can be counted.
A data set can also have two modes, or more. If two values tie for most frequent the set is bimodal, and reporting only one of them hides half the pattern. If every value occurs once, there is no mode at all, and inventing one describes a pattern the data does not have.
Reading the gap between mean and median
Comparing the two says something about the shape of the data before you plot anything.
- Mean above the median: the data is skewed towards high values, with a long right tail.
- Mean below the median: it is skewed towards low values.
- Mean and median close together: the data is roughly symmetric.
Averages alone are not enough
Two data sets can share a mean and be nothing alike. 49, 50, 51 and 0, 50, 100 both average 50, but one is tightly clustered and the other is spread across the whole range.
That is what the range and standard deviation are for. An average without a measure of spread is half a description, and the difference between the two data sets above is the half that was left out.
Questions about mean median and mode
When should I use the median instead of the mean?
Whenever the data contains outliers or is visibly skewed. One extreme value pulls the mean towards it while the median stays where most of the data actually is.
Can a data set have more than one mode?
Yes. If two or more values tie for most frequent, the set is bimodal or multimodal and all of them should be reported.
What if no value repeats?
Then there is no mode. That is a valid result, and naming one anyway implies a pattern the data does not contain.
Is the mean always the best average?
No. It is the most informative when the data is symmetric and free of outliers, and the most misleading when it is not. The right choice depends on the shape of the data, not on habit.