Mean, median, and mode are three ways to find the “center” of a set of numbers. The mean is the average, the median is the middle value, and the mode is the value that appears most often. Each one tells you something slightly different about your data, and knowing when to use which one is just as important as knowing how to calculate them.
How to Calculate the Mean
The mean is what most people think of as the average. Add up all the values in your data set, then divide by how many values you have. If your five test scores are 78, 82, 90, 85, and 95, you’d add them to get 430, then divide by 5 to get a mean of 86.
The mean uses every single number in the calculation, which makes it powerful but also sensitive. One unusually high or low value can pull the mean away from what feels like the “typical” number. This is why the mean has a reputation for being misleading when data is lopsided, something we’ll get into below.
How to Calculate the Median
The median is the middle value when you line up all your numbers from smallest to largest. If you have an odd number of data points, it’s simply the one sitting in the center. With the scores 78, 82, 85, 90, 95, the median is 85 because two values fall below it and two fall above it.
If you have an even number of data points, there’s no single middle value. In that case, you take the two values closest to the center and average them. So for the set 78, 82, 85, 90, the median would be halfway between 82 and 85, which is 83.5.
The key thing about the median is that it only cares about position, not size. It measures where the middle of your data falls without being dragged around by extreme numbers at either end.
How to Calculate the Mode
The mode is the value that shows up most frequently. In the set 4, 7, 7, 9, 12, the mode is 7 because it appears twice while everything else appears once. If no value repeats, there is no mode. If two values tie for the most appearances, you have two modes (called a bimodal data set). Some data sets have even more than two modes.
The mode stands apart from mean and median in one important way: it’s the only measure of central tendency you can use with non-numerical data. Imagine a survey asking people their favorite color. You can’t average “blue” and “green,” and there’s no way to find a middle value between them. But you can count which answer came up most often. That’s the mode, and for categories like colors, job titles, or performance ratings (“needs improvement,” “meets standards,” “exceeds standards”), it’s your only option for summarizing what’s typical.
Why Outliers Change Everything
Outliers are values that sit far away from the rest of your data, and they affect the mean far more than the median. A Khan Academy example illustrates this clearly: a golfer’s scores included one round of 80 alongside several rounds in the low-to-mid 90s. With the 80 included, the mean dropped to 90.4. When that score was removed, the mean jumped to 93, an increase of 2.6 points. The median, by contrast, only shifted from 92 to 93, a change of just 1 point.
This happens because the mean factors in the actual size of every number. One extreme value gets blended into the total and pulls the average toward it. The median just looks at which value sits in the middle of the lineup, so swapping out an extreme number at the edge barely moves it. The more extreme the outlier, the bigger the gap between what the mean and median tell you.
When to Use Each One
The choice between mean, median, and mode depends on the shape of your data and what kind of information it contains.
The mean works best when your data is roughly symmetrical, with no extreme values pulling things to one side. It’s the most commonly used measure in statistics because it has mathematical properties that make it easier to work with in further analysis. If you’re looking at something like daily temperatures over a month or exam scores in a class where nobody bombed or aced it dramatically, the mean gives you a reliable picture of the center.
The median is your better choice when data is skewed or contains outliers. The classic example is household income. In 2016, the median household income in the United States was $59,039. If you used the mean instead, a small number of extremely high earners (think tech executives and hedge fund managers) would drag the average well above what a typical household actually brings in. The median resists that pull, which is why economists and the Census Bureau report it as the standard measure of “typical” income.
The mode is most useful for categorical data or when you want to know the single most common outcome. It’s also the fastest to find in a large data set since you’re just looking for the value with the highest frequency. In smaller numerical data sets, the mode often isn’t very informative because every value might appear only once. But in large data sets, the mode, median, and mean tend to cluster close together, so the mode can serve as a quick approximation.
How They Shift in Skewed Data
When data is perfectly symmetrical (a bell curve), the mean, median, and mode all land in the same spot. Things get more interesting when the data is skewed, meaning it has a longer tail stretching out in one direction.
In a right-skewed distribution (where the tail stretches toward higher values), the mean gets pulled furthest to the right, the mode stays at the peak, and the median falls between them. Income data is a textbook example: most households cluster in a middle range, but a long tail of very high earners stretches to the right, pulling the mean above the median.
In a left-skewed distribution (tail stretching toward lower values), the pattern reverses. The mean gets pulled furthest to the left, the mode stays at the peak, and the median sits between the two. The consistent rule is that the mean always chases the longer tail. Moving from the peak of the data outward toward the tail, the order is always mode, then median, then mean.
This relationship is useful as a quick diagnostic. If someone tells you the mean of a data set is noticeably higher than the median, you immediately know the data skews right, likely because of some large values pulling the average up.
A Quick Reference
- Mean: Add all values, divide by the count. Best for symmetrical data with no extreme values. Sensitive to outliers.
- Median: The middle value when data is ordered. Best for skewed data or when outliers are present. Resistant to extreme values.
- Mode: The most frequently occurring value. The only option for categorical (non-numerical) data. A data set can have zero, one, or multiple modes.