The mean is the most common way to measure the center of a dataset. You calculate it by adding up all the values and dividing by how many values there are. If five students scored 80, 85, 90, 75, and 70 on a test, the mean score is (80 + 85 + 90 + 75 + 70) ÷ 5 = 80.
That simple calculation sits at the foundation of nearly all statistical analysis. But the mean has specific strengths, limitations, and properties worth understanding if you want to use it well.
How to Calculate the Mean
The process has two steps: sum all the values, then divide by the count. If you have a set of numbers like 12, 15, 20, 22, and 31, you add them to get 100, then divide by 5 to get a mean of 20. This works the same way whether you have five data points or five million.
In formal notation, statisticians distinguish between two versions. The population mean (written as the Greek letter μ, “mu”) describes the true average of an entire group, like the mean house value of every home in a state. The sample mean (written as x̄, “x-bar”) describes the average of a smaller subset you actually measured, like the mean house value from a sample of 200 homes. In practice, you rarely have data on an entire population, so you calculate x̄ from your sample and use it as your best estimate of μ.
Why Every Value Matters
Unlike some other measures of center, the mean is affected by every single number in your dataset. Change one value, even slightly, and the mean shifts. This makes the mean highly informative when your data is well-behaved, but it also makes it sensitive to extreme values.
A few useful properties follow from this sensitivity. The sum of the differences between each data point and the mean always equals zero. In plain terms, the values above the mean and the values below it balance out perfectly. The mean also minimizes the total squared distance from itself to every point in the dataset, which is why it serves as the starting point for so many statistical methods that rely on squared differences, like standard deviation and regression.
The mean is always unique for a given dataset (there’s only one correct answer), and it always falls somewhere between the smallest and largest values.
How Outliers Pull the Mean
Imagine five employees at a small company earn $40,000, $42,000, $45,000, $48,000, and $300,000. The mean salary is $95,000, a number that doesn’t reflect what most people at the company actually earn. That single high salary dragged the mean upward.
This is the mean’s most important limitation. In skewed data, the mean gets pulled toward the tail. When data skews right (a few very high values), the mean ends up higher than the median. When data skews left (a few very low values), the mean drops below the median. The median, which simply picks the middle value, stays planted near the center regardless of extremes. This is why you’ll often see median household income reported instead of mean income: a handful of billionaires would distort the average.
Mean, Median, and Mode Compared
These three measures of center each capture something different. The mean uses every value in the calculation. The median finds the middle value when all numbers are lined up in order. The mode identifies whichever value appears most often.
When data follows a perfectly symmetrical, bell-shaped (normal) distribution, the mean, median, and mode are all identical. In real-world data, they typically diverge. The direction they split tells you about the shape of your distribution. If the mean is noticeably higher than the median, your data likely has a right-side tail of high values. If the mean is lower, there’s probably a left-side tail dragging it down.
The Weighted Mean
Sometimes not all data points deserve equal influence. A weighted mean lets you assign more importance to certain values. The most familiar example is a college GPA: a four-credit course counts more heavily than a one-credit course. You multiply each grade by its credit hours, sum those products, then divide by the total credits.
The same logic applies in finance, where portfolio returns are weighted by the amount invested in each asset, and in energy markets, where gas prices are weighted by trading volume. Whenever you calculate an average from a frequency table, you’re computing a weighted mean, because values that appear more often contribute more to the total.
The Mean in Larger Statistical Work
The mean isn’t just a summary number. It plays a central role in one of the most important ideas in statistics: the central limit theorem. This principle says that if you take many random samples from any population and compute each sample’s mean, the collection of those sample means will form a bell-shaped (normal) distribution, regardless of how the original data was shaped. The center of that bell curve lands right on the true population mean.
This is why the mean is so powerful in practice. Even if the underlying data is messy or skewed, you can still make reliable probability statements about the mean of a large enough sample. It’s the mathematical basis for confidence intervals, hypothesis tests, and most of the inferential tools that let researchers draw conclusions from limited data.
When to Use the Mean vs. the Median
Use the mean when your data is roughly symmetrical and free of extreme outliers. It’s the best summary in that scenario because it incorporates every value and pairs naturally with standard deviation and other common statistical tools.
Use the median when your data is skewed or contains outliers. Income, home prices, and hospital wait times are classic cases where the median gives a more representative picture of what’s “typical.” You can always report both: the gap between the mean and median itself tells you something useful about how lopsided your data is.