The median is important because it gives you the most accurate picture of what’s “typical” in a dataset, especially when a few extreme values would otherwise distort the average. It’s the middle value when all numbers are lined up from smallest to largest, and that simple position makes it remarkably useful across economics, medicine, and everyday decision-making.
How One Extreme Value Breaks the Average
The mean (what most people call the “average”) adds up every value and divides by the count. That means a single unusually high or low number can drag the result away from where most of the data actually sits. The median sidesteps this problem entirely by looking only at the value in the middle position.
Here’s a classic example. Imagine five households in a small rural area earning $25,000, $30,000, $28,000, $32,000, and $1,200,000. The mean income is $263,000, a number that describes nobody in the group. The median is $30,000, which immediately tells you what a typical household actually earns. The millionaire’s income gets the same positional weight as everyone else’s: one spot in the lineup. That resistance to outliers is the median’s core strength.
Why Governments Report Median Income
When the U.S. Census Bureau and international organizations like the OECD publish income statistics, they lean heavily on the median rather than the mean. The reason is practical: income distributions are skewed. A relatively small number of very high earners pull the mean upward, making a country or region look wealthier than most of its residents actually are.
The Missouri Census Data Center puts it plainly: “an area of dire rural poverty that contained the residence of a millionaire could have a respectable average income that would not reflect the ‘typical’ income of the area.” Median household income avoids this trap. It tells policymakers and program administrators what the middle family actually brings in, which is far more useful for setting poverty thresholds, qualifying communities for aid, and tracking whether living standards are genuinely improving over time.
The OECD uses similar median-based metrics to compare material well-being across countries. Because income is skewed in roughly the same direction everywhere (a long tail of high earners), the median provides a stable, internationally comparable benchmark. Disparity ratios built on these figures show the degree of income inequality in ways that a simple national average never could.
Skewed Data and Where the Median Sits
Understanding skew makes the median’s value even clearer. In a perfectly symmetrical distribution (a classic bell curve), the mean and median are the same number. But most real-world data isn’t symmetrical. Income, home prices, hospital stays, and insurance claims all tend to have a long tail stretching in one direction.
In a right-skewed distribution, where the tail stretches toward higher values, the mean gets pulled to the right. The median stays closer to the peak, sitting between the most common value (the mode) and the mean. Moving from the peak toward the tail, the order is always mode, then median, then mean. For a left-skewed distribution, the pattern flips: the mean is pulled left of the median. This consistent positioning is what makes the median a reliable indicator of center regardless of the shape of the data.
Median Survival in Medical Research
Clinical trials measuring how long patients survive after a treatment almost always report median survival time rather than mean survival. The logic is similar to the income example: survival data is typically skewed. Most patients in a study may live a similar length of time, but a few may survive far longer (or shorter), which would distort a simple average.
Median survival tells patients and doctors the point at which half the group was still alive. That’s a concrete, interpretable number. If a cancer treatment has a median survival of 14 months, it means half the patients lived longer and half lived shorter. It’s not a prediction for any individual, but it gives a grounded sense of what to expect. Mean survival, by contrast, could be inflated by a handful of long-term survivors and paint an unrealistically optimistic picture for the typical patient.
There is a tradeoff. The median focuses narrowly on the middle of the survival curve and ignores what happens at the extremes. In trials with few events or limited follow-up periods, the median survival sometimes can’t even be calculated because fewer than half the patients have reached the endpoint. In those situations, other measures become necessary.
When the Mean Is Actually Better
The median isn’t always the right choice. Its biggest limitation is that it throws away information. The mean uses every data point in its calculation; the median only cares about position. That distinction matters whenever you need totals or aggregates.
If a company wants to know its total salary costs, the mean salary multiplied by the number of employees gives the answer. The median can’t do that. Similarly, if a government wants to estimate total tax revenue, it needs mean income figures (often weighted), not medians. The median tells you about the typical individual; the mean tells you about the whole pool.
The mean is also preferred when data follows a normal, symmetrical distribution, because in that case it captures the center more precisely and serves as the foundation for most standard statistical tests. Confidence intervals, hypothesis testing, and regression analysis all rely on means. If you’re doing formal statistical comparisons between groups, the mean is usually the starting point.
Choosing Between Them in Practice
The choice between mean and median comes down to what question you’re trying to answer and what your data looks like. If you want to know what a typical person, household, or patient experiences, and the data has a long tail in one direction, the median is almost always more honest. If you need to calculate totals, allocate resources, or work with symmetrical data, the mean is the better tool.
In many real-world reports, you’ll see both numbers presented together. The gap between them is itself informative. When median household income is $75,000 but mean household income is $105,000, that spread tells you the distribution is skewed right, with high earners pulling the average up. A large gap signals inequality; a small gap signals a more even distribution. Used together, the two measures reveal not just the center of the data but its shape.