Mean, Median and Mode: Which Average Should You Use?

“The average” sounds like a single, well-defined number, but there are three common ways to find the typical value in a set of data, and they can give very different answers. A report that quotes the wrong one can be technically correct and still misleading. This guide explains what each average measures, how to calculate it, and how to choose between them, with examples you can check in the average calculator.

Three ways to describe the middle

Take the delivery times, in minutes, of ten food orders: 18, 21, 22, 22, 24, 25, 27, 29, 31 and 88.

The mean adds every value and divides by the count. The total is 307, and 307 ÷ 10 = 30.7 minutes. It’s the answer to “if every order had taken the same time, how long would each take?”

The median is the middle value once the list is sorted. With ten values there are two in the middle, the 5th and 6th (24 and 25), so the median is their average, 24.5 minutes. Half the orders arrived faster than that and half slower.

The mode is the value that appears most often. Only 22 appears twice, so the mode is 22 minutes.

Same data, three averages: 30.7, 24.5 and 22. None of them is wrong. They answer different questions.

Why the mean is sensitive to outliers

The 88-minute order is an outlier: a value far from the rest. It accounts for 88 of the 307 total minutes, which is why the mean (30.7) is higher than eight of the ten delivery times.

Remove that one order and the mean drops to 219 ÷ 9 = 24.33 minutes, a fall of more than 6 minutes. The median barely changes: with nine values, the middle one is the 5th, 24 minutes, only half a minute lower than before.

This is the central trade-off. The mean uses the size of every value, which makes it informative but easy to drag around. The median uses only the order of the values, which makes it robust to outliers but blind to how extreme they are.

Outliers aren’t always mistakes. A genuinely late delivery matters to the customer who waited 88 minutes. The question is whether you want your average to reflect it (use the mean) or to describe the typical experience (use the median). Reporting both is often the honest answer.

Skew: reading the gap between mean and median

When data has a long tail on one side, it’s called skewed. The gap between the mean and median tells you which way.

Right-skewed data has a tail of high values, so the mean sits above the median. Incomes are the classic example. Imagine ten employees: nine earn between $38,000 and $64,000, and one earns $410,000. The mean salary is $85,600, but the median is $49,500, and nine of the ten people earn less than the mean. This is why government income statistics and home-price reports usually quote the median. Wait times, house prices and website session lengths tend to be right-skewed too.

Left-skewed data has a tail of low values, so the mean sits below the median. It’s common when there’s a ceiling. On an easy test where most students score in the 90s, one or two low scores pull the mean down. For the scores 55, 88, 90, 92, 94, 95, 96 and 98, the mean is 88.5 but the median is 93.

Symmetric data has the mean and median close together, and often the mode too. Heights of adults of the same sex, or measurement errors from a well-calibrated instrument, are roughly symmetric. In that situation it doesn’t much matter which average you quote, and the mean is usually preferred because every value contributes.

As a rule of thumb, if the mean and median differ by more than a few percent of their size, look at the data before choosing which one to report. The percentage change calculator can express that gap as a percentage if you want a quick measure.

When the mode is the right choice

The mode is the only average that works for categories. You can’t take the mean of shoe styles or the median of favorite colors, but you can find the most common one. Retailers stocking sizes, pollsters reporting the most popular answer and anyone asking “what do most people pick?” want the mode.

The mode also helps with ratings. Suppose ten customers rate a product 5, 5, 5, 5, 4, 4, 3, 1, 1 and 1 stars. The mean is 3.4, the median is 4 and the mode is 5. The mean suggests a mediocre product, but almost nobody gave it a middling score. The ratings are split between people who loved it and people who hated it. Seeing all three averages together, plus the spread, reveals that split. A single mean hides it.

The mode has limits. A list can have no mode (every value appears once) or several modes tied for most frequent. For continuous measurements like weights recorded to the gram, exact repeats are rare, so the mode often says little unless you group the values into ranges first.

Choosing an average: a quick guide

• Values are roughly symmetric with no extreme outliers: use the mean.
• The data is skewed or has outliers, and you want the typical value: use the median.
• The total matters, such as splitting a bill or estimating a budget: use the mean, because mean × count = total. The median doesn’t have that property.
• The data is categories or you want the most common choice: use the mode.
• The data is ratings or rankings where the gaps between levels aren’t necessarily equal: the median and mode are safer than the mean.
• You’re unsure: report the mean and the median, and say how far apart they are.

Other averages worth knowing

Weighted mean. When some values count more than others, multiply each by its weight before adding, then divide by the total weight. Course grades are usually weighted: a final exam worth 50% counts more than homework worth 20%.

Trimmed mean. Drop a fixed share of the highest and lowest values, then take the mean of the rest. Removing the fastest and slowest delivery above (18 and 88) leaves eight values totaling 201, a trimmed mean of 25.125 minutes. Trimmed means are used in some judged sports and in certain inflation measures because they resist outliers while still using most of the data.

Geometric mean. For growth rates that compound, such as investment returns, the arithmetic mean overstates performance. A gain of 50% followed by a loss of 50% averages to 0%, but $100 becomes $150 and then $75. The geometric mean of the growth factors, √(1.5 × 0.5) ≈ 0.866, shows the true typical change of about −13.4% per period. The guide on how to calculate percentage change covers compounding in more detail.

Common mistakes

• Averaging averages. If two groups have different sizes, the average of their means isn’t the overall mean. Combine the raw totals and counts instead.
• Calling the median “the average” without saying so. If you report a median, label it. Readers usually assume “average” means the mean.
• Removing outliers without a reason. Drop a value only if it’s a genuine error, such as a typo. Otherwise, report the median alongside the mean rather than quietly deleting data.
• Ignoring spread. The lists 49, 50, 51 and 0, 50, 100 have the same mean and median. Quote the range or another measure of spread so readers know how much values vary.
• Using the mean for ratings. A 3.4-star average can describe a product almost no one rated 3 stars.

Checklist

1. Sort the data and look at it before choosing an average.
2. Compare the mean and median. A large gap signals skew or outliers.
3. Choose the mean when totals matter or data is symmetric, the median for skewed data, and the mode for categories.
4. Report the spread, such as the range, alongside the average.
5. Label which average you used.

Paste your own numbers into the average calculator to see the mean, median, mode and range side by side. For more everyday math tools, browse the math calculators.