Mean vs Median vs Mode: Definitions, Formulas, and When to Use Each
Guide · Updated
The mean is the arithmetic average (sum divided by count), the median is the middle value when data is ordered, and the mode is the most frequently occurring value. While "average" colloquially refers to the mean, the median is often more representative of typical values in datasets with extreme outliers, such as salary distributions where one executive salary can inflate the mean far above what most workers earn.
Definitions and Basic Formulas
Mean, median, and mode are the three primary measures of central tendency—numbers that represent the center or typical value of a dataset. Each answers the question 'What is typical?' but in a different way.
The mean (or arithmetic average) is calculated by summing all values and dividing by the number of values: Mean = (Sum of all values) / (Number of values). The median is the middle value when all observations are arranged in order from smallest to largest. For an odd number of values, it is simply the middle number; for an even number, it is the average of the two middle values: Median = ((n+1)/2)th value for odd n, or average of (n/2)th and (n/2+1)th values for even n. The mode is the value that appears most frequently in the dataset; a dataset can have one mode, multiple modes, or no mode at all if all values appear equally.
These three measures often differ, especially when data is skewed or contains outliers. Understanding which to use depends on the nature of your data and what question you're trying to answer.
| Measure | What it is | Example: 2, 3, 3, 6, 11 |
|---|---|---|
| Mean | Sum Ă· count (the average) | (2+3+3+6+11) Ă· 5 = 5 |
| Median | Middle value when sorted | 3 |
| Mode | Most frequent value | 3 |
Worked Example: Company Salary Data
Consider a small company with 10 employees earning the following annual salaries: $32,000, $35,000, $38,000, $42,000, $45,000, $48,000, $51,000, $54,000, $58,000, and $520,000 (the CEO). Let's calculate each measure.
Mean: ($32,000 + $35,000 + $38,000 + $42,000 + $45,000 + $48,000 + $51,000 + $54,000 + $58,000 + $520,000) Ă· 10 = $923,000 Ă· 10 = $92,300. This figure grossly misrepresents the typical employee salary.
Median: When ordered, the 10 salaries sit between the 5th value ($45,000) and 6th value ($48,000). Median = ($45,000 + $48,000) Ă· 2 = $46,500. This is far more representative of a typical employee's pay.
Mode: Looking at frequency, each salary appears exactly once, so this dataset has no mode. In cases where one salary tier is repeated (e.g., multiple interns at $32,000), that value would be the mode.
This example shows why the median is often called the 'typical' value—it splits the data in half and is unaffected by the CEO's extremely high salary. The mean, pulled upward by that single outlier, makes the company's pay structure look far more generous than it truly is for most workers.
How Outliers Affect Each Measure
Outliers are extreme values that differ greatly from most observations. According to NIST's statistical handbook, outliers have dramatically different effects on mean, median, and mode. The mean is highly vulnerable: extreme values in the tails of a distribution pull the mean upward or downward, distorting the estimate of location. Conversely, the median is robust against outliers because it depends only on the rank position of values, not their magnitude. Whether one value is 10 or 10 million, its effect on the median is identical—it occupies the same position in the ordered list.
The mode is entirely unaffected by outliers because a single extreme value is, by definition, infrequent. An outlier will never be the most common value in a dataset, so it cannot change which value is the mode.
For distributions with extreme values, NIST concludes that 'the median provides a better estimate of location than the mean.' This is why financial analysts, economists, and demographers often report median income, median home price, and median rent—these statistics tell the story of what a typical person experiences, not what a handful of billionaires or luxury properties distort the average to suggest.
When to Use Each Measure
Use the mean when your data is roughly symmetric with no extreme outliers—such as test scores in a class, heights in a population, or daily temperature fluctuations. The mean is easy to calculate and interpret, and it uses all the information in your dataset.
Use the median for skewed distributions, data with outliers, or when the 'typical' experience matters more than the arithmetic average. Real-world examples include salaries, house prices, medical costs, and income distribution, where a few very high (or very low) values are common and don't represent the majority experience.
Use the mode when you need to know the most common value, especially with categorical data (What's the most popular color? What's the most frequent shoe size?). Mode is also useful for identifying which value or range occurs most in your dataset, complementing the other measures. Our Average Calculator tool can compute mean, median, and mode instantly for any dataset, and our Percentage Calculator helps contextualize how typical values represent parts of a whole.
The Word 'Average' and Common Misconceptions
In everyday speech, 'average' almost always means the arithmetic mean—the sum divided by count. However, in statistics, 'average' can technically refer to any central tendency measure. This ambiguity is the source of much confusion and, sometimes, unintentional (or intentional) manipulation in media and marketing.
A common misconception is that the mean and median are interchangeable. They are not. When you hear a news headline stating 'Average American worker earns $55,000,' verify whether they mean the mean or median. For income data, if the median is significantly lower than the mean, it signals that high earners are pulling the mean upward, and the median better reflects what most people actually earn.
Another misconception: 'There's always a mode.' In fact, when all values occur with equal frequency (as in our salary example), there is no mode. It's important to check for this when analyzing data, as the absence of a mode tells you something meaningful about your distribution.
Frequently asked questions
What does 'average' really mean?
In common usage, 'average' refers to the mean (sum of values divided by count). However, in statistics, the term can refer to mean, median, or mode. Always ask which measure is being used when someone cites an 'average,' especially for sensitive topics like income or prices.
Why is the median often reported for salary and housing data?
Because salary and housing price data contain outliers (very high earners, luxury homes) that skew the mean upward. The median tells you what a typical worker or homebuyer experiences, making it more representative and honest than the mean for skewed distributions.
Can a dataset have more than one mode?
Yes. If two or more values appear with the same highest frequency, the dataset is bimodal (two modes) or multimodal (more than two). For example, in a survey asking favorite colors, both red and blue might be chosen by 15 people each, while all other colors are chosen fewer times—making both red and blue modes.
How do I choose between mean and median?
Use the mean for symmetric data with no outliers. Use the median for skewed data or data with extreme values. If the mean and median are very different, suspect outliers and investigate your data further.
Does an extreme value change the mode?
No. By definition, an outlier is infrequent, so it cannot be the most common value. The mode remains unchanged regardless of how extreme any single value is, making it the most robust measure to outliers.
What if my dataset has no mode?
This happens when every value appears exactly once (or when all values appear with equal frequency). In such cases, report that there is no mode rather than forcing one. It tells you your data distribution is uniform rather than concentrated at any particular value.
Try the tools
Sources & references
This guide is general information to help you understand the topic and use the tools — it is not professional (financial, medical, legal, or tax) advice. Verify anything important before relying on it. See our Disclaimer.