Mean vs Median: What's the Difference and When to Use Each
By
Liz Fujiwara
•
The single value you report changes everything. A startup claiming an average salary of $180,000 might sound competitive, but that figure can hide the reality that two co-founders earning $2 million each are pulling the average upward while most workers make around $85,000. Both the mean and the median are measures of central tendency used across school statistics, business dashboards, and policy reports. They both attempt to describe what is “typical” in a data set, yet they can tell vastly different stories depending on the distribution.
If you’re a founder reviewing compensation benchmarks, a manager analyzing customer data, or a student interpreting survey results, understanding the difference between mean and median matters. The rest of this article walks through simple formulas, step-by-step examples, and concrete datasets to make the distinction clear.
Key Takeaways
The mean is the arithmetic average calculated by summing all values and dividing by the number of values, while the median is the middle value when data is arranged in ascending order.
The mean is highly sensitive to extreme values, such as a $5 million salary, which can skew results, whereas the median remains stable and often better represents a typical value.
In symmetric distributions like adult heights, mean and median are nearly identical, but in skewed distributions like 2026 house prices or startup valuations, they can differ significantly, affecting how readers interpret income reports, real estate listings, and business metrics.
What Is the Mean?
The mean is the arithmetic average of a dataset, found by summing all data values and dividing by the total number of observations.
The general formula looks like this:
Mean = (x₁ + x₂ + … + xₙ) / n
Where each x represents an individual value and n is the count of values.
Example: Consider test scores of 72, 80, 88, 90, and 95.
Sum: 72 + 80 + 88 + 90 + 95 = 425
Divide by count: 425 / 5 = 85
Conceptually, the mean is the “balance point” of your data. If you placed all the numbers on a seesaw, the mean is where it would balance perfectly, with the positive and negative deviations from the average value canceling out to zero.
Because the mean uses every single value in its calculation, it is extremely sensitive to outliers. One $1,000,000 income in a list of $50,000 salaries will dramatically pull the arithmetic mean upward.
What Is the Median?
The median is the middle value when all data points are arranged in ascending order (or descending order; the result is the same).
How you find it depends on whether you have an odd number or even number of observations:
Odd case: The median is the middle one directly.
Dataset: 12, 15, 18, 21, 24 (5 values)
Middle position: 3rd value = 18
Even case: The median is the average of the two middle values.
Dataset: 12, 15, 18, 21 (4 values)
Two middle numbers: 15 and 18
Median: (15 + 18) / 2 = 16.5
Notice that 16.5 doesn’t even appear in the original data set; that’s perfectly normal for even-numbered datasets.
The median depends only on order, not exact magnitudes. This makes it robust to extreme values. For skewed data like 2026 home prices in major cities, the median value often gives a better sense of what a “typical” buyer actually pays.
Mean vs Median
The following table summarizes the core differences to help you quickly determine which measure suits your needs.
Aspect | Mean | Median |
Definition | Arithmetic average of all the numbers | Middle number in sorted data |
Formula | Sum ÷ count | Middle score (or average of two middle values) |
Sensitivity to outliers | Highly sensitive and pulled by extreme values | Robust and unaffected by other values at extremes |
Best for | Symmetric, normal distribution data | Skewed data with outliers |
Example dataset: 3, 3, 4, 5, 50 | Mean = 65/5 = 13 | Median (3rd value) = 4 |
Real-world use | Heights of adults, exam scores | Mean income reports, house prices, net worth |
Behavior in symmetric data | ≈ same as median | ≈ same as mean |
Behavior in right skewed distribution | Pulled toward tail | Stays near center |
In the example row, notice how the outlier 50 drags the mean to 13, more than triple the median of 4. This single example shows why the most appropriate measure depends entirely on your data’s shape.
How to Calculate Mean and Median Step by Step
Let’s walk through two complete examples to reinforce the calculations.
Classroom Test Scores
Dataset: 65, 70, 78, 80, 84, 90, 95 (already in ascending order)
Calculating the mean:
Sum all the values: 65 + 70 + 78 + 80 + 84 + 90 + 95 = 562
Divide by count: 562 / 7 = 80.3
Calculating the median:
Count observations: 7 (odd)
Find the middle position: (7 + 1) / 2 = 4th value
Median = 80
With roughly symmetric scores, the mean and the median are nearly identical.
Monthly Incomes (Skewed)
Dataset: $30k, $32k, $35k, $38k, $40k, $45k, $200k
Calculating the mean:
Sum: 30 + 32 + 35 + 38 + 40 + 45 + 200 = 420
Divide: 420 / 7 = $60k
Calculating the median:
Already sorted, 7 values
4th value = $38k
The $200k outlier pushes the mean 58% higher than the median. If you reported “average income is $60k,” you’d mislead most people who actually earn around $38k.
Try these calculations yourself in a spreadsheet using AVERAGE and MEDIAN functions to reinforce understanding.
Outliers and Skew: When Mean and Median Tell Different Stories
An outlier is any value far from the rest; think of a $5 million exit among seed investments that are mostly $5k to $50k. Skewed data means the distribution has a long tail in one direction.
Example: Consider the dataset 30, 35, 40, 45, 50, 55, 1000.
Mean: 1255 / 7 = 179.3
Median: 4th value = 45
The mean is nearly four times the median because of that single 1000.
What a right-skewed histogram looks like:
Most data clusters on the left
A long tail extends to the right
The mean sits closer to the tail
The median stays near the peak where most people fall
In left-skewed data (e.g., exam scores where most students scored high but a few people bombed), the mean gets pulled toward the low tail while the median remains more representative of the majority.
This is why analysts, policymakers, and business leaders commonly report medians for incomes, rents, and house prices. The median income better reflects what most workers earn, not what a few billionaires bring in.
Choosing Between Mean and Median in Practice
Use this practical checklist to determine which measure fits your situation:
Scenario | Recommended Measure | Why |
Employee salaries at a small startup | Median | A few high earners (founders) skew results |
Customer satisfaction ratings (1-5 scale) | Mean | Bounded scale, typically symmetric |
Server response times with occasional delays | Median (or both) | Rare extreme delays inflate the mean |
Exam scores with normal distribution | Mean | Symmetric data, mean = median anyway |
2026 housing prices in a major city | Median | Luxury properties create right skew |
For large, roughly normal datasets, such as the heights of thousands of adults measured in recent surveys, the mean and median are almost identical, so the arithmetic mean is typically used for statistical analysis.
Best practice is to report both the mean and median when possible, especially in public reports, as this provides a more complete picture and signals when the distribution may be skewed.
No single value is “best” in all cases, and the right choice depends on the question you are asking and the shape of the underlying data.
Conclusion
The mean is the arithmetic average sensitive to every value in your dataset, while the median is the middle value that is less affected by outliers and skew. Real-world decisions, such as interpreting market reports, evaluating salary offers, or comparing numerical data across groups, change dramatically depending on which measure you use.
Take a look at your own data this week. Whether it is your team’s salary bands or product metrics from last quarter, compute both measures to see how they differ, as the gap between them reveals important information about your distribution.
Next topics to explore include the mode, which is the most frequent value, how mean, median, and mode work together, trimmed means that remove extreme values, and visual tools like box plots, bar charts, and histograms to better understand your data in data science contexts.
FAQ
Why are mean and median sometimes the same value?
Which is better for describing average income: mean or median?
Can the median be a value that is not actually in the dataset?
Is the mean always more accurate than the median?
Do I need special software to compute mean and median?
