Four small statistics describe most datasets well enough for everyday decisions. Mean is the arithmetic average. Median is the middle value. Mode is the most common value. Range is the spread from low to high. Each tells a different story, and the gap between them often matters more than any single one.
Key Takeaways
- Mean = sum of values ÷ count. Sensitive to outliers.
- Median = middle value when sorted. Robust to outliers.
- Mode = most frequent value. Useful for categorical and discrete data.
- Range = maximum − minimum. Simple measure of spread, ignores everything in between.
- When mean and median diverge, the dataset is skewed.
The Mean
Mean = Sum of all values / Number of values
Also called the arithmetic average. It is what most people mean when they say "average."
Example: Test scores 72, 85, 90, 78, 95. Sum: 420. Count: 5. Mean: 420 / 5 = 84.
The mean uses every value, so it reflects all the data. That is also its weakness: extreme values pull it heavily in their direction.
Two cousins of the mean appear in specific contexts:
- Weighted mean: Σ(value × weight) / Σ weights. Used when some values count more than others (course grades, portfolio returns).
- Geometric mean: the nth root of the product of n values. Used for averaging rates and ratios; investment returns are properly averaged geometrically.
The Median
The median is the middle value when the data is sorted. For an odd count, it is the single middle entry. For an even count, it is the average of the two middle entries.
Example (odd count): 72, 78, 85, 90, 95. Sorted; middle value is 85. Median = 85.
Example (even count): 60, 72, 78, 85, 90, 95. Two middle values are 78 and 85. Median = 81.5.
Median is robust; it does not move much when extreme values are added or removed. This makes it the better measure of "typical" for skewed data like income, home prices, and response times.
The Mode
The mode is the most frequent value. A dataset can have one mode (unimodal), two (bimodal), more (multimodal), or no mode if every value appears equally often.
Example: Shoe sizes sold: 8, 9, 9, 10, 10, 10, 11, 11, 12. Mode = 10.
For continuous data, the mode is rarely useful directly; instead, the most common range or bin is reported. For categorical data (favorite colors, product categories, survey responses), the mode is often the only meaningful measure of central tendency.
The Range
Range = Maximum − Minimum
The simplest spread measure. Easy to calculate, easy to interpret.
Example: Daily temperatures 62, 68, 71, 75, 80. Range = 80 − 62 = 18 degrees.
Range tells you the total span but says nothing about the distribution between the extremes. Two datasets can have the same range and very different shapes. For richer spread measurement, use variance and standard deviation (see Standard Deviation Explained).
When Each Measure Tells the Truth
Mean works well when:
- The data is symmetric (close to bell-shaped)
- No extreme outliers
- You care about totals (mean × count = sum)
Median works well when:
- Data is skewed (income, prices, wait times)
- Outliers are present
- You want the "typical" value, not the average
Mode works well when:
- Data is categorical or discrete
- You care about the most likely outcome
- Several distinct clusters exist (bimodal/multimodal data)
Range works well when:
- A quick sense of spread is enough
- The dataset is small
- You need to communicate variability to a non-technical audience
Why Mean and Median Diverge
The gap between mean and median reveals the skew of the distribution.
Right-skewed data (long tail on the high side): mean > median. Common in income, real estate, response times.
Left-skewed data (long tail on the low side): mean < median. Less common but appears in things like age at retirement, exam scores with a ceiling.
Symmetric data: mean ≈ median.
Example: Income of 9 employees in thousands of dollars: 42, 45, 48, 50, 52, 55, 60, 65, 290.
Mean: 78.6 Median: 52 Mode: none (no repeats) Range: 248
The mean is dragged up by the single high earner. The median better describes what a "typical" employee earns. Reporting the mean here would mislead; the typical employee makes $52k, not $79k.
Worked Example: Real Estate Comparison
A small town sold five houses last quarter, prices in thousands: 280, 320, 340, 380, 1,200.
Mean: 504. Median: 340. Mode: none. Range: 920.
The mean is heavily influenced by the $1.2M sale. Anyone hearing "the average home sold for $504k" would form a wrong impression. The median ($340k) is the honest description of the typical sale.
This is exactly why real estate reports use median sale price, not mean: to avoid the distortion from luxury outliers.
Comparison Table
| Statistic | Formula | Best For | Weakness |
|---|---|---|---|
| Mean | Sum / Count | Symmetric data, totals | Pulled by outliers |
| Median | Middle value | Skewed data, "typical" value | Ignores magnitudes |
| Mode | Most frequent | Categorical, discrete data | May not exist or be unique |
| Range | Max − Min | Quick spread snapshot | Sensitive to extremes |
Common Mistakes
Using mean for skewed data. Income, home prices, web traffic, response times: all are skewed. Median is almost always better.
Reporting "the average" without specifying which one. Mean, median, and mode are all averages. Be precise.
Treating range as a substitute for variability. Range only sees extremes. Standard deviation or interquartile range is usually more informative.
Computing mode for continuous data. Without binning, every value is unique, so there is no mode. Bin first, then identify the most common bin.
Forgetting that mean is sensitive to scale. Convert all values to the same units before computing.
Conflating averages with predictions. "The average is 50" does not mean any individual will be 50. Half might be 10 and half might be 90.
Practical Scenarios
Scenario 1: Performance review. A team of 10 has weekly outputs: 4, 5, 5, 6, 6, 6, 7, 7, 8, 12. Mean: 6.6. Median: 6. Mode: 6. The 12 is a high performer; the median and mode agree that typical output is 6.
Scenario 2: Customer wait times. Wait times in minutes: 1, 2, 2, 3, 4, 5, 8, 12, 18, 45. Mean: 10. Median: 4.5. Range: 44. The median tells the typical customer story; the mean is dominated by a single 45-minute wait that probably reflects a system error, not normal behavior.
Scenario 3: Survey results. Favorite color responses: red (45), blue (60), green (28), yellow (17). Mode: blue. Mean and median are not meaningful for categorical data.
Scenario 4: Investment returns. Five-year returns: +20%, −10%, +15%, +8%, −5%. Arithmetic mean: +5.6%. Geometric mean (the proper one): about +5.0%. Compounded math reduces the apparent average return slightly.
When to Report All Four
For any non-trivial dataset, reporting mean, median, and range together (and mode where it makes sense) gives the audience much more information than any single number. The pattern:
- Mean and median similar → symmetric, mean is fine
- Mean > median → right-skewed, lead with median
- Mean < median → left-skewed, lead with median
- Big range, small median-to-mean gap → spread without skew (use SD too)
- Big range and big median-to-mean gap → skewed with outliers
FAQ
What is the difference between mean and average? "Average" is colloquial and can refer to any measure of central tendency. "Mean" specifically refers to the arithmetic average (sum divided by count). To avoid confusion in writing, use "mean" or "median" rather than "average."
Can a dataset have more than one mode? Yes. Bimodal data has two modes; multimodal has more. A dataset where every value appears the same number of times has no mode.
When should I use median instead of mean? When data is skewed or has outliers: income, home prices, response times, wait times. Anywhere a few extreme values would distort the picture.
Is the mean always pulled toward outliers? Yes. The mean uses every value with equal weight, so any extreme value moves the result in its direction. The median is largely unaffected.
What does the range tell me? Only the total spread from minimum to maximum. It says nothing about the distribution between the extremes.
Can mean equal median? Yes, in perfectly symmetric distributions. The further apart they are, the more skewed the data.
How is weighted mean different from regular mean? Each value in a weighted mean is multiplied by its weight before summing, then divided by the sum of weights. Used when some values count more than others, like course grades or weighted index calculations.
Related Tools
The Mean Calculator and Median Calculator handle the basic statistics. The Statistics Calculator computes all four measures plus standard deviation and quartiles in one pass.
Related Articles
Final Thoughts
Any single number that summarizes a dataset is hiding something. Mean hides outliers; median hides magnitude; mode hides spread; range hides distribution. Used together they paint a far more honest picture than any one of them alone. The most useful habit when looking at data is to ask: which one of these would tell me the most about the question I actually care about, and is the article I'm reading using the right one?