Find the median of any list — sorted, middle-picked, outlier-proof, with steps shown.
House prices, salaries, and wait times are summarized by the median — the middle value — because one billionaire or one disaster shouldn’t rewrite the typical story. Our free Median Calculator sorts any list, picks the middle (or averages the two middles), and shows the sorted list so the answer is verifiable.
The median is the outlier-proof average: in {1, 2, 3, 4, 100}, the mean is 22 but the median is 3 — and 3 describes the group far better. Whenever “average” feels misleading, the median is usually the number actually wanted.
Enter numbers separated by commas, spaces, or semicolons and get the median with the sorted list and a mean-for-comparison. Analysts, homebuyers, and students get the robust middle with its working shown.
To find the median: sort the values, then take the middle. With an odd count, it’s the single center value — in {1, 1, 2, 3, 4, 5, 9}, the 4th of 7 is 3. With an even count, average the two middle values — in {1, 2, 3, 4}, that’s (2+3)/2 = 2.5. Sorting first is non-negotiable; the “middle” of unsorted data is meaningless.
The median’s superpower is resistance: change the largest value to anything larger and the median doesn’t budge. The mean chases every outlier; the median ignores them. That’s why “median home price” and “median income” are the standard headlines — they describe the typical case.
Median vs. mean also diagnoses skew: if the mean exceeds the median, the tail stretches right (a few huge values); if it trails, the tail stretches left. The widget reports both and flags notable disagreement — a one-line skewness test.
Key terms:
No sorting, no median — the middle of unsorted data is just a random element. The widget always shows the sorted list as proof.
Find the median of 3, 1, 4, 1, 5, 9, 2.
Note the mean is 3.57 — close here, because this data isn’t badly skewed.
Salaries (thousands): 45, 52, 48, 51, 500. Median?
The mean is 161.25 — the median’s 50 tells the true typical story. This is exactly why salary statistics use medians.
Fatima researches a neighborhood: recent sales are 310, 325, 299, 340, 315, and 1,200 (thousand). The mean is 465 — alarming. The median: sorted 299, 310, 315, 325, 340, 1200 → (315+325)/2 = 320. The typical home costs 320k; one mansion skewed the mean. She budgets to the median.
A second scenario: a support team’s ticket resolution times (hours): 1, 2, 2, 3, 48. Mean: 11.2 h — looks bad. Median: 2 h — typical. The 48-hour outlier was a vendor delay; the median tells the team’s real story and the manager investigates the one ticket, not the team.
A third: a teacher’s quiz scores: 55, 68, 72, 75, 78, 81, 84. Median 75, mean 73.3 — close, symmetric class. The agreement itself is information: no curve needed.
A fourth: a runner’s mile times over two weeks: seven values, median 8:12/mile. One disastrous 11:40 (a twisted ankle) doesn’t move it — the median tracks fitness, not mishaps.
Real estate: “median home price” is the industry standard because a few luxury sales would corrupt the mean. Every housing report leads with the median.
Economics: median income describes the typical worker; mean income describes the total divided by heads — different stories, and the gap between them measures inequality.
Medicine: median survival time in trials — half the patients live longer. Means would let a few long survivors distort the typical prognosis.
Tech: latency percentiles are medians’ cousins — p50 is the median response time. “Typical” latency is a median; the mean is polluted by rare timeouts.
Sports: median salaries describe the typical player better than means inflated by superstars — unions quote medians in negotiations.
Quality control: median defect sizes and lifetimes resist the occasional catastrophic outlier that would warp a mean-based spec.
Everyday judgment: “how long does it usually take?” is a median question. The mean answers “what’s the total divided by trips?” — rarely what was asked.
One pattern worth internalizing: when the data can surprise you, median first. The mean is for well-behaved data; the median is for the real world.Home affordability: price-to-median-income ratios define “affordable” — the median household, not the mean, is the policy target.
Transit: median commute time describes the typical rider; the mean is polluted by a few epic commutes. Planners build for the median.
Streaming: median watch time per title — half the viewers watch longer. The mean would let binge-watchers distort “typical.”
Startups: median time-to-funding, median seed round — founders benchmark against the middle, not the unicorn-inflated mean.
Ecology: median species lifespan vs. mean — a few ancient bristlecones don’t make the “typical” tree 5,000 years old.
The median’s democratic virtue: the median is the only “average” that a majority can agree describes them — half the data sits on each side by construction. This is why democracies’ statistics use it: median income is the income of the middle voter, the number that elections turn on. The mean, by contrast, lets the top 1% outvote the bottom 99 arithmetically. When Piketty-era inequality debates cite “average wealth,” the median-vs-mean choice is the entire argument: America’s mean household wealth dwarfs its median because wealth is brutally right-skewed. The widget reports both and flags disagreement — a one-line inequality detector hiding in a calculator.
Worked scenario — the latency SLA: a SaaS team promises “average response under 200 ms.” Their measurements (ms): 120, 140, 150, 160, 180, 200, 220, 8000. The mean is 1,146 ms — SLA blown, panic. The median is 170 ms — typical experience fine; one 8-second outage (a deploy gone wrong) poisoned the mean. The team rewrites the SLA around percentiles — p50 (median) under 200 ms, p99 under 2 s — and the on-call rotation stops getting paged for arithmetic artifacts. The median didn’t hide the outage (the range still shows 8000); it correctly assigned it to the tail instead of the typical. Measure the typical with the median; measure the tail with percentiles; never let one number do both jobs.
Edge cases and pro tips: with an even count, the median averages the two middle values — which can produce a value no observation holds (median of {1,2} is 1.5); that’s fine, the median is a statistic, not a sample. For tiny samples (n < 5), report the raw data alongside — any “typical” claim from three numbers is fragile. The median of a bimodal distribution (two humps) sits in the valley between them, describing neither — plot the data when the range looks suspicious. And remember the median’s breakdown point: it resists up to 50% contamination, the best possible — which is exactly why it’s the default “typical” for hostile, real-world data.
Median: sort values; odd count = middle value, even count = average of two middles. Outlier-resistant; mean-median gap diagnoses skew. Source: Sigma Calculator, Median Calculator, https://sigmacalculator.net/math-calculators/median-calculator/
Cite this page as: Median Calculator, Sigma Calculator, https://sigmacalculator.net/math-calculators/median-calculator/
Last updated: October 12, 2026