The short answer
The most common "average" is the arithmetic mean: add every number together and divide by how many there are. For 10, 20, 30, 40, 50 that's (10+20+30+40+50) ÷ 5 = 30. This calculator also computes the median, mode, geometric mean, harmonic mean, and quadratic mean (RMS), because each one answers a slightly different question about the same data.
Key takeaways
- The mean uses every value, so a single outlier can pull it far from where most of the data actually sits — the median is more resistant to extreme values.
- Growth rates and investment returns should be averaged with the geometric mean, not the arithmetic mean, or the result overstates the true compounded return.
- Rates over a fixed distance or amount — like average speed on a round trip — call for the harmonic mean, not the arithmetic mean.
- Mode is the only average defined for non-numeric (categorical) data, but if every value in a numeric set is unique, there is no mode at all.
Which average should I use?
| Situation | Best average | Why |
|---|---|---|
| Typical data, few outliers | Mean | Uses every data point equally |
| Skewed data (income, home prices) | Median | Resistant to extreme values |
| Most common item, categorical data | Mode | Only average defined for non-numeric data |
| Growth rates, investment returns | Geometric mean | Correctly compounds multiplicative change |
| Rates like speed or price per unit | Harmonic mean | Correct for averaging rates over equal amounts |
| Emphasizing larger deviations (error analysis) | Quadratic mean (RMS) | Squares amplify the effect of larger values |
Worked example: six averages, one data set
Take the numbers 4, 8, 6, 5, 3, 8. Every average type below is mathematically valid — they just weight the data differently:
| Average | Formula | Result |
|---|---|---|
| Quadratic mean (RMS) | √(Σx² / n) | 5.97 |
| Arithmetic mean | Σx / n | 5.67 |
| Median | middle of sorted list | 5.50 |
| Geometric mean | ⁿ√(x₁ × x₂ × ... × xₙ) | 5.33 |
| Harmonic mean | n / Σ(1/x) | 5.00 |
| Mode | most frequent value | 8 |
Notice the ordering: RMS ≥ mean ≥ geometric mean ≥ harmonic mean. This isn't a coincidence — it holds for every positive data set and is a useful sanity check when you're not sure whether a calculation went wrong.
Common mistakes to avoid
- Averaging percentage changes with the arithmetic mean. A -50% change followed by a +100% change does not average to +25% — $100 that drops to $50 then doubles back to $100 has a true (geometric) average change of 0%.
- Averaging rates (speeds, prices per unit) without weighting them by distance or quantity — this calls for the harmonic mean, covered in the FAQ below.
- Assuming mean and median are interchangeable. On skewed data like household income or home prices, a handful of very high values can push the mean well above where most people's actual numbers sit.
- Expecting a mode on data with no repeated values — if every number in the set is unique, there is no mode, not a mode of zero.
Related calculators
- Standard Deviation Calculator — measure how spread out your data is around the mean.
- Mean, Median, Mode & Range Calculator — a focused version of the four most-requested statistics.
- Statistics Calculator — get a full statistical summary including variance and quartiles.
- Percentage Calculator — work out percentage changes correctly before averaging them.