Trimmed Mean Calculator
Compute a trimmed (truncated) mean by dropping a percentage of the highest and lowest values — a robust average that resists outliers.
How it works
Sort the data · remove ⌊(trim% × n)⌋ values from each end · average what remains
A trimmed mean throws away a set percentage of the smallest and largest observations before averaging, which stops a few extreme outliers from dragging the result around. It is a robust compromise between the ordinary mean (which uses everything but is sensitive to outliers) and the median (which ignores nearly all the data). Trimmed means are used in Olympic scoring, price indices and any setting where a handful of extreme readings should not dominate.
Worked example
For 2, 4, 5, 6, 7, 8, 9, 100 (n = 8) a 10% trim removes ⌊0.8⌋ = 0 values, so it equals the ordinary mean of 17.6 — the 100 still dominates. A 25% trim removes ⌊2⌋ from each end, keeping 5, 6, 7, 8 for a trimmed mean of 6.5, far closer to the bulk of the data.
Sources: Truncated mean — Wikipedia
Frequently asked questions
How much should I trim?
Common choices are 5%, 10% or 20% from each end. More trimming means more robustness to outliers but discards more data; the 25% trimmed mean (the interquartile mean) is a well-known robust estimator.
Why did a small trim percentage change nothing?
The number removed is the floor of trim% × n. With few data points that can round down to zero — for example 10% of 8 is 0.8, which floors to 0 — so no values are cut until the percentage is large enough to remove at least one from each end.
Is the trimmed mean the same as the median?
Not quite. The median is the extreme case of trimming almost everything so only the one or two middle values remain. A trimmed mean keeps a band of central values, so it uses more information than the median while still resisting outliers.
Related calculators
This calculator is for educational purposes. Double-check important results before acting on them.