About the Trimmed Mean Calculator
This trimmed mean calculator gives you an average that is not dragged around by outliers. Paste your data, choose what percentage to trim, and it sorts the values, removes the same number from each end and averages what is left. It also reports the winsorized mean (extremes replaced by the nearest remaining value instead of dropped), the ordinary mean and median for comparison, the winsorized standard deviation and the standard error of the trimmed mean.
A trimmed (or truncated) mean sits between the mean and the median: it uses most of the data but ignores the tails. It is used in judged sports such as diving and figure skating, in economics (the Federal Reserve Bank of Dallas trimmed-mean PCE inflation rate), in benchmarking response times, and anywhere a few freak values would distort an average. A 10% to 25% trim is common; trimming 50% from each end gives the median.
You can enter the trim either as a percentage from each end or, like Excel’s TRIMMEAN function, as the total percentage removed from both ends combined. The number trimmed from each end is rounded down to a whole number of observations, which is how Excel, SciPy’s trim_mean and R’s mean(x, trim) behave.
With the default inputs, the trimmed mean is 220.6875. Change any value above to recalculate instantly.
How to use the trimmed mean calculator
- 1Paste or type your data, separated by commas, spaces or new lines.
- 2Enter the trim percentage (10% to 25% is typical).
- 3Choose whether the percentage applies to each end or to both ends combined, as in Excel.
- 4Read the trimmed mean and compare it with the ordinary mean, median and winsorized mean.
Formula and method
Sort the data from smallest to largest. With a trim proportion γ from each end, g = ⌊nγ⌋ observations are removed from the bottom and the same number from the top, and the remaining n − 2g values are averaged. The winsorized mean keeps all n observations but pulls the extremes in to the nearest kept value, so it is a little closer to the ordinary mean.
The standard error of a trimmed mean is estimated from the winsorized standard deviation s_w with Yuen’s formula shown above, where h = n − 2g is the number of values kept; it is the standard error used in Yuen’s trimmed-mean t-test and is close to the Tukey-McLaughlin form s_w ÷ ((1 − 2γ)√n). Excel’s TRIMMEAN takes the total proportion removed from both ends, so a 10% trim from each end corresponds to TRIMMEAN(range, 0.2).
- n
- Number of observations
- γ
- Proportion trimmed from each end (e.g. 0.10)
- g
- Number of observations trimmed from each end
- x₍i₎
- i-th smallest value (order statistic)
- s_w
- Sample standard deviation of the winsorized data
- h
- Number of values kept, n − 2g
Worked examples
Web response times with two spikes (10% each end)
Twenty response times include two spikes of 980 and 1,250 ms that push the ordinary mean to 307.4 ms. A 10% trim removes 2 values from each end (189, 198, 980 and 1,250) and averages the other 16, giving 220.69 ms — close to the median of 220.5 ms.
Five-judge panel: drop the highest and lowest score
Five judges score a dive. A 20% trim of 5 scores removes exactly one from each end (7.0 and 9.5), as in a five-judge diving panel. The remaining scores 7.5, 8.0 and 8.5 average 8.0, while the untrimmed mean is 8.1.
Excel TRIMMEAN with 20% total
This is the example from Microsoft’s TRIMMEAN help page. Removing 20% of 11 values is 2.2, rounded down to 2 — one from each end (1 and 7). The remaining nine values sum to 34, so the trimmed mean is 34 ÷ 9 ≈ 3.778.
Frequently asked questions
What is a trimmed mean?+
A trimmed mean is the average after removing a fixed percentage of the smallest and largest values. A 10% trimmed mean drops the lowest 10% and the highest 10% and averages the middle 80%, which makes it resistant to outliers.
What is the difference between a trimmed mean and a winsorized mean?+
A trimmed mean discards the extreme values. A winsorized mean keeps the same number of observations but replaces the extremes with the nearest remaining value before averaging, so it stays slightly closer to the ordinary mean.
How do I calculate a trimmed mean in Excel?+
Use =TRIMMEAN(range, percent), where percent is the total fraction removed from both ends. For a 10% trim from each end, use =TRIMMEAN(A1:A20, 0.2). Excel rounds the number removed down to an even number.
How much should I trim?+
Trimming 10% to 20% from each end is a common compromise between efficiency and robustness; Wilcox recommends 20% for robust inference. Trimming 25% from each end gives the interquartile mean, and trimming almost 50% gives the median.
Why does a small trim percentage not change my result?+
The number trimmed is rounded down to a whole number of values. With 8 values, a 10% trim from each end is 0.8 values, which rounds down to 0, so nothing is removed until the trim reaches 12.5%.