Trimmed Mean Calculator
The trimmed mean removes the most extreme values from both ends of a sorted dataset before computing the average. This makes it more resistant to outliers and skewness than the ordinary arithmetic mean. A 10% trimmed mean, for example, discards the lowest 10% and highest 10% of values and averages the remaining 80%. The trimmed mean is used in Olympic judging (scores from the highest and lowest judges are dropped), the Consumer Price Index (trimmed-mean CPI), and robust statistical analysis. Enter your dataset and the trim percentage to compute the trimmed mean alongside the arithmetic mean for comparison.
Trimmed mean formula
k = floor(n × trim% / 100)
Trimmed mean = mean of x[k+1] ... x[n-k] (sorted)
Where n is the total number of values and k is the number trimmed from each end. The remaining n minus 2k values are averaged. When k = 0 the result equals the arithmetic mean; when k = floor(n/2) the result approaches the median.
Effect of trimming
- With data 5, 7, 8, 10, 12, 14, 15, 18, 100, 200 (n = 10), the arithmetic mean is 38.90, heavily influenced by 100 and 200.
- A 10% trim removes the lowest 1 (value = 5) and highest 1 (value = 200), giving a trimmed mean of the remaining 8 values: about 20.50.
- A 20% trim removes 2 from each end, giving the mean of 8, 10, 12, 14, 15, 18, which equals 12.83.
- Larger trim percentages are more robust but use less of the available data, reducing statistical efficiency.
Frequently asked questions
What is the trimmed mean?
The trimmed mean (or truncated mean) is calculated by removing a fixed percentage of the smallest and largest values from a dataset and then computing the arithmetic mean of the remaining values. It is a robust measure of central tendency that is less sensitive to outliers than the ordinary arithmetic mean.
What percentage should I trim?
Commonly, 5%, 10%, or 20% are trimmed from each end. A 10% trimmed mean removes the lowest 10% and highest 10% of values and averages the middle 80%. The Olympic scoring system famously trims the highest and lowest judge scores before averaging.
How does the trimmed mean compare to the median?
The median is the extreme case of a 50% trimmed mean (only the middle value(s) remain). A 0% trimmed mean equals the arithmetic mean. Trimmed means at intermediate percentages offer a balance between robustness and efficiency, depending on the distribution of data.
When is the trimmed mean preferred over the arithmetic mean?
Use the trimmed mean when data may contain outliers or errors at the extremes, when the distribution is heavy-tailed or skewed, or when you want a result less influenced by a few extreme values. It is standard in robust statistics and widely used in economics (trimmed CPI) and sports judging.
Is the trimmed mean the same as the Winsorized mean?
No. The trimmed mean discards the extreme values. The Winsorized mean replaces extreme values with the values at the trimming threshold rather than discarding them. Both are robust alternatives to the arithmetic mean, but the Winsorized mean uses all n observations.
Official sources
- NIST/SEMATECH e-Handbook of Statistical Methods: Measures of Location.
- NIST/SEMATECH e-Handbook: Engineering Statistics Handbook.
Reviewed by the CalculatorHub team, edited by James Graham, 15 June 2026. See our methodology.