Trimmed Mean Calculator

A trimmed mean (also called a truncated mean) is the average of a dataset after a specified percentage of the smallest and largest values have been removed. For example, a 10% trimmed mean removes the bottom 10% and top 10% of values before computing the average. This makes it more robust to outliers than the standard mean.

First, sort the dataset in ascending order. Then calculate g = round(trim_percentage × n / 100), where n is the total number of values. Remove the g lowest and g highest values. Finally, compute the arithmetic mean of the remaining values. Our calculator handles all of these steps automatically.

Common choices are 5%, 10%, or 25%. A 25% trimmed mean is also known as the interquartile mean. The right percentage depends on how many outliers you expect in your data — higher trim percentages remove more extreme values but also reduce your effective sample size.

Both methods handle outliers, but differently. A trimmed mean removes the extreme values entirely and averages what remains. A winsorized mean replaces the extreme values with the nearest boundary values instead of discarding them, then averages the full (modified) dataset.

No. If you trim 50% or more from each end, you would remove all or more values than the dataset contains, making the mean undefined. The valid range is 0% (no trimming) up to just under 50%.

The trimmed mean is preferable when your dataset contains outliers or is heavily skewed. For instance, income distributions, reaction times, and sports scores often have extreme values that distort the standard mean. Trimming those extremes gives a better sense of the typical value.

The number of values removed from each end is g = round(trim_percentage / 100 × n). So for a 10% trim on a dataset of 20 values, g = round(0.10 × 20) = 2 values are removed from the bottom and 2 from the top, leaving 16 values for the mean calculation.

Yes, but interpret results carefully. With very small datasets (fewer than 10 values), even a modest trim percentage can remove a significant portion of your data, reducing reliability. For best results, use trimmed means with datasets of at least 10–20 values.