TRIMMEAN in Collabora Office: Average Data After Symmetric Count Trimming
Use Collabora TRIMMEAN with a total trimming fraction and understand how whole observations are removed from both ends.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: for 0, 1, 2, 3, 4, 5, 6, 7, 30 and 100, =TRIMMEAN(A2:A11;0.2) has independently expected result 7.25. It removes one value from each end before averaging the remaining eight.
Syntax and trimming fraction
=TRIMMEAN(Data; Alpha)
Data is the numeric dataset, usually a range. Alpha is the total fraction of observations to exclude across both ends, not the fraction from each end. A value of 0.2 therefore requests 20% total trimming.
The dedicated Collabora help describes marginal trimming but includes an imprecise sentence about percentages. The current implementation resolves the rule: sort usable numbers, approximately floor Alpha*n, reduce an odd result by one, and remove half that even count from each end.
The trimming is based on observation count. It is not a percentage of the sum, mean or numeric magnitude, and it does not subtract 20% from every value.
Example 1: inspect the retained observations
Place 0, 1, 2, 3, 4, 5, 6, 7, 30 and 100 in A2:A11. With Alpha=0.2 and n=10, the requested total count is two, so one smallest and one largest value are removed.
The retained values are 1, 2, 3, 4, 5, 6, 7 and 30. They sum to 58, and 58/8 gives 7.25. The ordinary mean of all ten observations is 15.8.
The function returns only the trimmed mean. It does not delete or visibly mark the excluded cells. Keep a sorted copy or an explicit retained-value list if readers need to audit which observations contributed.
Example 2: show the even-count rule
For the same ten values, Alpha=0.1 requests a floored total of one. The source reduces this to zero so that both ends are treated equally. The independent result remains the ordinary mean 15.8.
Alpha=0.3 produces a floored total of three, reduced to two, so its result is also 7.25. Different Alpha values can therefore produce the same trimmed mean over a finite dataset.
These steps are why a smooth-looking percentage change does not necessarily change the included observations. Report both the input fraction and actual number removed when that distinction matters.
Advanced application: increase the trimming level
With Alpha=0.4, four total observations are removed: 0, 1, 30 and 100. The remaining values 2, 3, 4, 5, 6 and 7 average 4.5.
With Alpha=0.9, the floored total nine is reduced to eight. Four observations are removed from each end, leaving 4 and 5, again averaging 4.5.
The equal result at 0.4 and 0.9 does not mean the calculations used the same data. The retained counts are six and two respectively, a substantial difference hidden by the identical mean.
Advanced application: distinguish trimming from winsorizing
TRIMMEAN omits extreme observations from the average. A winsorized calculation would instead replace extreme values with selected boundary values and retain a different denominator. Those are separate methods.
Sorting the original dataset into another order leaves the independent trimmed mean unchanged because the selection is by magnitude. Removing arbitrary first and last worksheet rows without sorting would not reproduce the rule.
A trimming percentage is an analysis choice. The function does not establish that extreme observations are mistakes or that excluding them is appropriate for a particular dataset.
Errors and troubleshooting
The current source requires 0<=Alpha<1. Alpha=0 retains all usable observations; Alpha=1 is outside the accepted domain. An empty numeric dataset follows a no-value path. Exact native error wording is not claimed.
The cited documentation says text and empty cells in a range are ignored. Numeric zeros remain observations and can be among the trimmed values. Missing data therefore affect both the usable count and the count-based trimming threshold.
Use fractions or percentage literals, such as 0.2 or 20%, rather than the number twenty. Preserve the original dataset and explicitly state the symmetric count-trimming method when reporting a result.
Availability and verification
The exact function is documented in Collabora Office 26.04 desktop help. The cited help snapshot was retrieved on 2026-10-13 and hash-verified October 15, 2026. Desktop documentation does not establish Online or mobile support. First introduction and last functioning release are unknown in this evidence.
No native Collabora Office execution occurred. All numerical results are independent mathematical expectations, with English names and semicolon argument separators. Native error displays, all coercions and extreme numerical behavior are not certified.
Implementation statements are corroborated by Collabora current engine source, retrieved October 15, 2026. This is moving-main source evidence, not a pinned desktop 26.04 binary or execution.