CONFIDENCE.T in Collabora Office: Calculate a Student t Margin of Error
Use Collabora CONFIDENCE.T to calculate a t-based mean-interval half-width with explicit sample size, degrees of freedom and interpretation.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: =CONFIDENCE.T(0.05;1.5;100) has independently expected result approximately 0.297632543. It is the half-width of the corresponding two-sided t-based construction.
Syntax and what the result contains
=CONFIDENCE.T(Alpha; StDev; Size)
Alpha is the total probability outside the two tails; 0.05 corresponds to a 95% construction. StDev is the positive standard-deviation input. For the usual one-sample t interval for a mean, use the sample standard deviation and Size equal to the number of sample observations.
The dedicated Collabora help labels StDev and Size using population terminology. The corroborating current implementation clarifies its actual calculation: StDev multiplied by the two-tail t critical value at Alpha with Size-1 degrees of freedom, divided by SQRT(Size).
There is no sample mean argument. The function returns a margin, not a complete interval, an estimated mean or a probability about a hypothesis. Statistical interpretation requires an appropriate sampling model in addition to correct arithmetic.
Example 1: reproduce the documented margin
Put 0.05, 1.5 and 100 in B2:B4, with labels Alpha, Sample standard deviation and Sample size. In B5 enter =CONFIDENCE.T(B2;B3;B4).
The independent standard error is 0.15, degrees of freedom 99, and critical t magnitude approximately 1.984216952. Their product is approximately 0.297632543, matching the official example.
For a transparent calculation, use =T.INV.2T(B2;B4-1)*B3/SQRT(B4) in B6. Do not halve Alpha before passing it to T.INV.2T, because that inverse already interprets its input as the combined two-tail probability.
Example 2: add the sample mean separately
Put sample mean 10 in B7. Enter =B7-B5 in B8 and =B7+B5 in B9. Independent endpoints are approximately 9.702367457 and 10.297632543.
The full width is approximately 0.595265085, twice the returned margin. A displayed confidence level of 95% does not mean the function should receive the number 95 or the probability 0.95 as Alpha.
The margin has the same unit as StDev. Keep units consistent when combining it with a mean; a standard deviation measured in seconds cannot be added directly to a mean expressed in minutes.
Advanced application: compare normal and t margins
At Alpha=0.05, StDev=1.5 and Size=100, the normal-based margin is approximately 0.293994598, slightly smaller than the t margin approximately 0.297632543.
For Size=25, the t margin is approximately 0.619169568, compared with the normal margin approximately 0.587989195. For Size=400, the t margin is approximately 0.147444547, close to the normal margin approximately 0.146997299.
Unlike a fixed-normal-quantile formula, increasing Size also changes the t degrees of freedom. Quadrupling the sample count therefore does not produce an exact halving of the t margin when all other inputs stay fixed.
Advanced application: distinguish the target of inference
This calculation can support an interval for a mean when the assumptions of that t-based procedure are appropriate. It does not estimate an interval containing a specified fraction of individual observations, a prediction interval or a nonparametric confidence interval.
Dependence between observations, complex survey weights or a strongly unsuitable small-sample model can invalidate the usual interpretation even if the displayed arithmetic matches the formula. The function does not inspect the original observations or assess those conditions.
Errors and troubleshooting
The current source approximately floors Size, rejects effective counts below one, requires StDev>0 and 0<Alpha<1, and uses a division-by-zero error path at Size=1. Use an integer sample size greater than one rather than relying on fractional conversion.
A constant sample with zero supplied standard deviation reaches a source guard rather than automatically yielding a zero margin. Native localized error displays, all coercions and extreme inverse convergence remain unverified.
The confidence percentage concerns repeated-sampling coverage under assumptions. It does not assign a posterior probability to a fixed mean lying inside the particular interval. Retain the original sample size, spread and mean alongside the margin so the calculation can be audited.
Availability and evidence
The exact function is documented in Collabora Office 26.04 desktop help. That documentation snapshot was retrieved on 2026-10-13 and hash-verified on October 15, 2026. It does not establish Online or mobile support. First introduction and last functioning release are unknown in this evidence.
Additional implementation statements come from Collabora’s current engine source, retrieved October 15, 2026. This moving-main source is corroboration, not a pinned desktop 26.04 binary or a native execution. No native Collabora Office run occurred. Worked outputs are independent mathematical expectations.
Formulas use English names and semicolon separators. The help identifies namespace COM.MICROSOFT.CONFIDENCE.T, outside its cited OpenFormula 1.3 standard; that does not establish identical behavior in another application.