For a standard deviation computed from n readings, the degrees of freedom are n minus 1; more generally they measure how much independent information supports an uncertainty estimate. A component with few degrees of freedom is itself uncertain, which widens the coverage factor needed for the final result. Type B components are often given infinite degrees of freedom, or a finite value reflecting trust in the bound, such as about 12 for a figure judged reliable to 20 %.
Three repeats gives you two degrees of freedom, that's why k blew up to 4.5.
Also heard as
- DOF
- nu
Related terms
Type A evaluation
Putting a number on an uncertainty component by running statistics on your own repeat readings.
Welch-Satterthwaite formula
The equation that estimates effective degrees of freedom for a combined uncertainty, used to pick the coverage factor.
Student t-distribution
A bell-shaped distribution with heavier tails than the normal, used to set coverage factors when data are limited.
Coverage factor
The multiplier applied to combined standard uncertainty to obtain expanded uncertainty, usually k = 2.