Introductory notes

When reporting the measurement result of a physical quantity, one must give a quantitative indication of the quality of such measurement result, so that those who will use it can estimate its reliability. If such an indication is missing, the result will be impossible to compare these measurement results, either to one another, or to reference quantity values set by a specification or norm.

The concept of uncertainty as a quantifiable attribute of a measurement result's quality is rather new in the history of measurement, although error and error analysis have long been part of metrology. It is actually considered that once all the known or suspected components of an error have been evaluated, and once the appropriate corrections have been made, there will still remain an uncertainty regarding the validity of the expressed result, that is to say a doubt on the extent to which the measurement result correctly accounts for the measured quantity's value.

The official definition given by the VIM (§ 3,9) for the term « measurement uncertainty » is thus the following:

From a practical point of view, the uncertainty of a measurement will be expressed under the form of a standard deviation (in a statistical sense), and will be referred to as standard uncertainty u(Y).