= Measurement result
A \b[measurement result] comprises the values attributed to a <measurand> together with the information needed to interpret those values. It commonly takes the form of a <measured quantity value> and its <measurement uncertainty>, but the result can contain a fuller <probability distribution>. A displayed number alone need not supply this information.<International vocabulary of metrology by JCGM>[VIM3entry2.9, definition and notes1–3]{ref}
For an original <electrical resistance> example, report a corrected value $R=100.020\,\Omega$ at $20\,{}^{\circ}\mathrm C$, with <standard uncertainty> $u(R)=0.010\,\Omega$. Specify that the uncertainty is a <standard uncertainty>, identify the resistor and the temperature condition, and retain the <measurement procedure> and <measurement calibration> records. Writing only “100.020” omits both the unit and the meaning of the last digits. Writing $100.020\pm0.010\,\Omega$ without stating what the second number represents is also incomplete.
An alternative report might use <expanded uncertainty> $U=0.020\,\Omega$ with <coverage factor> $k=2$. This is a different uncertainty convention for the same central estimate; it is not a newly observed change in the resistor. A probability interpretation requires the distributional assumptions accompanying that convention.
An instrument's indication, a corrected <measured quantity value> and the complete <measurement result> therefore play different roles. Corrections from the <measurement model>, the specified <measurand>, environmental conditions and <metrological traceability> can all matter when comparing results. Extra displayed digits cannot replace the omitted <measurement uncertainty>.
Back to article page
