Measurement Uncertainty: Reporting More Than a Single Number
Start with the quantity
Before interpreting uncertainty, establish what the number describes. A proportion of chromatographic peak area, an amount of material and a concentration are different quantities. Units and the reporting basis matter as much as the digits. Uncertainty attached to one quantity cannot simply be carried over to another.
For example, a result expressed relative to the received material may need different interpretation from one expressed on a stated dry basis. Ask the laboratory to explain the basis in ordinary language if the certificate uses shorthand that is unfamiliar to your team.
Ask what the interval represents
Measurement uncertainty describes the dispersion of values reasonably associated with a measured quantity under the stated evaluation. It is not a catalogue of every possible mistake. The Eurachem guide to quantifying uncertainty provides the analytical framework for this subject.
If uncertainty is reported, ask whether it is standard or expanded, what coverage factor or probability statement accompanies it and which contributions were included. Do not assume that a bare plus-or-minus figure has a universally fixed meaning. A report without an uncertainty statement also needs context; absence is not evidence of zero uncertainty.
Consider the sample as well
Instrument performance is only part of an experiment. Selection of the test portion and its preparation may contribute additional variation. Eurachem’s sampling-uncertainty guide specifically considers these stages. Ask whether an uncertainty estimate applies only to the laboratory measurement or to a broader sampling process.
This distinction becomes important when one result is used to describe multiple containers. A precise measurement of the submitted portion does not establish that every other container has the same composition. Record the population to which you intend to extend the result and the evidence supporting that extension.
Use uncertainty in the decision
Near a specification boundary, identify the decision rule before declaring that a requirement has been met. Greater displayed precision does not solve the issue. Report only the interpretation justified by the stated method and decision arrangement.
When reviewing documents in the COA library, keep the measured quantity, units and uncertainty information together. If a critical decision depends on information that is missing, ask a focused question rather than calculating an unsupported interval yourself.
For a UK research project, the useful endpoint is a transparent statement of what the measurement supports. That makes uncertainty an aid to choosing and interpreting experiments, rather than a decorative statistic appended to a certificate.
Related research reading
Explore the research blog and original batch-document library for further context.