Look at both endpoints
NIST’s reference-material guide includes measurement intervals that are not centered on the stated value. Section 7.3 explains how asymmetric probability distributions can lead to unequal distances on either side. A familiar plus-or-minus format is therefore not always the clearest way to preserve what a certificate reports.
A small arithmetic example
Imagine a purely illustrative, unitless estimate of 10 with an interval from 8 to 13. The lower distance is 2 and the upper distance is 3. Rewriting it as 10 ± 2.5 would instead give 7.5 to 12.5. Using 10.5 ± 2.5 preserves the endpoints but changes the stated estimate. Neither rewrite preserves all three original numbers.
Why this matters for concentration math
When a calculation uses measured inputs, copying only the central number can discard useful information. Before comparing two laboratory reports, keep their units, interval endpoints and stated coverage information together. Do not silently supply a probability level or an uncertainty model that the report does not provide.
Read the certificate on its own terms
NIST provides worked methods for handling and propagating such uncertainty. Our example only demonstrates bookkeeping; it is not a laboratory recipe, a statistical fit or evidence about any vendor’s product. The official 2020 guide is newly featured today. The artwork is symbolic rather than a plotted dataset.