Measurement uncertainty: what the number leaves open
Measurement uncertainty characterises the spread of values that could reasonably describe the quantity being measured. It accompanies the reported estimate so the central number is not treated as exact. [1]
KEY TAKEAWAYS
A value and its interval
NIST represents a result using an estimated value, y, and an expanded uncertainty, U, written as y ± U. The interval provides information beyond the estimate alone. It is not an instruction to correct the number upwards or downwards, and it does not mean every conceivable value must lie inside that interval. [2]
Why the coverage factor matters
NIST relates expanded uncertainty to combined standard uncertainty through U = k × uc. Under its stated conditions of an approximately normal distribution and a reliable estimate of uc, k = 2 corresponds to approximately 95% coverage. “Two means 95%” is therefore not a universal rule detached from those assumptions. [2]
Interval endpoints and relative uncertainty
For a hypothetical y = 10 and U = 1 in the same unit, y − U is 9 and y + U is 11. NIST defines relative expanded uncertainty as U/|y| when y is not zero. Here it is 0.1, or 10%. That percentage describes uncertainty relative to the estimate; it is not the coverage probability. The latter still needs the stated distribution assumptions and coverage factor. These numbers are a notation example, not product measurements. [2]
What belongs beside y ± U?
Our reading: A useful comparison keeps the measured quantity, units, method, conditions and uncertainty statement together. The nominal result and its interval answer different parts of the question. Neither a slightly larger central number nor overlapping intervals alone supply a complete verdict about which device or material performs better.
Article Sources
- National Institute of Standards and Technology (NIST) — Basic definitions of uncertainty
Primary source
Uncertainty (of measurement) definition and notes: dispersion of values attributed to the measurand, best estimate and systematic components. Separate entries define expanded uncertainty and coverage factors.
- National Institute of Standards and Technology (NIST) — Expanded uncertainty and coverage factors
Primary source
Expanded uncertainty definition, U = k × uc, and the passages on coverage factors and distributions; also Relative expanded uncertainty. No universal confidence level is asserted.
Written and reviewed with AI assistance using the linked sources. No hands-on product testing or independent specialist review was performed.
