Separate the reading from the verdict
Consider a hypothetical decorative workshop panel. Its stated width requirement permits values from 499.5 to 500.5 millimetres, and an inspection reports 500.4 millimetres. The reported value lies inside those limits. That arithmetic is easy; understanding the acceptance statement requires more information. These invented dimensions illustrate a report, not a recommended specification or a real inspection result.
NPL's introductory guide distinguishes a measurement from a test's pass-or-fail conclusion: measurements can contribute to the decision without being identical to it. In our example, the width is one reported result. A label declaring the panel acceptable is a further statement. Readers should be able to see how the report moved from one to the other.
Tolerance describes the requirement
JCGM 106 distinguishes the permitted values of a property from the measured values that lead to acceptance under a decision rule. These intervals can coincide, but need not. The requirement for the panel does not change merely because a different measurement process produces greater uncertainty. What changes is the evidence available for the acceptance decision.
NPL also distinguishes uncertainty from a known measurement error: uncertainty expresses the remaining doubt about the result. It is not an extra manufacturing allowance to add to the specified width. For a reader, keeping these labels separate prevents a tolerance, an uncertainty statement and an instrument specification from being presented as though they described the same thing.
Acceptance reflects an identified rule
A NIST paper on conformity assessment explains that evaluating uncertainty is a technical task, while selecting a decision rule involves consequences and costs. Incorrectly accepting an unsuitable item and incorrectly rejecting a suitable one are different errors. A report's conclusion becomes easier to assess when the chosen rule and its intended balance are visible.
JCGM 106 describes both simple acceptance and rules that place acceptance limits apart from tolerance limits, using guard bands. These are alternatives with different implications, not evidence that one universal rule governs every inspection. Our panel example does not select one. A reader should look for the applicable rule in the report or referenced requirements instead of inventing it from the colour of a dashboard indicator.
An interval needs its meaning
Suppose the panel report also states an expanded uncertainty of 0.2 millimetres around the measured value. The corresponding interval runs from 500.2 to 500.6 millimetres and crosses the upper tolerance limit. That observation helps explain why uncertainty matters near a boundary. It does not, without the relevant decision rule, establish the report's acceptance or rejection outcome.
NIST's historical uncertainty-reporting guidance calls for the coverage factor accompanying expanded uncertainty and a basis for any probability interpretation. Consequently, the hypothetical interval here is not a claim of a particular confidence level or a calculated probability of conformity. A report should explain those meanings where they are used; the plus-or-minus notation alone does not supply them.
Preserve the result beside the conclusion
For this panel, a useful summary would retain the measured width and uncertainty alongside the named requirement and decision rule. That lets a later reader understand what the acceptance label covered. A general declaration that the panel passed should not silently become evidence about its flatness, finish or every other characteristic that the width measurement never addressed.
This is our editorial recommendation for reading inspection claims, not a procedure for accepting components. Ask which property was assessed, what result supports the statement and which rule connects them. Keeping those pieces visible makes a conformity claim more intelligible without pretending that a short public report contains the full technical or contractual basis for the decision.
Primary sources reviewed 18 October 2026. JCGM 106 is the October 2012 guide; the NPL PDF is Issue 2 with March 2001 amendments. NIST TN 1297 is historical guidance on a page marked as no longer maintained. The 2014 NIST paper is used only for its public abstract. The panel and numerical values are hypothetical, imply no confidence level, and do not select a decision rule or authorize acceptance of any component.
Read the evidence.
- 01JCGM 106:2012 — The role of measurement uncertainty in conformity assessmentJoint Committee for Guides in Metrology
- 02A Beginner's Guide to Uncertainty of Measurement, Good Practice Guide 11, Issue 2National Physical Laboratory
- 03Assessment of Conformity, Decision Rules and Risk AnalysisNational Institute of Standards and Technology · 1 Apr 2014
- 04NIST Technical Note 1297, Section 7: Reporting UncertaintyNational Institute of Standards and Technology