How do we know a thermometer’s number means what it says? Not by trusting a label or a single comparison. A calibration connects the instrument’s readings to a reference, and that reference may itself have been compared with another. The chain lets different laboratories relate measurements to shared standards. It also carries uncertainty from link to link, so the result is useful only when its limits are understood.
A number needs a reference
Imagine a thermometer placed in a carefully controlled bath. A calibration compares its displayed temperature with a reference thermometer whose own measurement is known under specified conditions. The difference can reveal, for example, that the instrument reads slightly high at a particular temperature. Calibration is therefore a comparison that characterizes how an instrument behaves; it does not make every later reading exact.
But why trust the reference thermometer? It, too, must be compared with a more authoritative standard. That standard may be linked through further calibrations to a national or international realization of a measurement unit. Metrological traceability names this relationship: a measurement result is connected to a reference through a documented, unbroken chain of calibrations, each contributing to uncertainty [1].
What the chain preserves
Think of passing a carefully marked length down a line of people, each comparing one ruler with the next. If every comparison is recorded, a measurement made with the last ruler can be related to the shared reference at the start. In real laboratories, records identify the standards, procedures, and results involved; the chain is not merely a claim that an instrument was checked somewhere.
The documentation makes results comparable across places and over time. A hospital laboratory, for instance, may need measurements from different instruments to be consistent enough for their intended use. Traceability supports that common frame of reference, but the responsibility for supporting a traceability claim rests with the provider of the measurement result. A reference organization does not automatically vouch for every downstream measurement that cites it [1].
Uncertainty travels too
No comparison is perfectly exact. The reference has limits; the instrument has resolution and repeatability limits; conditions such as temperature can affect the result. Each calibration link contributes uncertainty, which is a quantified description of the doubt associated with a measured value. When results depend on several components, those contributions must be evaluated and combined according to the relevant measurement model [2].
Picture a chain of ruler comparisons in which each person can align marks only within a small range. The final length is still connected to the original standard, but the possible spread in its value reflects the whole sequence. A calibration certificate can report a result together with uncertainty, making clear that the reported number is an estimate rather than a perfectly sharp boundary. Uncertainty is part of interpreting the measurement, not evidence that the comparison failed [2].
Fit for purpose is a separate test
Suppose a thermometer is traceable and its stated uncertainty is a few degrees. That might be adequate for checking a broad environmental range, yet inadequate for a task that depends on distinguishing two temperatures only a fraction of a degree apart. The same documented chain can support both results, while only one meets a particular need.
That is why traceability alone does not guarantee fitness for purpose. The user must ask whether the uncertainty is sufficiently small for the decision or measurement task at hand [1]. A familiar unit and a long chain are valuable, but they do not replace that judgment. Calibration gives a result a route back to a reference; uncertainty describes how much room remains around it; purpose determines whether that room is acceptable.
Sources & further reading
Prepared for the independent PSI Center reading library. About edition dates and source checking.
This explainer describes scientific principles. It does not assess individual treatment choices.
