A metrological traceability chain prevents circular reasoning in measurement design

Abstract

In previous Editorials [1, 2], we described how the establishment of ‘metrological traceability’ [3] of a ‘measurement result’ [4], enables to walk back to the ‘reference’ [5] along the ‘‘trace’’ through which a measurement result did come to us: the ‘metrological traceability chain’ [6]. A characteristic of a chain is that it is one-dimensional. Does that also apply to a metrological traceability chain? In principle, yes. It may have ‘‘side chains’’ visualizing various secondary input quantities and their values (with associated uncertainties), which may influence the measurement result (and its associated uncertainty), but we will not here be looking at these secondary effects. We choose the simplest case of a metrological traceability chain: the chain consisting of one link [1], which runs from the value of the measurand to the value embodied in the working calibrator used to calibrate the measuring system used. In fact, there are two ‘‘sub-links’’ involved: the first from the value embodied in the working calibrator to establishing the calibration curve of the measuring system to be used in the measurement and the second associated with the measurement, result for the unknown value of the measurand obtained by using the calibrated measuring system. Walking backwards along the ‘‘trace’’ from the result to the chosen ‘reference’ enables to see which steps have been taken to obtain the measurement result and see how the link was established from an unknown ‘quantity value’ [7] of the ‘measurand’ [8] to a known value of the same quantity embodied in the working ‘calibrator’ [9] chosen by the analyst to calibrate his/her ‘measuring system’ [10]. The calibration of the analyst’s measuring system by means of the calibrator enables to operate that system and obtain a meaningful, i.e. calibrated, measurement result. The whole operation ‘‘realizes’’ a link between the unknown value of the measurand and the known quantity value embodied in the calibrator. We also showed that the metrological traceability chain is established in the office prior to the measurement in the laboratory: it is a decision on the choice of the calibrator. This decision must be taken, of necessity, before the measurement is carried out: it is needed to know which calibrator must be purchased or obtained otherwise and then used. Clearly, there is a natural sequence of decisions involved: the definition of the measurand by the analyst (possibly in consultation with the requestor of the measurement result) and the choice of the reference, i.e. the choice of a suitable working calibrator, to which metrological traceability of the measurement result will be established. The combination of these (definition of measurand and choice of working calibrator) enables to perform: