Measurement results in a Proficiency Testing programme need to be evaluated against independent criteria

Abstract

There exist many interlaboratory comparisons (ILCs). Frequently, they are used to detect, study, and compensate for (un)known effects affecting a measurement result. Another frequent use of ILCs is to evaluate measurement capability. However, an ILC can only be used for one intended use at a time. For the purpose of this Column, we define a PT (Proficiency Testing) programme as an interlaboratory comparison study used for the sole and exclusive purpose of evaluating the measurement capability (proficiency) of one or more participating laboratories. In a previous Column [1], the declaration by participating laboratories of their measurement results (defined in VIM entry 2.9 [2]) in a Proficiency Testing (PT) programme was discussed. For a definition of measurement capability, see entry 7.3-1 in the IUPAC study referred to in [3]. The usefulness of comparative graphs as easy means to evaluate this measurement capability was stressed. However, a measurement result must comprise its measurement uncertainty in order to be a valid result (see definition in VIM entry 2.9). Sadly enough, that is not done in many cases. Only the measured values are given and compared. Measurement results which do not contain a statement of measurement uncertainty ‘‘cannot be compared’’, as explicitly stated in section 0.1 in [4], yet being ‘‘comparable’’ is ‘‘a worldwide need’’ for measurement results as specified in section 0.3 in [4]. In addition, the metrological traceability (defined in VIM entry 2.41 in [2]) of a result must be indicated. An absence of these two characteristics casts a twofold doubt on such incomplete pictures and, therefore, onto the very validity of any conclusion of the programme. Measurement uncertainty and metrological traceability are intrinsic characteristics of any measurement result and hence also of any participant’s measurement result in a PT programme. They are conditions for the very metrological comparability of such results (defined in VIM entry 2.46 [2]). In addition, this metrological traceability must be established to the same reference (defined in VIM entry 2.41 NOTE 1 [2]). That logically leads to the question against which criterion this evaluation must be performed. A possible answer is: through the assignment to the programme by the organizer, of a ‘‘reference value’’ (short for ‘‘reference quantity value’’) with a clearly demonstrated metrological traceability. The definition of ‘‘reference value’’ is given in VIM entry 5.18 and its NOTE 2 [2]. Several reference values are possible (VIM entry 5.18 and its NOTE 2 in [2], as made available from: