“Quality of a measurement result” is established by means of metrological criteria
Paul De Bie`vre
0
0
P. De Bie`vre (&) Kasterlee,
Belgium
''Quality'' has become one of the buzzwords of our time. Just look at the number of times that ''quality control'' and ''quality assurance'' can be found in the literature. What is quality? Several attempts for definitions circulate. A few examples: quality is said to mean: degree of excellence [1], the totality of features and characteristics of a product or service that bear on its ability to satisfy stated or implied needs [2], the property of a product or service being proven correct by subsequent evaluation [3], and any of the features that make something what it is [4].
-
In the context of measurement, quality obviously
refers to quality of measurement results obtained for the
measurand (entry 2.3 in [5]) as decided by the analyst
(obviously prior to the measurement). This specification of
the measurand is essential to subsequently enable to talk
about quality of results obtained for this measurand.
Two main questions immediately do come to the mind:
where does the result come from (i.e., what is its metrological
traceability?) and: how doubtful should we feel about the
result? (i.e., what is its measurement uncertainty?). In
communication worldwide, quality of a measurement result must
be understood in the same way in the main cultures of that
world. It therefore requires common understanding of the
concepts metrological traceability (entry 2.41 in [5]) and
The author is a member of the Joint Committee on Guides for
Metrology (JCGM), Working Group 2 (VIM). The opinions expressed
in this Column do not necessarily represent the view of the Working
Group or of ACQUAL.
measurement uncertainty (entry 2.26 in [5]) as properties of
a measurement result. Common understanding by means of
mutually agreed definitions of these concepts is simply a
condition sine qua non.
Fortunately, such commonly accepted definitions are
availablesince 2008in an International vocabulary of
metrology (VIM) [5], patronized by eight international
organizations (BIPM, IEC, IFCC, ILAC, ISO, IUPAC,
IUPAP, and OIML). After 10 years of study work, two
periods of 6 months of formal consultation and a formal
vote according to previously established procedures, these
definitions were unanimously approved in the final version
and are available/downloadable for free on the Internet [5].
In the very recent version (released in 2012), various
editorial and typographical corrections were published [5].
When considering the property metrological
traceability, the question automatically arises: to what are
these results trace-able? In other words: what are
references for this trace-ability? They are given in the VIM in
entry 2.41 Note 1 in [5]:
1. a measurement unit (entry 1.9 in [5])
2. a reference measurement procedure (entry 2.7 in
[5])
3. a certified reference material (entry 5.14 in [5])
Metrologically speaking, it is difficult to see how values
can be trace-able to anything else than values. That is
consistent with the definition of reference given under
concept 2.6-1 in [6], a 10 years IUPAC study having been
subjected to formal examination procedures similar to
those given above. We note that values are indeed central
in this definition:
specification of kind-of-quantity and description of
how to obtain one or more quantity values of that
kind-ofquantity (not italicized in the original).
The three references given in entry 2.41 in VIM, Note 1
can now be interpreted as:
It would constitute a great leap forward if every
measurement laboratory institute, organization, research group,
university, and especially National Measurement Institutes
would use these commonly agreed references for metrological
traceability: they create the very basis for establishing
metrological comparability of measurement results (entry 2.46
Note 1 in [5]) on a regional, national, international, global
(intercontinental) scale, or within an association of
professionals in the same field. Such common references should
always be agreed before measurements are carried out the
results of which are intended to be compared. They should also
be accessible to all parties involved. Comparability enables
us to validly compare results other across time and space, the
very purpose of measuring, and the very reason for the
requirement metrological traceability.
The other property measurement uncertainty of a
measurement result presupposes an established
metrological traceability chain (entry 2.42 in [5]), and for evaluating
measurement uncertainty, we also have an international
guide [7]. We will not dwell on that here, but rather look at
a related matter.
In the practice of measurement, we like to distinguish
good results from medium and bad. That requires
setting a goal (prior to the measurement) for the maximum
permissible measurement uncertainty to quantify the
fitness-for-intended-use [8] of that result. A tool for that is
available from the VIM: target measurement uncertainty
(entry 2.34 in [5]). A (...truncated)