Accred Qual Assur

DOI 10.1007/s00769-014-1093-0

DISCUSSION FORUM

Trueness, precision and accuracy: a critical overview

Received: 21 July 2014 / Accepted: 18 November 2014

Abstract Many parameters are defined by international result of measurement can be assessed by way of the mea-
scientific committees to describe the performances of a surement uncertainty (uncertainty approach), which can be
measuring system functioning according to a given mea- quantified by constructing the uncertainty budget and cannot
surement procedure. A careful inspection of several be neither considered nor used as an expression of accuracy.
international and national documents in matters of mea-
surement reveals inconsistencies in the definition of some Keywords Trueness
Precision
Accuracy

fundamental terms. This paper discusses the meaning cur- Terminology
Measurement quality

rently attributed to trueness, precision and accuracy Measurement uncertainty
assuming as a reference the International Vocabulary of
Metrology 2012 (VIM). We support that accuracy cannot be
intended as composed by trueness and precision (as declared Introduction
by various standards of international—as ISO—or national
organisms), since trueness and precision require large Many parameters are defined by international scientific
number (infinite number, according to VIM 2012) of repli- committees to describe the performances of a measuring
cate measured quantity values to be assessed while accuracy system functioning according to a given measurement
refers to a single measured quantity value, according to procedure. A careful inspection of several international and
widespread current definitions. As to VIM and other scien- national documents in matters of measurement reveals
tific organisms, accuracy can be intended only in a discrepancies and inconsistencies in the definition of
qualitative fashion, avoiding to associate numbers with it. some fundamental terms. Although explicit definitions of
Hence, a measurement result unbiased and precise is accu- trueness, precision and accuracy are provided in the
rate. We propose to intend the term trueness only in an ideal International Vocabulary of Metrology 2012 (henceforth
meaning and to introduce the concept of exactness (error VIM), the understanding of these basic terms still proves
approach) correctly describing the matching between a difficult, especially in view of the additional specific noise
measurement result—calculated from a large number of brought by single languages. Current meaning of accuracy,
replicate measured quantity values—and an accepted ref- trueness and precision is debated in the papers of Patriarca
erence quantity value. The range of variability of a single et al. [1] and of Menditto et al. [2]. Conclusions of the
paper of Menditto et al. [2] allow some latitude to dis-
cussion, because authors recognize that these three basic
Papers published in this section do not necessarily reflect the opinion
of the Editors, the Editorial Board and the Publisher.
concepts are not yet uniquely defined.
A critical and constructive debate in the Discussion Forum or a Letter This paper discusses the meaning currently attributed by
to the Editor is strongly encouraged. various scientific organisations to trueness, precision and
accuracy, assuming as a reference the International
E. Prenesti
F. Gosmaro (&)
Vocabulary of Metrology 2012 by the Joint Committee for
Dipartimento di Chimica, Università degli Studi di Torino,
Via Pietro Giuria 5, 10125 Turin, Italy Guides in Metrology (JCGM) [3], which is largely inter-
e-mail: [email protected] nally consistent and provide a solid base in the field of

123
 Accred Qual Assur

metrology. Moreover, BIPM (Bureau International des and results—part 1: General principles and defini-
Poids et Mesures), IEC (International Electro technical tions’’] [4], we have:
Commission), IFCC (International Federation of Clinical
• Trueness: ‘‘Closeness of agreement between the
Chemistry), ILAC (International Laboratory Accreditation
average value obtained from a large set of test
Cooperation), ISO (International Organization for Stan-
results and an accepted reference value’’;
dardization), IUPAC (International Union of Pure and
• Precision: ‘‘Closeness of agreement between inde-
Applied Chemistry), IUPAP (International Union of Pure
pendent test results obtained under stipulated
and Applied Physic) and OIML (Organisation Internatio-
conditions’’;
nale de Métrologie Légale) are members of JCGM, and this
• Accuracy: ‘‘Closeness of agreement between a test
would ensure a wide and correct dissemination of metro-
result and an accepted reference value.’’
logical definitions in the world, albeit an examination of
specific documents (standards, guides and so on) still Although VIM [3] is the reference document in
reveals discrepancies and inconsistencies damaging the metrology, it must be evidenced that ISO standards are
standardization process of terms and procedures. Espe- largely followed worldwide, specially in view of labora-
cially in view of the necessity of validation of measurement tories accreditation procedure. We can cite ISO
methods (fields of metrology and of analytical chemistry), 3534-1:2006 [5], ISO/IEC 17025:2005 [6] and ISO
explicitly expressed by international organisms for stan- 15189:2007 [7] of particular importance for measurement
dardization, and also requested in specific national and in chemistry. For this reason, the ISO 5725-1:1994 stan-
European regulations, a revision of some basic concepts dard must be carefully examined in this critical overview of
still seems opportune, if not strictly necessary. Moreover, metrological concepts.
to promote a full harmonization between reports from
3. According to the DIN 55350-13:1987-07 standard [8]
different scientific organisms, all documents should be
(DIN = Deutsches Institut für Normung), as an exam-
tuned to the International Vocabulary of Metrology.
ple of national definitions, we have:
Attention must also be paid to consider a strict contextu-
alized language, since many terms can assume different • ‘‘Trueness is the qualitative term for the closeness
meaning in different areas of science, while we need to of agreement between the expected value (the
tune the reasoning on a circumscribed field with a well- arithmetic mean obtained from a large series of test
defined vocabulary. The paper focuses the attention on results) of the test results and an accepted reference
contextualized terminology to avoid inconsistencies and value’’;
confusion of meanings of specific widespread basic terms. • ‘‘Precision is the qualitative term for the closeness
of agreement between independent test results
obtained under stipulated conditions. Precision
Current definitions of trueness, precision and accuracy depends only on the distribution of random errors
and does not relate to the true value or the accepted
Let us now examine relevant definitions—by international reference value. The measure of precision is
scientific institutions—of trueness, precision and accuracy expressed as the standard deviation of the test
starting by these provided by JCGM 200:2012 [3]. results’’;
• ‘‘The term accuracy consists of two criteria, the
1. According to the VIM [3], we have: precision and the trueness.’’
• Measurement trueness: ‘‘Closeness of agreement From these definitions, we can infer that:
between the average of an infinite number of
replicate measured quantity values and a reference • Trueness deals with the systematic error of
quantity value’’; measurement,
• Measurement precision: ‘‘Closeness of agreement • Precision deals with the random (or accidental or casual
between indications or measured quantity values or indeterminate) error of measurement,
obtained by replicate measurements on the same or • Accuracy deals with the total (systematic and random)
similar objects under specified conditions’’; error of measurement.
• Measurement accuracy: ‘‘Closeness of agreement Being:
between a measured quantity value and a true
quantity value of a measurand.’’ • XR the accepted reference value of the quantity X,
• XM the mean value of repeated measurements of X,
2. According to the ISO 5725-1:1994 standard [‘‘Accu- • X1, X2, X3,…Xi, …, Xn-1, Xn the single values of
racy (trueness and precision) of measurement methods repeated measurements of X,

123
Accred Qual Assur

From the previous definition of each quality parameter, of Menditto et al. [2] (Fig. 1: Relationships between type
we need to define an operational quantity to obtain a of error, qualitative performance characteristics and their
number as quality descriptor, namely: quantitative expression), does not exist. As a consequence,
the link between the accuracy and the measurement
• for Trueness: the bias (also called measurement bias),
uncertainty must be critically re-examined. This fact will
• for Precision: the standard deviation (or the variance or
be further discussed, but, prior to examine in detail the state
the coefficient of variation),
of the art, we propose to subdivide the scientific route of a
• for Accuracy: the single deviation,
measurement method into three steps, since this will be
and, then, we can calculate each operational quantity as: used for rationalization of concepts and related quantities:
• Bias = XM-XR; • Design level: The measurement method is designed by
• Accidental deviation = Xi-XM from which the stan- a scientist (or by a team) and technically optimized,
dard deviation is calculated with the usual common usually with respect to a specific matrix and level of
formula; concentration; this step is chemically centered at a
• Single deviation = Xi-XR. research level of action;
• Validation level: The measurement method is tested
Since the true value of a measurement is always
(according to a specific protocol which includes various
unknown (VIM item 2.11 in [3]), in practice an accepted
repetitions of measurement and various specific exper-
reference value of a measurand (also named conventional
imental designs for each validation parameter selected)
true value or assigned value, namely XR) is used when
as to its quality performances to define (under statistical
available with respect to a specific matrix, analyte and level
control and having established a level of probability, or
of concentration. The accepted reference value is usually
of risk) qualitative and quantitative margins of proper
established by repeatedly measuring a NIST (National
application; this step is quality centered at a metrolog-
Institute of Standards and Technology) or a JRC-IRMM
ical level of action;
(Joint Research Centre, Institute for Reference Materials
• Application level: The measurement method is daily
and Measurements) or others traceable standard (eventu-
used by a chemical laboratory which employs in its
ally certified) reference materials (SRMs). Nevertheless,
routine to provide results, usually obtained by way of a
various experimental scenarios induce further necessities to
single measurement; this step is customer/business
clarify the concept and its operational achievement. In fact,
centered at a social level of action.
when dealing with a method-dependent result (as for an
operationally defined measurand), a reliable reference At each of the last two levels, one can find specific
value (from a proper certified reference material, CRM) is modus operandi and, then, those quality parameters strictly
often unavailable. Let us start to develop specific reflec- coherent with the predefined goals.
tions for each quality parameter.

The source of inconsistencies Precision

A brief history of the parameters that discussed and defined Precision indicates how close independent measurement
by international scientific committees might help to results obtained by replicated measurements are to one
understand the inconsistencies being examined. Before another and is usually quantitatively expressed by way of
1994, only accuracy—related to systematic error—and the standard deviation, which describes the spread of
precision—related to random error—were used to describe results obtained under a specific measurement protocol.
the quality of a measurement result. A shift in the meaning Both precision and trueness are defined starting by the
of these terms appeared with the publication of the ISO mean value of repeated measurements; hence, they can be
5725-1:1994 series of standards—‘‘Accuracy (trueness and evaluated only executing a wide number of measurements
precision) of measurement methods and results’’ [4]—in under a specific experimental design typical of a method
which was introduced the trueness and a relationship was validation procedure. Moreover, precision can be correctly
established between accuracy, trueness and precision. evaluated only under strictly defined measurement condi-
According to VIM, items 2.13 (Note 1) [3], ‘‘The concept tions (so, repeatability, intermediate precision and
‘‘measurement accuracy’’ is not a quantity and is not given reproducibility arise). Precision is the quality parameter
a numerical quantity value. A measurement is said to be better defined by current standards, and we believe it does
more accurate when it offers a smaller measurement error,’’ not require further discussion, at least at a general level of
then such a quantitative relationship, remarked in the paper thinking at which this paper aims to contribute.

123
 Accred Qual Assur

Trueness Accuracy and measurement uncertainty

The VIM definition of trueness is uncomfortable, since it Especially for accuracy, rough and conflicting definitions
involves a limit concept intractable to the finite (2.14: can be revealed in the literature: an inspection is reported,
‘‘infinite number of replicate measured quantity values’’) as an example, in the paper of medical area in Ref. [10].
[3]. Moreover, such an absolute and pure definition should Accuracy and trueness are both terms related to the devi-
match with a true value and not with ‘‘a reference quantity ation from a reference value.
value’’ (by 2.14 in [3]). The definitions of both DIN [8] and According to ISO 5725 guide [4], the term accuracy is
ISO [4] are, on the contrary, manageable, since they simply used to refer to both trueness and precision. Moreover,
(and realistically) consider large series of measurement according to ISO 3534-1:2006 guide [5], ‘‘An increasingly
results (a wide number of measurement replicates). common expression of accuracy is the so called measure-
In 2.14, Note 1 of Ref. [3], one can read: ‘‘Measurement ment uncertainty, which provides a single figure expression
trueness is not a quantity and thus cannot be expressed of accuracy.’’ Also according to DIN 55350-13:1987-07
numerically’’: The estimate for the closeness of agreement [8] standard, both trueness and precision are explicitly
is, in fact, the bias (or the measurement bias); JCGM [3] invoked to define the accuracy of a measurement result;
refers to ISO 5725 standard for the numerical expression of hence, accuracy would be linked to a quantity related to the
the systematic error (2.14 and 2.18 of VIM in [3]). total measurement error (both systematic and random).
Trueness can be assessed by using the difference (XM- According to the DIN 55350-13:1987-07 standard [8] or to
XR) between the measured value (as average value, XM) Eurachem document [11], the systematic component of the
and a reference one (XR). When assessing the trueness, a measurement error is related to the trueness while the
significant difficult is often finding the reference values to random one to the precision. As to NMKL (Nordic Com-
compare measurements with. Two basic techniques are mittee on Food Analysis) [12], accuracy is ignored while
available to evaluate the trueness: Checking against a ref- only trueness and precision are considered as validation
erence values for a specific material (matrix) or from a parameters of a measurement method.
result obtained by a well-described measurement proce- According to the GUM [13], measurement uncertainty
dure. If no suitable standard material exists, other estimate takes account of all recognized experimental
techniques must be employed [9]. Trueness is considered to effects operating on the measurement result; the uncer-
be the closeness of agreement between the average value tainties associated with each effect identified are combined
obtained from a large series of measurement results and a according to established procedures (various models of
reference accepted one. The terminology is very similar to calculations are available to construct an uncertainty bud-
that used for accuracy, nevertheless: get). Measurement uncertainty leads to a parameter
characterizing the dispersion of the values attributed to a
• trueness applies to the average value of a large number
measured quantity due to both type A and type B uncer-
of measurement results, while
tainties. It is noteworthy that a measurement uncertainty is
• accuracy applies to a single result of measurement.
expressed in terms of standard deviation (including a
Bias (the number associated with the trueness, the esti- coverage factor of statistical meaning leading to a confi-
mate of a systematic measurement error, item 2.18 of VIM dence interval): In our opinion, this is in contrast to ISO [4,
in [3]) is the difference between the average value of the 5] definition of accuracy and to the definition of mea-
large series of measurement results and the reference surement uncertainty [3]. Clarity on the topic can arise
accepted. Bias is, then, the quantity associated with the invoking the useful distinction between the Error approach
trueness. Bias, distinguished in laboratory and method bias, (systematic and random errors) and the Uncertainty
is equivalent to the total systematic error in the measure- approach (type A and type B uncertainties) to the evalu-
ment, and a correction to negate the systematic error can be ation of measurement quality [13]. In fact, it is incoherent
made by adjusting for the bias. ISO 5725-1 standard also that the same quantity, the accuracy, is defined, including
avoids the use of the term bias, because it has different and combining:
connotations outside the fields of science and engineering,
• a single deviation from a reference value (a simple
as in medicine and law.
difference), as for trueness, and
Finally, we want state that a measurement procedure can
• a standard deviation from an average value (a disper-
be labeled with a binary judge yes/no as to the trueness
sion index), as for precision, or
(confirming its qualitative nature). This goal can be
• expressed as dispersion of the values attributed to a
reached applying a simple t test (considering a Gaussian
measured quantity due to both type A and type B
population of data) of comparison between a measured
uncertainties.
(averaged) and a reference value.

123
Accred Qual Assur

This concept is also sustained in the papers of Hubert result (validation field) with a reference value adopted for a
et al. [14–17], in which authors wrote ‘‘In fact, one cannot certain measurand.
measure in only one parameter, difference compared to a Regarding to accuracy, the statement by DIN [8] ‘‘The
reference value and a dispersion of the results.’’ Some term accuracy consists of two criteria, the precision and the
conceptual aspects are still controversial and deserve fur- trueness’’ (also reflected by ISO standard in [4]) is, in our
ther discussion of trueness, precision and accuracy together opinion, wrong and misleading. Moreover, distinguishing
and attempt of clarification and agreement between theory trueness (exactness, in our proposal), accuracy and preci-
(definitions) and practices (numerical quantities derived sion only on the basis of the type of error, random and
from theoretical definitions). systematic, is neither sufficient nor adequate, and in fact, a
combination of systematic and random errors can be done
only in terms of the mean squared error on a variable. The
mean squared error is the sum of the squared bias and the
Discussion observed variance, and this formulation is of only theo-
retical interest since, of course, the true value is unknown.
In spite of the apparent simplicity, trueness, precision and Table 1 collects a synthesis of approaches to the eval-
accuracy are not yet uniquely defined in one document for uation of measurement quality and related qualitative and
the whole world. As fully reported above, precision is the quantitative indicators. Those based on error or uncer-
quality parameter better defined by current standards and tainty are different approaches in assessing the quality of
we believe it does not require further discussion, at least at measurements [13]; hence, since:
a general level of thinking at which this paper aims to
• the uncertainties are the effects of both random and
contribute. On the other hand, trueness, defined as
systematic errors,
‘‘Closeness of agreement between the average of an infinite
• the uncertainties are derived by a function y = f(x) typ-
number of replicate measured quantity values and a refer-
ical of the measurement model adopted,
ence quantity value’’ (VIM item 2.14 in [3]) represents an
• ‘‘There is not always a simple correspondence between
inconsistent definition. As a proposal, trueness could be
the classification into categories A or B and the
considered only an idealized concept and defined as
previously used classification into ‘‘random’’ and
‘‘Closeness of agreement between the average of an infinite
‘‘systematic’’ uncertainties,’’ (‘‘Introduction’’, item 0.7
number of replicate measured quantity values and a true
of ref. [13])
value of a measurand.’’ Note the matching between the
unrealizable ‘‘infinite number of replicate measured quan- they must never be confused or mixed neither in the
tity’’ and the nonexistent (unknowable, VIM item 2.11 in conceptualization nor in the calculation.
[3]) ‘‘true quantity value’’ (another idealized concept). Joining concepts of exactness, precision and accu-
Then, the corresponding real concept could be better racy (error approach) in a unique picture, we have that:
explained by way of a new term, i.e., exactness, defined as
• Exactness deals with the quality of a large number of
‘‘Closeness of agreement between the average of a large
measurement results—repetitions required—estab-
number of replicate measured quantity values and a refer-
lished during a validation process of a measurement
ence quantity value of a measurand.’’ Note the matching
method with respect to the systematic error of
between the realizable ‘‘average of a large number of
measurement;
replicate measured quantity’’ and the often (not always)
• Precision deals with the quality of a large number of
existent ‘‘reference quantity value’’ (see the definition of
measurement results—repetitions required—estab-
trueness by the ISO 5725-1:1994 standard [4], in this, our
lished during a validation process of a measurement
proposal adopted to define the exactness). GUM [13] in
method with respect to the random error of
Annex G, item G.1.6 uses the term exactness writing ‘‘a
measurement;
great deal of exactness.’’ Hence, ‘‘exact’’ is not opposed to
• Accuracy deals with the quality of a single result of
‘‘wrong’’ as in mathematics (contextualization of language)
measurement—repetitions unrequired—obtained in
and can be properly used in the context of metrology to
routine conditions of the measurement method
manage, without ambiguity, the experimental variability
application.
(gradual scale of judgement with respect to a recognized
reference) not existing in mathematics (absolute binary In this connection, and against the statement of both ISO
judgement exact/wrong). Then, a result of measurement 5725-1:1994 series of standards [4] and DIN
can be exact (or not), according to a pre-established degree 55350-13:1987-07 norm [8], it seems us incorrect to
of agreement with an accepted reference value. Finally, interpret the accuracy of a single result of measurement as
exactness refers to a degree of matching of a measurement composed by trueness and precision, since these three

123
 Accred Qual Assur

Table 1 Approaches to the evaluation of measurement quality and related qualitative and quantitative indicators
Evaluation of Indicators of measurement quality Quantitative aspects (numerical expression of the
measurement quality indicator)
quality

Error approach Trueness: idealized concept For exactness: bias (systematic error)
Exactness: related to systematic error, quantitative attribute existent; For precision: standard deviation (random error)
reference value necessary
Precision: related to random error, quantitative attribute existent;
reference value unnecessary
Total measurement error: mean squared total error calculated from the
combination of variances associated with random and systematic errors
Accuracy: qualitative attribute
Uncertainty Measurement uncertainty: index of dispersion of the measurement result; Type A uncertainty: based on data of statistically
approach calculation based on a function related to a measurement model; independent observations
reference value unnecessary Type B uncertainty: based on insight from
experience and general knowledge from various
sources
Uncertainty budget resulting by the combination of
both type A and B uncertainties

attributes refer to different fields of application of a mea- • that detected by repeated measurement results under
surement method. In addition, by NIST TN 1297 (sec. defined conditions: type A uncertainty (evaluated by
D.1.1.1), we have: «‘‘Accuracy’’ is a qualitative concept. statistical means),
Because ‘‘accuracy’’ is a qualitative concept, one should • that detected evaluating intrinsic characteristics of
not use it quantitatively, that is, associate numbers with it; instruments and procedures employed: type B uncer-
numbers should be associated with measures of uncertainty tainty (evaluated by other means with respect to the
instead» [18]. Qualitatively speaking, to obtain low value statistical ones),
of measurement uncertainty, both type A and type B
From GUM [13], Annex D, item D.5.1.: ‘‘Thus the
uncertainties have to be reduced but this does not mean that
uncertainty of a result of a measurement is not necessarily
measurement error is small. On the other hand, ‘‘accurate’’
an indication of the likelihood that the measurement result
derives from the Latin term ‘‘accuratus,’’ which means
is near the value of the measurand. It is simply an estimate
‘‘made with care,’’ and ‘‘care’’ is a typical qualitative
of the likelihood of nearness to the best value that is con-
concept. Hence, definitions of accuracy as those expressed
sistent with presently available knowledge.’’ The
in terms of closeness of agreement by VIM, ISO and other
measurement uncertainty can be evaluated—with a com-
scientific institutions can be eliminated, since they are free
plex procedure based on different model approaches—even
of practical address. Furthermore, as in NIST TN 1297
if an accepted reference value is actually unavailable and a
[18], also in ISO 3534-1:2006 guide [5], we can read ‘‘An
single deviation is, hence, unachievable.
increasingly common expression of accuracy is the so
called measurement uncertainty, which provides a single
figure expression of accuracy.’’ Also this affirmation, in Conclusions and proposals
our opinion and according to Huber et al. [14–17], is wrong
and misleading. A dispersion index (as the measurement With the goal to reach rational, sharp and unambiguous
uncertainty is) cannot result as sum of a range (accuracy, definition of trueness, precision and accuracy, without
2.13 of VIM in [3]) and of a standard deviation (precision, recourse to specific notes and footnotes to clarify ambig-
a dispersion index, Note 1 in 2.15 of VIM in [3]): This is a uous texts based on fuzzy concepts, we propose to
mathematical shame. Simply, measurement uncertainty abandon, or to redefine, the term trueness. A measurement
(‘‘Non-negative parameter characterizing the dispersion of result can be more or less exact with respect to a reference
the quantity values being attributed to a measurand, based value; therefore, exactness might be an effective term,
on the information used,’’ item 2.26 of VIM in [3]) rep- while trueness is simply a term corresponding to an ideal
resents a tool to evaluate the range of variability of a single concept. We state that a measurement result unbiased and
measurement result with respect to the two types of precise (according to pre-definite requirements and under
uncertainty contributions currently defined and accepted by stipulated conditions) is accurate (or, made with care), but
the scientific community, namely: the accuracy cannot be mixed with trueness and precision

123
Accred Qual Assur

neither at a conceptual nor at an operational (quantitative) Table 2 Relationships between type of measurement error, perfor-
level, as also sustained by NIST TN 1297 (sec. D.1.1.1) mance attribute of a measurement method (validation activity) and
their numerical expression
[18] comment and also reflected in the IUPAC Gold Book
(‘‘Accuracy is a qualitative concept’’) [19] and in the VIM Type of Performance attribute Numerical expression
2012 (‘‘measurement accuracy is not a quantity’’) [3]— measurement of the performance
error attribute
leading to intend accuracy only in a qualitative fashion, Qualitative term Quantitative term
without any quantity linked to. Besides trueness (exactness,
in our proposal) and precision, metrological and analytical Systematic Exactness of the measurement Measurement bias
methoda
communities can use the measurement uncertainty—a
composite standard deviation (including a variety of con- Random Precision of the measurement Standard deviation
method under defined
tributions combined)—as quality indicator of uncertainties experimental conditions
of type A and B (uncertainty approach to measurement (many scenarios)
quality) related to single results of measurement as output Total –b Mean squared error
of routine activity of a laboratory. a
Our proposal, see ‘‘Conclusions and proposals’’
Starting by the mentioned considerations and definitions b
The term accuracy is absent in this Table, see ‘‘Conclusions and
of international organisms, and trying a rational synthesis, proposals’’
we collect this list of concepts:
1. we propose to intend the current term trueness only in
use is circumscribed to measurement results quality
an idealized meaning, defining it as ‘‘Closeness of
assessment (large number of measurement results
agreement between the average of an infinite number
required, and the quality of single results is therefore
of replicate measured quantity values and a true
excluded); hence, accurate is only an adjective used to
quantity value of a measurand,’’ trueness is related to a
indicate the quality of an unbiased and precise
nonexistent true value and could be abandoned as term
measurement result (under defined scenarios of mea-
in metrology or employed specifying its theoretical
surements); in this point of view, the VIM definition of
meaning;
accuracy (item 2.13) as ‘‘Closeness of agreement
2. we also propose to introduce the term exactness,
between a measured quantity value and a true quantity
defined as ‘‘Closeness of agreement between the
value of a measurand’’ [3] must be abandoned, since
average of a large number of replicate measured
‘‘a closeness of agreement’’ brings to a calculation to
quantity values and a reference quantity value of a
obtain a numerical quantity value, which is out of
measurand,’’ correctly and lucidly describing the
sense and it is also incoherent with the text in Note 1
matching between a measurement result—calculated
underlining the qualitative nature of accuracy;
from a large number of test values (under specified
6. the quality of a single result of measurement can be
experimental conditions)—and an accepted (conven-
assessed and expressed by way of the measurement
tional) reference quantity value;
uncertainty (quantitative term) that can be quantified
3. measurement bias (or bias), the estimate of the
by constructing the uncertainty budget according to a
systematic measurement error, will be used as term
specific calculation model adopted (uncertainty
indicating the quantitative expression of exactness;
approach) to coherently combine type A and type B
4. either exactness (the qualitative term just proposed)
uncertainties properly expressed as standard devia-
or precision (qualitative term) refers to the quality
tions; no relationships between accuracy and
of a large number of measurement results and
measurement uncertainty can be sustained according
cannot be mixed with quantities used to describe the
to this scheme of terminology.
quality of a single result of measurement (namely,
accuracy); hence, ‘‘Accuracy (trueness and preci- Table 2 synthesizes into an organic and comprehensive
sion) of measurement methods and results’’ by ISO scheme our view regarding quantities implied in the error
5725-1:1994 series of standards [4] is a wrong and approach to the measurement quality evaluation.
misleading definition (a title of a guide) and must Probably, the next metrological documents with a
be abandoned; role of vocabulary will be asked to identify a series of
5. we believe that the term accuracy is finally eliminable idealized concepts with the corresponding meaning
in a quantitative meaning, while it can be properly attributed to and, then, to provide the meaning of the
used only according to a qualitative meaning, avoiding corresponding real concepts mirroring the ideal ones.
to associate numbers with it; then, we propose to adopt This will clarify the general landscape of the measure-
exclusively a qualitative meaning of accuracy whose ment science and will help scientists to distinguish what

123
 Accred Qual Assur

is practically addressed from what is only resulted of a 9. Bland M (1995) An introduction to medical statistics. Oxford
limit mental model. Medical Publications, Oxford
10. Ranstam J, Ryd L, Önsten I (1999) Acta Orthop Scand
70(4):319–321
11. Eurachem (1998) The fitness for purpose of analytical methods.
References A laboratory guide to method validation and related topics, 1st ed
12. NMKL Protocol No. 4 (2010) Manual for NMKL peer-verifica-
1. Patriarca M, Menditto A, Bettinelli M, Minoia C (2004) G Ital tion (interlaboratory verification) of methods
Med Lav Ergon 26(2):102–107 13. JCGM 100 (2008) Evaluation of measurement data: guide to the
2. Menditto A, Patriarca M, Magnusson B (2007) Accred Qual expression of uncertainty in measurement
Assur 12:45–47 14. Hubert P et al (2004) J Pharm Biomed Anal 36:579–586
3. JCGM 200 (2012) International vocabulary of metrology: basic 15. Hubert P et al (2007) J Pharm Biomed Anal 45:70–81
and general concepts and associated terms (VIM), 3rd edn 16. Hubert P et al (2007) J Pharm Biomed Anal 45:82–96
4. ISO 5725-1 (1994) Accuracy (trueness and precision) of mea- 17. Hubert P et al (2008) J Pharm Biomed Anal 48:760–771
surement methods and results: part 1: general principles and 18. NIST technical note 1297 (1994) Guidelines for evaluating and
definitions expressing the uncertainty of NIST measurement results.
5. ISO 3534-1 (2006) Statistics: vocabulary and symbols—part 1: Appendix D ‘‘Clarification and Additional Guidance’’, section
general statistical terms and terms used in probability D.1.1.1
6. ISO, IEC 17025 (2005) General requirements for the competence 19. IUPAC (1997) Compendium of chemical terminology. In:
of calibration and testing laboratories. ISO, Geneva McNaught AD, Wilkinson A (compiler) The ‘‘Gold Book,’’ 2nd
7. ISO 15189 (2007) Medical laboratories: particular requirements edn. Blackwell Scientific Publications, Oxford
for quality and competence
8. DIN 55350-13 (1987) Concepts in quality and statistics: concepts
relating to the accuracy of methods of determination and of
results of determination

123