The outcome of an analytical measurement (application of the
chemical measurement process), or
value attributed to a
measurand. This may be the
result of direct observation, but more commonly it is given as a
statistical estimate derived from a set of observations. The
distribution of such estimates (estimator distribution) characterizes
the chemical measurement process, in contrast to a particular
estimate, which constitutes an experimental result. Additional
characteristics become evident if we represent
as follows:
The
true value,
, is the value
that
would result if the chemical measurement process were
error-free. The
error,
,
is the
difference between an observed (estimated) value and the true value;
i.e.
(signed quantity). The total error
generally has two components, bias
() and
random error(), as indicated above. The limiting mean,
, is
the asymptotic value or population mean of the distribution that
characterizes the measured quantity; the value that is approached as
the number of observations approaches infinity. Modern statistical
terminology labels this quantity the expectation value or expected
value,
.
The bias,
, is the difference between the
limiting mean and the true value; i.e.
(signed quantity).
The
random error,
, is the difference between an observed
value and the limiting mean; i.e.
(signed quantity).
Source:
PAC, 1995, 67, 1699
(Nomenclature in evaluation of analytical methods including detection and quantification
capabilities (IUPAC Recommendations 1995))
on page 1705