It should be noted that good precision does not necessarily produce good
accuracy (analyst A) and poor precision does not necessarily produce poor
accuracy (analyst B). However, confidence in the analytical procedure and the
results is greater when good precision can be demonstrated (analyst D).
Accuracyis generally the more important characteristic of quantitative data to
be assessed, although consistency, as measured by precision, is of particular
concern in some circumstances. Trueness is a term associated with accuracy,
which describes the closeness of agreement between the average of a large
number of results and a true or accepted reference value. The degree of accuracy
required depends on the context of the analytical problem; results must be shown
to be fit for the purpose for which they are intended. For example, one result may
be satisfactory if it is within 10% of a true or accepted value whilst it may be
necessary for another to be within 0.5%. By repeating an analysis a number of
times and computing an average value for the result, the level of accuracy will be
improved, provided that no systematic error (bias) has occurred. Accuracy
cannot be established with certainty where a true or accepted value is not known,
as is often the case. However, statistical tests indicating the accuracy of a result
with a given probabilityare widely used (vide infra).
Precision, which is a measure of the variabilityor dispersionwithin a set of
replicatedvalues or results obtained under the same prescribed conditions, can
be assessed in several ways. The spreador range (i.e. the difference between the
highest and lowest value) is sometimes used, but the most popular method is to
estimatethe standard deviationof the data (vide infra). The precision of results
obtained within one working session is known as repeatabilityor within-run
precision. The precision of results obtained over a series of working sessions is
known as reproducibilityor between-runs precision. It is sometimes necessary
to separate the contributions made to the overall precision by within-runand
B2 – Assessment of accuracy and precision 27
Correct
result
A
B
C
D
19.70 20.00
Titer (cm^3 )
20.30
Fig. 1. Plots of titration data to distinguish accuracy and precision.