7.4 Trueness and recovery
Measurement trueness is the closeness of agreement between the average of an infinite number of replicate measured quantity values and a reference quantity value. Trueness can be evaluated on the basis of the observed measurement bias (estimate of a systematic measurement error) as:
or in percentage units as 
Bias shall not be confused with recovery (R - the amount measured as a percentage of the true amount of analyte in the sample), a term
commonly used in analytical chemistry to assess the yield of a particular complex sample treatment procedure (digestion,
concentration, etc.). Recovery is estimated after the analysis of reference samples treated with the whole sample preparation procedure as
.
As many EDXRF methods do not imply a complex sample preparation procedure, it is advisable to avoid the use of the recovery term, but instead to evaluate the trueness by means of the observed bias, which can be done in different ways.
The analysis of certified reference materials (CRMs) for which the element weight fractions are provided with a stated reference uncertainty (see uncertainty in section 7.7). CRMs must be selected as to have a matrix similar to that of the analysed samples and in an amount that would cover the interval of expected concentrations in the unknowns. The comparison of average results is meaningless without providing the measurement uncertainty, which must overlap with the stated uncertainty of the reference value.
In the absence of suitable CRMs the analytical results can be compared with those obtained by using an alternate reference method (e.g. a method with proven superior accuracy). The comparison of the results obtained by the two methods must incorporate an evaluation of the variances of the results as well.