Arc. Hill et C. Von Holst, Factor transformation to produce statistics describing the uncertainty of analytical data, ANALYST, 126(11), 2001, pp. 2053-2060
Analytical uncertainty produced by random error has a positively skewed dis
tribution and accuracy and precision have non-linear scales. Compared with
conventional statistics, factor transformation of the data allows more appr
opriate interpretation of results and facilitates graphical inspection of d
ata. Statistics are compared for practical examples of performance in profi
ciency tests and of repeatability and reproducibility in collaborative stud
ies. Factor transformation is shown to be applicable to wide ranges of anal
yte concentration and measurement precision.