INTERLABORATORY STUDIES FOR THE PURPOSE OF ANALYTICAL QUALITY-CONTROL

For interlaboratory studies, which are carried out not for validation of analytical procedures, but for the purpose of external analytical quality control, a new method of evaluation is described. Data originating from measurements performed for the same analyte in different concentration levels can be evaluated using a set of completely identical criteria. It provides the data base free of outliers together with some information concerning the statistical variability of acceptable data and its dependence on concentration. The method presented ist based on the normal respectively log-normal distribution of relative deviations between set and measured values. The method is suited especially for the evaluation of large data sets with several hundreds of single values and have already been applied successfully at many interlaboratory studies concerning different parameters of chemical water analysis.