Detecting errors in micro and trace analysis by using statistics

By assigning a standard deviation to each step in an analytical method it is possible to predict the standard deviation of each analytical result obtained by this method. If the actual variability of replicate analytical results agrees with the expected, the analytical method is said to be in statistical control. Significant deviations between analytical results from different laboratories reveal the presence of systematic errors, and agreement between different laboratories indicate the absence of systematic errors. This statistical approach, referred to as the analysis of precision, was applied to results for chlorine in freshwater from BCR certification analyses by highly competent analytical laboratories in the EC. Titration showed systematic errors of several percent, while radiochemical neutron activation analysis produced results without detectable bias.