Test statistici in Excel preliminari alla validazione e al controllo di qualità di misure chimico-fisiche.

As part of the research contract "Validation and quality control of chemical-physical analysis of compounds that affect the environment and health" between Intertek Italy Spa and the National Research Council (CNR) Pisa 2015, we have developed some software procedures for automatic recognition of large errors (outliers) in chemical-physical measures. Three different types of statistical tests1 were considered: IQR, Grubb and Hampel test which, together with the normality test, the Chi-square test of Pearson1, contribute synergistically to give more reliability to the subsequent statistical predictions to be applied to numerical series. These preliminary tests, although not visible, have been implemented and made executable in the same Microsoft Excel(r) spreadsheets which were stored algorithms responsible for the validation and quality control. An example of measures of a compound of the family of phthalates is described and used to display the effectiveness of these preliminary tests.