Abstract: | Abstract Data, obtained in quality control programs in automated clinical chemistry systems are often not normally distributed. Falsely assuming that the data is Gaussian produces errors in classifying patients, and in statistical inferences drawn from parametric tests. |