[Numerical synthesis of statistical data].

The paper describes the methods currently used for numerical synthesis of statistical data. A review is made of the main properties of the various statistical indices required to describe the central tendency, dispersion and shape of distribution of values. Mode, median, arithmetic mean, harmonic and geometric mean, range, quantiles and standard deviation, asymmetry and kurtosis indices are described and discussed using data drawn from practical experience. Special attention is paid to the errors most frequently incurred using these methods. An appendix supplies details of some of the mathematical properties of the operators described.