3. Characteristics of a distribution. Central tendency and dispersion

Up to now, we've only been concerned with the representation of statistical data. However, while it's true that the various tables and graphs defined above "summarize" the distribution, they don't allow any quantification. The aim of this paragraph is therefore to define, for each type of statistical distribution, a certain number of characteristics (or indicators), i.e. a few numbers that enable us to summarize each distribution in a quantitative (rather than qualitative) way. Of course, not just any quantity can be an indicator. In 1950, the statistician Yule set out a number of "common sense" properties that statistical indicators must, a priori, verify. According to him, they must :

• be objectively defined (and therefore independent of the observer);

• use all observations ;

• ...