The statistic comes from the Latin ” status “: state. We can distinguish different phases in the evolution of the statistics that we detail below.
In antiquity, the first traces of the statistic were found in Egypt with the recording of the floods of the Nile. But we will also find the Chinese or the Indians1. Greek philosophers, like Aristotle in Antiquity, approached in their time the beginnings of what will be descriptive statistics2 :
« At the same time, following the policy inaugurated by Aristide, or ensured to the multitude largely his livelihood. It came to pass that by the extraordinary contributions, by the rights and taxes, by the Allies, more than twenty thousand men were fed. There were indeed six thousand judges, a thousand six hundred archers and in addition twelve hundred riders; The council had five hundred members, the guards of the arsenals were five hundred, and the guards in the city numbered fifty; About seven hundred men exercised judiciaries in the country; About as much, outside the country. Later, when Athens began the war, there were two thousand and five hundred hoplites, twenty cruise ships, and other ships for the collection of the tributes on board the two thousand men designated by the fate. Let’s add the Prytanée, the orphans, the jailers. All this world was drawing its livelihood from public incomes. »
These examples show the universality of the statistical approach and the desire to recognise, compare… This overall aspect of the statistical approach by States is found in another element present in the early days of statistics, namely the measurement of soils and properties as they appear in the Constitution of the various cadastres.
It appears that the first position parameters that have been used are the mode, the most frequently occurring value, and the ” middle of the interval defined by extreme values “. The arithmetic mean appears clearly in the work of the Danish astronomer Tycho Brahe (1546-1601), who, by constituting a dataset on the movement of the planets, allowed Kepler to formulate his laws.
However, the first real traces of statistics as we know it today, are found from the 17thcentury via reflections on the chance carried out by Pascal, Fermat or even Huygens. These search come in parallel to the creation of comma, trigonometric tables and logarithm.
It will be in the 18th century that Professor G. Achenwall is given the creation of the word Statistik in 1746. This term is based on the following etymological path3 :
Latin: status = State >> Italian: Stato = State >> XVIth Statista = statesman >> 1633: Statistica = relative to the state >> 1672: Statistic in French and Statistik in German meaning at the time: methodical study by numerical processes of social facts which define the state
The name changes its value, first in England, by becoming ” statistics ” (1798), then in France in 1832, and takes the meaning of: “A set of mathematical interpretation techniques applied to phenomena. “
The name then refers to (1862) The purpose of the statistics: ” numerical data set for the same fact category “.
The 17th and 18th centuries will be key in the development of statistics with the profusion of research, theorem… Then, it will be with the ” Prince of Mathematics “, Carl Friedrich Gauss (1777-1855), with the law that bears his name, that the statistics that we know today begin.
From the 19th century until the middle of the 20th century, mainly thanks to the English school (Galton, Pearson, Gosset, Yule, Fisher), these are developed new statistical tools whose statistics inferential and especially the Understanding of the practical usefulness of statistics in many areas other than the state’s use. Overall, the development of this science has been done in the wake of the development of many other sciences in mathematics or physics.
Since the years 1960, with the development of knowledge and the trivialization of computer tools, statistics have evolved much, especially with multidimensional descriptive statistics (graph…). Now, statistics are everywhere to understand our consumption needs that predict the results of elections…
The two families of statistics
The oldest statistic is the description of a data set. This description can be done in a numerical or graphical way. In this objective, there is therefore no probabilistic hypothesis on this data, it is a gross representation.
By definition, “an inference is an operation by which one passes from one truth to another, judged as such according to its connection with the first“. In other words, inferential statistics, unlike descriptive statistics, aim to induce general behaviour according to the individual (example of surveys)5.
1 – J. C. Oriol (2007) – A historical approach to statistics: from the first steps to the “Geometry of Chance” and the “Masters of Error”
2 – Aristotle (330 BC)-Constitution of Athens
3 – The Robert-Edition 1993
4 – A. Gomez (2004) – Statistics for historians
5 – C. P. Dancey, J. Reidy (2007)-Statistics without maths for psychologists