Survey (statistical)

A survey is a method Statistique aiming at evaluating the proportions of various characteristics of a population starting from the study of a part only of this population, called sample. The proportions are given with margins of error, in which the proportions sought with such or such probability are.

By Métonymie, the word survey also indicates the document having the results of the study by survey.

The most known surveys of the general public relate to human populations. They are in particular the opinion polls carried out by the survey institutes. But it is only one particular application of the technique of the survey, of which the use is much more general.

General principles of the surveys

To carry out a survey right and adapted to the population concerned, it is necessary to choose the adequate technique of sampling. There exists indeed of many methods making it possible to create a representative sample of the population mother. There are initially the probabilistic techniques then the nonprobabilistic techniques. In the first case, the individuals of the population are drawn randomly and have consequently a whole a probability, equalizes or not, to appear in the sample. In the case of a nonprobabilistic sampling, all the individuals do not have inevitably a nonnull probability to appear in the sample.

History of the technique of the survey

cf: Dictionary of Economy and social sciences of Jean-Yves Capul and Olivier Garnier with the Hatier Editions.

Various applications

The surveys are thus not reduced to the opinion polls, and this method can apply to all kinds of populations, human or not.

Random links: