The statistical sample is the part of a population that is considered for the development of a study.
Sample is a concept with several uses. It may be the part of a whole that is considered representative or characteristic of it; of an example to imitate; or an exhibition. Statistics , meanwhile, is the area of mathematics that uses figures to generate inferences from the study of probabilities. The term also refers to the analysis of quantitative data .
A statistical sample , in this framework, is a portion of a population that is taken for the development of a study . From the examination of this sample, it is possible to know some characteristic or quality of the larger group.
What is a statistical sample
To understand the concept of a statistical sample, it is first necessary to focus on other terms. Population is called, in this context, the set of elements that constitute the object of interest of statistics.
The population is made up of multiple individuals (which can be people or things). Depending on the type of analysis carried out, there is a difference between an exhaustive study (covering the entire population) and a sample study (focused on a portion of the population in question).
In this way, the statistical sample is the portion of the population that is used to carry out the study . We talk about sample size to refer to the number of individuals in the population that are considered for research.
It is important that the statistical sample is representative of the population.
How to determine
It must be considered that an exhaustive statistical study is very expensive and time- consuming. An example of these works is a national census , which is usually done every ten years. Therefore, a sample study is generally used.
The determination of the sample is key for the study to yield valid or useful conclusions. The chosen individuals, therefore, have to represent the different strata of the population.
Suppose you want to investigate Internet access in a country. If the statistical sample is made up of inhabitants of a single city, the inferences will not be appropriate to understand the reality of the nation in general. The correct thing would be to extract individuals from many locations, with various characteristics (social, economic, etc.).
Main properties of a statistical sample
This search for representativeness of the statistical sample is based on two central factors. On the one hand, the sample must have a size that is large enough: if the population has 5,000,000 individuals , a sample cannot be assembled with 20 or 30 .
The other basic issue is the random aspect. Although the selection must have an initial orientation (so as not to extract individuals from a single social group, to mention one possibility), then randomness needs to be guaranteed so that the sample is not biased.
At this point, a distinction can be made between a probabilistic statistical sample (all individuals have identical probabilities of being selected) and a non-probabilistic statistical sample (the choice is linked to the type of procedure chosen, which can be discretionary, by quotas ). or of another kind).
With the statistical sample already defined based on these criteria , we can proceed to the study that makes the inference of the metrics possible. This step is carried out working specifically with the data obtained from the sample.