When studying the properties of an object, the characteristic is determined not all of the units of the general population, but only some of its parts. This allows, performing selective observation of the selected area, to sufficiently fully characterize all objects as a whole by some specific distinguished feature or property.
In order for the selected objects to be able to represent all the units under study with a sufficient degree of probability, the work must be organized in a certain way.
There are several traditional methods that allow you to extend a sample observation to the entire population, the main ones are the following: method of coefficients and direct conversion. The first option is used in cases where it is necessary to verify or clarify the data obtained as a result of continuous observation. In the second case, the product of the average value of the trait under study is determined by the entire volume of the general population. In this case, it is necessary to take into account the fact that not all factors make it possible to fully apply the point estimate when disseminating the results obtained for the sample to the general population. Therefore, in practice, the interval estimation has found widespread use, with which you can take into account the size of the marginal error of part of the population and calculate it for the average, and, if necessary, for the share of the attribute.
When the study is not conducted on the entire population, but only on its part, inevitably errors of selective observation can occur, which cause the deviation of the obtained results from the true values. At the same time, there are two main reasons that can provoke distortions of the final result:
- imperfection of the technical conditions of the study, the so-called registration error;
- violation of the rules, which can be both random and systematic in the selection of units for research.
Selective observation in statistics is quite widely used. In this case, reliable information about the objects of study can be obtained by significantly saving money and reducing costs. The guarantee of the representativeness of the research is scientifically based selection of respondents for the study.
In practice, selective observation is most often carried out at facilities that were randomly or regionally selected. In the first case, the same probability is ensured that one or another unit will fall into the sample. If the object of research after receiving information about it returns to the general population, then the random sample is called repeated. Otherwise, it is unique. A zoned sample is formed as a result of dividing the general population into a number of areas (groups) depending on the nature of the trait being studied. The use of a suitable selection technique ensures the presence of representatives from each group in the selected population.
If we talk about how selective observation should be organized, then in this case it is necessary to observe the following conditions:
- the number of units in the sample should be maximum to ensure the identification of the desired patterns;
- it is necessary to adhere to the principle of equiprobable hit of each object in the sample;
- It is worth using all, without exception, parts and areas of the studied population.
This will allow you, minimizing the volume of research, to obtain reliable information on issues of interest to you.