Depending on the purpose of the study, the object of study may be a single object or an entire class of objects. For example, we write the history of the organization in which we work. This study by its nature will be purely ideographic. The peculiarities of studying the individual case will be considered later. But most often the researcher deals not with a single object, but with a whole class of objects.
If the number of objects is not too large, it is in principle possible to conduct a continuous survey, that is, to collect information about each object. Let’s say that we are interested in large families living in this neighborhood. If their number does not exceed two or three dozen, then it is quite realistic to organize a general examination. It is only necessary to specify in advance the exact criterion of referring to this category, that is, to specify how many children should be in the family in order to consider it to be of many children.
However, very often the class of objects studied is very numerous. To embrace it is completely too difficult, and the purpose of the study does not foresee this. In this case, they limit themselves to researching a certain subclass (sample), on the basis of which they judge the class as a whole. It is clear that the sampled sample should be representative, that is, it is enough to represent this class, to be typical. Then we will be able to spread the conclusions obtained on a limited sample to the whole population. It is clear that any generalization is done with a greater or lesser degree of certainty. The practice of scientific research has developed a number of techniques to ensure sufficient reliability of the conclusions obtained on the basis of sampling.
Sociologists involved in the study of public opinion made a significant contribution to the development of the technique of correct sampling. The central task on which they perfected these techniques was the task of predicting the results of voting on the scale of the whole country on the basis of sample surveys. The technique of conducting mass polls by now is so improved that it is possible to predict with a reasonable degree of reliability the results of voting of millions of voters, based on the answers of a relatively small (two or three thousand people) group. But the logic on which the idea of selective research is based is of a generally scientific nature. For example, it is widely used in the theory of reliability of measurements.