The limits of the permissible error

To quantify the similarity between the sample and the general population use rate, which is called sampling error, the difference (in percentage) between the characteristics of a sample and the General sovokupnostey.Predely allowable error depends on the purpose of the study. Depending on what level of reliability we seek in this case, may be permissible error,%: high reliability, and 3, mild-3-10, approximately 10-20; indicative - 20-40 ...
More ...

The actual error, is still talked about the so-called random error caused by the probabilistic nature of the sample. But there are so-called systematic errors caused by a failed structure vyborki.Dopustim, we are interested in the demand of city residents for a certain product, we decided to interview 400 people. Technically, the easiest-to agree with the command of a military unit to distribute questionnaires and 400 soldiers. However, if we do so, we know the views of 400 young men, mostly unmarried, certain occupations, and the temporary residents of this city. There is no reason to extrapolate that view on other categories of people we net.V Ideally, it is desirable that the sample to match exactly with the structure of the general population. But it is not primarily because of the incompleteness of (usually) of information about the characteristics of the General sovokupnosti.Dalee, providing the sample and form its structure, we are inevitably based on subjective indicators. For example, we need to explore limited aggregate demand, for example, residents of one town or university students. ...
More ...

In analyzing the primary information, the researcher reduces the data obtained in the table on which the output indicators such as the allocation of the frequency, the average levels, the relationship factors, etc. If necessary, the data obtained are processed using advanced statistical methods and decision-making models used in the analysis of marketing informatsii.V a result of the study confirmed a preliminary hypothesis or refuted these hypotheses (the least likely option), or it turns out that they are generally correct, but need to be adjusted (most likely). Such studies should be repeated regulyarno.Pri conducting marketing research techniques are not questioned the entire mass of potential consumers, and a sample-sample. For marketing the use of sampling is crucial, because the market system probability, its inherent properties are uncertainties and risks posed by it. ...
More ...