FEATURES OF THE USE OF DESCRIPTIVE AND ANALYTICAL STATISTICS IN SOCIOLOGICAL RESEARCH

Zamyatin Vladimir Mikhailovich, Molchanova Elena Ivanovna

The purpose of the article is to identify the features of the analysis of sociological distributions of indices, linearly expressed based on data from large random samples. The relationship between the general drug addiction risk indicator and the normal type of distribution is considered. A data matrix for 75 indicators was obtained by testing 2109 ISTU students using the ASSOL IS. It is shown that, contrary to the opinion of Russian authoritative scientists, methods of descriptive and analytical statistics based on the normal distribution law can be applied to indices calculated from groups of factors consisting of ordinal and binary indicators. Using the 3  criterion, respondents with an increased risk of drug addiction were identified. Estimates of pair correlation according to Pearson, according to Spearman and using the phi coefficient (for binary scales) were compared. If there are related ranks in the data sample, the calculation of the Spearman coefficient should be carried out taking into account corrections. All three types of correlation coefficients gave similar results. It has been established that when analyzing large samples there are no clear criteria for assessing the strength of the connection. It has been shown that identifying strong pairwise correlations between indicators makes it possible to reduce the size of the questionnaire without loss of reliability.

Keywords:sociology, Pearson correlation coefficient, Spearman correlation coefficient, phi correlation coefficient, analytical statistics, descriptive statistics, standard normal distribution, random quantity, indicator, index, scale