In the second part of the article the author presents one of the implementations of his approach to analysis of public survey data, namely the statistical technique and results of its application, the necessity of which is proved in the first part of the article. The object of the statistical analysis is a network of nodes which represent the nominal values of sociological variables, and the links between nodes are statistically significant dependencies between the values of the variables, represented by the values of standardized residuals in the cells of the contingency tables. We study the local properties of such network in the nodes, which are variants of answers to similar questions about the trust in the four public institutions, given during the polls conducted on the representative nationwide samplings in 1998, 2001 and 2015. It is shown that the various characteristics that reflect of the centrality these nodes carry important information about the state of public opinion. The author demonstrates a qualitative change in the state of public opinion during the study period. It is shown that such changes are not homogeneous, but rather depend on the objects that are measured by the public opinion. The author substantiates the dependence of these changes on the state of the political system and society, and suggests a hypothesis, according to which, in some cases, the standard simple approaches to the analysis of the results of public survey data cannot provide relevant results, interpretable in a simple and familiar way.

 

 

trust; political sociology; public opinion; opinion survey; consistency; mathematical and statistical methods; social network analysis; data mining.