Methodology for constructing a factorial system of personality traits based on popular personality typology

This article describes a methodology for constructing a factorial system of personality traits based on statistical data obtained within the framework of studying popular personality typology. The results of the analysis of data from the socionic «Talanov Personality Questionnaire» for 2015–2016 are presented. Since there is no single authoritative source that defines the meaning of concepts used in popular typology, it is rather a subculture that develops its own language for describing personality characteristics. It is assumed that due to the «wisdom of the crowd» effect, popular typology contains important personality traits in the averaged images of types. Thus, information about the types indicated by the respondents of the questionnaire can be used to validate and optimize the system of personality factors. A factor analysis of the socionic questionnaire was conducted with subsequent simplification of factors, and for optimization and control, the quality metric of logistic regression classification on the constructed factors was used. Thirteen simple factors with weight in the classifiers of socionic types were obtained and ranked. Linear models by factors of socionic types were constructed, and threshold values for classification were determined, at which the distribution of factors in the models of types is closest to the distributions of factors by the declared types. Using the large linguistic model Gemini, asubstantive interpretation of the factors was obtained, and a comparison with the factors of the «Big Five» was made. The methodology described in this article can be used to analyze questionnaires of popular typologies of interest to a wide audience, while the collected information about personality qualities will be represented in a factorial form, which is more accurate and better applicable in research than information encoded in types.