Ferramentas multivariadas para triagem psicológica de condutores

Abstract

The psychological evaluation is a prerequisite for obtaining a national driver's license (CNH) in Brazil. Psychological tests are intended to investigate, evaluate, classify, identify, and standardize psychic functions considered to be fundamental for drivers, and psychological examination mainly takes into account aspects related to behavioral actions of attention, memory, and reasoning. Studies suggest that these aspects are related, among other factors, to age. In this way, this work aimed to investigate tools capable of performing the pattern recognition between the performance in the psychological tests of attention, memory, and reasoning, age and gender, among other categorical variables such as education, marital status, religion, dominant hand, intended CNH and category, paid activity by the candidate, state of birth and city of residence. The study was carried out in a group of male and female candidates that sought to obtain CNH using Principal Component Analysis (PCA) and Independent Component Analysis (ICA). The results of ICA and PCA allowed concluding that there is no distinction between male and female individuals in the performance of the psychological tests applied to obtain CNH. The PCA results indicate an inverse relationship between age and psychological memory tests and logical reasoning. On the other hand, the ICA results show statistical independence between these variables. According to this result, performance in memory and logical reasoning tests will not always be inversely related to age. The application of unsupervised pattern recognition tools allowed for the identification of a direct relationship between age and performance in the psychological tests of attention. The Correspondence Analysis (CA), applied to the evaluation of the categorical variable, showed that the intended categories are closely related to the candidate's age and educational level. The supervised pattern recognition tool by Partial Least Squares with Discriminant Analysis (PLS-DA) allowed concluding that the psychological tests applied in the context of obtaining or renewing CNH (TEACO, TEADI, TEALT, TEPIC M and BETA III) do not correlate with schooling, gender, CNH or category desired, marital status, religion, dominant hand, city of residence or state of birth, or with the remunerated activity on the part of the individual.