Personality and secure & risky behaviors: a hybrid SEM and machine learning approach

Abstract This study aims to assess the relationship between personality traits and cyber security behaviors using a hybrid approach augmenting Covariance-based Structural equation modeling (CB-SEM) with Machine Learning (ML). Data were collected from 919 participants through an online survey, and the relationships of personality traits with secure and risky behaviors were tested in two models. In cyber secure behavior model, Agreeableness, Conscientiousness, Openness, Machiavellianism and Narcissism are positively, while Neuroticism is negatively associated with cyber secure behavior, explaining 17% variance. In cyber risky behavior model, only one factor from both BF-5 and DT, conscientiousness and psychopathy, were associated with cyber risky behavior, explaining 18% of the variance in cyber-risky behavior. The application of machine learning analysis further revealed that the personality-based cyber secure model had a predictive accuracy of 74% while cyber risky model had a predictive accuracy of 80%. The implications of this study call for the design and development of online sites based on personality traits and the development of educational programs to increase cyber security.