From Legal to Illicit Substances: Unveiling Personality Influences with Machine Learning

The rapid increase of drug consumption in recent years can be linked with the personality and demographic traits of the user. This social phenomenon could be considered to stem from legal narcotic substances whose systematic use may lead to illicit use due to personality factors of the user. Although the treatment of the problem focuses on drug rehabilitation, a deeper understanding of the demographic factors influencing substance use could help identify drug consumption risk as well as in its most effective treatment. In this paper, two experiments were performed. The first experiment uses two different Machine Learning Algorithms to predict transition from nicotine to illicit drug consumption based on user’s personality traits. The second focuses on the investigation of personality and demographic factors that influence transition from legal drug use to illicit regarding users’ education level. A dataset of 1885 participants was examined using features derived from the Five Factor Model, a widely used method in the framework of psychology for analyzing personal traits and behaviors. The results demonstrate significant associations between specific personality traits and illicit drug consumption. Targeting nicotine as the legal drug, the drug-personality associations were determined concerning both nicotine and non-nicotine users based on their education level. The results of the research could provide valuable information on the role of personality in drug consumption with a view to more effective, user-tailored, rehabilitation as well as for future research to assess the risk of drug abuse.