Crime risk prediction, focusing on integrating sociodemographic, behavioral, and psychological data to forecast first-time offenses and recidivism, is a cornerstone of criminal psychology. Assessments typically follow the "80-20" principle, targeting high-risk groups such as individuals with mental illnesses, who are statistically overrepresented in violent crimes. However, individual-level prediction remains difficult due to the limited ecological validity of traditional tools like questionnaires and Situational Judgment Tests, which poorly replicate real-world social and interpersonal pressures.To improve accuracy, actuarial methods employing statistical algorithms and diverse data sources have been developed, yet they often underperform at the individual level. Consequently, research has integrated psychological theories, emphasizing traits like impulsivity, low self-control, and reinforcement sensitivity—the latter influencing decision-making and moral disengagement. Comparative offender studies and advanced analytics are further refining predictive models by analyzing multi-scale psychological data. Simultaneously, virtual reality has emerged as a promising tool, offering immersive, ecologically valid simulations that enhance presence and enable embodied interaction. VR can advance risk assessment by:First,simulating crime-relevant scenarios to better elicit offense-related psychological processes;Second,capturing multi-modal behavioral and physiological data for richer computational analysis; Thrid,leveraging efficient modeling techniques to improve scalability and cost-effectiveness. These pathways position VR as a transformative approach for overcoming the limitations of conventional assessment tools.
Jiao et al. (Fri,) studied this question.