Automated Meta-Analysis of Risk Factors Associated with Violence and Recidivism

Violence and delinquency have been central topics in criminology, with a growing body of research examining risk factors for violent behavior and recidivism. The increasing volume of studies highlights the need for systematic, scalable meta-analytic approaches to identify patterns across risk factors, behavioral outcomes, and psychosocial domains.This study introduces an automated text-mining approach to analyze the relationships between predefined risk factor terms and their associations with violence and recidivism. We curated dictionaries of terms spanning familial, cognitive, peer, personality, and substance-related factors, among others. Using these, we systematically collected scientific articles and measured co-occurrence probabilities to generate data-driven profiles of risk factors. This large-scale dataset enables the exploration of key associations, comparative analyses, and novel insights into how risk factors interact across behavioral and social contexts. By providing an openly available dataset, this study serves as both a pedagogical tool and a foundation for advancing criminological ontology, emphasizing the need for standardized terminology and a structured framework for risk factor classification.