Presentation Master's Thesis - Julia Broderick-Hale - Brain & Cognition

Youth criminality is a serious societal issue, however the processes that lead a child into it are poorly understood. Many risk factors have been identified, whereas protective factors (which lower an individual’s likelihood of delinquency) have received less attention. A comprehensive understanding of youth delinquency requires a model that considers both risk and protective factors, but integration of protective factors into prediction models is conceptually and methodologically challenging. Bayesian Network Modelling (BNM) allows quantitative estimation of the influence of various factors on the likelihood of an outcome (e.g., delinquency). It can also estimate the likelihood of the outcome given the presence/absence of different factors. 

BNM is typically applied to risk factors. The current research is a proof-of-concept study exploring the viability of expressing protective factors in BNM. Statistics representing associations between protective factors and delinquency were gathered from published literature and used to build four small Bayesian Network Models. Analysis demonstrated that, using these models, the influence of different (combinations of) factors on the likelihood of the outcome could be quantified. Additionally, the models were validated through sensitivity analysis, which deemed them robust. Exploratory analyses showed that expected characteristics of protective factors could be seen in the models. In conclusion, the viability of BNM of protective factors was supported. To provide a roadmap for research following this exploratory proof-of-concept, procedural recommendations are outlined. Applications for this research include the creation of assessment tools and the development of interventions tailored to the factors most relevant for a particular individual.