A Combined Theory Data-Driven Approach to Predicting Delinquent Risk in the FFCWS

• In Event: Advancing Methodological Approaches to Biopsychosocial Research

Abstract

Machine learning offers benefits such as the ability to analyze multiple domains, as demonstrated by Chen et al. (2023). Their study used a feed-forward neural network to predict conduct disorder, incorporating sociological, psychological, and biological domains, outperforming single-domain models. Extending this approach, our study will assess delinquency risk using data from the Future of Families and Child Wellbeing Study (FFCWS) to capture various risk factors across socio-environmental, psychological, and genetic domains, promising a more comprehensive understanding of delinquency risk. The FFCWS comprises a sample of 4,898 children born in major U.S. urban areas between 1998 and 2000. This sampling strategy intentionally oversampled births to unmarried mothers to ensure comprehensive representation across different family structures and socio-economic backgrounds. The findings of this study will be preliminary and will serve as the first chapter of a dissertation.

Author

• Nicholas Vietto, University of Nebraska Omaha