Back to search

University of Nevada - Reno

Risk, Need, and Racial Inequality: A Machine Learning Analysis of Rearrest in Juvenile Drug Treatment Courts and Traditional Juvenile Courts

Abstract

dc:description.abstract

Juvenile justice system involvement has many impacts on the lives of youth. This often includes negative outcomes for youth who receive highly punitive treatment rather than more rehabilitative approaches. One approach to reforming the juvenile justice system to be rehabilitative is the use of diversion options, such as Juvenile Drug Treatment Courts (JDTCs). JDTCs are intended to offer more personalized interventions for youth based on their risk and need factors as compared to Tradition Juvenile Court (TJC) settings. To better understand the complex interactions of tailored programming and individual factors for justice-involved youth, an integrated theoretical approach, including the Risk-Need-Responsivity framework and Disproportionate Minority Contact, was used to frame the current study. This study applied machine learning analysis techniques (random forests and logistic regression models) to a rigorous, longitudinal secondary dataset of youth in JDTCs and TJCs to determine which risk and protective factors were most important in predicting rearrest up to 1 year following court intake. The sample included 415 youth from JDTCs and TJCs in 10 jurisdictions across the US. Results revealed that both random forest and logistic regression models performed similarly for each court type as well as the combined sample, and that models were most accurate for the JDTC sample and least accurate for the TJC sample. Highly influential risk factors associated with higher likelihood of having at least one rearrest during the study period included higher scores on the family ineffectiveness scale, social risk scale, and crime and violence screener. Alternatively, highly influential protective factors associated with higher likelihood of not having any rearrests during the study period included not having an assessed risk level assigned to youth and being of Hispanic ethnicity. Race and previous juvenile justice system involvement were not important features in preliminary models and therefore were excluded from final models. Implications for future research, data-driven decision-making practices, and the ethics surrounding the use of machine learning models for juvenile justice involved youth are discussed.

Degree

thesis:*
Level thesis:degree_level
Doctorate Degree
Department thesis:degree_department
Social Psychology
Year dc:date.issued
2022

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Lyons, Courtney
Advisors dc:contributor.advisor
  • Evans, William P
  • Marsh, Shawn C
Committee members dc:contributor.committeemember
  • Berthelot, Emily
  • Hand, Emily
  • Yelderman, Logan

Subjects

dc:subject × 6

Rights

dc:rights
Statement dc:rights
  • Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 United States
  • Author(s)

Identifiers

dc:identifier.*
Handle dc:identifier.uri
http://hdl.handle.net/11714/8387

Chain of custody

source
Harvested from
University of Nevada - Reno
Base URL
scholarwolf.unr.edu/server/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Lyons, Courtney. Risk, Need, and Racial Inequality: A Machine Learning Analysis of Rearrest in Juvenile Drug Treatment Courts and Traditional Juvenile Courts. Doctorate Degree thesis, 2022. http://hdl.handle.net/11714/8387