Reading Remediation and Outcomes in Detention

Learning disabilities (LD) are among the most common types of disabilities in juvenile offenders that have been linked to delinquency. Nationwide, children and youth with special education needs are overrepresented in the US justice systems. Reports estimate delinquent juveniles with a disability to comprise about 30% to 60% of the entire delinquent population. A national survey in the US states an average prevalence rate of 33.4% of incarcerated juveniles with disabilities in correctional facilities. Moreover, concerns have long been raised on the recidivism rates of youth with disabilities and special education backgrounds. In general, regarding educational performance, academic deficits such as a lack of basic skills in reading, writing, and mathematics have been associated with recidivism. However, these studies largely neglect the dynamic nature of delinquent and criminal behavior that has been documented within the field of developmental criminology with a focus on the onset, continuity, and extinction of deviant behavior. In light of this research, the identification and remediation of LD as a risk factor for repeat offending has been a persistent challenge given the accumulation of and overlap with other risk factors such as poverty, familial patterns of criminality, influence of delinquent peers, and the differential impact of risk factors across an individual's developmental trajectory. Altogether, there is a challenge around the implementation of sophisticated methodology to model the complex longitudinal and reciprocal links between juvenile delinquency and educational problems such as LD, and how they relate to other risk factors over time. This challenge is intensified by the required large samples to detect robust and interpretable patterns and predictive relationships for groups of youth with severe LDs that, by definition, are small in size and censored with regard to various educational outcomes (e.g., academic performance). This research is conceived to contribute to the field's understanding of the connection between LD and delinquent behavior. We hope to generate unique findings capitalizing on the availability of the relevant big data, the clinical strengths of its members and their capacity to develop and administer educational therapy to juvenile offenders, and its embedment within communities empowering the creation and processing of multi-level longitudinal datasets, merging sociological (i.e., criminological), behavioral, neurophysiological, and genetic/genomic data. Big Data: The Big Data component seeks to address three objectives: (1) to conduct big data analytics and data mining to quantify/qualify the complexities of severe LD among juvenile justice (JJ)-involved youth using statistical methods applied to big datasets that capture multiple indicators (i.e., offending and educational outcomes) cross-sectionally (for one-time offenders) and longitudinally (for repeat offenders); (2) to utilize big data analytics to identify a group of youth with severe LD who will serve as a comparison group for individuals participating in a reading intervention; and (3) to evaluate the findings from the first two objectives based on the statistical assumptions underlying the models applicable to the big data at hand. Hypotheses. Extending previous research, Big Data seeks to identify factors that relate LD to recidivism, frequency of re-offending and re-arrest, and the time to post-involvement recidivism. We hypothesize that (a) overall recidivism rates (i.e., regardless of the specific type of offense) will be higher among JJ-involved youth with severe LD (operationalization based on standardized achievement tests and special education status) than in youth without LD; (b) youth with severe LD will show more incidents of school-related problems after their first involvement (e.g., attendance problems, disciplinary incidents such as in-school and out-of-school suspensions and expulsions) than youth without LD; (c) youth with severe LD who had school-related problems prior to their first JJ involvement will be at higher risk of recidivism than youth without LD; (d) lower levels of reading achievement will be associated with shorter periods of time to repeated offending; and (e) participation in the intervention will significantly decrease recidivism risk compared to other youth with severe LD who did not participate in the intervention. Intervention: The intervention component seeks to address three objectives: (1) to improve the reading skills of an understudied, high-risk population (JJ-involved youth) using an educational therapy that has been specifically designed for JJ-involved youth, provided in a 1:1 setting, and expanded with technology, including material relevant to daily living skills and appropriate reinforcement that will extend the program into the post-release period; (2) to reinforce skills and maintain reading gains made during the educational therapy, and increase the automaticity of skill use by using a high-interest "gamified" learning tool that youth will play on a SmartPhone given to them when they leave detention; and (3) to investigate the individual response to therapy with respect to other variables (e.g., attention, executive functioning, impulse control, learning ability, other academic achievement) collected at baseline. Hypotheses. The key hypotheses of the intervention component seek to identify the factors associated with response to therapy for severe LD relative to pre-therapy functioning. We hypothesize that (a) overall improvement in reading will be related to pre-therapy academic skills, attention, and executive functioning; (b) inclusion of a metacognitive component in the therapy will provide carry-over effects, improving other academic skills; and (c) the gamification of the learning tool will result in robust technology usage, with a dose-response relationship between time spent on task and improvement in skills trained (and measured) via game play.