Testing Effects of Social Support and Service Utilization on Criminal Justice Involvement Using Structural Equation Modeling

Background & Purpose: The adolescent population is one of the fastest growing age groups and are more likely to be charged with a violent felony than adults. Risky behaviors are closely linked to criminal justice involvement (CJI), particularly substance use and risky sexual behaviors. Research substantiates that engagement in these risky behaviors among adolescents can be considered normative given brain underdevelopment and critical periods of risk that extend into the mid- to late 20’s. During this period of risk, social support and behavioral health services can act as protective factors to decrease risky behaviors leading to CJI and reincarceration. However less is known of how social support and behavioral health service utilization impact high-risk adolescents in early stages of CJI. Thus, this study asks: Does social support and/or service utilization mediate CJI among adolescents involved in high-risk behaviors? We hypothesize that presence of social support and high service utilization will decrease CJI.

Methods: Using structural equation modeling (SEM), four waves of the National Longitudinal Study of Adolescent to Adult Health (Add Health) data were utilized to examine the mediating role of social support and high service utilization (Wave 2) connecting risk behaviors (Wave 1) and CJI (Wave 3 & 4). Guided by the Comprehensive Health Seeking and Coping Paradigm, we constructed three latent variables, namely risky behaviors (alcohol use, drug use, and risky sexual behaviors), service utilization (substance abuse and mental health services), and social support (parents, friends, and teachers). The outcome variable, CJI, was dichotomous (yes/no). The weighted least square (WLSMV) estimation corrected for categorical indicators. For missing data, we used multiple imputation with 20 imputed data sets. The Bootstrap procedure aided in obtaining the mediation effect. We controlled for gender, race, birth year, and parental income in all analyses.

Results: The structural model indicated that model fit the data well (CFI = 0.935, TLI = 0.902, RMSEA = 0.031). Results showed that risk behaviors were positively associated with service utilization (b=0.71) but negatively associated with social support (b=-0.49). Further, risk behaviors (b=0.752; exp(b)=2.12) and service utilization (b=0.125; exp(b)=1.13) significantly increased the odds of CJ involvement, but the effect of support was not significant (b=-0.019; exp(b)=0.98). A significant indirect effect (b=0.092) was observed on service utilization but not social support, suggesting that risky behaviors increase service utilization, which in turn, lead to higher CJI. The indirect effect explained 12% of the total effect connecting risk behaviors to CJI.

Conclusions & Implications: Adolescence marks a key developmental stage where prevention efforts could be targeting risky behaviors and CJI. While the results of this study do not fully support the proposed hypothesis, results suggest an opportunity to advance existing prevention models to increase naturally occurring social supports for high-risk youth, as well as improve service utilization models to protect against CJI. More research is needed given the preliminary and foundational results of this study to further understand the impact social support and service utilization can have to decrease adolescent CJI.