Characteristics of Juvenile Delinquent Behavior and Their Association with Adult Arrest: A Latent Class Approach

Background/Purpose:Juvenile delinquency is often researched in terms of criminal rule violations, which reduce youth to the mutually exclusive groups of juvenile delinquent or non-delinquent juvenile. However, research has demonstrated that juvenile delinquent behavior does not consistently transition to adult criminal behavior. Because delinquency does not accurately predict adult criminal behavior, better methods of identifying subtypes of delinquency are needed. We hypothesize that delinquency has patterned behavioral characteristics that can be separated into unique classes.  Identifying behavioral characteristics that more accurately describe subtypes of delinquency would more accurately inform early interventions for the purpose of reducing later-life socially deviant behavior.

Methods:  Latent class analysis (LCA) and logistic regression were used to conduct analyses of data from the National Longitudinal Survey of Youth 97. The analytic sample included 8905 participants. Juvenile delinquency was measured in 1997 using twelve indicators of social deviance (i.e., smoking under age, stealing, assaulting others, etc.). Adult arrest was developed from subsequent interviews where the participant indicated an arrest, and was between the ages of 18-23. After identifying the delinquency classes from the LCA, logistic regression was used to assess the relationship between delinquency classes and adult arrest, while controlling for sex and race/ethnicity.  

Results:Results of the LCA indicated five distinct classes of juvenile delinquency: Class 1 - little to no probability of social deviance; Class 2 - statutory deviance (i.e., behaviors that are only delinquent because the youth is under age); Class 3 -property deviance (i.e., behaviors such as theft or damaging property); Class 4 - high indications of both statutory and property deviance; and Class 5 - high levels of social deviance across all indicators, including potentially violent offenses.

Results of the logistic regression indicated that compared to the non-deviant group, all four of the deviant classes had higher odds of adult arrest. The most socially deviant group had the greatest odds of adult arrest (OR = 9.37; 95% CI: 7.09,12.34) followed by the combined property and statutory offenses group (OR = 4.40; 95% CI: 3.72-5.21), the statutory offenses group (OR = 2.28; 95% CI: 1.94, 2.68) and the property offenses group (OR = 2.65; 95% CI: 2.28-3.08).

Conclusions/Implications:This study extends knowledge about the relationship between delinquent behavior and adult arrest by suggesting that adult arrest does not occur simply as a result of juvenile deviance, but from the subtype of characteristics associated with delinquency. Because deviant groups had higher odds off arrest as compared to non-deviant groups, one interpretation could be that these findings actually reflect how we typically dichotomize people into delinquent or not. Further, this research suggests that during adolescent development, greater odds of adult criminality occur as youth expand deviant activities, suggesting the need for additional research to 1) identify aspects of youth development that increases the probability of property or statutory offenses, 2) examine if juvenile co-occurring statutory and property deviance develops concurrently or consecutively, and 3) identify best practice interventions for each class. Implications for intervention and policy are discussed.