Identifying patterns of delinquent trajectories and testing stability of self-control over time among South Korean youth using multivariate latent growth curve modeling

Wanhee Lee


The South Korean yearly national report revealed that since 2006 there has been a steady increase in juvenile crimes (Seoul Police Department, 2009). In addition, the report demonstrated that South Korean juveniles' age of onset in delinquent activity has been continuously decreasing. In South Korea, the age-crime curve sharply peaks at age 16, holds constant until 19, and then begins to decline. Thus, this "peak" within the age-crime curve has been a frequently researched topic. This has resulted in some empirical support demonstrating that the age in which criminal involvement peaks is considered the most dynamic period in individuals' life-course (Wiesner & Windle, 2004). Thus, juveniles are most susceptible to criminality when they are experiencing the most physical, emotional, and academic changes. With such a lack of stability, juveniles may seek control through other means; thus committing deviant acts. The purpose of this study is to identify distinctive trajectories of delinquent behaviors during adolescence, and based on observed patterns, examine the association between variables representing control theories of crime, demographic variables, and their relationship with identified developmental trajectories of delinquency. This study will be conducted through use of the Korea Youth Penal Study (KYPS), a six-year longitudinal study (from 2003 to 2008) of South Korean youth. This data will be employed to examine how social control affects delinquency involvement throughout the life-course by examining developmental trajectory patterns. Additionally, this study will examine the stability of self-control as a time-variant variable, as well as how levels of self-control relate to offender groups across a five-year period among South Korean youth. The method of analysis consists of two stages. First, it aims to identify distinctive patterns of juvenile delinquency by applying the method of dynamic classification of the offender model as first implemented by Loeber, Stouthamer-Loeber, Von Kammen, and Farrington (1991) in their study on juvenile offending. Then, contingent on observed patterns, a series of latent growth modeling (LGM) will be used to examine the trajectory of delinquent youths' individual growth or change curves, as well as the influence of the levels of self-control on juvenile delinquency over time. The current study will provide information on the developmental trajectories of South Korean youth and how those behavioral patterns/trajectories significantly affect various offender groups. The results from this analysis will be examined in light of previous findings and policy implications discussed.