Developmental Trajectories of Delinquent Peer Association Among Korean Adolescents: A Latent Class Growth Analysis Approach to Assessing Peer Selection and Socialization Effects on Online and Offline Crimes

Date01 August 2021
Published date01 August 2021
Subject MatterArticles
Journal of Contemporary Criminal Justice
2021, Vol. 37(3) 379 –405
© The Author(s) 2021
Article reuse guidelines:
DOI: 10.1177/10439862211001617
Developmental Trajectories
of Delinquent Peer
Association Among Korean
Adolescents: A Latent Class
Growth Analysis Approach to
Assessing Peer Selection and
Socialization Effects on Online
and Offline Crimes
Sujung Cho1, Brett Lacey1, and Youngsik Kim2
The relationship between peers and delinquency has been taken as evidence for
selection and socialization effects in the etiology of adolescents. Accumulating evidence
suggests that both effects are involved. This study examines whether adolescents’
aggressive propensities and behaviors predict their peers (selection) and whether
peers’ propensities and behaviors predict adolescents’ behaviors (socialization). The
latent class growth analysis approach revealed three distinct subgroups: an early-
onset group (0.9%); a late-peak group (3.37%); and a normative group (95.73%).
Both selection and socialization effects were supported using a longitudinal Korean
adolescent self-report. The results showed that adolescents with less self-control
who are online more frequently and exhibit higher levels of traditional bullying and
delinquency were more likely to be members of both the early-onset and late-peak
groups compared with the normative group. Also, the aggressive behaviors fully
mediated the link between aggressive propensities and delinquent peer associations.
Furthermore, adolescents in the late-peak group (but not those in the early-onset
group) were associated with a greater likelihood of online and offline delinquency,
1Southern Illinois University Carbondale, IL, USA
2Seowon University, Chungju-si, Republic of Korea
Corresponding Author:
Sujung Cho, Department of Criminology and Criminal Justice, Southern Illinois University Carbondale,
1000 Faner Drive Faner Hall, Mail Code 4504, Carbondale, IL 62901, USA.
1001617CCJXXX10.1177/10439862211001617Journal of Contemporary Criminal JusticeCho et al.
380 Journal of Contemporary Criminal Justice 37(3)
but cyberbullying and traditional bullying in late adolescence levels were high in both
groups’ members.
peer selection, peer socialization, latent class growth analysis, delinquent peer
association, online and offline bullying and delinquency
With recent advancements of technology, new forms of criminal and delinquent acts
have emerged, including the now widespread phenomenon of cyberbullying.
Cyberbullying can be defined as “using information and communication technolo-
gies to repeatedly harm, harass, hurt, and/or embarrass a target” (Peter & Petermann,
2018). Similar to other types of offending, it is hypothesized that those who partici-
pate and are victimized in cyberbullying have peers with similar behaviors. Previous
studies have shown that children and adolescents form and maintain associations
with peers who demonstrate like behaviors including bullying (Espelage et al.,
2007). The homophily hypothesis has been tested using multiple frameworks includ-
ing the selection hypothesis or population heterogeneity (Espelage et al., 2007;
Turanovic & Young, 2016), the socialization model or state dependence (Birkett &
Espelage, 2015; Poteat & Espelage, 2007), and an integration of both models called
the interaction model (Wong et al., 2013) . Research aiming to determine whether
selection or socialization is more related to delinquency in gangs showed support for
both models (Wong et al., 2013). Other empirical studies have found support of
theoretical propositions for state dependence and population heterogeneity, as well
as an interaction model integrating both (Cho & Wooldredge, 2016; Ttofi et al.,
2012; Young & Rees, 2013).
Despite the mounting volume of research on traditional forms of bullying (i.e.,
physical, verbal, social), cyberbullying research maintains a narrow scope with little
empirical efforts having been made to identify patterns of cyberbullying developmen-
tal trajectories (Olweus & Limber, 2018). Also, the growing tendency in developmen-
tal criminology is to subdivide the offender population and examine distinct causal
influences at different stages of the criminal career (Sampson & Laub, 2017). Little is
known about person-centered approaches that abstain from predetermined cutoff
points and instead depend upon observed response patterns. For these reasons, this
study investigates distinct patterns of group-based trajectories of cyberbullying using
a 6-year longitudinal panel dataset of Korean adolescents between the ages of 14 and
19. It also uses a person-centered approach better suited to identify particular response
patterns compared with variable-based approaches such as mean-based cutoff values.
Thus, it is hypothesized that the person-centered approach is less prone to distortion
and misclassification as this approach relies upon response patterns of observed vari-
ables to assign to unobserved classes or groups (Ryoo et al., 2015). In other words,

To continue reading

Request your trial

VLEX uses login cookies to provide you with a better browsing experience. If you click on 'Accept' or continue browsing this site we consider that you accept our cookie policy. ACCEPT