From Criminological Heterogeneity to Coherent Classes

Date01 July 2018
DOI10.1177/1541204017699257
Published date01 July 2018
Subject MatterArticles
Article
From Criminological
Heterogeneity to Coherent
Classes: Developing a Typology
of Juvenile Sex Offenders
Bryanna Fox
1,2
and Matt DeLisi
3
Abstract
Although juvenile sex offenders (JSOs) are a pressing topic among researchers and juvenile justice
practitioners, empirically driven typologies of JSOs using U.S. data are lacking. Here, we develop the
first statistical typology of male and female JSOs using data from the United States selected from a
sample of 4,143 JSOs referred to the Florida Department of Juvenile Justice. Significant predictors of
juvenile sex offending (age of criminal onset, criminal history, impulsivity, empathy, depression,
psychosis, and childhood sexual abuse) derived from the literature were used as grouping covariates
to develop a profile of male and female JSOs using a latent class analysis (LCA). Results of the LCA
show four unique subtypes of male JSOs and two subtypes of female JSOs exist within the data.
These groups had differential compositions for key features such as criminal history and onset,
psychopathologies, empathy and impulsivity, and sexual abuse victimization. These differences may
be critical toward developing more tailored and effective correctional and treatment responses that
balance containment and therapeutic approaches depending on the individual needs of the JSOs
based upon their profile. Other practical and theoretical implications of these findings are discussed.
Keywords
juvenile sex offenders, latent class analysis, typology, profile, criminal careers
Given the impact of juvenile sex offenders (JSOs; Barbaree, Hudson, & Seto, 1993; Becker & Abel,
1985; Federal Bureau of Investigation, 2014; Harris & Socia, 2016), our understanding of these
offenders remains limited. On a variety of topics, there remains conflicting evidence on the con-
tinuity between juvenile and adult sex offending (Lussier & Blokland, 2014; Piquero, Farrington,
Jennings, Diamond, & Craig, 2012; Sipe, Jensen, & Everett, 1998),
1
reliable methods to identify and
treat juvenile sex offending (Caldwell, Ziemke, & Vitacco, 2008; Hargreaves & Francis, 2014), and
understanding of the personality traits, psychopathologies, and subtypes of JSOs (Cale, Lussier,
1
Department of Criminology, University of South Florida, Tampa, FL, USA
2
Department of Mental Health, Law, and Policy, University of South Florida, Tampa, FL, USA
3
Department of Sociology, Iowa State University, Ames, IA, USA
Corresponding Author:
Bryanna Fox, University of South Florida, 4202 E. Fowler Avenue, SOC 107, Tampa, FL 33620, USA.
Email: bhfox@usf.edu
Youth Violence and JuvenileJustice
2018, Vol. 16(3) 299-318
ªThe Author(s) 2017
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1541204017699257
journals.sagepub.com/home/yvj
McCuish, & Corrado, 2015; Fox, 2017). However, as a growing body of research suggests that
JSOs are a largely heterogeneous population comprised of a variety of unique personality traits,
criminal behaviors, demographic traits, psychopathologies, and childhood traumas (Calley, 2007;
Drury et al., 2017; Fox, 2017; Gordon & Porporino, 1990; Hunter, Hazelwood, & Slesinger, 2000;
McCuish, Cale, & Corrado, 2017; McCuish, Lussier, & Corrado, 2016; Rice & Harris, 1997;
Robertiello & Terry, 2007; Seto & Lalumie`re, 2010; Worling, 2001), it would appear that a single
understanding of these offenders’ recidivism rates, causes and types of offending, and method of
treatment is not the ideal response. Instead, identifying the unique causes, behaviors, and treatments
for the unique subtypes of JSOs would seem a far better approach. In other words, developing a
discrete typology or “profile” of JSOs is needed to allow for a more efficient and effective way to
predict, manage, and treat young sex offenders.
Key Predictors of Juvenile Sex Offending
JSOs encompass a multitude of unique personality traits, psychopathologies, criminal behaviors,
demographic features, and childhood histories. Seto and Lalumie`re’s (2010) meta-analysis aimed to
systematically evaluate the most significant of these predictors by examining the impact of age of
criminal onset, criminal history, conduct problems, antisocial tendencies, substance abuse, child-
hood abuse, exposure to household violence and substance abuse, interpersonal problems, sexuality,
psychopathology, cognitive abilities, and impression management in successfully predicting JSOs.
Results indicated that criminal history, deviant sexual interests, sexual and physical abuse in child-
hood, and certain psychopathologies were strong and significant predictors of juvenile sex offending
based upon findings from the 59 studies in their meta-analysis (Seto & Lalumie`re, 2010).
While Seto and Lalumie`re’s analysis was extremely beneficial on updating the field on the state
of the research on JSOs, the limitations of many studies included in the meta-analysis give reason to
temper some of the findings. For example, no study included all, or even most, of the measures used
as the key predictors of juvenile sex offending, meaning it was not possible to control for other
measures’ influence in these studies or in Seto and Lalumie`re’s meta-analysis. Only 9 of 59 studies
used age of criminal onset as a predictor of juvenile sex offending in their analyses, and just 17
studies controlled for criminal history. Additionally, most of the studies (46 of 59) utilized samples
with less than 60 offenders, raising issues regarding internal and external validity (Worling, 2001).
Finally, the meta-analysis was based upon the studies conducted entirely on male offenders, so
evaluating any variation in the predic tors of juvenile sex offending due to fema le offenders is
impossible (see Matthews, Mathews, & Speltz, 1991).
Recently, Fox (2017) aimed to address many of the limitations faced by prior research by
examining the key predictors of juvenile sex offending by males and females using 16 covariates
and a sample of 64,329 juvenile offenders referred to the Florida Department of Juvenile Justice
(FDJJ). The predictors used in the analysis were selected due to their significance from prior
research on juvenile sex offending.
2
These measures were (1) age of criminal onset, (2) criminal
record, (3) antisocial peer association, (4) empathy, (5) impulsivity, (6) anger/irritability, (7) aggres-
sion, (8) depression, (9) psychosis, (10) emotional abuse, (11) physical abuse, (12) sexual abuse,
(13) witnessing household violence, (14) household substance abuse, (15) race/ethnicity, and (16)
gender. These measures were evaluated and were evaluated in their ability to predict juvenile sex
offending, while controlling for the impact of all the other measures. Results of Fox’s (2017) study
suggest that out of the many theoretically and research-supported measures included in the multi-
variate analysis, only gender, age of criminal onset, criminal record, impulsivity, empathy, depres-
sion, psychosis, and sexual abuse were significant predictors of sexual versus non-sexual offending
among the male and female juvenile offenders. Specifically, Fox (2017) found that having an earlier
age of criminal onset, more felony arrests, experiencing sexual abuse, having low empathy, high
300 Youth Violence and Juvenile Justice 16(3)

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