Although there is a vast body of research on juvenile sexual offending with regard to its precursors, courses, and consequences, and despite the considerable interest in the role of ACEs in these developments, the pertinent empirical basis is limited by a wide range of quali- fications. This is due to specific study features but also to the nature of the phenomenon of juvenile sexual offending that is characterized by limited sample sizes and the great hetero- geneity of JSOs, e.g., regarding offense characteristics and ACE burden.
4.1 Study procedures and prevalence rates
Most studies on juvenile sexual offending have included rather small JSO samples, which poses a particular problem for the investigation of risk factors and risk prediction of persistent crime, i.e., due to the low prevalence of sexual reoffending among JSOs (Aebi et al., 2011; Fanniff & Letourneau, 2012; Hempel et al., 2013; Lussier, 2017; Miccio-Fonseca, 2016; Parks & Bard, 2006; Wijetunga et al., 2016; Worling et al., 2012). Concerning recidi- vism rates, studies that have only relied on officially registered crime data may have failed to include the considerable number of offenses that are perpetrated in the dark field; thus, stud- ies may benefit from including non-registered offenses as well (e.g., Maier et al., 2013; Wolff & Baglivio, 2016). The deduction of conclusive findings has been further complicated by the fact that studies have included JSOs with very diverse recidivism periods. Considering the potential changes of risk factors in the courses of the adolescents’ developments, studies should examine JSOs with comparable recidivism periods, and best include different recidi- vism periods to examine longitudinal differences (e.g., Fanniff & Letourneau, 2012; Hempel et al., 2013; Miccio-Fonseca, 2016; Ralston & Epperson, 2013; Schlank, Matheny, & Schilling, 2016; Viljoen, Mordell, & Beneteau, 2012). Moreover, the comparability of differ- ent studies is hindered when findings are based on JSOs from different sampling conditions, e.g., JSOs from correctional facilities (e.g., Martinez et al., 2015) versus JSOs from consecu-
tive samples (e.g., Aebi et al., 2011). Including consecutive samples appears beneficial over selective JSO samples, however, because those samples may include a wider range of JSOs with regard to their personal and offense characteristics (e.g., Aebi et al., 2011).
With respect to the examination of ACEs, findings from previous studies that were only based on self-reports have been challenged regarding their validities, especially in of- fender samples (e.g., Hardt & Rutter, 2004; Seto & Lalumière, 2010). Furthermore, studies that have only investigated single ACEs may have failed to consider the effects of their coex- istence (e.g., Charak & Koot, 2015). A common approach to consider the influence of multi- ple ACEs is to create a cumulative sum score as done in the original ACE study (Felitti et al., 1998). However, only considering the sum of different ACEs is deficient as ACEs are com- monly mutually dependent and the cumulative nature of the ACE score does not respect the actual types of ACEs or the effects of specific ACE patterns (Berzenski & Yates, 2011; Teicher & Samson, 2016).
As outlined above, the varying conceptualization of ACEs across studies additionally hinders the deduction of conclusions for research and practice (e.g., Kalmakis & Chandler, 2014). Lastly, although research has found time-dependent effects of ACEs on general juve- nile delinquency and other maladaptive outcomes (e.g., Thornberry et al., 2001), no study has yet investigated potential timing effects of ACEs in the field of juvenile sexual offending.
4.2 The heterogeneity of JSOs
A major factor that complicates universal conclusions about the occurrence, mainte- nance, and prediction of crime in JSOs is their heterogeneity (e.g., Andrade, Vincent, & Saleh, 2006; Becker & Hicks, 2003; Fanniff & Kimonis, 2014; Van Wijk et al., 2006). Not only may JSOs differ widely with regard to their experiences of ACEs (see above), but they are also heterogeneous on personal features as well as on offense and victim characteristics (Aebi, Vogt, Plattner, Steinhausen, & Bessler, 2012; Bijleveld & Hendriks, 2003; Gunby & Woodhams, 2010; Hunter, Hazelwood, & Slesinger, 2000; Miccio-Fonseca & Rasmussen, 2009; Veneziano & Veneziano, 2002). Therefore, comparisons of ACEs between JSOs on the whole and adolescent nonsexual offenders may fall short on detecting subtle differences re- lated to specific JSO-subtypes (Fanniff & Kolko, 2012; Van Wijk et al., 2006).
Various studies have compared theoretically derived subtypes of JSOs with regard to offense characteristics and ACEs. As further described in Study 3 of the present thesis (Barra, Mokros, Landolt, Bessler, & Aebi, 2017), most prominently, JSO with child victims (JSO- Cs) were contrasted to JSO with adolescent or adult victims (JSO-As; e.g., Leroux, Pullman, Motayne, & Seto, 2016). For instance, compared to JSO-As, JSO-Cs have been found to less
often conduct offenses in groups, to more likely perpetrate against male victims, to be of younger age at the onset of their offenses, to less often show antisocial behavior problems, and to be more likely to have been exposed to ACEs (Hart-Kerkhoffs, Doreleijers, Jansen, van Wijk, & Bullens, 2009; Hendriks & Bijleveld, 2004; Kjellgren, Wassberg, Carlberg, Långström, & Göran Svedin, 2006; Skubic Kemper & Kistner, 2010). More specifically, JSO-Cs were - compared to other JSOs - more frequently burdened with sexual abuse (Aebi et al., 2012; Hart-Kerkhoffs et al., 2009; Van Wijk et al., 2006) and peer bullying (Gunby & Woodhams, 2010; Hendriks & Bijleveld, 2004). Gunby and Woodhams (2010) further found that JSO-Cs had more often experienced caregiver changes, whereas JSO-As were more fre- quently burdened with witnessing intra-familial violence, with having criminal family mem- bers, and with family financial deprivation. In addition, JSO-As have been characterized by lacks of parental supervision (Fanniff & Kolko, 2012). Concerning criminal persistence, JSO- Cs were found to be at lower risk of general criminal recidivism compared to JSO-Cs, where- as subgroups have not differed with regard to sexual reoffending (Fanniff & Kolko, 2012).
In addition, JSOs have been distinguished according to whether they have engaged in other, nonsexual crime (JSO+) or not (JSO-; e.g., Aebi et al., 2012). For example, JSO who had both committed sexual and other violent offenses appeared to be particularly burdened with behavioral peculiarities as well as family dysfunction and ACEs (Murphy et al., 2016). Furthermore, JSO+ were more likely to engage in general reoffending than JSO-, but did not differ to JSO- with regard to sexual recidivism (e.g., Chu & Thomas, 2010).
Differences have also been found between JSOs who perpetrated on their own com- pared to JSOs who perpetrated in groups. Whereas solo offenders were more likely to perpe- trate against children, group offenders were more likely to perpetrate against coeval or older victims (e.g., Höing, Jonker, & van Berlo, 2010; Kjellgren et al., 2006). Concerning ACEs, solo offenders appeared to be more frequently burdened with experiences of sexual victimiza- tion and peer bullying than group offenders (e.g., Bijleveld & Hendriks, 2003; Höing et al., 2010). Group offending (rather than solo offending) has been related to ongoing criminality (Hart-Kerkhoffs et al., 2009).
In sum, subtyping approaches may be preferable over studies that consider JSOs as a homogeneous group in order to derive conclusions about the nature of juvenile sexually of- fending (e.g., Murphy et al., 2016). However, assignments of JSOs to a priori defined (theo- ry-driven) subtypes have been criticized due to the potential overlaps among the characteris- tics of these subtypes (e.g., JSO-C & solo-offender subtypes), and rather dimensional, empir- ical subtyping approaches have been considered as beneficial (e.g., Aebi et al., 2012).