During the past 15 years, researchers have repeatedly pointed out that interpersonal experiences of violence tend to co-occur across children and adolescentsâ lives (Turner, Finkelhor, & Ormrod, 2006), meaning that individuals are rarely victims of an isolated type of victimization. Studies have highlighted the importance of assessing a wide range of experiences of violence rather than focusing on a single form (Finkelhor, Ormrod, & Turner, 2007), in order to provide an accurate explanation of child victimization (Hamby & Grych, 2013).
Several frameworks have been designed to analyze this co-occurrence (e.g., multi-type maltreatment, see Higgins & McCabe, 2000; complex trauma, see Cook et al., 2005; polytraumatization, see Gustafsson, Nilsson, & Svedin, 2009), in attempts to assess the complexity of child and adolescent experiences of violence. Poly-victimization, defined as the experience of multiple types of victimization in different episodes during the course of a childâs life (Finkelhor, Ormrod, Turner, & Hamby, 2005), constitutes another framework for addressing this phenomenon, and it seems to affect a high percentage of children and youth across the globe (see Chan, 2013, in China; Cyr et al., 2013, in Canada; Pereda, Guilera, & Abad, 2014, in Spain; Radford, Corral, Bradley, & Fisher, 2013, in the UK). Studies have shown that poly-victimization is higher in countries with lower income levels (Le, Holton, Romero, & Fisher, 2016), but also in children and adolescents at higher risk for victimization, such as sexual minority adolescents (Sterzing, Ratliff, Gartner, McGeough, & Johnson, 2017), adolescents with mental health issues (Ălvarez-Lister, Pereda, Abad, Guilera, & the GReVIA, 2014), children cared for by child welfare systems (Cyr et al., 2012) or those involved in the juvenile justice system (Ford, Grasso, Hawke, & Chapman, 2013).
However, few instruments have tried to measure victimization experiences in childhood from a multidimensional, comprehensive perspective that includes different forms of interpersonal violence in different contexts, avoids fragmentation, and acquires the information directly from the children themselves (Hamby & Finkelhor, 2000). Some of these instruments are the Childhood Trauma Questionnaire (CTQ; Bernstein, Ahluvalia, Pogge, & Handelsman, 1997), the ICAST instruments created by the International Society for the Prevention of Child Abuse and Neglect (ISPCAN) (Zolotor et al., 2009), and the Childhood Experiences of Violence Questionnaire (CEVQ) by Walsh, MacMillan, Trocmé, Jamieson, and Boyle (2008). However, these measures focus only on experiences of maltreatment, mainly by caregivers, and therefore do not include all the possible victimization experiences a child may suffer, nor all the contexts in which these incidents may happen (Hamby & Finkelhor, 2000).
Today, most studies addressing the overlap of victimization in childrenâs lives use the poly-victimization approach and apply the Juvenile Victimization Questionnaire (JVQ). For both lifetimes and past-year timeframes, this questionnaire offers a comprehensive assessment of five general areas of child and adolescent victimization: conventional crime, child maltreatment, peer and sibling victimization, sexual victimization, and witnessing/exposure to indirect victimization. The JVQ thus gives a complete profile of child victimization (Finkelhor, Hamby, Ormrod, & Turner, 2005).
However, current methodological methods to the definition of poly-victimization depend on the specific objectives of the research and its time period (i.e., lifetime or past year), the method used (i.e., victims above the mean, the top 10% of child victims, or clustering techniques), the version of the JVQ applied in a particular country (which may include different numbers of items and consider different victimization modules) and the characteristics of the sample (e.g., community, clinical juvenile justice, welfare). All these variables may affect the rates of prevalence recorded. This means that it is difficult to know whether all studies have identified the same at-risk group, namely poly-victims. One of the key aspects in its analysis is the time perspective we apply â that is, over the lifetime, or over the past year. Finkelhor, Ormrod, and Turner (2009) suggested that both lifetime and past-year poly-victimization have advantages and drawbacks. In this regard, they argued that focusing on past-year victimization can guide clinicians towards a more accurate assessment of victimization and can also prevent retrospective biases (Widom, Raphael, & DuMont, 2004); however, a lifetime assessment provides a more complete description of the victimization profile (Finkelhor et al., 2009).
Another key aspect is the approach used to analyze the data. Two approaches have been widely applied to assess lifetime and past year poly-victimization, using three different methods. The first approach sums the variety of victimization experiences lived by a child and focuses on the most victimized adolescents. From this approach, the first method selected children and adolescents who had experienced at least one victimization more than the mean number among the victim group as a whole (Finkelhor et al., 2005). Using this one-above-the-mean method, Finkelhor et al. (2005) considered 22% of the community sample of children interviewed to be past-year poly-victims. For the lifetimeframe, this method has also been applied in some studies, with percentages ranging from 14% to 17% (e.g., Chan, 2013; Dong, Cao, Cheng, Cui, & Li, 2013). Inside the same approach, Finkelhor et al. (2009) proposed another method of identifying poly-victims, consisting in selecting the top 10% of the community sample of children who experienced the highest number of victimizations, both lifetime and past-year. The JVQ lifetime poly-victimization cutoff point for the top 10% of the sample establishes poly-victimization in community samples at more than 11 (Turner, Finkelhor, & Ormrod, 2010), 12 (RadRadford et al., 2013), or 13 different types of victimization (Finkelhor et al., 2009). Finally, the second approach has used clustering methods such as traditional cluster analysis with community and high-risk samples (Ălvarez-Lister et al., 2014; Holt, Finkelhor, & Kaufman, 2007) or latent class analysis (LCA) (Kretschmar, Tossone, Butcher, & Flannery, 2016; Reid & Sullivan, 2009; Turner, Shattuck, Finkelhor, & Hamby, 2016). The authors used the responses to the JVQ items (yes/no) as observed categorical variables to identify subgroups of victims with different victimization profiles or combinations of victimization experiences (i.e., clusters or latent classes), among which one or more groups of poly-victims are identified. Using clustering methods, the authors identified the groups who report a high mean of multiple types of victimization experiences as poly-victims.
The present study
Today, researchers apply a range of methodologies to identify the most victimized children and adolescents. This variety may prevent us from consistently selecting the same group of poly-victims ac...
