It is estimated that 1.7 million youth run awayFootnote 1 from home every year in the United States (Castillo et al., 2022; Fernandes-Alcantara, 2018; Sedlak et al., 2017). Although most runaway episodes last less than one week, those who run away for more than a day report more problems with parents, peer relationships, internalizing and externalizing behaviors, and substance use compared to youth who do not runaway (Bao et al., 2000; Castillo et al., 2022; Sanchez et al., 2006; Sill, 2018). Runaway youth are at an increased risk of experiencing commercial sexual exploitation (CSEFootnote 2), or what is sometimes referred to as survival sex, domestic minor sex trafficking (DMST), or youth prostitution (Chen et al., 2004; Estes & Weiner, 2005; Flowers, 2001; Pullmann et al., 2020; Reid, 2016; Roe-Sepowitz, 2012). Research on runaway youth has indicated that certain patterns and characteristics exist that predispose youth to go on the run and subsequently engage in other forms of high-risk behavior (Development Services Group (DSG), 2015; Edinburgh et al., 2015; Jeanis et al., 2019; Mello et al., 2018). For example, runaways have higher rates of prior victimization, committing theft/property damage, binge drinking and substance use, depression and suicidal behavior, and problems at school compared to youth who do not run away (Development Services Group (DSG), 2015; Thompson & Pillai, 2006; Tyler & Bersani, 2008).

Runaways and CSE victims were historically treated as delinquents by the juvenile justice system (JJS), and as such, they were often arrested and detained (e.g., Development Services Group (DSG), 2015; Levin & Cohen, 2014; Trejbalová et al., 2021). The legal status of runaways and minor CSE victims has changed dramatically since the 1970s, which has affected how such youth are processed by the JJS. The Runaway Youth Act (1974) was passed in conjunction with the Juvenile Justice and Delinquency Prevention Act (P.L. 93–415), which authorized funding for programs providing shelter, counseling, and other services to runaway youth. It also decriminalized status offenses and encouraged states to do the same (Coalition for Juvenile Justice, 2014). Running away is classified as a status offense rather than a delinquent act, because it is only considered a law violation due to a youth’s age as a minor (similar to underage drinking, truancy or curfew violations; Development Services Group (DSG), 2015). Youth involved in prostitution were reclassified as minor sex trafficking victims in 2000 by the Victims of Trafficking and Violence Protection Act (TVPA; 22 U.S.C. §7102, Sec. 103), which is the foundation for anti-trafficking policy in all 50 states (Branscum et al., 2021). Thus, language such as, “youth involved in prostitution” or “youth sex worker” is inappropriate, as minors are not legally capable of consenting to the sale of sex (Gies et al., 2020; Mitchell et al., 2010; Smith et al., 2009; Swaner et al., 2016). Thus, the Runaway Youth Act and TVPA provided a catalyst for changing how states deal with these youth; from historically treating them offenders, to treating them as minors in need of help rather than punishment/detention (Gies et al., 2020; Development Services Group, (DSG) 2015; Jafarian & Ananthakrishnan, 2017; Mir, 2013).

However, juveniles continued to be arrested and incarcerated for running away or being a CSE victim even after federal laws decriminalized status offenses and youth involved in prostitution (Adams et al., 2010; Gies et al., 2020; Levin & Cohen, 2014; McMahon–Howard, 2017; Office ofJuvenile Justice and Delinquency Prevention, 2022). The JJS has a history of continuing to detain youth for decriminalized offenses, a process referred to as carceral protectionism,Footnote 3 usually with the intent of keeping them safe or getting them access to services (Development Services Group (DSG), 2015; Musto, 2010; Nichols et al., 2022). For example, runaways may be detained to protect them from the risks of being away from home, or their placement (e.g., sex trafficking, substance abuse), or to ensure they appear at their court date (Development Services Group (DSG), 2015; Subramanian, 2020). Minor CSE victims may be detained because they are often misidentified as offenders, or if they are correctly identified, they may be detained to prevent re-trafficking (i.e., when a victim goes right back to their trafficker/pimp) or being homeless, as there is a huge lack of services available for these victims (Halter, 2010; Nichols et al., 2022; Pullmann et al., 2020; Reid & Piquero, 2014a). Yet, detaining runaways and CSE victims is problematic because a substantial body of literature suggests this exacerbates their problems, interrupts their education, delays appropriate therapeutic responses, exposes them more serious youth offenders, stigmatizes (and potentially traumatizes) them, and increases JJS costs (Clawson et al., 2009; Dank et al., 2015; Holman & Ziedenberg, 2011; Jafarian & Ananthakrishnan, 2017; Lanctôt et al., 2020; Levin & Cohen, 2014; Naramore et al., 2017; Nichols et al., 2022; Swaner et al., 2016; Trejbalová et al., 2021).

This study seeks to describe the risk factors among runaway and minor CSE victims who end up detained in the justice system so that policy and programs recommendations are more reflective of the unique needs of these vulnerable youth. The current study uses a nationally representative sample of incarcerated youth to describe the characteristics and risk factors of youth whose most serious offense (MSO) is running away or being a minor CSE victim (referred to as “involved in prostitution” in the data). This study then compares these youths’ characteristics to incarcerated youth whose MSO is more serious than running away or CSE victimization. Comparing the risk factors of youth with official runaway and prostitution histories to other justice-involved youth has not been done with a nationally representative sample of incarcerated youth. This matters because knowing the risk factors among this group of detained youth may improve the implementation of early intervention efforts, lead to a collaborative criminal justice response, and lead to more responsive policy that will reduce harm and improve outcomes for the affected youth (Fedina et al., 2016; Naramore et al., 2017; Nichols et al., 2022; Reid & Piquero, 2014b; O'Brien et al., 2017; Pullmann et al., 2020).

Background

Runaway Demographics and Risk Factors

National surveys suggest between five and eight percent of adolescents run away from home or their placement (e.g., foster home) in any given year, and estimates vary by gender, age, and race/ethnicity (Castillo et al., 2022; Minnesota Department of Education, 2016; Ringwalt et al., 1998; Sanchez et al., 2006; Sedlak et al., 2017). Prevalence estimates also vary depending on study definitions, sampling and methodology. For example, the Third National Incidence Studies of Missing, Abducted, Runaway, and Thrownaway Children (NISMART-3), estimated that 413,000 children ran away or were thrownawayFootnote 4 in 2013, at a rate of 5.3 per 1,000 children (Sedlak et al., 2017), which is comparable to earlier NISMART studies (Greene et al., 1999; Hammer et al., 2002).

It can be difficult to describe the risk factors for running away, as studies vary on certain characteristics depending on the type of subpopulation being examined (e.g., general population, justice-involved youth, those in the welfare system, youth who seek help while on the run). Nonetheless, several patterns emerge from prior research regarding runaway risk factors. The evidence regarding gender differences is mixed, though many studies found a greater rate of running away among girls (Benoit-Bryan, 2011; Branscum & Richards, 2022; Castillo et al., 2022; Courtney & Zinn, 2009; Kempf-Leonard & Johansson, 2007; Latzman & Gibbs, 2020; Sanchez et al., 2006; Tyler et al., 2011); although some studies indicated higher likelihood of runaway episodes among males (Fernandes-Alcantara, 2016; Pergamit et al., 2010). Other studies did not find any gendered differences among runaway youth (Hammer et al., 2002; Milburn et al., 2007). Some common risk factors for runaway girls include family instability (e.g., abandonment, child maltreatment, poor parenting), a history of physical or sexual abuse, and being a CSE victim (Edinburgh et al., 2015; Hershberger et al., 2018; Saewyc et al., 2019; Tyler et al., 2008). Runaway boys are more likely to have an early onset of running, be in a gang, have antisocial peers, engage in substance abuse, and not live with their parents (Jeanis et al., 2019; Kempf-Leonard & Johansson, 2007; Tyler et al., 2011).

Victimization is a major risk factor for running away and varies by gender. Mainly, females have higher rates of sexual victimization and polyvictimization while males report higher rates of indirect violence (Baglivio et al., 2014; Finkelhor et al., 2014; U.S. Department of Health & Human Services, 2023). Being a direct victim of physical or sexual abuse, including CSE, is a predictive factor for running away (Branscum & Richards, 2022; Greene et al., 1999; Hammer et al., 2002; Hershberger et al., 2018; Kempf-Leonard & Johansson, 2007; Saewyc et al., 2019; Tyler et al., 2001, 2008). Additionally, indirect victimization, such as witnessing neighborhood or domestic violence, increases the frequency of running away by 75 percent (Tyler & Bersani, 2008). Some scholars have found that prior abuse was more common among female runaways than males (Kempf-Leonard & Johansson, 2007; Tyler & Bersani, 2008). It could be that females are more likely to run away because they are more likely to experience sexual abuse, and running away may be seen as an escape from more abuse (Jeanis et al., 2020; Saewyc et al., 2019; Thrane et al., 2006; Tyler & Bersani, 2008).

Regarding age, youth who are at least 15 years old are more likely to run away than kids between the ages of 10 and 14 (Courtney & Zinn, 2009; Lin, 2014; Sunseri, 2003). Similar to gender, racial/ethnicity differences are mixed, though most studies indicate that minorities are at an increased risk compared to whites (Benoit-Bryan, 2011). As an exception, Hockenberry and Puzzanchera (2015) found that whites had a higher rate of runaway offenses than blacks when examining case records from 2,400 juvenile courts. Several studies have found that blacks and Hispanics in foster care/group homes are more likely to run away than whites, though rates vary by age, gender, and context (Branscum & Richards, 2022; Courtney & Zinn, 2009; Lin, 2012; see Wulczyn, 2020 for more details). Attar-Schwartz, (2013) argued that cultural affiliations may also affect gender differences among runaway youth who reside in foster care or residential institutions.

In addition to demographic predictors, several risk factors predict runaway behavior among youth, including: living situation, lack of parental support, school disengagement, substance abuse, prior runaway episodes; lesbian, gay, bisexual, transgender, queer/questioning (LGBTQ) identity, and mental illness (Branscum & Richards, 2022; Courtney & Zinn, 2009; Dank et al., 2015; Latzman et al., 2019; Lin, 2012; Moskowitz et al., 2013; Sunseri, 2003; Thompson & Pollio, 2006; Tucker et al., 2011; Tyler & Bersani, 2008; Whitbeck et al., 2004). For youth who live at home, those who live with both biological parents are least likely to run away, followed by those who lived with at least one biological parent, and those in other family structures (Sanchez et al., 2006; Tyler et al., 2011). Youth who live in foster care and residential institutions have an increased risk of running away during their placement (Attar-Schwartz, 2013; Branscum & Richards, 2022; Kempf-Leonard & Johansson, 2007). More specifically, youth involved in foster care are about twice as likely to run away than youth in the general population (Benoit-Bryan, 2011; Jeanis et al., 2020; Sedlak et al., 2005, 2017), and youth with repeated moves between placements are more likely to run away than those with residential stability (Bowden & Lambie, 2015; Branscum & Richards, 2022; Courtney & Zinn, 2009; Lin, 2012).

Consequences of Running

Runaway youth may be picked up by the police, taken to a juvenile justice intake office, and some are detained even though federal law prohibits status offenders from being placed in secure detention (Development Services Group (DSG), 2015). As previously noted, federal legislation for runaway youth was nearly non-existent prior to the 1970s but emerging empirical focus on youth deviance and runaway behaviors forced Congress to formally address the issue, resulting in the Runaway Youth Act of 1974 as part (Title III) of the Juvenile Justice and Delinquency Prevention Act (JJDPA). A core component of this Act was the deinstitutionalization of status offenders (DSO) core requirement that said youths charged with status offenses, and/or abused and neglected youths, shall not be placed in secure detention or locked confinement (Development Services Group (DSG), 2015; Fernandes-Alcantara, 2018). This requirement encouraged states to divert status offenders away from the JJS and place them in less restrictive, community-based programs, or risk losing federal funding from OJJDP (Development Services Group (DSG), 2015). However, many states continued detaining status offenders by circumventing the DSO requirement with a valid court order (VCO) exception, wherein judges and other actors could place adjudicated status offenders in secure detention if they violated a VCO (e.g., stop running away from home, attend school regularly; Development Services Group (DSG), 2015). Additionally, runaways in many states were often detained to prevent youth from being a risk to themselves and to ensure they attended judicial hearings (Development Services Group (DSG), 2015). Despite the DSO requirement, many runaway youths continued to be detained throughout the nation, which generally begins their involvement in the JJS (Holman & Ziedenberg, 2011).

Runaway youth typically experience complex and co-occurring issues in multiple life domains which, at least in part, cause them to run away (Benoit-Bryan, 2011; Castillo et al., 2022; Pergamit et al., 2010; Thompson & Pillai, 2006; Tucker et al., 2011). Most runaways do not leave their home state, return to their parents or guardian, and are gone less than a week (Hammer et al., 2002; Milburn et al., 2007; Pergamit et al., 2010). Nonetheless, being on their own for even a short period of time may exacerbate the problems that caused them to run away in the first place and may increase their likelihood of engaging in high-risk behavior or being exploited by others (Branscum & Richards, 2022; Chen et al., 2007; Hammer et al., 2002; Thompson & Pillai, 2006; Tyler & Bersani, 2008). For example, youth who run away report higher rates of substance abuse (Courtney & Zinn, 2009; Johnson et al., 2005), clinical depression (Fernandes-Alcantara, 2018; Tucker et al., 2011), self-harm and suicidal behavior (Moskowitz et al., 2013; Yoder, 1999), sexually transmitted diseases (Lacoursiere & Fontenot, 2012; Saewyc et al., 2019), and are more likely to be suspended or expelled from school (Crosland & Dunlap, 2015; Tucker et al., 2011) compared to non-running youth. Runaway youth are also at an increased risk of experiencing victimization, including CSE (Chen et al., 2004; Cohen et al., 1991; Flowers, 2001; Latzman et al., 2019; O'Brien et al., 2017; Pullmann et al., 2020; Reid, 2016; Roe-Sepowitz, 2019; Saewyc & Edinburgh, 2010).

Prevalence and Risk Factors of Youth CSE Victims

It is difficult to estimate the prevalence of minor CSE victims due to difficulties measuring and identifying sex trafficking since, by nature, it is a hidden crime (Farrell & Reichert, 2017; Kulig, 2022). Thus, there are dramatic differences among prevalence estimates and descriptions of adolescent sex trafficking victims depending on the type of data used (e.g., convenience samples, official data from law enforcement, justice systems, child welfare; multiple systems estimation; Farrell & Reichert, 2017; Kulig, 2022; Institute of Medicine (IOM) & National Research Council (NRC), 2014; McMahon–Howard, 2017; Reid, 2012). However, there are common characteristics and risk factors of minor CSE victims from these studies. Most youth tend to fall victim to CSE for the first time between ages 12 and 14 (Clarke et al., 2012; Clayton et al., 2013; Cobbina & Oselin, 2011; Havlicek et al., 2016; Kramer & Berg, 2003; Reid, 2016; Roe-Sepowitz, 2012).

Minor CSE victims have much higher rates of prior abuse (physical, sexual, emotional, neglect) than non-exploited youth (Ahrens et al., 2012; Clayton et al., 2013; Cobbina & Oselin, 2011; Estes & Weiner, 2005; Fedina et al., 2016; Reid, 2011; Reid & Piquero, 2014a; Roe-Sepowitz, 2012; Smith et al., 2009). In fact, many minor CSE victims fit the definition and description of polyvictims (Finkelhor et al., 2007, 2011), meaning they have an extensive history of distinct types of victimization (Landers et al., 2017; Reid et al., 2015). Studies using adverse childhood experiences (ACEs)Footnote 5 have also found that youth formerly charged with trading sex were more likely to report nearly every ACE compared to youth charged with other crimes (Naramore et al., 2017). Relevant to this study, justice-involved youth have higher rates of ACEs compared to non-delinquent youth, and minor CSE victims have higher rates of ACEs than justice-involved youth who have not been sex trafficked (Baglivio & Epps, 2016; Baglivio et al., 2015; Fox et al., 2015; Reid et al., 2017, 2019).

Many scholars have also found that youth who have poor home functioning or come from a dysfunctional family are more likely to be victims of CSE (e.g., Clarke et al., 2012; Clawson et al., 2009; Landers et al., 2017). Additionally, involvement in foster care or the child welfare system increases youths’ risk of being a victim of CSE (Ahrens et al., 2012; Bounds et al., 2015; Estes & Weiner, 2005; Havlicek et al., 2016; Landers et al., 2017; Pullmann et al., 2020). In fact, several studies have found that child welfare systems are a recruitment site for sex traffickers (Institute of Medicine (IOM) & National Research Council (NRC), 2014; Latzman et al., 2019; Lillie, 2013; Smith et al., 2009; Reid, 2016). As noted above, scholars using different samples have found that minor CSE victims are more likely to be runaways and/or homeless (Murphy, 2017; Pullmann et al., 2020; Reid, 2012; Roe-Sepowitz, 2012). Once youth are away from their home, they may engage in survival sex in exchange for food, shelter, safety or other necessities (Dank et al., 2015; Edwards et al., 2006; Murphy, 2017; Tyler et al., 2004).

Previous research has also noted other risk factors that were associated with juvenile CSE, including gang membership, intellectual disabilities, and conflict with parents (Dorias & Corriveau, 2009; Estes & Weiner, 2005; Hickle & Roe-Sepowitz, 2016; Kennedy et al., 2007; Institute of Medicine (IOM) & National Research Council (NRC), 2014; Reid, 2012, 2016, 2018; Reid & Piquero, 2014b, 2016; Reid et al., 2015; Roe-Sepowitz, 2019). Substance abuse or dependency issues (drugs or alcohol) also heightens youths’ risk of being a CSE victim (Chohaney, 2016; Clarke et al., 2012; Edwards et al., 2006; Martin et al., 2010; Roe-Sepowitz, 2012). Youth with problems at school (e.g., truancy, falling behind a grade, being suspended or expelled) are also at an increased risk for minor CSE victimization (Chohaney, 2016; Clarke et al., 2012; Clawson et al., 2009; Kramer & Berg, 2003). Lastly, juveniles with prior justice system involvement (e.g., arrests, charges, probation, detention) are at an increased risk for CSE victimization (Chohaney, 2016; Dank et al., 2015; McMahon–Howard, 2017; Naramore et al., 2017; Perkins & Ruiz, 2017; Pullmann et al., 2020; O'Brien et al., 2017; Showden & Majic, 2018).

Consequences of Minor CSE Victimization

The amount of trauma that minor CSE victims experience is well documented, and this trauma often results in short- and long-term problems (Institute of Medicine (IOM) & National Research Council (NRC), 2014; Lanctôt et al., 2020; Reid et al., 2019; Roe-Sepowitz, 2012). As noted above, CSE victims have already suffered abuse and many are polyvictims, and CSE only exacerbates the level of trauma they experience, resulting in high levels of trauma symptoms (Gambon et al., 2020; Lanctôt et al., 2020; Naramore et al., 2017; Reid, 2012; Reid et al., 2015, 2017). For example, Landers et al., (2017) found extensive histories of abuse (i.e., physical, sexual, emotional, neglect, witnessing criminal acts, and family and community violence), and high rates of mental and behavioral health problems (i.e., anxiety, anger control, depression, substance use, impulsivity/ hyperactivity, oppositional behavior) among a sample of 87 minor CSE victims in a specialized treatment program. Sometimes trauma from CSE may lead to other negative effects, such as mental health issues or justice-system involvement (Chapple & Crawford, 2019). For example, CSE trauma is linked to the use of alcohol or drugs, which increases the odds that youth may encounter and be arrested/detained by law enforcement because of drug-related charges (e.g., Reid & Piquero, 2014b).

Additionally, human traffickers often manipulate their victims and coerce them to engage in criminal activities (referred to as forced criminality; Reid, 2016; Roe-Sepowitz et al., 2020; U.S. Department of State, 2014). Forced criminality may result in youth being arrested by police for other offenses, such as, drug offenses, property theft, financial crimes [stolen checks, identity theft, welfare benefits fraud, scams], or “trick rolling” (robbing sex buyers), which can result in lengthy arrest and criminal records (National Survivor Network, 2016; Roe-Sepowitz et al., 2020; U.S. Department of State, 2014). Moreover, many youth CSE victims may be identified by law enforcement as offenders rather than victims for “selling sex” and become entangled in the JJS (Chapple & Crawford, 2019; Farrell et al., 2019; Lanctôt et al., 2020; Levin & Cohen, 2014; Naramore et al., 2017; Swaner et al., 2016; Trejbalová et al., 2021). When this occurs, law enforcement may charge youth with masking charges (e.g., loitering, running away, truancy, minor in possession) rather than with “prostitution” so they can detain them since after the U.S. Victims of Trafficking Victims Protection Act (TVPA) (2000) passed, youth engaged in prostitution were legally sex trafficking victims (Adams et al., 2010; Finkelhor & Ormrod, 2004; Smith et al., 2009).

States began creating legislation against human trafficking in 2003, and by 2014 every state had some type of anti-trafficking statute, though the quality of laws varied from state to state (Branscum et al., 2021; Polaris Project, 2014). However, minor CSE victims continued to be arrested, charged and detained for prostitution even after federal and state laws said that it was not a crime and that youth engaged in such acts were minor sex trafficking victims (Bejinariu et al., 2021; Clayton et al., 2013; Farrell et al., 2019; Geist, 2012; Gies et al., 2020; Trejbalová et al., 2021). So, in 2008, states began enacting safe harbor lawsFootnote 6 to help with the conflicting treatment of juvenile CSE victims by police and the justice system (Institute of Medicine (IOM) & National Research Council (NRC), 2014; Polaris Project, 2015; Wasch et al., 2016). Not every state has a safe harbor law and evidence that such laws reduce how many youth CSE victims are treated criminally is mixed (Geist, 2012; Wasch et al., 2016). For example, some scholars found that safe harbor laws had no or little effect on how law enforcement or the JJS treated youth CSE victims (McMahon–Howard, 2017; White et al., 2017), while Gies et al., (2020) found safe harbor laws may be responsible for the decline in the number of juvenile arrests for prostitution offense from 2005–2015 using the Uniform Crime Reports (UCR).Footnote 7

The Current Study

Numerous scholars and professionals working with at-risk youth have argued that even brief periods of incarceration are counterproductive and damaging for youth who are detained for running away or being sexually exploited (e.g., Development Services Group (DSG), 2015; Pullmann et al., 2020). They argue that being detained may be traumatizing, interrupts youths’ education, exposes them to youth who commit actual crimes, and keeps them away from community-based solutions that have been shown to be more effective (Coalition for Juvenile Justice, 2014; Harper et al., 2015; Lanctôt et al., 2020; Levin & Cohen, 2014; Naramore et al., 2017; Nichols et al., 2022; Swaner et al., 2016; Trejbalová et al., 2021). The problem of carceral protectionism (or detaining runaway youth and CSE victims) may also affect certain groups disproportionately, such as females or minority youth because they tend to score higher on risk assessments due to other disadvantages they face (e.g., single-parent households, higher rates of sexual victimization, welfare system involvement). The problem of disproportionate minority confinement (DMC) has been an important research topic in many jurisdictions because of the injustice that may occur if youth are treated differently by the JJS based on extra-legal factors, like race/ethnicity (Cabaniss et al., 2007; Kempf-Leonard, 2007; McCarter, 2009; Rovner, 2021; Sickmund et al., 2011). Additionally, detaining youth who are not a risk to public safety or themselves puts an undue financial burden on taxpayers because secure detention is more costly than alternatives, such as probation (Casey Foundation [AECF], 2014; Holman & Ziedenberg, 2011). Thus, it is important to understand the characteristics of detained youth whose MSO is running away or being a CSE victim so the JJS and youth-serving agencies can offer alternatives to detention to meet the needs of these youth who, by law, should not be detained at all (Casey Foundation [AECF], 2014; Development Services Group (DSG), 2015).

The purpose of this study is to add to the growing literature describing the characteristics and risk factors (e.g., abuse histories, school problems, gang involvement, living arrangement growing up) of incarcerated youth whose MSO is running away or being a CSE victim. This study will also compare the characteristics of youth incarcerated for running away or CSE victimization to youth incarcerated for all other MSOs, which has not been done with a nationally representative sample of justice-involved youth. This is important because to date, there is limited knowledge about how youth’s characteristics who are incarcerated for running away or CSE victimization compare to youth incarcerated for more serious offenses, as most studies are conducted on samples of a single or few cities, counties or states, or they include a single sex rather than both males and females (e.g., Dank et al., 2015; Naramore et al., 2017; O'Brien et al., 2017; Reid et al., 2017; Salisbury et al., 2015; Trejbalová et al., 2021). There are two research questions in this study: (1) What are the characteristics and risk factors of incarcerated youth whose MSO is running away or CSE victimization? (2) Are the characteristics of youth detained for running away or CSE victimization different than youth detained for more serious offenses?

The purpose of this study is to add to the growing literature describing the characteristics and risk factors (e.g., abuse histories, school problems, gang involvement, living arrangement growing up) of incarcerated youth whose MSO is running away or being a CSE victim. This study will also compare the characteristics of youth incarcerated for running away or CSE victimization to youth incarcerated for all other MSOs, which has not been done with a nationally representative sample of justice-involved youth. This is important because to date, there is limited knowledge about how youth’s characteristics who are incarcerated for running away or CSE victimization compare to youth incarcerated for more serious offenses, as most studies are conducted on samples of a single or few cities, counties or states, or they include a single sex rather than both males and females (e.g., Dank et al., 2015; Naramore et al., 2017; O'Brien et al., 2017; Reid et al., 2017; Salisbury et al., 2015; Trejbalová et al., 2021). There are two research questions in this study: (1) What are the characteristics and risk factors of incarcerated youth whose MSO is running away or CSE victimization? (2) Are the characteristics of youth detained for running away or CSE victimization different than youth detained for more serious offenses?

Methods

Data

This study uses secondary data from the Survey of Youth in Residential Placement (SYRP; Sedlak, 2003), sponsored by the OJJDP. The SYRP was combined with two other census surveys, the Census of Juveniles in Residential Placement (CJRP) and the Juvenile Residential Facility Census (JRFC), which collects facility-level data from the detention facility administrators. The SYRP is unique in that it is the only nationally representative survey to date to gather information directly from the youth in custody rather than using official data or surveying detention administrators (Sedlak et al., 2012). The SYRP drew a nationally representative sample from all youth in state and local facilities that were identified by the CJRP and the JRFC (Sedlak et al., 2012). The SYRP interviewed pre- and post-adjudicated youth between the ages of 10 and 20 housed in juvenile facilities in a multi-stage cluster sampling procedure. The study used a stratified two-stage, probability proportional-to-size sample design, in which they first sampled the facilities and then youths from each facility. Facilities were sampled using the facility offender count as the size measure and then clusters of youth were sampled from each selected facility in the second stage. A total of 290 facilities were randomly sampled across 36 states, of these 204 juvenile justice facilities participated in the survey (70.3% response rate). Of the 9,495 eligible youths within these 204 facilities, 7,073 youths (74.5%) completed the surveys (see Sedlak et al., 2012 for a detailed methods).

The SYRP asked youth about their backgrounds, victimization experiences, offense histories, conditions of confinement, drug/alcohol use, and expectations for the future (Sedlak, 2010). The surveys were electronic and used an audio computer-assisted self-interview (ACASI) system to ask questions and record answers. Youth indicated their response choice by touching it on the screen and the computer program automatically navigated to the next appropriate question based on the youth's earlier answers, storing all the data anonymously and securely. Youths’ survey responses were never associated with their identities and their facility identifiers were removed before data were unencrypted for analysis (Sedlak et al., 2012).

Due to the complex design, the data is weighted so the sample of 7,073 youth reflect the sampling probabilities of both the facility and youth and adjusts for nonresponse at both levels. Survey weights must be used in all analyses of the SYRP data to compute valid totals and proportions and to guard against underestimating standard errors (Sedlak et al., 2012). In this way, the sample of 7,073 provided accurate estimates of the size and characteristics of the national youth offender population in custody, which is estimated as more than 100,000 youth. For this study, 1,097 (15.5%) youth were removed from the sample. Specifically, 833 were removed due to missing data on either the MSO variable (n = 452) or a victimization measure (n = 264), and 381 youth whose MSO was a status offense other than running away or prostitution were removed in order to compare the RunCSE group to youth incarcerated for more serious acts.Footnote 8 This left an unweighted sample of 5,976 youth, but a weighted full SYRP sample of 84,328 youth. Additionally, I completed an attrition analysis, comparing the full sample (n = 7,073) to the reduced sample (n = 5,976) and although there were significant differences on many of the variables due to the large sample size, none were meaningful (see also Sedlak et al., 2012, p. 108–112, regarding rates of missing data for these items and reasons for missing). Readers can request the attrition analysis from the author. Access to the data was granted by approvals from the Florida Atlantic University institutional review board and the National Archive of Criminal Justice Data.

MeasuresFootnote 9

The creators of the SYRP data ran extensive tests to gauge the potential influence of exaggerated or extreme responsesFootnote 10 (Sedlak et al., 2012), which the author replicated to ensure there were no univariate or multivariate outliers. Most serious offense: Multiple questions formed a scale categorizing youths’ current offense by the most severe offense youth had committed. Respondents were coded 1 “Yes” for the MSO type that led to their current detention, and 0 “No” for all other offense types. These categories were then recoded into seven dichotomous indicators of the MSO for which the youth was currently detained. Violent offense included murder, kidnapping, robbery, and assault. Rape included only the indicated offense. Property offense included arson, burglary, auto theft, theft, vandalism, and trespassing; and Drug offense included selling drugs, drug possession, and testing positive for drugs. Carrying a weapon, Running away, and Prostitution included only the indicated offenses. Prostitution was defined to youth as “being paid for having sexual relations with someone” (Sedlak, 2003, p. 106). The sub-sample for this study includes youth whose current MSO was either running away or prostitution (referred to as CSE) and youth incarcerated for all other MSOs. There were too few youths incarcerated for CSE (0.2% of sample) compared to running away (3.5%), that the analyses for CSE youth were unstable when examined alone. Thus, the groups were combined. It should be noted that 54% of the CSE youth had either a prior running conviction (46%) or a current charge for running away (25%). The RunCSE group included 3,117 youth (3.7%), while the MSO_Else group included 81,211 youth (96.3%).

Female was a dichotomous variable coded to indicate (0 = male; 1 = female). Race/ethnicity was measured with five dichotomous variables, one for each of the following: white (non-Hispanic), Black (non-Hispanic), Hispanic, Native American/Asian/ Hawaiian, and Other, or ≥ two races. All dichotomous variables in the study were coded (0 = No; 1 = Yes). Age at interview represents the number of years old when youth were surveyed, not the age they were when they entered the facility.

Prior victimization was assessed using five dichotomous variables and one variable for polyvictimization. All the questions inquired about youths’ victimization experiences prior to their current incarceration. Responses to the five victimization variables were coded as dichotomous variables, with 1 “Yes” and 0 “No.” Physical abuse was based on the question: “When you were living with your family or in another household, did a grown-up in your life hit, beat, kick, or physically abuse you in any way?”. Molestation was based on a survey question that asked: “While you were living with your family or in another household did a grown-up ever touch your private parts when you didn't want them to, or make you touch their private parts?” Youth had forced sex growing up asked: “While you were living with your family or in another household did a grown-up ever force you to have sex?”Footnote 11Emotional abuse was based on one question that asked youth: “While you were living with your family or in another household did you ever get scared or feel really bad because grown-ups called you names, said mean things to you, or said they didn't want you?” Witnessed serious violence was based on one question that asked: “Have you EVER in your whole life seen someone severely injured or killed (in person, not in the movies or on TV)?” Polyvictimization was a dichotomous variable based on the number of victimization types that youth experienced growing up, ranging from zero to five (physical abuse, molestation, forced sex, emotional abuse, and witnessed serious violence). Youth who reported three or more types of victimization were coded 1 (Yes), and those with less than 3 were coded 0 (No), based on Finkelhor et al., (2007, 2011).Footnote 12

School related issues were measured with four dichotomous variables that represent the various issues youth reported, including whether they were below their modal grade, had a learning disability (expert-diagnosed by clinician), and were suspended or expelled from school during the year before they were taken into custody). Four dichotomous variables represent youths’ family/living situation before their arrestFootnote 13 including the following: lived with parent(s) before arrest, lived in group home or with foster parents before arrest, lived with friends before arrest, and homeless or living on own before arrest. A dichotomous variable was also included to measure whether youth had been in foster care or a group home (ever).

Prior juvenile justice involvement was measured a few ways allowed by the data. Three dichotomous variables indicated the type of prior criminal involvement youth had prior to their current incarceration, including whether youth had ever experienced: prior custody, prior probation, and prior conviction. The number of times youth had been in custody prior to their current stay was measured with a dichotomous variable: 3 or more times prior custody. Gang member at time of offense was also included as a dichotomous variable. Additionally, substance use during offense was coded with two dichotomous variables (using drugs only, and using both alcohol & drugs).

Finally, mental health was measured with seven variables representing depression and suicide ideation/behavior. Above average depression symptoms is a scale within the data set that measured depression and anxiety indicators using six items from the Massachusetts Youth Screening Instrument (MAYSI) because of its well established reliability and validity as a screening tool in juvenile justice settings (Grisso & Barnum, 2006; Grisso et al., 2001; Sedlak, 2010). The items asked youth about their feelings and experiences over the past few months, using yes/no response options. Two items asked youth whether they had “felt angry a lot” and “had a lot of bad thoughts or dreams about a bad or scary event that happened” to them. Two items assessed anxiety by asking whether youth had nervous or worried feelings that kept them from doing things they wanted to do, and if they EVER in their whole life had something very bad or terrifying happen to them. Two items assessed isolation by asking youth if they had “felt lonely too much of the time” and if they felt they “don’t have fun with friends anymore.” Scores ranged from 0 to 6 (m = 2.48, SD = 1.77), with higher scores indicating more depressive/anxious symptoms. Above average depression symptoms was coded 1 (Yes) if youth scored a 3 or more on the depressive/anxiety scale and 0 (No) if they scored less than 3.

Suicide ideation was measured a couple different ways, starting with four dichotomous variables (0 = No; 1 = Yes) that asked about youths’ recent suicidal feelings and thoughts, and were taken directly from the MAYSI 5-item Suicide Ideation Scale (Grisso & Barnum, 2006; Grisso et al., 2001). All four items start with “In the past few months have you,” and continue to ask: “wished you were dead”, “felt like life was not worth living”, “felt like hurting yourself”, and “felt like killing yourself”. A fifth overarching variable any suicide ideation (recent) was coded yes (1) and no (0) for a youth who responded yes to the previous four suicide ideation questions. Finally, the variable tried to kill self, asked youth about their lifetime history of suicide attempts (this item is included in several standardized assessment instruments; Goldston, 2000; see Sedlak & McPherson, 2010, p. 14). This variable asked, “Have you EVER in your whole life tried to kill yourself,” using a yes/no response option. This measure was included in addition to the suicide ideation questions because having a previous attempt is one of the strongest predictors of suicide risk (Hayes, 2009; National Action Alliance for Suicide Prevention, 2013).

Results

All analyses were completed in IBM SPSS Statistics v28. Due to the complex sampling design noted above, the data were assigned weights that are required for use in any analysis to yield accurate estimations (see Sedlak et al., 2012). First, univariate statistics were used to describe the characteristics of the full weighted sample of this sample of incarcerated youth. Table 1 includes the means/percentages, standard deviation, and range of MSO variables, and Table 2 includes these statistics for all other variables for the 84,328 respondents included in this study. As noted above, this full sample was then split into two subsamples: youth whose MSO was running away or CSE victimization (group is referred to as RunCSE), and youth whose MSO was something other than running or prostitution (group referred to as MSO-Else). Means and standard deviations of all variables are given for these two subsamples, along with x2 tests (and t-test for age at interview) to determine whether there were significant differences between groups (see Table 3). The significance tests used a Bonferroni adjustment to reduce the odds of a Type I error due to the many comparisons.

Table 1 Most serious offense (MSO) currently detained for (n = 84,328)
Full size table
Table 2 Descriptives for full SYRP sample (N = 84,328)
Full size table
Table 3 Test for differences between youth whose MSO is running away or CSE victimization and youth whose MSO is something else
Full size table

Characteristics and Risk Factors of All Youth in the SYRP Sample

Nearly 4% of the full sample reported their MSO was running away (3.5%) or being a victim of CSE (0.2%). The majority of youth (40.2%) in the sample were incarcerated for a violent offense (i.e., murder, kidnapping, robbery, assault), followed by a property offense (26.9%), drug offense (18.3%), rape (8%), and a small but notable number were incarcerated for carrying a weapon (2.9%).

The average age of respondents was 16.1 years, and the sample was predominantly male (85.5%), white (33.1%), followed by Black, non-Hispanic (31.5%), Hispanic (24.2%), and other or 2 or more races (8.4%). Over 80% of the sample reported experiencing any of the five types of victimization prior to their current detention. Witnessing serious violence was the most prevalent (68.5%) type of victimization youth reported experiencing, followed by physical abuse as a child (34.2%), emotional abuse as a child (28.5%), and molestation (10.1%). About one in five youth were polyvictims, or experienced 3 or more types of victimization prior to their current incarceration.

Looking at the school risk factors among the full sample, nearly half of them were below modal grade, while almost a third (30.4%) had an expert-diagnosed learning disability, and over half (58%) were suspended from school in the year prior to custody. Three out of four youth reported living at home with at least one parent before arrest, while nearly 10% lived with friends, 5.5% were in a foster/group home, and 6% were homeless or living on their own prior to arrest. Over 15% reported living in a foster/group home at any time in the past.

Youth in the full sample had a lot of prior involvement with the JJS, as 84% reported prior probation, 85% had a prior conviction, and 67% had been in custody before, with 40% of youth reporting they had been in custody 3 or more times. About a third of youth (29%) said they were a gang member at the time of their offense. Regarding substance use during their offense, over half said they were not using alcohol or drugs (52.4%), while 19% used drugs only, and 23% used both alcohol and drugs. Almost half of the sample had above average depression symptoms (47.8%), while nearly a third reported having recent suicide ideation (30%), and 22.8% had attempted suicide (ever).

Differences Between Youth Detained for Running or CSE Victimization and Youth with MSOs

The first research question asked about the risk factors for incarcerated youth whose MSO was running away or CSE victimization, and the second question asked if these youth are significantly different than youth detained for more serious offenses. As shown in Table 3, there were many differences between the samples of youth whose MSO was Running or CSE victimization and the MSO-Else group on gender, race/ethnicity, victimization experiences, school-related issues, living situations, justice system involvement, substance use and mental health issues. As expected in a nationally representative sample, males composed the majority of detained youth (85.5%), but when looking at the RunCSE sub-sample, the disparity shifts so that females make up 35% of this sample, which is significantly more females than youth in the MSO-Else group (13.4%). There were no age differences between the groups, but a couple of racial/ethnic differences existed. Most notably, youth whose MSO was running or CSE had a higher rate of youth who were other or ≥ two races (11%) than youth in the MSO-Else group (8.3%). Youth in the MSO-Else sample had significantly more Black (non-Hispanic) youth (31.5%) compared to the RunCSE sample (26%).

Prior Victimization & Polyvictimization

Four out of five youth in both groups experienced some type of prior victimization, but the RunCSE group had higher rates of each victimization type compared to the MSO-Else group, except for witnessed serious violence, wherein 70% of MSO-Else youth reported this type compared to 61% of RunCSE youth (this was the most common victimization type for both groups). About 45% of RunCSE youth reported experiencing physical abuse as a child compared to 34% of MSO-Else youth. A fifth (18.8%) of RunCSE youth were molested as a child, which was almost double what the MSO-Else group reported (9.9%). About 11% of RunCSE youth were raped and 43% were emotionally abused as a child, compared to 7.5% of MSO-Else youth who were raped and 28% who were emotionally abused. More RunCSE youth also experienced polyvictimization, as 29% reported ≥ 3 victimization types compared to 20% of MSO-Else youth.

School-Related Issues

Unlike victimization, MSO_Else youth reported higher rates of all four school-related issues compared to youth whose MSO was running away or CSE victimization. More specifically, MSO-Else youth had higher rates of school suspensions (60%) and expulsions (30%) in the year before custody, a third (31%) had an expert-diagnosed learning disability, and half were below their modal grade. This is compared RunCSE youth, of which 45% reported being suspended and 20% expelled in the year before custody, 27% were diagnosed with a learning disability, and 44% were below their modal grade. Although there is no comparison group of youth who are not detained, these findings are consistent with research that finds juvenile delinquents and youth who run away or are CSE victims experience problems at school at a higher rate than their counterparts (Clarke et al., 2012; Clawson et al., 2009; Development Services Group (DSG), 2015; Tucker et al., 2011).

Family/Living Situation

Like the victimization measures, RunCSE youth had higher rates of risk factors in the family/living situation domain compared to the MSO-Else group. For example, MSO-Else youth were significantly more likely have lived with at least one parent(s) before their arrest (76%) compared to RunCSE youth (62%). Before their arrest, RunCSE youth were more likely to report living in a foster/group home (11%), with friends (16%), or living on own or homeless (11%), compared to MSO-Else youth, of which 5.5% lived in a foster/group home, 10% with friends, and 6% on own or homeless. Finally, 26% of RunCSE youth reported prior foster/group home living (ever) compared to 15% of MSO-Else youth.

Juvenile Justice Involvement

Youth in both sub-samples had high rates of prior contact with the justice system and there was no significant difference on prior custody (ever), as about two thirds of both groups had been in custody before their current detention. However, 46.5% of the RunCSE group reported prior custody 3 or more times, compared to 42.8% of MSO-Else youth. Significantly more RunCSE youth reported prior probation (92.1%) than MSO-Else youth (82.9%), yet there were no group differences on prior convictions, as 85% of both groups reported a prior conviction. RunCSE youth were significantly less likely to be a gang member at the time of their offense (21%) compared to MSO-Else youth (30.5%).

Substance Use During Offense

Substance use was present in commission of many youths’ current offenses, although MSO-Else youth were more likely than RunCSE youth to be under the influence of drugs and/or alcohol at the time of their offense (48.5% and 23.2%, respectively). MSO-Else youth were most likely to report using both alcohol and drugs (23.4%), followed by using drugs only (20%), compared to 13.2% of RunCSE youth using both substances, and 6.6% using drugs only.

Mental Health

An alarming rate of youth reported above average depression symptoms, and although MSO-Else youth had higher rates (47.9%) of this measure than RunCSE youth (45%), these differences were insignificant. The SYRP questions on depression-anxiety include 6 of the 9-items on this MAYSI scale (Sedlak & McPherson, 2010), and MAYSI designers classify scores of 3 or higher in the “caution” range, while scores of 6 or higher are in the “warning” range (Grisso & Barnum, 2006; Grisso et al., 2001). Sedlak & McPherson, (2010) developed a conservative (minimum) estimate of the prevalence of problem scores using only the abbreviated 6 items included in SYRP, 42% of youth have “caution” scores (3–5 indicators) and another 6% have scores at the “warning” level (report all 6 indicators). However, it should be noted the MAYSI guide advises that high scores on the depression-anxiety measure may reflect youth’s reactions to being in trouble with the law and placed in custody (Grisso & Barnum, 2006). Turning to suicide ideation, RunCSE youth reported higher rates of all the suicide ideation indicators, with almost 35% having any recent suicide ideation, compared to 30% of MSO-Else youth. Additionally, significantly more RunCSE youth (31%) reported they had tried to kill themselves (ever), compared to 22.4% of MSO-Else youth. These prevalence rates of attempted suicide are alarming but mostly consistent with prior research, wherein lifetime prevalence rates range from 11 – 27% among detained youth (Abram et al., 2008; Abrantes et al., 2005; Archer et al., 2004).

Discussion

The purpose of this study was to gain a better understanding of the risk factors of detained youth whose MSO was running away or CSE victimization, and to determine whether these youth are different from youth detained for MSOs on these risk factors. This article presents data collected on a nationally representative sample of justice-involved youth. Neither running away or being a CSE victim should result in secure detention according to federal (and many state) laws on status offenders and DMST victims (Casey Foundation [AECF], 2014; Development Services Group (DSG), 2015; Farrell et al., 2019; Jafarian & Ananthakrishnan, 2017; Swaner et al., 2016). Yet, as previously noted, youth continued to be detained for these “offenses” long after state and federal laws decriminalized these acts (Adams et al., 2010; Development Services Group (DSG), 2015; Gies et al., 2020; Levin & Cohen, 2014; Mitchell et al., 2010; Office ofJuvenile Justice and Delinquency Prevention, 2022; Pullmann et al., 2020).

One contribution of this study is that it confirms and extends prior research on the characteristics of this population who end up in the depths of the justice system (i.e., detention). Consistent with previous studies, the youth incarcerated for running away or CSE victimization have high rates of prior victimization, including physical abuse (45% of the youth), emotional abuse (43%), witnessing serious violence (61%), and molestation (19%) and rape as a child (11%; Lanctôt et al., 2020; Levin & Cohen, 2014; Naramore et al., 2017; O'Brien et al., 2017; Reid et al., 2017). Youth detained for running away or CSE victimization had significantly higher rates of every type of abuse and polyvictimization compared to youth detained for more serious offenses, except for witnessing serious violence. This is significant because youth in the full SYRP sample already had high rates of prior victimization (as seen in Table 2), but it is safe to say from these findings that RunCSE youth have more extensive histories of victimization than youth incarcerated for more serious offenses. This finding is consistent with many studies that found justice-involved youth have higher rates of ACEs compared to non-delinquent youth (Baglivio & Epps, 2016; Baglivio et al., 2015; Ford et al., 2013; Fox et al., 2015), and minor CSE victims have higher rates of ACEs than justice-involved youth who have not been sex trafficked (Naramore et al., 2017; Reid et al., 2017, 2019).

While there is prior research examining incarcerated youth who are simply runaways or victims of CSE, no studies to our knowledge have examined the risk factors of detain youth with a nationally representative sample, or compared this population to youth detained for other crimes (as most studies utilize samples of a single or few cities or states; or they include a single sex, or have too few CSE victims in a sample to make meaningful comparisons; Chapple & Crawford, 2019; Dank et al., 2015; Naramore et al., 2017; O'Brien et al., 2017; Reid et al., 2017; Salisbury et al., 2015). Knowing the risk factors among this group of detained youth may improve the implementation of early intervention efforts and could lead to a collaborative response between the JJS and other entities to find alternatives to detention to keep these youth safe from returning to the streets or their trafficker (Clayton et al., 2013; Mitchell et al., 2010; Pullmann et al., 2020; Reid & Piquero, 2014a, b).

There were numerous differences between the characteristics of the two groups that should be considered when responding to each group. For example, youth incarcerated for more serious offenses than running away or being a CSE victim had higher rates of issues at school and above average number depression symptoms, though both groups had rates that should be a warning for early intervention efforts. Like prior victimization, youth detained for running away or CSE victimization reported higher rates of other risk factors examined in this study compared to the MSO-Else group, including poor living situations before arrest and growing up, prior probation, and suicidal ideation and prior attempts. It is crucial that communities and law enforcement have a more trauma-informed response to runaways and CSE victims, in addition to responsive policies that will reduce harm and improve outcomes for the affected youth (Edmond, 2018; Farrell et al., 2019; Institute of Medicine (IOM) & National Research Council (NRC), 2014; Reid et al., 2019; Scott et al., 2019).

As previous noted, disproportionate minority contact (DMC) has been an important research and policy issue because if youth are treated differently by the JJS based on extra-legal factors, like race/ethnicity, it is unfair and unjust (Casey Foundation [AECF], 2021; Cabaniss et al., 2007; Kempf-Leonard, 2007; McCarter, 2009; Sickmund et al., 2011). Many studies of incarcerated youth have found that minorities are disproportionately confined compared to their make-up in the general population (e.g., Casey Foundation [AECF], 2021; Jafarian & Ananthakrishnan, 2017; Rovner, 2021). Consistent with this body of research, this study also found that minorities were over-represented in both sub-samples compared to their makeup in the U.S. population in 2003 (when the SYRP was collected). For example, only 36% of youth in the RunCSE group and 33.4% in the MSO-Else group were white (non-Hispanic), which is significantly less than the rate of white (non-Hispanic) youth (aged 10–19) in the U.S. population (62.9%; National Adolescent Health Information Center National Adolescent Health Information Center, 2003). Blacks were over-represented in the current study compared to the general population (14.5%), and this was exaggerated for youth in the MSO-Else group (31.5%) compared to the RunCSE group (26%; National Adolescent Health Information Center, 2003). There was also a disparate rate of Hispanics in the current study for both groups compared to the youth population (19%), though the rate was slightly higher for RunCSE youth (25.1%) than MSO-Else youth (23.9%; National Adolescent Health Information Center, 2003).

Implications

Detaining youth who are not a risk to public safety or themselves puts an undue financial burden on taxpayers because secure detention is more expensive than alternatives, such as probation (Casey Foundation [AECF], 2021; Holman & Ziedenberg, 2011). Additionally, detaining youth who should not even be considered offenders (legally) and who have extensive histories of victimization and mental health issues only exacerbates their underlying vulnerabilities that may have led youth to run away or be sexually exploited in the first place (Chapple & Crawford, 2019; Henrichson & Delaney, 2012; Malvaso et al., 2018; Okwori et al., 2022; Peterson et al., 2018; Safe & Sound, 2020). Findings from this study have important implications for prevention and intervention services provided to youth at risk for running away or CSE victimization, especially those who end up in the JJS (e.g., Development Services Group (DSG), 2015; Geist, 2012; Showden & Majic, 2018;). According to Farrell et al., (2019), most CSE victims interact with police engaged in routine duties during their victimization, and analysis of arrest data reveals that police tend to view these youth CSE victims as offenders rather than victims (Bejinariu et al., 2021; Halter, 2010; Mir, 2013; Williams, 2015). Vanwesenbeeck, (2017) argued that the criminalization of CSE victims and police responses to them promote stigma and discrimination, placing these minor CSE victims at an increased risk of violence, while also limiting their access to victim services and police protection (see also Halter, 2010; and Smith et al., 2009).

Thus, education and training for law enforcement, as well as other professionals likely to come into contact with runaways or minor CSE victims, like teachers or healthcare workers (Institute of Medicine (IOM) & National Research Council (NRC), 2014; Gambon et al., 2020), is necessary so these professionals can identify sex trafficking victims and reduce misidentification as offenders, especially since most CSE victims do not identify themselves as victims (Bounds et al., 2015; Clawson et al., 2009; Smith et al., 2009). States without policies that mandate training for law enforcement, juvenile justice actors, and healthcare workers on how to identify and respond to victimized youth, especially runaways and CSE victims, should consider passing such legislation (Edmond, 2018; Scott et al., 2019). Additionally, JJS actors should utilize validated screening tools for all arrested youth regardless of current offense so they can identify youth who may be runaways or CSE victims who need access to services/treatment, rather than detention (Development Services Group (DSG), 2015; Naramore et al., 2017; Reid et al., 2017; Salisbury et al., 2015). Such screening tools could also help identify youth at risk for suicide and in need of mental health services, as youth in the current study reported high rates of suicide ideation/attempts and depression (see also Chapple & Crawford, 2019; Hayes, 2009; Yoder, 1999).

Additionally, local and state agencies who work with runaways and youth CSE victims should emphasize trauma-informed training and practices in their respective fields (Institute of Medicine (IOM) & National Research Council (NRC), 2014; Edmond, 2018; Gambon et al., 2020; Nichols et al., 2022; Reid et al., 2019; Scott et al., 2019; Showden & Majic, 2018). If states invested in providing treatment and residential care for runaway youth and minor CSE victims rather than detaining them to keep them safe, perhaps this would help youth exit their current trajectory, which increases many adverse effects noted above (Casey Foundation [AECF], 2021; Chapple & Crawford, 2019; Fox et al., 2015; Jeanis et al., 2020; Nichols et al., 2022).

Limitations

There are a few limitations to this study that warrant discussion. First, there is no group of run away or minor CSE youth who are not involved in the justice system to compare risk factors of those who are detained versus not detained. Similarly, there are likely many more youth among the SYRP (2003) sample who had run away or been a CSE victim, but they did not have an official record (i.e., arrest, charge, or conviction) of the act, or were charged with more serious offenses and thus were not included in the RunCSE sub-sample this study examined more closely.

Another limitation is that most of the measures were based on youths’ self-reports, which may be subject to poor memory or recall problems, such as an inability to remember/comprehend what happened to them, or an unwillingness to admit something on the survey, such as victimization (Finkelhor et al., 2007; Tourangeau et al., 2000). Additionally, there are other types of victimization that would have been useful to have in the study, but were absent from the SYRP data (e.g., bullying, witnessing domestic violence). However, a strength of the SYRP data is that it asked youth about different types of victimization they experienced rather than just having an umbrella victimization measure. There were other risk factors were missing from the data that would have been beneficial to include in this study, such as LGBTQ status, poverty, peer delinquency, self-control, parental mental health and criminal history, and aggregate-level factors (to determine if the RunCSE youth were from just a few facilities/states; Dank et al., 2015; Franchino-Olsen, 2021; Landers et al., 2017). Some may also argue the SYRP data is outdated since it was collected in 2003 and many states have changed their policies and practices around the treatment of runaway youth and minor sex trafficking victims. Yet, even after safe harbor laws were enacted and there was an emphasis on alternatives to detention for status offenders, many of these youth continued to be arrested and detained (Casey Foundation [AECF], 2021; Development Services Group (DSG), 2015; Geist, 2012; Gies et al., 2020; Wasch et al., 2016; Williams, 2015).

Conclusions

This study adds to the growing body of research regarding the risk factors of runaway and CSE-victimized youth who end up incarcerated using a nationally representative sample of justice-involved youth. This study also helps distinguish the risk factors between these youth and youth incarcerated for more serious offenses. It is recommended that future research examine both the risk and protective factors of youth who runaway or are CSE victims, and look at these factors at various stages of justice-system involvement rather than just adjudicated or detained youth. Preventing direct and indirect victimization during childhood and adolescence is essential to reduce the number of youths at risk of running away, being sexually exploited, abusing drugs or alcohol, mental health and suicidal ideation, offending behavior, or further victimization. However, when prevention is too late, it is important for intervention and treatment programs to be available and fully funded in order help youth who have experienced victimization. The JJS would benefit from taking a trauma-informed approach to all youth they interact with, and especially runaways and youth CSE victims, so that those with traumatic experiences can receive services to help them cope with any adverse effects (Edmond, 2018; Harper et al., 2015; Holman & Ziedenberg, 2011; O'Brien et al., 2017; Scott et al., 2019). Although both prevention and intervention programs may cost a lot and it is difficult to determine their effectiveness, such programs may cost less than the dealing with effects of victimization on children and adolescents (in terms of health care, child welfare, and justice-system costs for victims who become delinquent; Henrichson & Delaney, 2012; Okwori et al., 2022; Peterson et al., 2018).