PSYCHOPATHY AND CHOICE OF VICTIMS: IMPLICATIONS FOR THE SUB-CATEGORIZATION AND TREATMENT OF JUVENILE SEXUAL OFFENDERS Except where reference is made to the work of others, the work described in this dissertation is my own or was done in collaboration with my advisory committee. The dissertation does not include proprietary or classified information. ______________________________________________________ Jason Francis Sikorski Certificate of Approval: _______________________________ ______________________________ William Buskist Barry Burkhart Professor Professor, Chair Psychology Psychology _______________________________ ______________________________ Alejandro Lazarte Frank W. Weathers Assistant Professor Associate Professor Psychology Psychology _______________________________ ______________________________ Gregory S. Kowalski Stephen L. McFarland Profesor Dean Sociolgy Graduate Schol PSYCHOPATHY AND CHOICE OF VICTIMS: IMPLICATIONS FOR THE SUB-CATEGORIZATION AND TREATMENT OF JUVENILE SEXUAL OFFENDERS Jason F. Sikorski A Dissertation Submitted to the Graduate Faculty of Auburn University in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy Auburn, Alabama December 16, 2005 iii PSYCHOPATHY AND CHOICE OF VICTIMS: IMPLICATIONS FOR THE SUB-CATEGORIZATION AND TREATMENT OF JUVENILE SEXUAL OFFENDERS Jason F. Sikorski Permission is granted to Auburn University to make copies of this dissertation at its discretion, upon request of individuals or institutions and at their expense. The author reserves all publication rights. ____________________________________ Signature of Author ____________________________________ Date of Graduation iv DISSERTATION ABSTRACT PSYCHOPATHY AND CHOICE OF VICTIMS: IMPLICATIONS FOR THE SUB-CATEGORIZATION AND TREATMENT OF JUVENILE SEXUAL OFFENDERS Jason F. Sikorski Doctor of Philosophy, December 16, 2005 (M.A., East Tennessee State University, 1999) (B.A., LeMoyne College, 1997) 137 Typed Pages Directed by Barry R. Burkhart Adolescent sexual offenders are responsible for a considerable portion of the sexual abuse victimization that takes place in American society today. Convicted sexual offenders report that they most often began their sexual offending proclivities in adolescence; a significant minority of these juvenile sexual offenders will continue sexually offending against multiple victims well into adulthood. Being able to identify those juvenile sexual offenders at the highest risk for developing an entrenched pattern of sexual deviancy into adulthood represents a social imperative, as failure to identify these youths might result in untold emotional and financial costs to victims, their families, and society as a whole. Most often, research pertaining to the assessment of juvenile sexual offenders has grouped all of these juveniles into a homogenous set. However, descriptive studies and comparative analyses to date, provide evidence that juvenile sexual offenders are a heterogeneous collection of individuals with varying treatment needs and associated risks for sexual and non-sexual criminal recidivism. Many investigators have called for v research to focus on ways of sub-categorizing juvenile sexual offenders in useful and reliable ways. Further, explicating the differences among varying categories of juvenile sexual offenders and non-sexually offending juvenile delinquents in general, has great potential utility for the individuals responsible for tailoring and managing treatment programs for delinquent youth. 381 juvenile delinquents housed in a secure detention center in the Southeastern United States served as participants. Youth were categorized based on the age and gender of their sexual offense victims (5 levels: i) juvenile sexual offenders who targeted females exclusively, and at least one victim younger than themselves by 4 years or more; ii) juvenile sexual offenders who targeted females who were peer-aged exclusively; iii) juvenile sexual offenders who victimized at least one male and at least one victim younger than themselves by 4 years or more; iv) juvenile sexual offenders who targeted peer-aged victims, at least one of whom was male; and v) non sexually offending juvenile delinquents with no documented history of sexual offenses). The sample was further grouped based on their scores on the Psychopathy Checklist: Juvenile Version (PCL:YV) (2 levels: i) at or above the sample median Psychopathy Total Score; and ii) below the sample median Psychopathy Total Score). 2-way Analyses of Variance and Pearson Chi- Square Non-Parametric tests were used to analyze for differences across the levels of the independent variables for multiple dependent measures. Dependent measures were obtained for the following general areas: Demographics, History of Abuse Victimization, History of Violence/Criminal Offenses, Violence Exposure, Sexual Offending Variables, Psychiatric History/Personality Functioning, and Substance Abuse. Results suggested that categorizations of juvenile sexual offenders based on the age(s) and gender(s) of their sexual offense victims and their adherence to behaviors consistent with a psychopathic personality pattern hold great promise for future researchers. Individuals high in psychopathy were shown to feature more extensive criminal histories, chaotic caregiver relationships, and difficulties obeying authority figures. They were also exposed to a considerable degree of violence from an early age, and tend to behave in an unruly and forceful fashion. Juvenile sexual offenders who targeted female peers exclusively were shown to be most similar to non-sexually offending juvenile delinquents in their cognitions, behaviors, and self-concepts. Juvenile sexual offenders who victimized young children, especially those who had at least one male sexual offense victim, showed the greatest number of signs and symptoms consistent with the development of sexual deviancy. They described being preoccupied with sexual issues, feeling like they needed professional help to control their impulses, and had fewer perceived friends and social supports to assist them in managing their behavior and mood. Juvenile sexual offenders rated high in psychopathy and featuring a sexual offense history against at least one male and at least one victim younger than themselves by 4 years or more were identified as a highly impulsive and difficult to treat sub-group of juvenile sexual offenders. Directions for future research, and the importance of further exploring this categorization structure, are discussed. vi ACKNOWLEDGEMENTS I can honestly say that this work would not have been completed properly without the assistance of my major professor, Barry R. Burkhart, Ph.D. I thank him for the generous amount of time he spent to ensure that this work was completed before beginning my post-doctoral fellowship. Further, I would also like to thank William Buskist, Ph.D. for reviewing several versions of this manuscript, and taking time to offer words of support and encouragement. My intention is to pass on the lessons that I have learned about mentorship from these two individuals to my future students. I would also like to mention Todd Smitherman, a friend in the truest sense of the word. He made sure that my graduate training was not all about work and stress. I?d also like to thank Ketti Horton, Jeremy Sikorski, and Jessica Swanker for offering hours of their time to ensuring that ?everything else? was taken care of while I saw 40 clients per week and worked to complete this manuscript while on internship. I will never forget the contributions they have made to my success. Finally, and most importantly, I would like to thank my parents, Francis and Jacqueline Sikorski. Without their emotional and financial support, it would have been impossible to complete my degree. A person could not ask for more loving parents. This degree is as much yours, as it is mine. Thank you from the bottom of my heart. Words cannot be used to articulate the degree of gratitude that you both deserve. And Pop, this one is for you!!! vii Style manual or journal used: American Psychological Association Publication Manual: 5 th Edition. (2001). American Psychological Association, Incorporated: Washington, D.C. Computer softwared used: 1) Statistical Program for the Social Sciences (SPSS), Version 10.0 for Windows. 2) Microsoft Office Word 2003 viii TABLE OF CONTENTS Section/Sub-Section Page LIST OF TABLES x INTRODUCTION 1 The Stae of the Art 3 Demographics, Family Variables, and Social Relationships 5 History of Delinquency 7 Psychiatric History 12 History of Sexual and Physical Abuse Victimization 17 Sexual Offense Related Variables 21 Sumary 25 The Presnt Study 25 Hypothes 3 METHOD 40 Particpants 40 Materials/Measures 47 Rating Scales 48 Self-Report Measures 49 Procedure 52 Study Design 54 ix TABLE OF CONTENTS (continued) Section/Sub-Section Page RESULTS 58 Demographics/Family Life/Friends 58 History of Abuse 61 History of Violence/Criminal Offenses 62 Violence Exposure Across Life Contexts 67 Sexual Ofending Varibles 71 Psychiatric History/Personality Functioning 79 Substance Abuse 87 DISCUSION 102 Juvenile sexual offending and non-sexually offending delinquents 103 Within-Group Comparisons of Juvenile Sexual Offender Types 107 Psychopathy-Based Group Differences Interaction Efects 15 Limitaions 16 Futre Directions 18 Conclusion 18 REFERENCES 12 x LIST OF TABLES Title of Table Page Table I: Dependent variables examined through 29 2-way Analyses of Variance (ANOVA) (continuous dependent variables) Table II: Dependent variables examined through 31 Pearson Chi-Square Analyses (categorical dependent variables) Table III: Percentages of Sample Featuring a Past 43 History of Specific Criminal Offenses Table IV: Percentage of Juvenile Sexual Offenders 45 Featuring specific index offenses Table V: Power Estimates, Effect Sizes, and P-Values 92 Associated with Significant ANOVA Main and Interaction Effects 1 INTRODUCTION The literature on assessment and treatment of juvenile sexual offenders is in its infancy. It has only now been recognized as an area of research distinct from the literature on adult sexual offenders over the last two decades. Rather than viewing juvenile sexual offending behavior as a serious problem, researchers and public officials once seemed to conclude that the incidence of juvenile sexual offending is low and that sexual acts perpetrated by youths are exploratory in nature and not as serious as those committed by adult sexual offenders (Becker, 1988). However, victim surveys have shown that 30% to 50% of all sexual abuse can be linked to adolescent perpetrators (Davis & Leitenberg, 1987). Moreover, approximately 50% of adult sex offenders reported committing their first sexual offense in adolescence and feature 2 to 5 times more offenses than apprehensions (Groth, Longo, & McFadin, 1982). The sexual offenses perpetrated by adults also appear similar in modus operandi and victim choice to the sexual offenses that they committed as juveniles (Groth, 1977), suggesting a developmental continuum in sexually deviant behavior exhibited by sexual offenders. Sexual offenses committed by juvenile sexual offenders are often dismissed by others, especially when the victim in question was a sibling, friend, or neighbor (Groth, 1977). Thus, although juvenile sexual offenders may show only one prior sexual offense on their criminal records, it is more likely that their deviant sexual history is more extensive. Because sexual offenses perpetrated by juveniles are often ignored, the risk 2 becomes that sexual offenders will continue perpetrating sexually deviant acts into adulthood, broadening the network of sexual offense victims in society (Groth et al., 1982). In a well-designed examination of the potential impact that sexual offenders can have on society if left unchecked, Abel, Becker, Mittelman, Cunningham-Rathner, Rouleau, and Murphy (1987) recruited non-incarcerated sexual offenders to provide information on the type and frequency of sexual offenses they committed during their lives. All participants were briefed that their reports would be kept strictly confidential provided that they omitted revelations involving specific information that might identify their victims. Although these sexual offenders featured only a few sexual offenses according to their official records, they collectively reported an array of victims and number of offenses that customarily spanned several years. The 561 adult sexual offenders engaged in behaviors that ranged from frotteurism and exhibitionism to rape and child molestation, and admitted to 291,737 paraphillic acts against 195,407 victims. Undoubtedly, the risk of failing to treat sexual offenders early in their developmental history is exceptional, in terms of societal costs, individual victims, the families of victims, offenders, and the families of offenders. Thus, an imperative for mental health professionals is to identify and intervene with juvenile sexual offenders before their sexual offending proclivities become more difficult to remediate. These statistics suggest that juvenile sexual offending represents a persistent problem that frequently intensifies in adulthood. However, the burgeoning developmental psychopathology perspective in mental health offers considerable hope. Cichetti and Rogosch (2002) suggested that children and adolescents may represent excellent 3 candidates to learn new skills and alter patterns of behavior before they become entrenched in their behavioral repertoire and psychosocial functioning. Featuring follow- up periods of up to 4 years, some studies have suggested that treated juvenile sexual offenders feature recidivism rates that are less than 15% (Nisbet, Wilson, & Smallbone, 2004; Rasmussen, 1999; Smith & Monastersky, 1986), and sometimes even lower (Kahn & Chambers, 1991). Thus, if identified early, juvenile sexual offender treatment appears to represent a potentially useful tool in preventing future sexual victimization and steering juvenile sexual offenders toward learning skills that increase their chances of improving. The State of the Art The majority of research on juvenile sexual offenders is descriptive in nature. This research has produced discrepant results based on the samples utilized (e.g., incarcerated youths, youths treated on an outpatient basis, etc.), where data were collected (e.g., detention center, the youth?s home, etc.), informants used (e.g., parent- completed measures, teacher-completed measures, self-report scales, rating scales, interview-based procedures, etc.), and the time at which the juvenile sexual offenders were assessed (e.g., pre-treatment assessment vs. post-treatment assessment). This literature includes all types of juvenile sexual offenders as being representative of a homogenous subset of delinquent youths. However, this perspective might result in all juvenile sexual offenders being treated in a regimented way, despite the glaring anomalies that each individual possesses. Failing to sub-categorize the different types of juvenile sexual offenders may result in rising sexual and non-sexual recidivism rates following treatment. 4 Below studies of the historical, individual, and contextual characteristics of juvenile sexual offenders are reviewed. However, one caveat is in order: without the use of appropriate delinquent comparison groups, it is impossible to ascertain whether the characteristics of juvenile sexual offenders identified in descriptive studies to date represent characteristics or risk factors for juvenile sexual offending, juvenile delinquency in general, or neither (Davis & Leitenberg, 1987; Spaccarelli, Bowden, Coatsworth, & Kim, 1997). Multiple types of adolescent sexual deviance, from forcible rape to voyeurism, exist in society. Sub-categorizing these different forms of sexual misconduct will likely enhance our ability to construct theories that guide our choice of assessment measures used to understand and predict the future behavior of many specific types of juvenile sexual offenders (Fagan, & Wexler, 1988). Thus, studies documenting the presence of differences and similarities across varying subtypes of juvenile delinquents are examined. Finally, studies in which researchers sub-classified juvenile sexual offenders and their relevant comparison groups are reviewed and used to delineate conclusions about the characteristic differences among categories of juvenile sexual offenders and other juvenile delinquent groups. Emphasis is placed on variables likely to inform the type and duration of treatment that individual juvenile sexual offenders receive to reduce their risk of sexual offense recidivism, general criminal recidivism, and psychiatric and psychological deterioration. Careful attention is applied to the types of samples utilized, design procedures employed, when data were collected, and the setting in which these data were obtained. Following this review, an argument is made for conducting additional research on sub-classifying juvenile sexual offenders based on offense characteristics such as victim 5 age and gender. In addition dispositional variables such as psychopathy also hold great promise in constructing useful typologies, as the construct of psychopathy has shown to predict general criminal and sexual offense recidivism in adult offenders (Hildebrand, de Ruiter, & de Vogel, 2004) and juvenile offenders (Gretton, McBride, Hare, Shaugnessy, & Kumka, 2001). Demographics, Family Variables, and Social Relationships The average age of juvenile sexual offenders participating in studies to date has generally ranged from 15 to 16 (Becker, Cunningham-Rathner, & Kaplan, 1986; Vinogradov, Dishotsky, Doty, & Tinklenberg, 1988). Incarcerated offenders have been shown to be older than outpatient juvenile sexual offenders, likely due to their longer history of general delinquent behavior leading to their incarceration (Aljazireh, 1993; Vinogradov et al., 1988). Sexual offenders are found throughout the country, in both urban and rural areas, and come from families with a wide range of socioeconomic statuses (Ryan, Miyoshi, Metzner, Krugman, & Fryer, 1996). Juvenile sexual offenders typically have been found to live as part of families where significant dysfunction is present (Aljazireh, 1993). For instance, studies have shown that juvenile sexual offenders often do not live with both birth parents (Becker et al., 1986, Ryan et al., 1996). A significant minority live on the streets, with foster parents, in group homes, or in detention centers (Becker et al., 1986). In fact, 50% of a sample including non-sexual juvenile delinquents, child sexual assaulters, and child molesters were separated from their mothers, and 70% were separated from their biological fathers (Awad & Saunders, 1991). Parents and caregivers of juvenile sexual offenders have been found to feature prominent substance abuse histories (Awad & Saunders, 1991; Becker et 6 al., 1986.; Fagan & Wexler, 1988; Hsu & Starzynski, 1990; Kahn & Chambers, 1991; VanNess, 1984), and personal involvement with the criminal justice system (Fagan & Wexler, 1988; Hsu & Starzynski, 1990). The juvenile sexual offender is most often not the only child in these turbulent home environments; one nationwide study has shown that as many as 97% of juvenile sexual offenders live in homes where other juveniles are present (Ryan et al., 1996). However, in comparison to other groups of juvenile delinquents, juvenile sexual offenders appear to come from families featuring less family dysfunction (Aljazireh, 1993). A heterogeneous sample of juvenile sexual offenders was found to have been more likely to live with their birthparents while violent, non-sexual offenders often lived in single parent homes (Fagan & Wexler, 1988). Another study, noted for its methodological sophistication, found that a demographically matched group of assaultive offenders was found to possess lower levels of family cohesion and adaptability than a heterogeneous group of juvenile sexual offenders (Blaske, Borduin, Henggeler, & Mann, 1989). Child sexual assaulters, child molesters, and non-sexual juvenile delinquents were found to possess comparably high levels of parental psychopathology. Approximately 25% of the mothers of a sample of child sexual assaulters (n = 49) had a history of depression and two had committed suicide (Awad & Saunders, 1991). Further, some studies have shown that as many as 26% of juvenile sexual offenders come from family situations in which close relatives were known to feature varying types of sexual deviance including prostitution, bigamy, and sexual assault (Awad & Saunders, 1991). 7 Fehrenbach, Smith, Monastersky, and Deisher (1986) found that 32% of a heterogeneous group of juvenile sexual offenders reported having no friends; approximately 65% exhibited serious signs of feeling socially isolated from others. Adolescent child molesters reported having very few close friends and rarely being able to keep friends for a period beyond 6 months (Awad & Saunders, 1991). In addition, juvenile child molesters were found to self-report significantly more feelings of social isolation and loneliness compared to a sample of adolescent sexual assaulters and a sample of non-sexual juvenile delinquents (Awad & Saunders, 1991). When compared to violent juvenile delinquents without a documented history of juvenile sexual offending behavior, the friends of juvenile sexual offenders were shown to feature significantly fewer delinquent behaviors (Blaske et al., 1989; Fagan & Wexler, 1988). Little research has been devoted to examining the influence of potentially important demographic characteristics like religion, socioeconomic status, and cultural influences on mediating or moderating one?s treatment and risk of future sexual offense recidivism (Aljazireh, 1993). In examining the general psychotherapy literature for adult and juvenile samples, these variables are posited as important issues to incorporate into effective treatment protocols. History of Delinquency Adult sexual offenders have shown to feature very high rates of non-sexual criminal behavior. Using an automated interview with adult sexual offenders committed to a sex offender treatment program, a sample of 99 participants admitted to committing over 19,500 nonsexual criminal offenses in the year preceding their incarceration (Weinrott & Saylor, 1991). Both rapists and child molesters were found to engage in 8 extremely high rates of criminal activity that most often included substance abuse, burglary, robbery, and assault. However, the criminal activity displayed by juveniles varies based on whether the sample is obtained from outpatient clinics, residential treatment facilities, or detention centers. Studies involving large and heterogeneous samples of juvenile sexual offenders have found that anywhere from around 40% (Fehrenbach et al., 1986) to approximately 60% (Awad & Saunders, 1991; Ryan et al., 1996) feature a prior history of committing non-sexual criminal offenses. In fact, Van Ness (1984) found that about 85% of a heterogeneous and small sample of adjudicated sexual offenders had featured at least four incidents of past violent behavior in their files alone. However, using delinquent and non- delinquent comparison groups allows researchers to develop more specific hypotheses about which types of juvenile delinquents are most likely to feature an extensive history of criminal involvement that does not include sexual offending. The majority of studies comparing heterogeneous samples of juvenile sexual offenders to heterogeneous samples of other types of juvenile delinquents with a history of extensive violent crime other than sexual offenses have resulted in null findings. For instance, Otnow-Lewis, Shankok, and Pincus (1979) compared a group of incarcerated non-sexual violent juvenile offenders charged with committing at least two serious crimes involving persons (e.g., murder, assault, robbery, armed robbery) to a small and heterogeneous sample of juvenile sexual offenders charged with rape, attempted rape, forcible sodomy, or sexual assault. The sexual offenders featured an extensive and severe history of non-sexual violent behavior that began in childhood. They committed an array of non-sexual violent crimes, were involved in multiple physical altercations with peers, 9 and witnessed a wide range of extreme violence in their homes and neighborhood. In short, the violence perpetrated and witnessed by this juvenile sexual offender sample was remarkably similar in frequency and seriousness to the non-sexual violent delinquent group (Otnow-Lewis et al., 1979). Fagan and Wexler (1988) compared 208 juveniles charged with a non-sexual violent felony and a sample of 34 juvenile sexual offenders charged with a range of serious sexual offenses. They found that juvenile sexual offenders featured family members who were less likely to be involved with the criminal justice system. In addition, they had less gang involvement, fewer arrests for nonviolent offenses, and were less likely to have documented problems with drugs and alcohol. Moderately sized samples of juveniles charged with a sexual offense, nonsexual violent offense against persons, or a nonsexual and nonviolent property crime were compared based on their scores on the Jesness Inventory (Oliver, Nagayama-Hall, & Neuhaus, 1993). The heterogeneous group of juvenile rapists and child molesters (30 charged with rape, 20 charged with gross sexual imposition) featured the least deviant background and personality characteristics amongst the three groups of juvenile delinquents. The sexual offending group featured fewer past mental health contacts, less remarkable personal and family criminal histories, and was most likely to feature un- elevated Jesness profile index scores. However, the sexual offenders in this study were found to be least likely to be classified at the I-4 level of the Jesness Classification System, which is denoted as being most indicative of the development of interpersonal maturity. Of note, the adolescent sexual offender sample was atypical in that African- Americans were over-represented, and that the most pathological adolescent sexual 10 offenders were likely being treated in a hospital or residential setting as opposed to remaining in the community (Oliver et al., 1993). Incarcerated samples of delinquent youths likely feature individual participants with the most severe degree of psychiatric difficulties, turbulent family histories, and criminal involvement. Caputo, Frick, and Brodsky (1999) broke a moderately sized sample of incarcerated delinquents into small groups of contact sexual offenders, non- sexual violent offenders who committed crimes like burglary and robbery, and non- violent offenders who featured crimes that did not involve physical contact (e.g., destruction of property, theft, and other property offenses). The contact sexual offender and violent contact offender groups featured extensive histories of witnessing serious domestic violence perpetrated against their mother or mother figures. The violence that these groups witnessed occurred most commonly for more than 3 years. The non-contact offenders featured a significantly less turbulent history of witnessing domestic violence in their homes. The groups were then compared based on their scores on the Psychopathy Screening Device, which represents a juvenile extension of the widely used and researched Psychopathy Checklist ? Revised version (PCL-R; Hare, 1991). Groups did not differ on the percentage of individuals from each category featuring elevated scores on the scale designed to gauge impulsivity, whereas significant group differences did emerge when examining scores on the Callous/Unemotional scale. Sexual offenders featured scores that were significantly higher than both the nonsexual violent offender group and the non-contact offender group. Further research on the development of callous and unemotional traits and their relationship to juvenile sexual offending was recommended by the authors. However, the use of an exclusively self-report based 11 assessment protocol likely limits the confidence that one might place in their results obtained (Caputo et al., 1999). Comparing a heterogeneous group of incarcerated juvenile sexual offenders to a mixed group of delinquents featuring low levels of violence in their past histories resulted in the emergence of robust and statistically significant differences (Spaccarelli et al., 1997). The juvenile sexual offenders were shown to have been exposed to more physical and domestic abuse in their homes, featured attitudes that were more supportive of an aggressive lifestyle, and commonly chose to cope with their own distressing emotions by causing pain and distress in others (Spaccarelli et al., 1997) However, these same authors found few differences between a heterogeneous sample of juvenile sexual offenders and a heterogeneous sample of violent juvenile delinquents who had committed acts of serious violence against others. The two groups featured statistically indistinct histories of violent behavior, attitudes toward physical and sexual aggression, self-reported social competence and coping strategies, and self-reported symptoms of anxiety, depression, post-traumatic stress disorder, and ruminative tendencies. The similarities between the juvenile sexual offenders and the violent non-sexual offenders included the fact that a significant proportion of the most violent non-sexual offenders admitted to acts of sexual aggression (19.7%) at levels that exceeded expected base rates of sexual abuse perpetration in the community (Spaccarelli et al., 1997). In a 32-year follow-up review of approximately 7400 individuals adjudicated for crimes in the United Kingdom, Soothill, Francis, Sanderson, and Ackerley (2000) found that individuals committing consensual and non-consensual crimes against males of varying ages were different from individuals committing consensual or non-consensual 12 crimes against females in several ways. Most notably, those offenders who chose female victims were more likely to be convicted of violent crimes against persons during the follow-up period. However, male victim sexual offenders in the United Kingdom were more likely than female victim sexual offenders to recidivate sexually during the follow- up period. All groups featured an alarmingly high rate of subsequent arrests for violent criminal activity that did not include sexual offense recidivism (approximately 40%). Smith (1988) suggested that those juvenile sexual offenders who had committed the ?most serious? types of juvenile sexual offenses were most likely to have a serious non-sexual criminal history. However, Smith chose to define a sexual offense as ?more serious? if the victim was a relative rather than a friend, acquaintance, or stranger. This type of ?seriousness distinction? for sexual offenses is not widely used in the literature, and represents an empirical question that has yet to be examined thoroughly. Thus, studies attempting to differentiate groups of juvenile sexual offenders on their prior criminal history, or any other variable, must be based on well-defined and widely used definitional criteria in order for progress to be made in the field. Psychiatric History Studies have shown that as many as 70% of heterogeneous samples of juvenile sexual offenders ?show signs of moderate to severe maladjustment? displayed across the home and school contexts (e.g., Awad & Saunders, 1991). In examining a very large sample of juvenile sexual offenders from across the country, approximately 45% had been engaged in psychotherapy at some point in their lives (Ryan et al., 1996). It could be hypothesized that the high rates of psychosocial difficulties observed in this sample of youths are related to their history of insecure attachment, as around 14% of the juvenile 13 sexual offenders had experienced the loss of a parental figure through death, abandonment, or out-of-home placement (Ryan et al., 1996). In a small-scale comparison study (Kavoussi, Kaplan, & Becker, 1988), outpatient adolescent sexual offenders, grouped according to the gender and age of their sexual offense victims, were assessed for the presence of signs and symptoms of DSM-III psychiatric disorders. Structured interview instruments, discussions with family members, and interviews with each child?s referral source (e.g., probation officer, social worker, etc.) were utilized. Kavoussi, et al. (1988) chose to document relevant signs and symptoms of psychiatric disorder, rather than attempt to provide mental disorder diagnoses that might have proven unreliable based on the little amount of time spent with each offender during the assessment (Kavoussi, Kaplan, & Becker, 1988). The sample was comprised largely of individuals who featured an unremarkable history of other violent and non-violent crimes, and had only been implicated in the commission of one sexual offense during their lifetime. In addition, African-Americans (61%) and Hispanic participants (28%) were over-represented in the sample (Kavoussi et al., 1988). Conduct Disorder (48.3%) and Cannabis Abuse (10.3%) were the only DSM-III disorders to be diagnosed in more than 10% of the sample of juvenile sexual offenders. However, many of the youths featured at least one symptom of the mental disorders assessed: 67% featured at least one sign or symptom of Conduct Disorder, 34.5% had at least one sign or symptom of Attention Deficit/Hyperactivity Disorder, and 20.7% of the sample had at least one sign or symptom of an Adjustment Disorder with Depressed Mood. In comparing individual juvenile sexual offenders who had raped adult women to those who committed another type of sexual crime (e.g., child molestation of a young boy or girl, 14 frotteurism, voyeurism), 75% of the small rapist sub-sample met the diagnostic criteria for Conduct Disorder, while only 38% of the other sexual offenders were diagnosed with the disorder (Kavoussi et al., 1988). Hsu and Starzynski (1990) also found high rates of Conduct Disorder in small samples of both adolescent rapists (victims were at least 12 years old) and adolescent child sexual assaulters (victims were 11 years old or younger). Juvenile delinquents have also been compared on their level of externalizing and internalizing behaviors, and results have varied based on the informants used and the research setting in which information was collected (e.g., Blaske, Borduin, Henggeler, & Mann, 1989; Kempton & Forehand, 1992). Small samples of incarcerated juvenile delinquents were grouped according to whether they had committed sexual offenses only, sexual offenses and other confrontational offenses, confrontational offenses only, or neither type of offenses. Teachers provided ratings using the Teacher Report Form of the Child Behavior Checklist. The sexual offender only group was judged to be less aggressive and feature fewer internalizing symptoms consistent with anxiety and depressive constructs compared to the other groups of juvenile delinquents (Kempton & Forehand, 1992). This result was entirely inconsistent with a similar examination in which Blaske et al. (1989) found their sexual offending group of juveniles had more severe and frequent symptoms of anxiety compared to the other groups of juvenile delinquents. The sexual offenders also exhibited a constellation of ruminative/paranoid symptomatology that statistically exceeded the degree of symptoms endorsed by the assaultive offenders. The Blaske et al. (1989) study compared demographically matched and well defined groups of juvenile sexual offenders, juvenile assaultive offenders, juvenile 15 nonviolent offenders, and juvenile non-delinquent controls who lived in father-absent homes across a range of individual, family, and peer group characteristics. This well- designed study featured reports from mothers and teachers, and also asked each youth to fill out relevant self-report measures. The use of appropriate non-violent, violent, and non-delinquent comparison groups allows one to determine, with greater confidence, whether information is linked to sexual offending, violent offending, or general delinquency in juveniles. However, similar to Kempton and Forehand (1992), the authors concluded that the group of sexual offenders appeared most similar to the non-delinquent control groups in individual, family, and peer functioning. Nonetheless, sexual offenders exhibited symptoms consistent with anxiety and a feeling of estrangement from others. Compared to the other groups of juvenile delinquents, child molesters choose their own friends and acquaintances, rather than being pursued by others. Violent offender groupa, however, have endorsed a preference to approach others and solicit their involvement in both delinquent and non-delinquent activities (Ford & Linney, 1995). Findings pertaining to the psychiatric and psychosocial functioning of juvenile sexual offenders differ when standardized and well-established self-report measures are utilized (Davis & Leitenberg, 1987). Becker, Kaplan, Tenke, and Tartaglini (1991), in a sample comprised largely of African-American and Hispanic juvenile sexual offenders being treated on an outpatient basis, found Beck Depression Inventory (BDI) scores indicative of mild depression. The sample was then differentiated based on their histories of sexual or physical abuse. Approximately 49% of abused juvenile sexual offenders featured BDI scores indicative of moderate to severe depression, while only about 39% of the non-abused sub-group featured similar scores (Becker et al., 1991). 16 Utilizing the Minnesota Multiphasic Personality Inventory (MMPI), small but well-defined groups of adolescent rapists (forced vaginal sex with victim(s)), sodomists (engaged in or attempted anal intercourse with victim), and sexual abusers (engaging in another type of sexual offense) were compared (Herkov, Gynther, Thomas, & Myers, 1996). When sexual offenders featured more than one type of sexual offense they were included in the group that was deemed inclusive of the most serious type of sexual offense, with the researchers ranking Sodomy, Rape, and Sexual Abuse in order from most serious to least serious. These subgroups and juveniles committed to an inpatient psychiatric facility for serious psychological and/or psychiatric problems were then compared (Herkov, et al., 1996). The Sodomy and Rapist subgroups featured higher scale elevations, and two point codes most often associated with serious psychopathology. More specifically, the Rape and Sodomy groups were most likely to feature scores reflecting social alienation, anger problems, and difficulties with reality testing (Herkov et al., 1996). Katz (1990) compared groups of adolescent child molesters, non-sexual juvenile delinquents, and a control group of adolescents from a local high school on their psychosocial functioning. The group of adolescent child molesters were more lonely, socially isolated, and anxious compared to the group of high school students. The child molesters also were shown to be less confident in hetero-social situations compared to the group of high school students. The strength of these findings is highlighted by the use of a more homogenous group of juvenile sexual offenders and two well-defined control groups (Katz, 1990). 17 History of Sexual and Physical Abuse Victimization In another study featuring a small participant sample of adjudicated juvenile offenders committed to a state youth services treatment facility, Brannon, Larson, and Doggett (1989) compared a heterogeneous sample of non-sexual offenders that included both status offenders and property offenders to a small group of individuals charged with sexual offenses. The sample of sexual offenders was not described further on the nature of the sexual offenses they perpetrated. As a whole, 70% of the sample of juvenile offenders featured a history of sexual abuse or child molestation. The perpetrators of the abuse were most commonly known to the offender, and the abuse usually occurred when the juveniles were less than 10 years of age and was of a very serious nature (e.g., sodomy and fellatio as opposed to fondling). Differences in the sexual victimization histories of juvenile sexual offenders and other types of juvenile delinquents were not discussed, and Brannon, et al. (1989) failed to describe the manner by which inquiries were made about an offender?s sexual offense history. Aljazireh (1993) explained that studies featuring very high prevalence rates of sexual abuse victimization often feature similar methodological weaknesses that lessen the degree of confidence researchers can place in findings. Other studies have found that only 11% of juvenile sexual offenders report a past history of sexual abuse (Smith & Monastersky, 1986). Studies that fail to use control groups of non-juvenile delinquents and/or non- sexually offending juvenile delinquents do not allow for any comparisons regarding the relationship between sexual victimization history and the development of specific types of juvenile delinquent behavior. In a comparison of a heterogeneous sample of court- referred adolescent sexual offenders and violent non-sexual juvenile delinquents, the 18 juvenile sexual offenders were more likely to come from families where they witnessed domestic violence, experienced physical abuse, or were sexually abused (Fagan & Wexler, 1988). Without comparisons among varying typologies of juvenile sexual offenders, no associations can be drawn between the nature and duration of sexual victimization and the offender?s future sexual offending behavior (Aljazireh, 1993). Benoit and Kennedy (1992) randomly selected 25 individuals per comparison group from the entire population of youths committed to a state-run juvenile detention facility. Groups formed included those deemed non-aggressive offenders, aggressive offenders; juvenile sexual offenders who chose only young, female victims; and juvenile sexual offenders who featured a history of abusing only male children or both male and female children. All of the offenders shared a common history of property offenses in their juvenile records. The authors found no differences among the groups in their incidence of physical or sexual abuse victimization. In fact, even the group of juvenile sexual offenders who featured the highest number of sexual abuse victims (the male/female molestation group) was no more likely to be sexually abused than the other groups. In fact, all four groups of incarcerated juvenile delinquents featured high rates of both physical and sexual abuse. The authors suggested that the link between abuse history and delinquent offending in general is an indirect one moderated by other factors (Benoit & Kennedy, 1992). Studies of adult sexual offenders indicate that the incidence of sexual offense victimization in child molesters has shown to be twice as high as that in a rapist sample (Seghorn, Prentky, & Boucher, 1987). However, Worling (1995) used homogeneous samples of adult or juvenile sexual offenders who featured incidence rates of sexual 19 abuse victimization that varied based on whether data were collected before (22%) or subsequent to treatment (52%). Studies comparing the sexual abuse histories of juvenile sexual offenders featuring victims of varying genders and ages run the risk of confounding the two variables, as offenses against children are likely to involve male or female victims while offenses against peers and adults almost exclusively involve female victims (Worling, 1995). In assessing an outpatient sample of juvenile sexual offenders after becoming engaged in regular clinical interactions, Worling (1995) found that 25% of offenders who victimized females and 75% of offenders who had victimized at least one male child featured a history of sexual abuse victimization. However, there were no significant differences in sexual victimization incidence between offenders who victimized female children and those who victimized female peers or adults. There also were no differences between youths who victimized only male children and those who victimized both male and female children (Worling, 1995). Becker and Stein (1991) obtained similar results in that 32% of their sample of juvenile sexual offenders who assaulted male children and only 18% of juvenile sexual offenders who assaulted only female children had endorsed a history of previous sexual abuse victimization. Worling?s (1995) result is also consistent with Hanson and Slater?s (1988) study featuring adult sexual offenders, where 39% of sexual offenders against male children and only 18% of offenders against female children reported a history of sexual abuse victimization. These findings are particularly salient in light of some studies that tentatively suggested a history of sexual abuse and choosing at least one male victim being correlated with deviant sexual arousal patterns (e.g., Murphy, DiLillo, Haynes, & Steere, 2001). 20 Ford and Linney (1995), examining a recently adjudicated sample of juvenile delinquents, compared juvenile sexual offenders who assaulted peer-aged victims (juvenile rapists), juvenile sexual offenders who assaulted victims at least 5 years younger than themselves (juvenile child molesters), violent offenders (e.g., charged with assault, involuntary manslaughter, robbery, etc.), and status offenders (e.g., runaways, truancy charges). Juvenile child molesters were significantly more likely to have been sexually victimized and to have witnessed or experienced domestic and physical violence in the home. Juvenile child molesters feature higher rates of sexual abuse victimization compared to non-sexual juvenile delinquents and adolescent sexual assaulters. Studies have not often shown incidence differences in sexual abuse victimization between adolescent sexual assaulters and non-sexual juvenile delinquents (e.g., Awad & Saunders, 1991). Awad and Saunders (1991) showed that their child molester group had been exposed more frequently and earlier to pornographic materials compared to adolescent sexual assaulters. They were also more likely to report learning about sexual issues from watching other people in comparison to the other delinquent groups. In reference to concerns about confounding age and sex of victim when sub-typing juvenile sexual offenders (Worling, 1995), these authors did not provide information on the percentages of child rapists and molesters who chose male or female victims. The incidence of physical abuse present in heterogeneous samples of juvenile sexual offenders has generally been gauged to be approximately 15% (Becker et al., 1988; Fehrenbach et al., 1986; Smith, 1988). However, other studies have found higher rates of physical abuse victimization in samples of juvenile sexual offenders being treated 21 on an outpatient basis (47%; Kahn & Chambers, 1991) and those housed in a secure unit for violent offenders (76.5%; Lewis, Shanok, & Pincus, 1979). Divergence in how researchers define the physical abuse criterion, in tandem with difficulties associated with obtaining accurate retrospective reports from adolescents, likely lead to inconsistency in research findings regarding the incidence of physical abuse victimization in juvenile sexual offenders and juvenile delinquents in general (Aljazireh, 1993). Using delinquent and non-delinquent control groups will likely allow for the positing of more specific hypotheses about the relationship between physical abuse history and specific types of juvenile delinquency, in general, and juvenile sexual offending in particular. Sexual Offense Related Variables In one study, juvenile sexual offenders were shown to present at intake with around 7-8 victims, with data collected from a wide range of sources including outpatient treatment programs and secure residential facilities (Ryan et al., 1996). Many of the referring offenses involved ?penetrating and/or oral/genital contact? (Ryan et al., 1996, p. 19), and the wide majority of offenders admitted to committing the wide range of sexual offenses for which they were referred (Ryan, et al., 1996). Large-scale descriptive studies have been consistent in noting that approximately 60% of a heterogeneous sample of juvenile sexual offenders offended against children under the age of 12 (e.g., Awad & Saunders, 1991; Fehrenbach et al., 1986). Other studies have shown that as many as 90% of the victims of juvenile sexual offenders are between the ages of 3 and 16 years of age (e.g., Ryan et al., 1996). However, Kahn and Chambers (1991) noted that 95% of juvenile sexual offender victims were between the 22 ages of 1 and 18, while around 35% of victims were between the ages of 3 and 4. The high percentage of child victims observed by Kahn and Chambers (1991) is not consistently observed in the literature (Davis & Leitenberg, 1987). Participants in this analysis were receiving outpatient or residential treatment. Adjudicated and incarcerated juvenile sexual offenders typically have lower rates of offending against young children. In fact, 28% of incarcerated juvenile sexual offenders chose victims between the ages of 13 and 17 (peer-aged victims), while another 32% chose victims that were adults between the ages of 18 and 35 (Fagan & Wexler, 1988). The majority of studies of juvenile sexual offenders to date have found that the majority of victims are females (Awad & Saunders, 1991; Becker et al., 1986; Groth, 1977; Hsu & Starzynski, 1990; Kahn & Chambers, 1991; Ryan et al., 1996; VanNess, 1984). However, when the victims of juvenile sexual offenders happen to be male, a very high proportion of these victims tend to be young children (Groth, 1977; Hsu & Starzynski, 1990; VanNess, 1984; Worling, 1995). A large scale descriptive study of juvenile sexual offenders that featured over 1600 participants found that approximately 39% of the victims of juvenile sexual offenders were relatives who lived in the same home, 10% of victims were peers, 6% were strangers, and only 4.5% of victims were adults (Ryan et al., 1996). Juvenile sexual offenders committing crimes of indecent liberties or rape involving children have been shown to choose victims that are either relatives or acquaintances, whereas juveniles who rape adults are far more likely to choose strangers for victims (Fehrenbach et al., 1986). In comparing small samples of adjudicated juvenile sexual offenders who offended against female victims at or above the age of 12 (adolescent rapists) with those 23 offending against male and female victims 11 years or younger (adolescent child sexual assaulters), a multitude of sexual offense-related distinctions are evident (Hsu & Starzynski, 1990). Adolescent rapists are more likely to offend against victims with at least one other assailant, and to use violence during the commission of their sexual offenses (Becker et al., 1986; Hsu & Starzynski, 1990). Interestingly, in comparing individuals who acted alone in offending against a female stranger, all seven of the adolescent rapists committed their offenses indoors, while all eight of the child sexual assault offenses were committed outdoors (Hsu & Starzynski, 1990). Although the difference was not statistically significant due to small sample sizes, the adolescent rapist group featured a higher incidence of alcohol and/or marijuana use prior to the offense (53%) compared to adolescent child sexual assaulters (24%) (Hsu & Starzynski, 1990). However, other studies have shown that alcohol and substance use immediately prior to committing sexual offenses has been self-reported in over 50% of some samples of juvenile sexual offenders (VanNess, 1984). Yet, it is unclear whether these high rates actually occurred or were reported to escape some responsibility for the offenses they committed (Davis & Leitenberg, 1987). Some large-scale descriptive studies have found that as few as 6% of a heterogeneous sample of juvenile sexual offenders had used alcohol and/or drugs at the time of their offense(s) (e.g., Fehrenbach et al., 1986). An investigation of 63 adolescents incarcerated for perpetrating a total of 67 rapes against female women found that the participants typically featured a range of prior arrests, committed their rapes during the evening on weekends, and commonly victimized others with another assailant (Vinogradov et al., 1988). In fact, many of the adolescent 24 rapists in this sample had perpetrated their offenses as part of completing another criminal offense (e.g., burglary), had used alcohol or drugs immediately prior to the commission of the rape (15%), committed their offense against individuals who they did not know of the same race, and perceived their victims as having done nothing to provoke them into perpetrating their offenses (88%) (Vinogradov et al., 1988). Some research has also suggested that juvenile child molesters, in comparison to juvenile child assaulters, are more likely to know their victims and offend against more than one victim (e.g., Awad & Saunders, 1991). In addition, child molesters have also been found to be less likely to use violence during the commissions of their offenses or to threaten their victims during the sexual offense (Awad & Saunders, 1991). Adolescent sexual offenders with a past history of sexual abuse victimization tend to be more likely to perpetrate sexually against at least some males. Perpetrators who were themselves sexual abuse survivors were shown to select victims that were younger than offenders who featured no history of sexual abuse victimization: incarcerated juvenile sexual offenders? ?methods of gaining victims? trust, ensuring compliance with sexual acts, and maintaining victims? silence were found to vary based upon victim, offender, and offense characteristics? (Kaufman & Hilliker, 1996, p. 22). Studies that sub-categorize juvenile sexual offenders add to the richness of clinical information available on these youths. They highlight the necessity of conducting research on sub-categories of juvenile sexual offenders. Grouping juvenile sexual offenders into homogenous groups confounds most empirical studies intended to inform the treatment that these youths receive to reduce their risk of sexual offense recidivism. 25 Summary Research findings outlining the characteristics of juvenile delinquents, and juvenile sexual offenders in particular, vary based on a number of factors. Sub-samples tend to be differentiated in varying ways and criterions are often defined in a subjective fashion. The nature of differences among groups of juvenile sexual offenders and non- sexually offending juveniles vary based on where samples are obtained (i.e., detention centers, outpatient clinics), and the nature and duration of the participants? past criminal histories. Rarely do studies distinguish groups of juvenile sexual offenders on important criterions based on victim characteristics. The importance of describing participant samples and sub-samples in great detail is obvious in examining the literature comparing juvenile sexual offenders to non-sexually offending juvenile delinquents. Little research to date has been conducted in which juvenile sexual offenders are broken into groups and compared to each other, and to non-sexually offending juvenile delinquents. The Present Study The present study seeks to extend previous studies that have examined characteristic differences between juvenile delinquents with a documented history of sexual offenses and juvenile delinquents without a documented or reported history of sexual offending behavior (Becker et al., 1986; Blaske et al., 1989; Brannon et al., 1989; Caputo et al., 1999; Fagan & Wexler, 1988; Katz, 1990; Kemptom & Forehand, 1992; Oliver et al., 1993; Otnow-Lewis et al., 1979; Spaccarelli et al., 1993). Our interest also lies in comparing different groups of juvenile sexual offenders. Many of the studies comparing juvenile sexual offenders have featured small sample sizes, and lumped 26 heterogeneous groups of both juvenile sexual offenders and non-sexual offenders into artificially homogenous subsets. The present study seeks to determine how the age and gender of the victims chosen by juvenile sexual offenders can serve as one useful categorization variable in 2-way Analyses of Variance (ANOVA) to inform specific types of treatment with specific types of juvenile sexual offenders. Further, the manner by which the levels of our variable are defined will allow for a comparison between juvenile sexual offenders and non-sexually offending juvenile delinquents (Those with no documented history of sexual offense victims will serve as one of five levels of this Victim Choice variable). It is these types of distinctions that have, after all, that have been shown to provide more useful information in working with heterogeneous sets of juvenile sexual offenders (Awad & Saunders, 1991; Benoit & Kennedy, 1992; Ford & Linney, 1995; Herkov et al., 1996; Hsu & Starzynski, 1990; Kaufman & Hilliker, 1996; Kavoussi et al., 1988; Smith, 1988; Soothill, Francis, Sanderson, & Ackerly, 2000; Worling, 1995) In reviewing the literature with adult criminals, the construct of psychopathy (Hare, 1991) is a robust predictor of sexual offense recidivism and general criminal recidivism across a range of studies (e.g., Hare, 1991; Hildebrand, de Ruiter, & de Vogel, 2004). Studies have also indicated that psychopathy can serve as a useful sub-grouping variable to use in predicting the success of specific treatment modalities (Seto & Barbaree, 1999). As a new measure of psychopathy has been developed for use with juveniles (Psychopahty Checklist Youths Version: PCL:YV; Forth, Kosson, Hare, 2003), the construct of psychopathy has also shown to be a robust predictor of conduct problems in children (Frick, O?Brien, Wootton, & McBurnett, 1994), and general and violent 27 criminal recidivism in adolescents (Catchpole & Gretton, 2003). Encouraging results have been obtained in initial attempts to utilize this measure as a potential predictor of sexual offense recidivism for juvenile sexual offenders (Gretton, McBride, Hare, Shaughnessy, & Kumka, 2001). Past studies have shown that the Total Score on the PCL:YV (Gretton, et al., 2001) is useful in differentiating among groups of juvenile delinquents. Thus, the Total Score on the PCL:YV (Psychopathy Total Score) will serve as the second variable in our first 2-way ANOVA. Although the Callous/Unemotional Factor Score has been shown to correlate .7 with the PCL:YV Total Score (Forth, Kosson, & Hare, 2003), there is some research to suggest that this variable could have utility in distinguishing among groups of juvenile sexual offenders (Caputo et al., 1999). Thus, this Callous/Unemotional Factor Score will also be analyzed in a 2-way ANOVA, with our categorization based on the age and gender of sexual offenses victims again serving as the second variable. The results of these analyses will only be reported if they are found to provide information that is in excess of those obtained from analyzing the Total Score on the PCL:YV (Callous/Unemotional Score). No hypotheses are offered for this variable. On the basis of previous literature pertaining to adult and juvenile sexual offenders, a variety of hypotheses can be posited. However, it is important to note that the present study is largely exploratory. There have been no studies conducted in which juvenile sexual offenders are grouped based on their level of psychopathy and their choice of victims (based on age and gender). Thus, offering any hypotheses on potential interaction effects would represent mere speculation. It is hoped, however, that the nature of the interaction effects observed in this large sample study will prove highly beneficial 28 to future researchers. Only main effect hypotheses are offered below. For a complete listing of dependent variables used in these analyses, please see Table I. Further, a series of Pearson Chi-Square analyses will also be run for a select number of categorical dependent variables (see Table II). No hypotheses are offered on these exploratory analyses. 29 Table I Dependent variables examined through 2-way Analyses of Variance (ANOVA) (continuous dependent variables) Demographics/Family/Peer Variables Age in Months Number of self-reported friends Number of marriages for biological father Number of marriages for biological mother Number of biological siblings Number of step-siblings Number of half-siblings Number of biological siblings living in the home Number of step-siblings living in the home Number of half-siblings living in the home Number of total people living in the home prior to incarceration Wechsler Intelligence Scales: Full Scale IQ Score Wechsler Intelligence Scale: Verbal IQ Score Wechsler Intelligence Scale: Performance IQ Score Criminal History/Juvenile Delinquency Variables Number of lifetime school suspensions (Out-of-school suspensions) Number of self-reported fights in the last year Number of self-reported fights in the last three years Number of past juvenile delinquency commitments 30 Table I (continued) Dependent variables examined through 2-way Analyses of Variance (ANOVA) Criminal History/Juvenile Delinquency Variables (continued) Number of past arrests Number of past adjudicated sexual offenses Total number of past sexual offenses committed All Jesness Inventory (JI) Personality Scale Scores and Subype Scale Scores Personality/Psychiatric History Variables Total number of inpatient psychiatric commitments All Millon Adolescent Clinical Inventory (MACI) Scale Scores Total Score on the Reynolds Adolescent Depression Scale (RADS) Violence Exposure Across Contexts All Scales from the Screen for Adolescent Violence Exposure (SAVE) Alcohol and Substance Abuse Histories Number of average self-reported days out of each week that participants used alcohol Number of average self-reported days out of each week that participants used illicit drugs All Substance Abuse Subtle Screening Inventory: Second Edition (SASSI2) Scores Sexual Offending Variables All Multiphasic Sex Inventory (MSI) Scale Scores 31 Table II Dependent variables examined through Pearson Chi-Square Analyses (categorical dependent variables) Demographics/Family/Peer Variables Racial Background (4 levels) 1 = Caucasian; 2 = African-American; 3 = Hispanic; 4 = Other Abuse History Variables History of Sexual Abuse Victimization (2 levels) 1 = Yes; 2 = No History of Physical Abuse Victimization (2 levels) 1 = Yes; 2 = No History of Neglect 1 = Yes; 2 = No Personality/Psychiatric History Variables Past History of Previous Psychological/Psychiatric Treatment 1 = Yes; 2 = No Past History of Being Prescribed Psychotropic Medications 1 = Yes; 2 = No Current Prescription for Psychoptropic Medications 1 = Yes; 2 = No 32 Table II (continued) Dependent variables examined through Pearson Chi-Square Analyses (categorical dependent variables) Sexual Offending Variables Location in which index sexual offense took place (8 levels) 0 = No information; 1 = Offender?s Residence; 2 = Victim?s Residence; 3 = Other Residence; 4 = Non-Residential Building; 5 = Outdoors; 6 = Victim and Offender?s Residence; 7 = None of the Above Physical Intrusiveness of index sexual offense (9 levels) 1 = Fondling; 2 = Penetration; 3 = Providing Oral Sex; 4 = Receiving Oral Sex; 5 = Fondling and Providing Oral Sex; 6 = Fondling and Receiving Oral Sex; 7 = Fondling and Penetration; 8 = Penetration and Providing Oral Sex; 9 = Penetration and Receiving Oral Sex Was the offender under the influence of alcohol or drugs at the time of their any sexual offenses (2 levels) 1 = Yes; 2 = No Were there any precautionary actions taken by the offender once the offense was committed (e.g., removing evidence, creating an alibi, threatening the victim) (2 levels) 1 = Yes; 2 = No 33 Hypotheses Demographic/Family Variables 1. It is hypothesized that a main effect difference will be observed for the Victim Choice variable, in that individuals with no history of sexual offending behavior will be older than all of the groups of our sexual offenders. We also expect that juvenile sexual offenders who offend against peer-aged or older victims and select female victims only will be older than juvenile sexual offenders who have offended against at least one victim younger than themselves by four years or more, whether they have chosen male, female, or a mixed pattern of victims. 2. A second main effect difference is hypothesized for the Psychopathy Total Score variable where individuals with high total scores on the PCL:YV (>25) will be older than individuals scoring low on the construct (<25). 3. It is hypothesized that a main effect difference will be observed in that groups of juvenile sexual offenders who have selected at least one victim that was younger than themselves by 4 years or more, whether the victim(s) were male, female, or mixed, will feature homes with the most people living in them. 4. Some studies have shown that juvenile sexual offenders of varying types possess deficits in self-confidence and other social skills that prevent them from establishing meaningful friendships or cause them to interpret themselves as being unskilled in these areas. Thus, it is hypothesized that non- sexual offending juvenile delinquents will self-report having the most friends. The group of juvenile sexual offenders who have offended against at least one 34 male victim and at least one victim younger than them by at least 4 years is expected to self-report the fewest number of friends. 5. Juvenile delinquents scoring high in psychopathy are likely to self-report having more friends than those scoring low in psychopathy. 6. Youths scoring low to moderate in psychopathy are expected to feature significantly higher Full Scale, Verbal, and Performance IQ scores on the Wechsler Intelligence Scales compared to those scoring high in psychopathy. Criminal History/Juvenile Delinquency Variables 7. Juvenile delinquents who have no documented history of sexual offending will feature the most arrests in comparison to our groups of juvenile sexual offenders. In addition, juvenile sexual offenders who have committed sexual offenses against only females who are peer-aged or older will feature more lifetime arrests than the other groups of juvenile sexual offenders. 8. A main effect difference is expected such that individuals scoring high in psychopathy will feature more lifetime arrests than individuals scoring low in psychopathy. 9. It is hypothesized that juvenile delinquents who were not charged with a sexual offense will feature the highest percentages of individuals who have committed violent crimes against persons. It is also hypothesized that our group of juvenile sexual offenders who have only committed sexual offenses against peer-aged or older female victims will be more likely to have committed a range of other types of violent crimes in comparison to the other groups of juvenile sexual offenders. Juvenile sexual offenders who have 35 committed at least one sexual offense against a male victim and at least one sexual offense against a victim younger than themselves by at least four years will feature the least remarkable histories of non-sexual violent offenses. 10. A main effect difference is expected such that individuals scoring high in psychopathy will be more likely to have committed a range of non-sexual violent offenses compared to juvenile delinquents scoring low in psychopathy. 11. It is hypothesized that juvenile delinquents scoring high in psychopathy will feature more fights in the last year, fights in the last 3 years, school suspensions, and lifetime juvenile delinquency commitments than those scoring low to moderate in psychopathy. 12. Juvenile delinquents without a history of sexual offending behavior are expected to feature more juvenile delinquency commitments, fights in the last year, fights in the last 3 years, and school suspensions than all of the groups of juvenile sexual offenders. 13. It is hypothesized that the groups featuring only female victims that are peer- aged and our non-sexual control groups will feature the highest scores on the following MACI scales: Unruly, Oppositional, Substance Abuse Proneness, Delinquent Predisposition, and Impulsive Propensity. 14. Main effect differences for our psychopathy factor are expected such that individuals scoring high in psychopathy will feature higher scores on the following MACI scales: Substance Abuse Proneness, Delinquent Predisposition, and Impulsive Propensity.. 36 15. A main effect difference is expected on the Victim Choice variable such that the non-sexual delinquents and the group of juvenile sexual offenders who report to have only offended against female peers will feature the highest scores on the following Jesness scales: Social Maladjustment, Manifest Aggression, and the Asocial Index. Individuals who have offended against at least one male victim and at least one victim younger than themselves by 4 years or more are expected to feature the highest scores on the following Jesness scales: Immaturity, Alienation, Withdrawal, Social Anxiety, Repression, and Denial. 16. It is hypothesized that a main effect difference will emerge on the psychopathy factor such that individuals scoring high in psychopathy will feature the highest scores on the following Jesnesss scales: Social Maladjustment, Manifest Aggression, and the Asocial Index. Abuse History and Violence Exposure 17. Juvenile sexual offenders who have only offended against peer-aged or older victims and the non-sexual offending juvenile delinquents will be most likely to have experienced the highest degrees of home, school, and neighborhood violence as gauged by the Screen for Adolescent Violence Exposure (SAVE). It is also expected that the group of juvenile sexual offenders who feature at least one victim who is younger than themselves by 4 years or more and one male sexual offense victim will feature the lowest levels of violence exposure across the home, neighborhood, and school contexts. 37 18. It is hypothesized that a main effect difference will emerge showing that juvenile delinquents scoring high in psychopathy will feature a more extensive history of exposure to violence across the home, school, and neighborhood contexts compared to those scoring low in psychopathy. 19. It is hypothesized that individuals featuring at least one male sexual offense victim, whether they are children, peer-aged, or adults, will feature the highest degrees of sexual abuse victimization Analyses examining group differences in histories of neglect are exploratory in nature, and thus no specific hypotheses are offered. In addition, it is unclear as to whether main effect differences should be expected to emerge for the psychopathy factor, as no research to date has examined the construct of neglect specifically in sufficient detail with a sample similar to this study. Personality Profiles and Psychiatric History 20. Main effect differences are expected such that juvenile sexual offenders featuring at least one male victim and one victim younger than the offender by at least four years will feature the highest scores on the following MACI scales: Introversion, Inhibited, Doleful, Submissive, Self-Devaluation, Peer Insecurity, Child Abuse, Anxious Feelings, Depressive Affect, and Suicidal Tendency. 21. It is hypothesized that the groups featuring only female victims that are peer- aged and the non-sexual control group will feature the highest scores on the following MACI scales: Social Insensitivity. 38 22. Main effect differences for the psychopathy factor are expected such that individuals scoring high in psychopathy will feature higher scores on the following MACI scales: Egotistic, Self-Demanding, Borderline Tendency, and Social Insensitivity. Analyses will also be conducted to determine whether any main or interaction effects will be observed across the groups in whether they have had previous contact with mental health professionals prior to being incarcerated. Further, any differences in the past or present use of prescribed psychotropic drugs, and previous inpatient hospitalizations for mental health related difficulties will be noted. Alcohol and Substance Abuse Histories It is hypothesized that all groups of juvenile delinquents categorized based on the nature of their committing offenses and choice of sexual offense victims will feature relatively high rates of alcohol and substance abuse as gauged by the MACI and SASSI2. However, there is little guidance in the literature for us to use in confidently asserting any potential group differences that might emerge. Thus, no specific hypotheses are offered Sexual Offense History/Sexual Offense Treatment Variables The hypotheses offered in this section only pertain to those individuals who have a documented history of sexual offending behavior. Analyses will be run to determine whether any main effect differences emerge based on the preferred location of sexual offenses and whether alcohol or drugs were used in the offense(s). There is little literature to examine in offering useful hypotheses, and thus these analyses are deemed exploratory in nature. 39 23. It is hypothesized that juvenile delinquents without a documented history of sexual offenses will feature lower scores on a wide range of MSI scale scores, because these scales are designed to assess for the presence of factors shown to be related to sexual offending behavior specifically. 24. Youths scoring high in psychopathy will feature higher scores on the following MSI scales: Treatment Attitudes; Cognitive Distortions/Immaturity; Justifications; Rape Scale; Bondage and Discipline; Sado-Masochism; and Sexual History. 40 METHOD Participants The present investigation was conducted through a state-funded program of research (The Accountability Based Juvenile Sexual Offender Assessment and Treatment Program) at the Mt. Meigs juvenile detention complex, a residential Department of Youths Services (DYS) facility in Alabama. Licensed clinical psychologists and clinical psychology graduate students from Auburn University, licensed social workers and social work graduate students from the University of Alabama, and administrators and employees of the Alabama Department of Youths Services collaborated on this ongoing project. All of the participants in this study were adjudicated delinquent by an Alabama county court and committed for varying sentence durations at the juvenile detention facility. The participants fell into one of two general categories. The first group included male adolescents charged and found guilty of committing a contact sexual offense in the State of Alabama. These youths were court-ordered to participate in am empirically- grounded sexual offender treatment program while incarcerated. The second group consisted of male adolescents charged and found guilty of committing a criminal offense that was not a sexual offense. These youths underwent treatment typically designed to assist them in managing their anger, learning to problem solve and control their impulses, and when applicable curbing their use of alcohol and illegal drugs. 41 Through the Accountability Based Juvenile Sexual Offender Assessment and Treatment Program, every youth convicted of a sexual offense and remanded to the Department of Youths Services is required to complete a 6.5 hour assessment protocol prior to beginning therapy for their sexual offending tendencies. The protocol includes a comprehensive clinical interview, two rating scales, one diagnostic interview, and nine self-report measures. The interview and measures were chosen based on a review of the empirical literature on juvenile sexual offender assessment and treatment. Given that the literature on juvenile sexual offenders is in its infancy, the protocol has been updated and revised on several occasions over the last four years. Each youth?s progress in group therapy is tracked throughout treatment. At the conclusion of the group therapy program, caseworkers and group therapists are queried about each youth?s strengths and weaknesses displayed during their commitment. A post-therapy assessment protocol is administered prior to release that includes a review of relevant file information, a comprehensive clinical interview, the administration of self-report measures, and monitoring each youth?s responses to questions constructed to provide insight into how they would respond in high-risk situations upon leaving the detention facility. Following each youth?s release, their aftercare involvement and future criminal activity is tracked through frequent contacts with mental health professionals and law enforcement officials throughout the State of Alabama. For the purposes of the current investigation, pre- therapy assessment data only will be presented. Three hundred and eighty-one incarcerated adolescent males convicted of a criminal offense in the State of Alabama served as participants. Their average age was 15.84 years (SD= 1.60 years; Range: 10.67 years to 19.17 years), and their mean grade 42 level was 8.91 (SD = 1.62; Range: No schooling to graduated high school/obtained GED). Those youths who obtained a GED were coded the same as youths who graduated high school: 13 years of formal education. Twenty-three participants were found to have met the complete DSM-IV criteria for Mental Retardation (6.0%), while another 63 (16.5%) exhibited signs and symptoms of intellectual limitations that were not severe enough to warrant a diagnosis of Mental Retardation without the collection of additional data on adaptive functioning, which could not be obtained at the time of assessment. Just over half of the sample (n = 198) was Caucasian (52.0%), and 45.7% (n = 148) were African-American. The remaining 2.3% of our sample was Hispanic (.5%) (n = 2) or of mixed racial descent (1.8%) (n = 7). Participants featured a median number of 3 criminal arrests (Range: 1 to 33) and a median number of one past juvenile delinquency commitment (Range: 0 to 33). Thus, it is clear that this sample of juvenile delinquents committed many criminal offenses in the past, and most had contact with law enforcement officials before their current incarceration. The median descriptive statistic is presented instead of the mean due to the extreme positive skew characteristic of the data pertaining to the variables measuring the number of past juvenile delinquency commitments and past criminal arrests. Table III documents the percentage of delinquents in our sample charged with particular types of past offenses. Table IV documents the index offenses that resulted in our sample of juvenile sexual offenders being incarcerated at the Mt. Meigs facility. Together, these data provide a picture of the sample as being a collection of serious juvenile delinquents with an extensive history of both status and contact offenses. 43 Table III Percentages of Sample Featuring a Past History of Specific Criminal Offenses The Criminal Offense Percentage of Sample Theft, Burglary, or Trespassing 36.2 % Probation Violation or Failure to Appear in Court 27.5 % Asault 17.5 % Drug-related charges (e.g., possession, manufacturing, selling) 14.1 % Disorderly Conduct 5.1% Resitng Arest 5.2 % Driving-related offenses (e.g., reckless driving, speeding, 4.4 % driving without a license) Weapon Posesion 5.4 % Badgering a Witness 3.8% Robbery, Armed Robbery, or Attempted Robbery 3.3 % Escape or Attempted Escape 3.0 % Assault on a police officer or obstructing a police officer 2.7 % Criminal Mischief 2.4 % Truancy 2.7 % Possession of Stolen Property 2.2 % Harsment 1.9 % Possession of Alcohol by a minor 1.6 % Destruction of Property 1.6 % Child in Need of Supervision (CHINS) 3.3 % 44 Table III (continued) Percentages of Sample Featuring a Past History of Specific Criminal Offenses The Criminal Offense Percentage of Sample Runaway 1.9 % Arson 1.9 % Criminal Trespas 1.% Contempt of Court .8% Murder .3% 45 Table IV Percentage of Juvenile Sexual Offenders Featuring specific index offenses Sexual Abuse, 1 st degree 27.7% Rape, 1 st degree 14.5% Sodomy, 1 st degree 13.5% Sexual Misconduct 10.6% Rape, 2 nd degree 6.5% Sexual Abuse, 2 nd degree 2.9% Sodomy, 2 nd degree 1.3% 46 Sexual offending youths featured a wide range of index sexual offenses that are defined according to the Alabama Center for Law and Civic Education (2002). Sexual Abuse, 1 st degree served as the index offense for 27.7% of our study participants. The charge can result when a person uses force to touch sexual or other body parts of another person for the purposes of sexual gratification. The charge also can be levied if an individual subjects a person who is physically or mentally incapacitated to sexual contact, or if someone older than 16 years of age subjects a person who is younger than 12 to sexual contact. Sexual Abuse, 2 nd degree is charged when an individual subjects another person to sexual contact when they are incapable of providing legal consent because of a factor other than being under the age of 16. Only 2.9% of our sample featured this uncommon index offense. About 14.5% of the participants were charged at intake with at least one count of Rape, 1 st degree. This charge is to be levied against a youths when: (a) an individual engages in sexual intercourse with another person using force or threat; or (b) when a youths engages in sexual intercourse with someone who is deemed incapable of providing consent due to their age (less than 12 years of age). Rape, 2 nd degree served as the committing offense for 6.5% of our study participants. Youths are charged with this offense when they are above the age of 16 and engage in sexual intercourse with a member of the opposite sex who is younger than 16 and older than 12 years of age. The offender is required to be at least 2 years older than the victim. Sodomy, 1 st degree served as the index offense for 13.5% of our study participants. State law specifies that this charge is filed when an individual uses force or threat to engage in sexual contact involving the mouth or anus of a person who is not the perpetrator?s spouse. Sodomy, 2 nd 47 degree is the charge when an individual who is at least 16 years of age engages in deviate sexual intercourse with someone who is younger than 16 and older than 12 years of age or incapable of providing consent by reason of mental incapacity. About 1% of our sample featured this index offense. Finally, 10.6% of our sample featured an index offense of Sexual Misconduct, a charge that can result from the commission of any of the aforementioned sexual offenses as well as other illegal sexual like frotteurism, placing obscene phone calls, or threatening to harm another person in a sexual manner. In reviewing the legal definitions for each of the sexual offenses perpetrated by our study participants, much overlap was observed among crimes. Thus, a legal charge is not commonly reflective of an exclusive category of offending behavior. Subsequent analyses on sexual offense related variables attend to the behaviors reported by either or both the perpetrator(s) and victim(s) of these sexual offenses, rather than unreliable legal terms. Materials/Measures Comprehensive Clinical Interview The pre-treatment clinical interview was a semi-structured protocol aimed at collecting historical information relevant to understanding the behavioral functioning and the residing environmental contexts of the juvenile sexual offender sample. Information derived included: relevant demographics, family history and adjustment issues, health screening issues, alcohol and drug use, educational and work history, abuse and trauma history (e.g., sexual abuse, physical abuse, and neglect), current stressors, detailed criminal history, history of psychological and psychiatric problems, and their history of sexual offending behavior. The interview was tailored in such a way as to allow the 48 clinician to ascertain those factors used to score the two rating scales in the assessment protocol; the diagnostic interview, and factors derived from the comprehensive clinical interview. The protocol ordinarily takes approximately 2-3 hours to complete. Rating Scales Hare Psychopathy Checklist: Youth Version (PCL:YV). The PCL:YV is a 20-item rating scale designed to assess personality traits or stable dispositions that are consistent with the development of a psychopathic personality pattern in adolescents (Forth, Kosson, & Hare, 2003). It is this pattern that has been linked to an increased likelihood of future criminal activity, the development of pronounced interpersonal deficits, and poor occupational and social functioning in adults (Hare, 1991). The measure is designed for males and females ranging in age from 12 to 18, and represents an extension of the Hare Psychopathy Checklist ? Revised version (Hare, 1991) that has been used to assess psychopathic tendencies in adults for over a decade. The measure offers individual items scored on a 3-point Likert-scale, a total score, and two important factor scores designed to gauge two distinct patterns of psychopathic personality development. The first factor, termed the Selfish, Callous, and Remorseless Use of Others Factor, gauges a collection of interpersonal and affective traits consistent with the construct of psychopathy in adults (Hare, 1991). The second factor score on the PCL:YV is termed the Chronically Unstable and Antisocial Lifestyle Factor. This factor score assesses how aimless, irresponsible, and impulsive individuals are during their daily lives. This rating scale requires the clinician to conduct a detailed clinical interview and thoroughly review multiple collateral sources of information across a range of life domains in order to rate items measuring the interpersonal, affective, and behavioral 49 features of a psychopathic personality pattern. Extensive training and expertise are required to score this measure (Forth, Kosson, & Hare, 2003). Psychometric data for the juvenile version of the Psychopathy Checklist suggests that the psychopathic personality pattern can be measured reliably in youths (Forth, Hart, & Hare, 1990). The PCL:YV has been shown to feature high internal consistency (Cronbach?s alpha indices ranged from .85 to .90), inter-rater reliability (.82 to .94) (Brandt, Kennedy, Patrick, & Curtin, 1997; Gretton et al., 2001), and single rater reliability (.90) (Catchpole & Gretton, 2003). This rating scale has also shown to be a good predictor of recidivism in general juvenile delinquent samples (Brandt et al., 1997), and with samples of juvenile sexual offenders (Gretton et al., 2001). Self-Report Measures Millon Adolescent Clinical Inventory (MACI). The MACI is a 160-item self- report inventory that assesses a wide range of adolescent personality characteristics and clinical syndromes (Millon, 1993). In fact, the measure features 31 scale scores gauging personality patterns, expressed concerns, clinical syndromes, and modifying indices. The measure has been normed for 13 to 19-year olds. The internal consistency of the MACI scales range from a low of .69 to a high of .90, and the median test-retest reliability coefficient for the scales was .82. The measure has also been shown to feature good convergent validity (Millon, 1993). The Jesness Inventory (JI). The JI is a 155-item self-report questionnaire designed to assess a multitude of traits, attitudes, and perceptions that are consistent with a criminal lifestyle (Jesness, 2002). The authors intended to create a measure that was predictive of future delinquency and antisocial behavior in adulthood, and have revised 50 the measure on several occasions since 1962 in order to reach these ends. The measure contains 10 personality scales and 9 subtype scales. The interpretation of scores on each of these scales allows clinicians to fit delinquents into typologies consisting of individuals that have been shown to respond best to varying types of treatment services to reduce subsequent criminal recidivism (Jesness, 2002). The internal consistency of the Jesness personality scales can be described as ranging from adequate to very good, except for the Immaturity scale, which features a Cronbach alpha indicative of low internal consistency (Jesness, 2002). The author suggested that this scale be interpreted with the utmost caution. The test-retest reliability of individual scales, in general, can be described as acceptable to quite good (Jesness, 2002). The reliability of JI sub-typing strategies can best be described as adequate. A median test-retest correlation coefficient of .65 was obtained for subtype scale scores after a one year follow-up (Jesness, 2002). Several studies have examined the construct validity of the JI by correlating specific personality and subtype scale scores with relevant measures of psychological functioning. The measure has also shown to have predictive validity, as anticipated subtype classifications (e.g., CFC, NA, AP) represented significant predictors of future probation referrals. The Screen for Adolescent Violence Exposure (SAVE). The SAVE is a 32-item self-report measure designed to assess the scope and severity of the individual adolescent?s exposure to traumatic violence, indirect violence, and physical/verbal Abuse across the home, school, and neighborhood contexts (Hastings & Kelley, 1997). The measure represents an improvement over other available measures of violence exposure in adolescents by including easy-to-read items, distinguishing the settings in which the 51 violence is experienced or witnessed, and taking pains to establish a psychometrically sound measure (Hastings & Kelley, 1997). The internal consistency of the measure is excellent, with setting scale alpha levels ranging from .90 to .94. During pilot work, the scales possessed adequate to good convergent, divergent, and construct validity (Hastings & Kelley, 1997). The Reynolds Adolescent Depression Scale (RADS). The RADS is a 30-item self- report measure designed to assess a wide range of depressive symptoms in youths ranging in age from 13 to 18. Using a 4-point Likert scale, individuals are assessed for the presence of cognitive, vegetative, somatic, and interpersonal type symptoms of depression. The measure has shown excellent internal consistency with alpha scores ranging from .91 to .94. Test-retest reliability has been estimated to fall in the .63 to .79 range. Convergent and concurrent validity indices have been shown to fall in the adequate to good range, while the construct and divergent validity of the measure was described by the author as adequate (Reynolds, 1987). The Multiphasic Sex Inventory (MSI). The MSI is a 300-item self-report inventory constructed to assess a range of psychosexual characteristics in juvenile sexual offenders. The measure features three validity scales, two accountability scales, three sexual deviance scales, five atypical sexual behavior scales, a sex knowledge scale, and a treatment attitudes scale (Nichols & Molinder, 2001). The measure asks youths to respond to questions in a True/False format. The MSI has been shown to feature good to excellent test-retest reliability. The convergent and concurrent validity of the measure, across a range of small sample studies conducted by the authors of the measure, can be assessed as good. Alpha coefficients ranged from .50 to .94 across four separate studies. 52 Alpha coefficients obtained using an adolescent sample exclusively ranged from .66 to .89 (Nichols & Molinder, 2001). The Substance Abuse Subtle Screening Inventory: Second Edition) (SASSI-2). The SASSI-2 is a 100-item self-report measure designed to assess for the presence of signs and symptoms characteristic of substance abuse and substance dependence. The measure represents an extension and revision of the original SASSI (Miller, Renn, & Lazowski, 1990), and authors report 94% overall accuracy rates in identifying substance abuse disorders in a large sample of adolescents from treatment programs and juvenile justice facilities (Miller, Renn, & Lazowski, 2001). No test-retest reliability or alpha coefficients were provided in the manual. Procedure Upon meeting with each youth initially, each was provided with a detailed assent form and information on the nature of the assessment that they would be completing (e.g., the purposes of the research, their rights as participants in psychological research). Youths then learned of the efforts taken by researchers to preserve their confidentiality by assigning them identification numbers to be used in place of any other information that could be used to identify their assessment materials in the future. The juvenile sexual offenders also were informed that they would complete an additional assessment protocol following their completion of the court-ordered sexual offender treatment program. Youths were free to withdraw from participation as research subjects or take a break from the intensive assessment at any time. Juvenile sexual offenders were told that completion of the pre-treatment assessment was a required part of their treatment. Prior to initiating the assessment, youths were encouraged to respond in an honest fashion, and researchers 53 were instructed to probe for inconsistencies in the youth?s report and information derived from a thorough review of relevant file material. When inconsistencies were observed, youths were questioned and instructed to provide further information to the examiner. Advanced graduate students in clinical psychology were responsible for conducting the clinical and diagnostic interviews. Undergraduate research assistants administered the self-report measures to each participant on a date different from that in which the interview was conducted. Approximately 66% of participants were interviewed prior to the administration of self-report measures, while the remaining 34% of the sample completed the self-report battery first. Due to the lengthy nature of the assessment protocol, no youths was asked to complete the entire assessment in one day. Those with the highest intellectual abilities and adequate attention spans were able to complete the entire protocol in two days (34.9%). The remainder of youths required three (49.9%) or four days (15.0%) to complete the entire protocol. All interviewers took part in a comprehensive training session prior to working with the detained youths. The program included advanced training in: building rapport with detained youths, basic interviewing skills, and scoring/coding self-report measures, interview questions, and rating scales. A licensed clinical psychologist with extensive experience in working with juvenile delinquents and juvenile sexual offenders was available if questions or problems arose. In addition, each week, all individuals working through the Accountability Based Juvenile Sexual Offender Assessment and Treatment team met for at least one hour to discuss their experiences and resolve any scoring discrepancies. Clinicians conducting the clinical and diagnostic interviews and scoring the rating scales frequently met with participants jointly, completed their scoring 54 independently, and then met to resolve any scoring discrepancies. These calibration exercises were employed to ensure that data were collected and scored in a reliable manner. Undergraduate assistants administering self-report measures received advanced training in: building rapport with detained youths, detecting reading and/or learning weaknesses that could affect the accuracy of the information collected, and assisting youths having difficulty understanding the nature of certain questions. Undergraduate assistants often read self-report questions to participants. Computer scoring was available for the Millon Adolescent Clinical Inventory (MACI) and Jesness Inventory (JI). The remaining self-report measures were scored manually by the undergraduate assistants. Detailed instruction sheets for scoring of these measures were constructed and provided to assistants. They also were instructed to check their scoring twice before documenting test scores for each youth. Graduate student clinicians checked the accuracy of our undergraduate assistants? scoring. Clinicians coded interview-based variables and test scores on a variable coding sheet prior to entering the information into the computer database. Random checks ensured that information was reliably coded from the interview and test protocols to the variable coding sheet, and from the variable coding sheet to the computer database. Study Design This juvenile delinquent sample was sub-categorized into groups of individuals who featured an index charge of a contact sexual offense and those juvenile delinquents who featured no documented or self-reported history of sexual offenses. The juvenile sexual offenders were further categorized based on the age(s) and gender(s) of their 55 victim(s). This operationally defined distinction extends upon research with adult and juvenile sexual offenders where age or gender of victim(s) only, defined in varying ways, served to sub-categorize heterogeneous samples of sexual offenders (Awad & Saunders, 1989; Awad & Saunders, 1991; Ford & Linney, 1995; Worling, 1995). Butz and Spaccarelli (1997) suggested that additional grouping variables be used, in tandem with victim characteristics, to shed light on historical, individual, and contextual differences within groups of juvenile sexual offenders, and between groups of juvenile delinquents. Thus, the present study features two main sets of analyses, featuring different independent variables. The first analysis features two factors; one factor is based on the gender and age of sexual offense victims (Victim Choice), and the other is based on the juvenile?s endorsement of behaviors consistent with a ?psychopathic personality pattern (Psychopathy Total Score on PCL:YV). The second set of analyses features two factors, our Victim Choice variable, and a variable based on the youth?s scores on the Callous/Unemotional Factor of the PCL:YV (Callous/Unemotional Factor Score). This second study design enabled the investigation of main effects for the Callous/Unemotional Factor, as well as any additional interaction effects. These results will only be reported if they provide information in excess of that obtained through examining main effect differences across the levels of the Psychopathy Total Score (PCL:YV Total Score). Considering that the Callous/Unemotional Factor score has been shown to be correlated .70 with the Total Score on the PCL:YV (Forth et al., 2003), it is unclear whether any additional information will gained through these analyses. The first factor for the first Two-Way Analysis of Variance (ANOVAS) featured 5 levels: (a) juvenile sexual offenders who only chose female victims and featured at least 56 one victim who was younger than the offender by 4 years or more (n = 149; 130 chose only female victims who were at least 4 years younger, 19 chose only female victims but offended against both peer-aged victims and those younger than themselves by 4 years or more); (b) juvenile sexual offenders who only chose female victims and only offended against victims who were peer-aged or older (n = 78); (c) juvenile sexual offenders who featured at least one male victim and at least one victim who was younger than the offender by 4 years or more (n = 73; 43 victimized only males at least 4 years younger, two victimized males who were both peer-aged and younger by 4 years or more, 15 offended against both males and females who were all younger than the offender by 4 years or more, 13 offended against both males and females who were both peer-aged or younger by 4 years or more); (d) juvenile sexual offenders who offended against at least one male victim and only offended against victims who were peer-aged or older (n = 17; 14 who offended against only males who were peer-aged or older, three who offended against both males and females who were peer-aged or older); and (e) incarcerated juvenile delinquents charged with a range of non-sexual offenses that featured no self- reported or documented history of sexual offending behavior (n = 58). Six youths featured incomplete data and were eliminated from subsequent analyses, leaving 375 for the participant sample pool. The second factor in the first Two-Way Analysis of Variance (ANOVA) featured two levels, and was based on each participant?s total score on the Psychopathy Checklist: Youth version (PCL:YV) (Forth et al., 2003). The score corresponding to the 75 th percentile in our sample was 24. Thus, individuals scoring 24 or higher on the PCL:YV were categorized high in psychopathy (n = 99), while individuals scoring below 57 25 were categorized as medium to low in psychopathy (n = 270). Seven individuals did not provide the data required to score the psychopathy measure and were eliminated from subsequent analyses. Some research suggests that the Callous/Unemotional Factor Score (Callous/Unemotional Factor Score) derived from the PCL:YV might prove particularly useful in differentiating groups of juvenile sexual offenders from non-sexually offending juvenile delinquents (Caputo, et al., 1999). Thus, the potential utility of this scale in sub- categorizing juvenile delinquents was examined in a second 2-way Analysis of Variance (ANOVA), with the Victim Choice variable again serving as the second independent variable. The Callous/Unemotional Factor Score featured two levels based on whether participants scored at or above (n = 113), or below (n = 256), the 75 th percentile of 8.. Dependent variables for this study included self-report and interview based measures of areas related to treating juvenile sexual offenders on an individualized basis in order to assist in reducing their risk of future sexual and non-sexual criminal recidivism. The areas measured included: Demographics/Family Life/Friends; History of Violence; Violence Exposure Across Life Contexts; Sexual Offending Variables; Psychiatric History/Personality Functioning; Abuse History; and Substance Abuse. When the dependent variable information represented nominal or ordinal level data, Pearson Chi-Square analyses were used to determine statistically significant differences across our aforementioned groups (Please see Table 1 and Table 2 for a complete listing of all dependent variables examined in the two studies). 58 RESULTS Significant results of two-way analyses of variance and chi-square nonparametric tests are presented below. The independent variables for the analyses presented below are: a) the categorical variable formed by examining the age and gender of any sexual offense victims (Victim Choice Variable; 5 levels); and b) the variable formed by categorizing participants based on the median total score of our sample on the Psychopathy Checklist: Juvenile version (PCL) (Psychopathy Total Score ; 2 levels) (Forth, Kosson, & Hart, 2003). An alpha level of .05 was used for all statistical tests. The Tukey post-hoc test was used to examine mean differences across the five levels of the Victim Choice variable when a significant main effect was observed. Demographics/Family Life/Friends Participants differed in their age based on the age and gender of their sexual offense victims (Victim Choice Variable) (F(4,369) =15.542, p < .0001). Tukey post-hoc testing revealed that non-sexual offending delinquents were older than every category of youths who had committed sexual offenses in the past. There were no age differences among the categories of sexual offenders. A chi square analysis determined that a significant association between the Psychopathy Total Score and a Race variable classification (Caucasian, African-American, Hispanic, Other) was present (? 2 (3) = 9.211, p < .028). Results suggested that more African-Americans than expected were classified in the high 59 psychopathy range of the Psychopathy Total Score categorization. A second chi-square analysis showed a significant association between the Race variable and the Victim Choice variable (X 2 (12) = 23.934, p < .021). Results suggests that a greater proportion than expected of individuals who committed sexual offenses against at least one male victim and one victim younger than the offender by four years or more were Caucasian. Further, a greater proportion than expected of youths without a history of sexual offenses and those youths who committed sexual offenses against peer-aged females exclusively were African-American. In addition, a higher than expected proportion of juvenile sexual offenders who targeted female victims, at least one of whom was younger than the offender by four years or more, were Caucasian. Youths rated low to moderate in psychopathic features (Psychopathy Total Score), had fathers (F(1,369) = 4.99, p < .03) and mothers (F(1,369) = 3.886, p < .05) who had significantly more marriages in their lifetime than those scoring high in psychopathy. There were no main effect differences observed for the number of biological or step-siblings that participants had, based on their Psychopathy Total Score or Victim Choice categorizations. However, a main effect difference was observed such that those scoring high in psychopathy had more half-siblings compared to those scoring low to moderate in psychopathy (F(1,369) = 4.379, p < .04). There were no main effect differences observed for number of biological siblings, half-siblings, or step-siblings living in the home. Yet, there was a main effect group difference for the number of people who lived in the home with the offenders prior to incarceration. Offenders differed based on their Victim Choice categorization, F(1,369) = 3.018, p < .021. Tukey post hoc testing revealed that youths featuring at least one male victim and offending exclusively 60 against peer-aged victims, lived with significantly more people in their home than youths who featured no documented or self-reported history of sexual offending behavior. For those youths in which IQ scores were available, main effect differences were observed. Youths scoring low to moderate in psychopathy featured a higher Full Scale IQ (F(1,243) = 6.606, p < .012), Verbal IQ (F(1,242) = 5.157, p< .024), and Performance IQ (F(1, 242) = 5.081, p <.026) than those scoring high in psychopathy. In addition, a main effect difference for the Victim Choice variable was observed when examining Performance IQ Scores (F(1,242) = 5.081, p < .028). Juvenile sexual offenders featuring a history of offending against at least one male victim and at least one victim younger than them by 4 years or more had a significantly higher Performance IQ Scores compared to those who committed sexual offenses against female peers only. In summary, juvenile delinquents rated low to moderate in psychopathy tended to have biological parents who were, on average, more likely to have multiple marriages. Those rated low to moderate in psychopathy also featured more advanced verbal, non- verbal, and overall intellectual abilities than those rated high in psychopathy. The high psychopathy group had more half-siblings than those rated low to moderate in psychopathy. Non-sexually offending delinquents were older than all groups of juvenile sexual offenders. They also featured fewer people living in their homes prior to incarceration compared to juvenile sexual offenders who victimized peer-aged victims exclusively, at least one of whom was male. A greater than expected proportion of non-sexually offending delinquents were of African-American descent. Juvenile sexual offenders who victimized children younger than themselves by 4 years or more, whether victims were 61 male or female, featured a greater than expected proportion of individuals who were Caucasian. Youths who sexually offended against at least one male and at least one victim younger than themselves by 4 years or more, had significantly higher non-verbal intellectual abilities compared to non-sexually offending delinquents. History of Abuse A series of chi-square analyses were run to determine whether a significant association existed between each of our independent variables (Victim Choice and Psychopathy Total Score), and our Sexual Abuse, Physical Abuse, and Neglect variables. Based on a comprehensive interview and a detailed review of relevant records, individuals were labeled as either featuring a history of the specified type of abuse or not. Thus each of our abuse variables featured only two levels. A significant chi square statistic was found when examining the association between our Neglect variable (2 levels) and our Psychopathy Total Score variable (2 levels) (X 2 (1) = 4.002, p < .045). A higher proportion than expected of youths scoring high in psychopathy had a documented history of neglect. Significant chi square analyses were discovered when examining the association between sexual abuse (2 levels) and the Victim Choice variable (5 levels) (X 2 (4) = 18.792, p < .017, and physical abuse (2 levels) and the Victim Choice variable (5 levels) (X 2 (4) = 11.636, p < .021). It was discovered that a significantly higher proportion of youths who sexually offended against at least one male victim and one victim younger than the offender by four years or more featured a history of sexual abuse victimization. In addition, a significantly lower proportion than expected of youths with no history of sexual offending and those offending against female peers exclusively featured a history 62 of sexual abuse victimization. Significantly more youths than expected who sexually offended against at least one victim younger than themselves by four years or more, whether victims were male and/or female, featured prominent histories of physical abuse victimization. In addition, youths with no documented sexual offenses in their records and those who offended against female peers exclusively had fewer youths than expected who had histories of physical abuse victimization. In general, youths scoring high in psychopathy featured a greater proportion of participants than expected who had a history of neglect. There were no differences in rates of sexual abuse victimization or physical abuse victimization across our levels of the Psychopathy Total Score variable. Juvenile sexual offenders who victimized at least one male and at least one victim younger than themselves by 4 years or more featured a significantly higher proportion of individuals who had histories of physical abuse victimization and sexual abuse victimization. Those juvenile sexual offenders who targeted females, at least one of whom was younger than the offender by 4 years or more, had a significantly higher than expected proportion of youths who had a history of physical abuse victimization. Non-sexually offending delinquents and juvenile sexual offenders targeting female peers exclusively featured a lower proportion of participants than expected who featured a history of sexual abuse victimization and physical abuse victimization. History of Violence/Criminal Offenses Individuals scoring high in psychopathy featured more school suspensions (F(1,367) = 13.812, p< .0001), more fights in the last year (F(1,368) = 14.534, p < .0001), and more fights in the last 3 years (F(1,368) = 8.117, p < .006) than those scoring 63 low to moderate in psychopathy. A significant main effect was also observed in examining Victim Choice variable group differences in number of lifetime school suspensions (F(4,367) = 2.544, p < .04). Tukey post hoc testing indicated that youths featuring no documented sexual offenses had significantly more school suspensions than individuals who had committed sexual offenses against exclusively female victims, one of whom was at least 4 years younger than the offender. A significant interaction effect was discovered when examining the number of school suspensions that youths reported (F(4,367) = 2.678, p < .033). Participants who had offended against at least one male victim and one victim younger than them by at least four years, and featuring high scores in psychopathy, featured the highest average number of school suspensions among all the cell means. There were no significant main effect differences in fights over the course of the last year, or over the course of the last three years, across the levels of the Victim Choice variable. Main effect differences were observed for the Victim Choice variable (F(4,369 = 15.471, p < .0001), and the Psychopathy Total Score variable (F(1,369) = 22.559, p < .0001) when examining mean differences in the number of juvenile delinquency commitments. Youths scoring high in psychopathy had significantly more juvenile delinquency commitments than those scoring low to moderate in psychopathy. Post hoc testing revealed that offenders who had no documented history of sexual offenses featured a significantly higher number of past juvenile delinquency commitments than all of our groups of juvenile sexual offenders. There were no differences across the sexual offenders in their number of past juvenile delinquency commitments. Main effect differences were found when examining the number of past arrests in the records of the 64 juvenile offenders. Youths scoring high in psychopathy had significantly more past arrests than those scoring low to moderate in this construct (F(1,368) = 34.582, p < .0001). In addition, differences in number of past arrests also emerged across the levels of the Victim Choice variable (F(4,368) = 38.582, p < .0001). Tukey post hoc results showed that youths without a history of sexual offending behavior had a significantly greater number of past arrests than all groups of juvenile sexual offenders. Analyses of Variance were run to determine if main effect differences were observed on all Personality and Subtype Scales of the Jesness Inventory (JI). Group differences that were statistically significant across the levels of the Victim Choice variable include the: Alienation Scale (F(4,345) = 4.572, p < .002); Asocial Index (F(4, 346) = 2.394, p < .051); Cultural Conformist/Group Oriented Subtype (F(4,346) = 4.358, p < .003); and Neurotic, Anxious/Introspective Subtype (F(4,346) = 5.522, p <.0001). Juveniles without a history of sexual offenses described being more hostile toward others, and distrustful and disobedient toward authority figures, than all the groups of juvenile sexual offenders except for the small sample group of youths who committed sexual offenses against peer-aged victim where at least one victim was male (JI - Alienation Scale). In addition, youths without a history of sexual offenses had significantly higher ratings on general maladjustment, distress, and psychological impairment compared to the groups of juvenile sexual offenders who offended against females, at least one of whom was younger than the offender by 4 years; and those who offended against at least one male victim and one victim younger than the offender by 4 years (JI: Asocial Index). In short, the non-sexual offenders featured a personality more 65 consistent with a delinquent lifestyle than juvenile sexual offenders who sexually offended against at least one young victim, regardless of the victim?s gender. Juvenile sexual offenders who offended against females, with at least one victim being younger than them by 4 years or more, and those who offended against at least one male and one victim younger by at least 4 years, featured a personality pattern characterized by low self-esteem, self-criticism, and worry compared to those who had no history of sexual offending behavior (JI - Neurotic, Anxious/Introspective Subtype Scale). In contrast, participants without a history of sexual offenses featured more characteristics consistent with being happy and content with their lives and unwilling or uninterested in changing their lifestyle compared to all groups of sexual offenders, excluding the small group of youths who sexually offended against peer-aged victims; at least one of whom was male (JI ? Cultural Conformist/Group Oriented Subtype Scale). Individuals rated high in psychopathy featured significantly higher scores than those scoring in the low to moderate range in psychopathy on the following Jesness scales: Social Maladjustment Scale (F(1,346) = 6.351, p < .013); Value Orientation Scale (F(1,346) = 8.435, p < .005); Autism Scale (F(1,345) = 4.124, p < .044); Manifest Aggression Scale (F(1,346) = 11.277, p <.002); Asocial Index (F(1,346) = 6.541, p <.012); Unsocialized, Aggressive/Undersocialized, Active subtype (F(1,346) = 12.327, p < .002); Cultural Conformist/Group Oriented subptype (F(1,346) = 5.299, p < .023); and Neurotic/Acting-Out/Autonomy Oriented subtype (F(1,346) = 11.494, p < .002). Individuals rated low to moderate in psychopathic features had significantly higher scores than those scoring high in psychopathy on the following scales: Immature Conformist/Conformist subtype (F(1,346) = 17.054, p < .0001); Situational Emotional 66 Reaction/Inhibited subtype (F(1,346) = 7.515, p < .007); Cultural Identifier/Adaptive subtype (F(1,346) = 3.895, p < .05). To summarize, individuals scoring high in psychopathy featured more school suspensions, more fights, more past juvenile delinquency commitments, and more arrests in their history compared to those scoring low to moderate in psychopathy. In short, youths scoring high in psychopathy had more prominent histories of past violence and trouble with authority figures than those scoring low to moderate in psychopathy. These youths were more likely to come from low socioeconomic backgrounds, and endorse values consistent with trying to survive in their lives through any means necessary. Youths judged high in psychopathy were more likely to feature three distinct types of personality patterns compared to those deemed low to moderate in psychopathy. First, they were more likely to feature a personality characterized by negative attitudes toward authority, family, and school. Their behavior was more unpredictable, nonconforming, and aggressive. In short, they were more likely to admit and take pride in possessing a hardened, delinquent-prone personality. Second, they were more likely to feature a personality pattern characterized by low motivation, low achievement, and negative attitudes toward societal conventions. This group had difficulty respecting authority figures, following rules, and making decisions independent of their delinquent peer group. Third, they were more likely to feature a personality pattern characterized by views of themselves as independent, and not requiring assistance through family members, school, or outside agencies of any kind. These individuals tend to be outspoken, provocative, and non-conforming. 67 The personalities of those judged low to moderate in psychopathy were more likely to be consistent with one of the following personality patterns, in comparison to those deemed high in psychopathy. First, they were more likely to have somewhat more positive attitudes toward family, school, and home. This group of individuals maintained a more positive self-concept, and tended to follow delinquent peers readily. Second, individuals scoring in the low to moderate range in psychopathy were more likely to feature personality structures characterized by a na?ve, confident, and conforming style of behavior. They have positive attitudes toward school and family, and tend to be more socially skilled when interacting with others. Third, they were more likely to feature personalities characterized by good verbal skills, motivation in school, confidence in interpersonal relationships, and relatively higher self-esteem. In short, those judged low to moderate in psychopathy, compared to those scoring high in psychopathy, showed a variety of characteristics consistent with a greater likelihood of treatment success (Jesness, 2002). Juveniles without a history of sexual offenses tended to have more indicators of past aggression and trouble with authority figures than all of our groups of sexual offenders. Non-sexually offending delinquents had more past arrests and juvenile delinquency commitments than all groups of juvenile sexual offenders, and more past school suspensions than juvenile sexual offenders who targeted females, at least one of who was younger than the offender by 4 years or more. Interestingly, those juvenile sexual offenders who victimized at least one male and at least one victim younger than themselves by 4 years or more, when rated high in psychopathy, featured the highest 68 mean number of past school suspensions compared to all other groups of juvenile delinquents distinguished across the levels of the Victim Choice variable. In examining significant results from the Jesness Inventory (JI), non-sexually offending juvenile delinquents featured significantly higher average scores on measures of distrust and disobedience of authority figures compared to all groups of juvenile sexual offenders. Also, these non-sexually offending delinquents described themselves as being more comfortable and content with their lives than groups of juvenile sexual offenders. These non-sexual offenders were also deemed to be more likely to feature characteristics of a delinquent personality structure when compared to juvenile sexual offenders who victimized at least one person younger than themselves by 4 years or more, whether the victims were male or female. Youths who sexually offended against victims younger than themselves by 4 years or more, regardless of the victim?s gender, were characterized as having less self-confidence and self-esteem than non-sexually offending delinquents. Differences in measures of delinquency were not found between non-sexually offending delinquents and juvenile sexual offenders who targeted peer-aged victims who were female. Violence Exposure Across Life Contexts The Screen for Adolescent Violence Exposure (SAVE) assesses the degree of Traumatic Violence, Indirect Violence, and Interpersonal Violence experienced by youths in the home, neighborhood, and school contexts. In examining Total Violence Exposure scores on the SAVE, two significant main effects emerged for the Victim Choice and Psychopathy Total Score variables. Youths scoring high in psychopathic personality traits has significantly higher Total Violence Exposure scores than those 69 having low to moderate levels of psychopathy (F(1,350) = 21.682, p < .0001). In addition, a main effect emerged on the Victim Choice variable (F(4,350) = 3.569, p < .008). Tukey post-hoc tests revealed that participants who had not committed a sexual offense featured significantly higher levels of total violence exposure than all the groups of sexual offenders. In general, participants rated high in psychopathy featured more exposure to varying types of violence across multiple life contexts. For instance, youths judged to be high in psychopathy featured higher scores than those deemed low in psychopathy on the following SAVE scales: Neighborhood Total Violence Scale (F(1,351) = 26.040, p < .0001), Neighborhood Interpersonal Violence Score (F(1,351) = 7.108, p < .009), Neighborhood Indirect Violence Score (F(1,351) = 20.308, p < .0001); Neighborhood Traumatic Violence Score (F(1,351) = 29.545, p < .0001); Home Total Violence Score (F(1,351) = 16.054, p < .0001); Home Indirect Violence Score (F(1,351) =14.249, p < .0001); Home Traumatic Violence Score (F(1,351) = 15.544, p < .0001); School Total Violence Score (F(1,351) = 7.552, p < .007); School Indirect Violence Score (F(1,351) = 6.652, p < .011); School Traumatic Violence Score (F(1,351) = 10.051, p < .003). Youths across the five levels of the Victim Choice variable differed in the level of total violence that they had been exposed to in their neighborhood contexts (F(4,351) = 3.971, p < .005), degree of Indirect violence in their neighborhood contexts (F(4,351) = 4.707, p < .002), and degree of traumatic violence in their neighborhood contexts (F(4,351) = 3.841, p < .006) . Tukey post-hoc testing revealed that for all neighborhood violence exposure variables, participants with no history of sexual offenses featured significantly higher scores than all of the sexual offender groups. 70 A main effect was also obtained for the Victim Choice variable pertaining to the degree of one specific type of violence exposure in the home context; Traumatic Violence (F(4,351) = 2.786, p < .028) . The Tukey post-hoc test revealed that youths without a history of sexual offenses reported significantly more exposure to traumatic violence at home than sexual offenders who victimized females, at least one of whom was younger than the offender by at least 4 years. For the school context violence exposure variables, main effects were obtained on the Victim Choice variable for The School Violence Total Score (F(4,351) = 3.391, p < .011), the School Interpersonal Violence Score (F(4,351) = 2.442, p < .048), the School Indirect Violence Score (F(4,351) = 4.030, p < .004), and the School Traumatic Violence Score (F(4,351) = 2.506, p < .043). Post hoc testing revealed specific differences across groups for each of the types of violence that participants were exposed to in the school context. Youths who had no history of sexual offending featured more total exposure to violence in school than: sexual offenders who offended against peer-aged youths with at least one male victim; sexual offenders who victimized females who were peer-aged, and sexual offenders who offended against female victims, one of whom was at least 4 years younger than the offender. Juvenile sexual offenders who victimized at least one male, and one victim younger than them by 4 years or more, were exposed to more interpersonal violence at school than those who offended against female peers only. Juvenile offenders with no history of sexual offenses featured significantly more exposure to indirect violence at school compared to all of the juvenile sexual offender groups. In addition, juvenile sexual offenders without a history of sexual offending behavior featured higher levels of traumatic violence exposure at school than juvenile 71 sexual offenders who victimized female peers, and those who chose female victims, at least one of whom was younger than the offender by 4 years or more. In summary, youths scoring high in psychopathy featured more exposure to violence across all life contexts. Further, juvenile delinquents without a documented history of sexual offending, featured more exposure to violence in the neighborhood context than all groups of juvenile sexual offenders. Most notably, the violence that they were exposed to in the neighborhood context was of an indirect nature; the kind of violence that you may come across by simply walking down the street in a high-crime neighborhood characterized by low socioeconomic status. Further, the violence that non- sexually offending delinquents were exposed to in the school context was significantly more extensive than all groups of juvenile sexual offenders. Compared to youth who sexually offended against female victims, at least one of whom was younger than the offender by at least 4 years, non-sexually offending juvenile delinquents were more likely to witness domestic violence in their homes. Juvenile delinquents who had sexually offended against at least one male and at least one victim younger than themselves by 4 years or more were significantly more likely to be exposed to, or be the victim of interpersonal violence at school compared to juvenile sexual offenders who victimized peer-aged females. Sexual Offending Variables As expected, a main effect difference in the number of past adjudicated sexual offenses was observed across the five levels of the Victim Choice variable (F(4.368) = 17.26, p <.0001). Tukey post hoc testing revealed that youths without a documented history of sexual offenses featured significantly fewer adjudicated sexual offenses than 72 all of the sexual offender groups. Of more interest is the fact that participants who had at least one male victim and at least one victim younger than them by 4 years had a greater number of past adjudicated sexual offenses compared to those who had sexually offended against only females who were peer-aged. A main effect difference across the levels of the Victim Choice variable was observed when examining the total number of sexual offense victims for study participants (F(4,347) = 9.429, p < .002). Tukey post hoc testing revealed that juvenile sexual offenders who chose at least one male victim and one victim younger than themselves by 4 years or more, featured significantly more sexual offense victims in their history than all the other sexual offender groups, except for the small sample of those who had sexually offended against only female peers. In addition, juvenile delinquents with no history of sexual offenses had significantly fewer sexual offense victims than all of our groups of sexual offenders. This was due to the fact that these youths had no sexual offense victims whatsoever. Main and interaction effects were observed in examining scale scores on the Multiphasic Sex Inventory (MSI). For the following scales, significant main effects were obtained showing that youths rated high in psychopathy featured significantly higher scores than those scoring low to moderate in psychopathy: MSI Cognitive Distortions/Immaturity Scale (F(1,347) = 4.181, p < .043); MSI Rape Scale (F(1,347) = 7.917, p < .0001); MSI Exhibitionism Scale (F(1,347) = 6.398, p < .013); MSI Voyeurism Scale (F(1,347) = 8.692, p < .004); MSI Obscene Phone Calls Scale (F(1,347) = 8.789, p < .004); MSI Sado-Masochism Scale (F(1,347) = 8.879, p < .004); MSI Physical Disabilities Scale (F(1,347) = 11.674, p <.002). 73 Together, these results suggest that those offenders scoring high in psychopathy were most likely to be a ?blitz? type sex offender (Nichols & Molinder, 2001, pg. 19) who usually targeted peer-aged victims and tried to scare and intimidate them. They tended to be sexually active from an early age, interested in trying new things sexually, possessed a range of cognitions that were degrading or dismissive of women, excused their own sexual misconduct, and/or showed a clouded perception of how sexual intimacy develops and can be maintained. Further, participants deemed to be high in psychopathy were shown to be more likely to place obscene phone calls, engage in voyeuristic behaviors, and have the potential to voluntarily hurt others when engaged in sexual contact. Interestingly, although rarely endorsed on the MSI (Nichols & Molinder, 2001), those participants judged to be high in psychopathy were more likely to endorse items claiming that physical problems impede them from enjoying or participating in sexual contact. Main effect differences emerged across the 5 levels of the Victim Choice variable for the following MSI scale scores: MSI Sexual Obsessions Scale (F(4,347) = 2.679, p < .033); MSI Cognitive Distortions/Immaturity Scale (F(4,347) = 8.003, p < .0001); MSI Justifications Scale (F(4,347) = 14.377, p < .0001); MSI Treatment Attitudes Scale (F(4,347) = 5.808, p < .0001); MSI Child Molest Scale (F(4,347) = , p <. 007); MSI Rape Scale (F(4,347) = 5.310, p < .0001); MSI Exhibitionism Scale (F(4,347) = 4.688, p < .002); MSI Voyeurism Scale (F(4,347) = 4.336, p < .003); MSI Obscene Phone Calls Scale (F(4,347) = 8.122, p < .0001); MSI Sexual Apprehension/Confidence Scale (F(4.347) = 13.666, p < .022; and the MSI Sexual History Scale (F(4,347) = 8.399, p < .0001. 74 Tukey post-hoc testing revealed that juvenile sexual offenders who offended against at least one male and one victim younger than the offender by 4 years or more featured significantly more deception and a greater degree of preoccupation with sexual issues compared to juvenile sexual offenders who offended against peer-aged females exclusively (MSI: Sexual Obsessions Scale). In addition, youths without a history of sexual offenses featured significantly fewer cognitions associated with a lack of accountability for sexual acting out compared to all the sexual offender groups, except for the small sample group who offended against peer-aged victims, with at least one victim being male (MSI: Cognitive Distortions/Immaturity Scale). Non-sexually offending juvenile delinquents featured significantly fewer justifications for sexual acting-out as well (MSI: Justifications Scale Score). Of course, these findings make sense considering that youths without a history of sexual offending likely did not have any sexual behavior that required justifications or accountability for their actions. The fact there were no differences observed among groups of sexual offenders means that cognitive distortions associated with sexual acting-out are not exclusive to sexual offenders who target specific types of victims. All juvenile sexual offenders featured high, and statistically indistinct, histories of thinking in ways that support their sexual offending proclivities. On the MSI: Treatment Attitudes Scale, the main effect emerged when youths who offended against at least one male and at least one victim younger than themselves by 4 years or more featured significantly more thoughts associated with wanting to cooperate in treatment and work to remediate their problem compared to non-sexual offending delinquents and juvenile sexual offenders who victimized peer-aged females only. Youth who victimized peer-aged females exclusively often do not feel that they 75 deserve to be arrested for their crime, believing that the sexual activity they engaged in was normal and appropriate for their age level. Non-sexually offending juvenile delinquents did not report feeling that they needed sexual offender treatment because they likely featured no sexual offense victims during their lifetime. Further, juvenile sexual offenders who offended against peer-aged victims, at least one of which were male, displayed scores on the MSI: Treatment Attitudes Scale that compared to delinquents without a documented history of sexual offenses. In short, these youths reported their perception that they did not need sexual offender treatment, because they did not see themselves as being sexual offenders. Rather, they saw themselves more as individuals who were wrongly accused and incarcerated. However, statistically significant differences did no emerge due to the relatively small sample size of this sub-group of participants. The Child Molest Scale of the MSI measures thoughts and behaviors associated with sexually offending against young children, thus it is not surprising that non-sexual offenders featured significantly lower scores on this measure than all groups of juvenile sexual offenders. Also not surprising is the finding that youths who offended against females, at least one of whom was younger than the offender by 4 years, featured significantly higher scores than youths who sexually offended against peer-aged females. A significant difference between youths who sexually offended against at least one young, male victim and those who sexually offended against peer-aged victims, at least one of whom was male, was not observed. This lack of a significant finding is likely due to the small sub-sample size of our group of juvenile sexual offenders who targeted peer- aged victims, at least one of whom was male. 76 Youths who sexually offended against peers, at least one of whom was male, and those who sexually offended against at least one male and one victim younger than the offender by 4 years or more featured significantly higher scores than non-sexual offenders on the MSI: Rape Scale. This finding suggests that juvenile sexual offenders who targeted at least one male victim, regardless of the age of the victims, were more impulsive sexually than non-sexual offending juvenile delinquents. Youths who sexually offended against at least one male victim and one victim younger by at least 4 years were significantly more likely to feature a history of exhibitionistic tendencies compared to those who offended against females, at least one of whom was younger by 4 years or more (MSI: Exhibitionism Scale). The group featuring at least one male sexual offense victim and one victim younger than themselves by 4 years or more featured significantly higher scores on the MSI: Voyeurism and MSI: Obscene Phone Calls Scales than all other groups; except for the small sample group that sexually offended against peer-aged victims, at least one of whom was male. Simply put, juvenile sexual offenders who chose at least one victim younger than themselves by 4 years or more, and at least one male victim, were judged to be most likely to feature voyeuristic tendencies and to place obscene phone calls to others. Juvenile sexual offenders who victimized females, at least one of whom was younger than the offender by 4 years, and the group that sexually offended against at least one male and one victim younger by at least 4 years, had significantly higher scores on the MSI: Sexual Apprehension/Confidence Scale compared to the non-sexual offender group. These two groups who sexually offended against at least one victim younger than themselves by four years or more, regardless of their gender, were significantly less 77 confident about sexual matters than were non-sexually offending juvenile delinquents. The two groups of juvenile sexual offenders who targeted peer-aged victims did no differ in their levels of confidence about sexual issues when compared to non-sexually offending juvenile delinquents. However, the group that offended against peers, at least one of whom was male, featured a small sample size that decreased the likelihood of obtaining a significant finding. Non-sexual offending adolescents featured significantly lower scores than youths who sexually offended against females, at last one of whom was younger than the offender by 4 years or more, and those who victimized at least one male and one victim younger than the offender by 4 years or more on the MSI: Sexual History Scale. In addition, juvenile sexual offenders who offended against at least one male and one victim younger than the offender by 4 years or more featured significantly higher scores on the MSI: Sexual History Scale than the sexual offenders who offended against peer-aged females and youths who offended against females, at least one of whom was younger than the offender by at least 4 years. In short, juvenile sexual offenders who victimized at least one male and at least one individual younger than themselves by four years or more had significantly more sexual experiences, were more likely to have a sexual abuse history of their own, and had tried more things sexually than all the other groups except for the small sample of individuals who targeted peers, at least one of whom was male. Three significant interaction effects were observed when examining group differences across levels of the Victim Choice variable and Psychopathy Total Score variable for the following MSI scales: MSI Exhibitionism Scale (F(4,347) = 3.135, p < .016); MSI Obscene Phone Calls Scale (F(4,357) = 3.841, p < .006); MSI Sexual 78 Apprehension /Confidence Scale Score (F(4,346) = 13.458, p < .023). On the MSI Exhibitionism and Obscene Phone Call Scales, the interactions emerged when the group high in psychopathy who sexually offended against at least one male victim and one victim younger than the offender by 4 years or more featured dramatically high cell mean scores. Those high in psychopathy featured a statistically greater likelihood of placing obscene phone calls and having exhibitionistic tendencies, but when these youths also featured at least one male sexual offense victim and one sexual offense victim younger than themselves by four years or more, the elevations in the mean scores observed were more dramatic than expected. For the MSI Sexual Apprehension/Confidence Scale, the interaction emerged when the high psychopathy group who sexually offended against peer-aged victims, at least one of whom was male, featured the highest cell mean scores. This is in direct opposition to main effect findings on this scale which showed that youths who offended against at least one victim younger than themselves by four years or more, regardless of the gender of victims, to be the most uncomfortable and apprehensive about their sexuality. A significant chi square analysis was found when examining the association between whether youths had a history of using alcohol and/or drugs immediately prior to committing sexual offenses (3 levels: yes, no, unclear) and our Psychopathy Total Score (2 levels) (X 2 (2) = 13.394, p < ..002). A significantly higher proportion than expected of youths high in psychopathy had used alcohol and/or drugs immediately prior to committing sexual offenses. Further, a significantly lower proportion of youths judged low to moderate in psychopathy had used alcohol and/or drugs immediately prior to committing their sexual offense(s). 79 When examining the association between whether youths had taken any precautions to ensure their victim?s silence or alter the environment to hide their actions after committing their sexual offense(s)(2 levels: yes, no) and the Victim Choice variable, a significant Chi-square emerged (5 levels) (X 2 (8) = 28.246, p < ..0001). Results suggested youths who sexually offended against peer-aged females exclusively and those who had sexually offended against at least one male victim and one victim younger than them by four years or more featured a significantly lower proportion than expected of individuals who had taken precautions to hide their sexual offenses. Of interest, it is likely that these groups of sexual offenders did not take precautions to hide their offenses for different reasons. The group targeting peer-aged females exclusively may not have thought that their offenses were a crime, or committed their actions in a highly impulsive manner. Meanwhile, the group that targeted at least one male victim and one victim younger than themselves by four years or more, are more likely to be driven by sexual obsessions and drives that cause them to take advantage of any opportunity to engage in sexually inappropriate behavior. The concern with this group is likely to be obtaining sexual satisfaction at all costs, no matter how society might view their actions. A significant chi square was found when examining the association between the location in which sexual offenses were committed (7 levels: no information, offender?s residence, victim?s residence, other residence, non residential building, outdoors, other) and the Victim Choice variable (5 levels) (X 2 (28) = 129.626, p < .0001). Results suggested that youths who offended against female victims, at least one of whom was younger than the offender by four years or more, featured a significantly higher proportion of individuals who committed their sexual offenses at the victim?s residence 80 or their own residences. Participants who committed sexual offenses against females who were peer-aged exclusively featured a higher proportion than expected of youths who committed their sexual offenses in a non-resident building or outdoors. A significantly higher than expected number of youths who sexually offended against at least one male victim and one victim younger than themselves by 4 years or more committed their sexual offenses at their own residence. A significantly lower proportion than expected of youths in this category committed their sexual offenses at another residence or in a non- residential building. Psychiatric History/Personality Functioning Analyses were run for all of the scales from the Millon Adolescent Clinical Inventory. The following scales featured significant differences across the levels of the Victim Choice variable: Inhibited Scale (F(4,338) = 2.844, p < .25); Submissive Scale (F(4,338) = 2.411, p < .05); Dramatizing Scale (F(4,338) = 4.249, p < .002); Egotistic Scale (F(4,338) = 3.095, p < .017;.Unruly Scale (F(4,338) = 4.192, p < .004); Forceful Scale (F(4,338) = 5.742, p < .0001); Body Disapproval Scale (F(4,338) = 2.863, p < .024); Sexual Discomfort Scale (F(4,338) = 5.975, p < .0001); Peer Insecurity Scale (F(4,338) = 3.236, p < .014); Social Insensitivity Scale (F(4,338) = 12.952, p < .0001); Family Discord Scale (F(4,338) = 2.785, p< .028); Child Abuse Scale (F(4,338) = 2.636, p < .035; Substance Abuse Proneness Scale (F(4,338) = 6.094, p < .0001); Delinquent Predisposition Scale (F(4,338) = 4.431, p < .003); Anxious Feelings Scale (F(4,338) = 3.578, p < .013; and the Desirability Scale (F(4,338) = 4.663, p < .002). Tukey post hoc testing revealed the nature of significant differences among the MACI scales. Youths who sexually offended against exclusively female victims, of 81 which at least one of the victims was younger than the offender by 4 years or more, were significantly more inhibited and shy than youths who sexually offended against females who were peer-aged only and those without a documented history of sexual offenses (MACI ? Inhibited Scale Score). Youths who sexually offended against exclusively female victims, with at least one victim being younger than them by at least 4 years, were deemed more submissive than youths without a documented history of sexual offenses (MACI ? Submissive Scale Score). In summary, these participants who targeted at least one female victim, and one victim younger than themselves by at least four years, were shown to be the most socially awkward, quiet, and unassuming. Their ability to relate to peers their own age in an effective fashion was shown to be very much in doubt. Youths without a documented history of sexual offenses featured a behavioral and emotional pattern that can be characterized as being more dramatic and attention-seeking than youths who had committed sexual offenses against exclusively females, with at least some victims being younger than the offender by 4 years or more; and youths who had offended against at least one male victim and at least one victim younger than themselves by 4 years or more. Further, those youths who had committed sexual offenses against females who were peer-aged were more dramatic and attention seeking than those who had offended against at least one male victim and one victim who was younger than them by at least 4 years (MACI ? Dramatizing Scale Score). In addition, participants who had no documented history of sexual offending behavior were far more egotistic and self- centered than youths who had committed sexual offenses against females, with at least one being younger than the offender by 4 years or more; and those offending against at least one male and one victim younger than them by at least 4 years (MACI - Egotistic 82 Scale Score. Sample participants who sexually offended against young children, regardless of the age of the victim, were shown to be significantly more unassuming, quiet, and withdrawn compared to non-sexual offenders. Youths with no documented history of sexual offenses featured significantly higher ratings of unruly behavior compared to juvenile sexual offenders in all groups, except those who offended against peers exclusively, with at least one victim being male (MACI - Unruly Scale). In addition, those offenders without a history of sexual offending behavior were deemed more forceful and aggressive in their interactions with others compared to juvenile sexual offenders who victimized females, at least one of whom was younger than the offender by four years or more (MACI - Forceful Scale Score). Juveniles with no documented history of sexual offenses also featured more behaviors consistent with a delinquent predisposition compared to those juvenile sexual offenders who offended against females, at least one of whom was younger than the offender by 4 years or more; and those who offended against at least one male and at least one victim younger than the offender by 4 years or more (MACI - Delinquent Predisposition Scale). Again, these findings provide supportive evidence to suggest that juvenile sexual offenders who target young victims are less prone to committing a range of violent and non-violent criminal offenses compared to incarcerated non-sexually offending juvenile delinquents. They appear to be more specialized in their sexual crimes, and tend to possess a range of insecurities and concerns that direct their behavior in specifically sexual ways. Juvenile sexual offenders who offended against at least one male victim, and at least one victim younger than them by four years or more, displayed the most glaring 83 concerns about their body shape and functioning compared to all other groups of juvenile delinquents, except for the small sub-sample group of youths who sexually offended against peer-aged victims, with at least one victim being a male (MACI - Body Disapproval Scale). Although the groups of juvenile sexual offenders did not differ significantly in their levels of sexual discomfort, they all featured significantly higher scores on sexual discomfort and uncertainty compared to those with not documented history of sexual offenses (MACI - Sexual Discomfort Scale). In general, juvenile sexual offenders of all types were shown to be significantly more uncomfortable with their sexuality and discussing sexual issues compared to non-sexually offending juveniles. In particular, juveniles who broke two societal taboos when committing sexual offenses, when they offended against at least one young victim and one victim who was of the opposite sex, were most concerned about their body, how it was functioning, and whether they had a problem that required remediation. Youths with a history of sexual offenses against at least one male and at least one victim younger than themselves by 4 years or more displayed more signs of being uncomfortable and ineffective around peers than those who offended against only female peer-aged victims and those without a history of sexual offenses. In addition, participants who offended against females, at least one of whom was younger than the offender by 4 years or more, were deemed less confident and effective around their peers compared to those who offended against female peers only (MACI - Peer Insecurity Scale). Of the groups of juvenile sexual offenders assessed, those who targeted peer-aged victims, at least one of whom was male, were significantly more secure in their relationships with peers, and were generally more comfortable relating to others. In fact, in many ways, 84 their social relationships were very consistent with those of non-sexually offending juvenile delinquents. Yet, no statistically significant differences were observed due to the small sample size of this sub-group of participants. Youths with no history of sexual offenses featured greater levels of social insensitivity and poor social skills compared to all the groups of juvenile sexual offenders. They were more careless, less ruminative, and generally more impulsive in their actions within society. Interestingly, juvenile sexual offenders who offended against females, at least one of whom was younger than the offender by at least 4 years, were deemed significantly less empathic and interested in forming social relationships with family and peers compared to those juvenile sexual offenders who victimized female peers exclusively (MACI - Social Insensitivity Scale). It is possible that this group that committed sexual offenses against young, female children, were more likely to be focused on sexual gratification and sexual exploration compared to juvenile sexual offenders who victimized female peers. Individuals with no history of documented sexual offenses reported greater levels of family discord than all the groups of sexual offenders, except for the small sample of youths who offended against peers, at least one of whom was male (MACI - Family Discord Scale). Specifically, juvenile sexual offenders who featured offenses against at least one male victim and one victim younger than them by 4 years or more, featured a greater degree of child abuse victimization (e.g., sexual abuse, physical abuse, neglect) than those who offended against peer-aged female victims only (MACI ? Child Abuse Scale). Thus, non-sexually offending adolescents were most likely to come from chaotic home environments often characterized by violence and loud arguments. However, those juvenile sexual offenders who victimized at least one male and one victim younger than 85 themselves by 4 years or more, were significantly more likely to be assaulted in the past than were those who sexually offended against female peers. Participants with no history of documented sexual offenses featured behaviors and thoughts that were more consistent with being at risk for current and future substance abuse compared to juvenile sexual offenders in all groups, except for the small group that offended against peers, at least one of whom was male (MACI - Substance Abuse Proneness Scale). Juvenile sexual offenders who offended against females, at least one of whom was younger than the offender by 4 years or more, and those who offended against peer- aged victims who were females exclusively, featured higher levels of anxiety than youths who had no history of sexual offenses (MACI - Anxious Feelings Scale). It is unclear why youth who sexually offended against at least one male and at least one victim younger than themselves by 4 years or more did not also endorse symptoms of anxiety to the degree that those who targeted females did. Clues were provided when it was shown that juvenile sexual offenders who had at least one male victim and at least one victim younger than themselves by 4 years or more, featured significantly fewer attempts to portray themselves in a socially desirable manner, compared to youths who offended against peer-aged female victims exclusively (MACI - Desirability Scale). The following scales featured significant differences across the Psychopathy Total Score variable, such that those featuring high levels of psychopathy had significantly higher scores than those scoring low to moderate in psychopathy, on the following scales: Unruly Scale (F(1,338) = 16.073, p< .0001); Forceful Scale (F(1,338) = 18.141, p < .0001; Oppositional Scale (F(1,338) = 5.194, p < .024); Social Insensitivity (F(4,338) = 21.115, p < .0001); Family Discord Scale (F(4,338) = 10.334, p < .002); Substance 86 Abuse Proneness Scale (F(1,338) = 9.850, p < .003); Delinquent Predisposition Scale (F(1,338) = 19.813, p < .0001); Impulsive Propensity Scale (F(1,338) = 15.361, p < .0001); and Disclosure Scale (F(1,338) = 5.123, p < .025). These results are entirely consistent with modern conceptions of the psychopathy construct, with those featuring high levels of psychopathy being more disruptive, aggressive, disobedient, disrespectful toward authority figures, and impulsive in their decision making. These characteristics likely stem, at least in part, from their exposure and active involvement in chaotic homes featuring frequent aggression and arguing. The following scales featured significant differences across the PCL variable, such that those featuring low to moderate levels in psychopathy had significantly higher scores than those featuring high levels of psychopathy. These scales included the: Submissive Scale (F(1,338) = 15.665, p < .001), Conforming Scale (F(1,338) = 10.748, p < .002); Sexual Discomfort Scale (F(1,338) = 10.390, p < .002); and Anxious Feelings Scale (F(1,338) = 11.215, p < .002). Youths scoring low to moderate in psychopathy were significantly more unassuming, rule-abiding, and ruminative compared to youth scoring high in psychopathy. High psychopathy youth were shown to live their lives in a more carefree manner, whereas low psychopathy participants tended to worry about their sexuality, future lives, and their current incarceration. Interaction effects were also observed for offenders when categorized based on the age and gender of their victims and their score in comparison to the sample median on the PCL Total Score. Juvenile sexual offenders who offended against males with at least one victim being younger than them by 4 years and those offending against peer-aged victims, at least one of whom was male, had the highest ratings on the Forceful Scale 87 (MACI - Forceful Scale Score (F(4,338) = 3.101, p < .017). In examining main effect results, individuals were expected to score highest if they were non-sexual offenders who scored high in psychopathy. Further, juvenile sexual offenders who offended against at least one male and had one victim who was at least 4 years younger than the offender, and who were high in psychopathy, featured the highest degree of substance abusing tendencies across all cells (MACI ? Substance Abuse Proneness (F(4,338) = 2.408, p < .05). These interaction effect findings cast doubt on the perception that youth who target young male victims are meek, socially awkward, and merely sexual offending specialists. It would appear that when these youths are also judged high in psychopathic traits, they represent high risks for future criminal activity, aggression, and self-destructive tendencies. In addition, these youths are also the most likely to avoid efforts at portraying themselves in a positive light (MACI ? Desirability Scale (F(4,338) = 2.750, p < .029). Substance Abuse Two main effect differences and one interaction effect were observed for the number of times that youths reported drinking alcohol per week. Those scoring high in psychopathy reported drinking alcohol more frequently during the week than those scoring low in psychopathy (F(1,399) = 4.374, p <.039. In addition, a main effect difference in the number of days that youths reporting drinking alcohol each week was observed for the groups of juvenile offenders differentiated based on the age and gender of sexual offense victims, if any (F(4,369) = 5.317, p < .0001). Youths who had not committed a sexual offense in the past drank more frequently during the week than all of the groups of juvenile sexual offenders. An interaction effect for number of days that youths drank alcohol per week was also observed (F(4,369) = 3.569, p < .008). 88 Participants who scored high in psychopathy and abused only female victims who were peer-aged, drank alcohol more frequently than all other groups distinguished on the Victim Choice Variable. A main effect difference for the number of times that youths self-reported using illicit drugs was observed for the Victim Choice variable (F(4,369) = 4.833, p < .002. Tukey post hoc tests revealed that youths who had not committed a sexual offense in the past used drugs more frequently during the week than youths who offended exclusively against peer-aged females; exclusively against females, with at least one victim being more than 4 years younger than the offender; and those who offended against at least one male victim and one victim who was at least 4 years younger. Main and interaction effects were examined for individuals who completed the Substance Abuse Suble Screening Inventory: Second Edition (SASSI2). As the SASSI2 was a relatively new measure, the number of participants administered this inventory was less than those who responded to other questions and measures in the protocol. Nonetheless, a number of significant main and interaction effects were observed on this measure. For the following scales on the SASSI2, main effects emerged that showed individuals high in psychopathy to feature significantly higher scores than those low in psychopathy: Face Valid Alcohol (F(1,147) = 7.020, p < .010), Face Valid Other Drugs (F(1,147) = 7.079, p < .010), Family/Friends Risk (F(1,147) = 11.791, p < .002), Attitudes (F(1,147) = 6.462, p < .013), Symptoms (F(1,147) = 10.037, p < .003), Obvious Attributes (F(1,147) = 4.194, p < .043), Subtle Attributes (F(1,147) = 13.978, p <.0001), and Supplemental Addiction Measure (F(1,147) = 5.081, p < .027). Results suggest that 89 youth judged high in psychopathy were significantly more likely than those judged low to moderate in psychopathy to use alcohol regularly, use illicit drugs regularly, abuse alcohol and/or illicit drugs, and have family and friends who used alcohol and illicit drugs. These youths tended to feature attitudes supportive of substance abuse, and were more likely to feature symptoms consistent with alcohol or drug dependence. Main effects also emerged when examining mean scores across the levels of the Victim Choice variable. For the Face Valid Alcohol score on the SASSI2, a main effect emerged showing that individuals who committed sexual offenses against female peers exclusively were less likely to self-report alcohol use compared to youths with no documented history of sexual offenses (F(4,147) = 3.743, p < .007). In addition, youths without a documented history of sexual offenses featured significantly more attitudes supportive of alcohol and illicit drug abuse than all of our sexual offender groups formed through the Victim Choice variable (SASSI2: Attitudes Scale) (F(4,147) = 5.929, p < .0001). On the Supplemental Addiction Measure of the SASSI2, individuals without a documented history of sexual offenses featured significantly higher scores than sexual offenders who committed their offenses against female victims with at least one victim being younger than the offender by four years, female peers exclusively, and those who committed offenses against at least one victim younger than them by four years and at least one victim who was male (F(4,147) = 3.678, p < .008). In short, non-sexual offenders were significantly more likely to feature behaviors and thought patterns consistent with alcohol dependence compared to all of our groups of sexual offenders, excluding our small group of juvenile sexual offenders who targeted peer-aged victims, at least one of whom was male. 90 Three interaction effects emerged when examining scales from the SASSI2. For the Face Valid Alcohol Scale. Results showed that youths who committed sexual offenses against peer-aged victims, at least one of whom was male, featured a pattern in which those scoring high in psychopathy had lower mean scores than those scoring low to moderate in psychopathy (F(4,147) = 5.292, p < .002). This finding is difficult to explain, and requires future replication with a larger sample of youth who sexually offended against peer-aged victims, at least one of whom was male. For the Face Valid Other Drug Scale, an interaction effect emerged showing that participants with no documented history of committing sexual offenses and those offending against peer-aged victims, at least one of whom was male, featured no significant difference in scale scores across the levels of their psychopathy scores (F(4,147) = 4.047, p < .005). For the other groups of juvenile delinquents categorized through the Victim Choice variable, youths scoring high in psychopathy featured higher scale scores than those scoring in the low to moderate range in psychopathy. Regardless of their psychopathy ratings, these two groups of youth were particularly at-risk for using and abusing illicit drugs; usually marijuana. On the Symptoms Scale, an interaction effect emerged suggesting that individuals without a documented history of sexual offenses featured no significant differences in scale scores between those scoring high in psychopathy and those scoring low to moderate in psychopathy (F(4,147) = 2.926, p < .024). This trend was not evident for other groups of juvenile delinquents, and is not easily explained. Summary Power estimates obtained for significant main and interaction effects most commonly fell in the range of .7- to 1.0. Yet, effect sizes were smaller than expected. 91 Table 5 provides a presentation of p-values, observed power estimates, and effect sizes (eta-squared) for all significant main and interaction effects observed in examining ANOVA results. An additional 2-way ANOVA was used to determine whether information gathered from examining Callous/Unemotional Factor Scores from the PCL:YV would provide information distinct from that obtained from examining results from the ANOVA featuring the Psychopathy Total Score variable. Main and interaction effects obtained through these analyses featured results that were remarkably similar to those obtained from investigating the Psychopathy Total Score variable. This finding is not surprising considering that the Callous/Unemotional Factor of the PCL:YV has been shown to correlate .7 with the PCL:YV Total Score in past research studies (Forth et al., 2003). Thus, these results will not be presented. 92 Table V Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Victim Choice Variable ? Significant Main Effects Variable P-value Eta-squared Power Age in Months .000 .148 1.00 Number of people living in the home .020 .033 .799 Performance IQ Score .027 .040 .694 Number of School Suspensions .039 .028 .718 Number of Juvenile Delinquency .000 .147 1.00 Commitments Number of past arrests .000 .301 1.00 SAVE Total Score .007 .040 .868 SAVE Total Neighborhood Score .004 .045 .905 SAVE Neighborhood Indirect Score .001 .052 .950 SAVE Neighborhood Traumatic Score .005 .043 .895 SAVE Home Traumatic Score .027 .032 .762 SAVE School Total Score .010 .038 .849 SAVE School Interpersonal Score .047 .028 .698 SAVE School Indirect Score .003 .045 .910 SAVE School Traumatic Score .042 .029 .710 Number of past adjudicated sex offenses .000 .162 1.00 93 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Victim Choice Variable ? Significant Main Effects Variable P-value Eta-squared Power Number of past sexual offense victims .001 .059 .945 MSI Sexual Obsessions Scale .032 .031 .743 MSI Cognitive Distortions/Immaturity Scale .000 .087 .998 MSI Justifications Scale .000 .146 1.00 MSI Treatment Attitudes Scale .000 .064 .982 MSI Child Molest Scale .006 .135 1.00 MSI Rape Scale .000 .059 .971 MSI Exhibitionism Scale .001 .053 .849 MSI Voyeurism Scale .002 .049 .931 MSI Obscene Phone Calls Scale .000 .088 .998 MSI Sexual Apprehension/Confidence Scale .021 .034 .787 MSI Sexual History Scale .000 .091 .999 MACI Inhibited Scale .024 .034 .771 MACI Submissive Scale .049 .029 .691 MACI Dramatizing Scale .001 .049 .925 MACI Egotistic Scale .016 .036 .810 MACI Unruly Scale .003 .049 .921 94 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Victim Choice Variable ? Significant Main Effects Variable P-value Eta-squared Power MACI Forceful Scale .000 .065 .981 MACI Body Disapproval Scale .023 .034 .774 MACI Sexual Discomfort Scale .000 .068 .985 MACI Peer Insecurity Scale .013 .038 .829 MACI Social Insensitivity Scale .000 .136 1.00 MACI Family Discord Scale .027 .033 .761 MACI Child Abuse Scale .034 .031 .735 MACI Substance Abuse Proneness Scale .000 .069 .986 MACI Delinquent Predisposition Scale .002 .051 .936 MACI Anxious Feelings Scale .012 .038 .830 JI Alienation Scale .001 .052 .944 JI Asocial Index .050 .028 .687 JI Cultural Conformist/Group Oriented .002 .049 .932 JI Neurotic, Anxious/Introspective .000 .062 .976 Number of days the youth drank alcohol .000 .056 .973 out of the week 95 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Victim Choice Variable ? Significant Main Effects Variable P-value Eta-squared Power Number of days the youth used illicit .001 .051 .955 drugs out of the week SASSI2 Face Valid Alcohol Scale .006 .099 .878 SASSI2 Attitudes Scale .000 .148 .982 SASSI2 Supplemental Addiction Measure .007 .097 .872 96 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Psychopathy Total Score Variable: Significant Main Effects Variable P-value Eta-squared Power Number of marriages for biological father .029 .014 .606 Number of marriages for biological mother .049 .011 .502 Number of half-siblings .039 .012 .551 Full Scale IQ Score .011 .028 .507 Verbal IQ Score .023 .022 .618 Performance IQ Score .025 .021 .612 Number of School Suspensions .000 .037 .960 Number of fights in the last year .000 .039 .967 Number of fights in the last 3 years .005 .022 .811 Number of past juvenile delinquency .000 .059 .997 commitments Number of past arrests .000 .088 1.00 SAVE Total Score .000 .060 .996 SAVE Neighborhood Total Score .000 .071 .999 SAVE Neighborhood Interpersonal Score .008 .020 .758 SAVE Neighborhood Indirect Score .000 .056 .994 SAVE Neighborhood Traumatic Score .000 .080 1.00 97 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Psychopathy Total Score Variable: Significant Main Effects Variable P-value Eta-squared Power SAVE Home Total Score .000 .045 .979 SAVE Home Indirect Score .000 .040 .964 SAVE Home Traumatic Score .000 .044 .976 SAVE School Total .006 .022 .782 SAVE School Indirect Score .010 .019 .730 SAVE School Traumatic Score .002 .029 .885 MSI Cognitive Distortions/Immaturity Scale .042 .012 .531 MSI Rape Scale .000 .023 .801 MSI Exhibitionism Scale .012 .019 .713 MSI Voyeurism Scale .003 .025 .836 MSI Obscene Phone Calls Scale .003 .025 .840 MSI Sado-Masochism Scale .003 .026 .844 MSI Physical Disabilities Scale .001 .033 .926 MACI Unruly Scale .000 .047 .979 MACI Forceful Scale .000 .052 .989 MACI Oppositional Scale .023 .016 .623 MACI Family Discord Scale .001 .031 .893 98 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Psychopathy Total Score Variable: Significant Main Effects Variable P-value Eta-squared Power MACI Substance Abuse Proneness Scale .002 .029 .879 MACI Delinquent Predisposition Scale .000 .057 .993 MACI Impulsive Propensity Scale .000 .045 .974 MACI Disclosure Scale .024 .015 .617 MACI Social Insensitivity Scale .000 .060 .996 MACI Submissive Scale .000 .046 .977 MACI Conforming Scale .001 .032 .905 MACI Sexual Discomfort Scale .001 .031 .895 MACI Anxious Feelings Scale .001 .033 .916 JI Social Maladjustment Scale .012 .019 .710 JI Value Orientation Scale .004 .024 .825 JI Autism Scale .043 .012 .526 JI Manifest Aggression Scale .001 .032 .918 JI Asocial Index Scale .011 .019 .723 JI Unsocialized, Aggressive/Under- .001 .035 .938 socialized,Active Subtype 99 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Psychopathy Total Score Variable: Significant Main Effects Variable P-value Eta-squared Power JI Cultural Conformist/Group Oriented .022 .016 .631 Subtype JI Neurotic/Acting-Out/Autonomy .001 .033 .922 Oriented Subtype JI Immature Conformist/Conformist .000 .048 .985 Subtype JI Situational Emotional Reaction .006 .022 .780 Inhibited Subtype JI Cultural Identifier/Adaptive Subtype .049 .011 .503 Number of times the youth used alcohol .037 .012 .550 out of the week SASSI2 Face Valid Alcohol Scale .009 .049 .749 SASSI2 Face Valid Other Drug Scale .009 .049 .752 SASSI2 Family/Friends Risk Scale .001 .079 .926 SASSI2 Symptoms Scale .002 .068 .882 SASSI2 Obvious Attributes Scale .042 .030 .529 SASSI2 Subtle Attributes Scale .000 .093 .960 100 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Psychopathy Total Score Variable: Significant Main Effects Variable P-value Eta-squared Power SASSI2 Supplemental Addiction Measure .026 .036 .610 SASSI2 Attitudes Scale .012 .045 .714 101 Table V (continued) Power Estimates, Effect Sizes, and P-Values Associated with Significant ANOVA Main and Interaction Effects Victim Choice and Psychopathy Total Score Variables: Significant Interaction Effects Variable P-value Eta-squared Power Number of School Suspensions .032 .029 .743 MSI Sexual Apprehension/Confidence Scale .022 .033 .780 MSI Obscene Phone Calls Scale .005 .044 .894 MSI Exhibitionism Scale .015 .036 .816 MACI Forceful Scale .016 .036 .811 MACI Substance Abuse Proneness Scale .049 .029 .690 MACI Desirability Scale .028 .032 .755 Number of days youth used alcohol .007 .038 .869 out of the week SASSI2 Face Valid Alcohol Scale .001 .134 .968 SASSI2 Face Valid Other Drug Scale .004 .106 .905 SASSI2 Symptoms Scale .023 .079 .775 102 DISCUSSION The present investigation, featuring a large sample size, moderately high to excellent power estimates, and multiple measures of varying constructs, replicated and extended studies conducted to examine characteristics of juvenile sexual offenders (Awad & Saunders, 1991; Becker et al., 1986; Kahn & Chambers, 1991; Nisbet et al., 2004; Rasmussen, 1999; Ryan et al., 1996; Smith & Monastersky, 1986; Van Ness, 1984; Vinogradov et al., 1988). Specifically, the study sample consisted of youths who were incarcerated for serious sexual offenses, and another group of youth who were incarcerated for committing serious crimes that were of a non-sexual nature. Importantly, this allowed for comparisons between juvenile sexual offenders and non-sexually offending juvenile delinquents (Becker et al., 1986; Blaske et al., 1989; Brannon et al., 1989; Caputo et al., 1999; Fagan & Wexler, 1988; Katz, 1990; Kemptom & Forehand, 1992; Oliver et al., 1993; Otnow-Lewis et al., 1979). These data were collected prior to treatment, and information was obtained through extensive interviewing of the youth, a review of relevant file information, and standardized self-report instruments and rating scales. In addition, juvenile sexual offenders were not merely grouped together into an artificially homogenous set. Rather, four groups of juvenile sexual offenders were formed based on the age(s) and gender(s) of their victim(s). Past research, featuring dramatically fewer dependent variables, have used similar distinctions to provide information on how individual juvenile sexual 103 offenders tend differ from a historical, behavioral, and personality-based standpoint (Benoit & Kennedy, 1992; Worling, 1995). Thus, comparisons between different types of juvenile sexual offenders were also examined. Finally, youth were compared on scores designed to gauge behaviors consistent with a psychopathic personality pattern (Hare, 1991). Recent research has shown that psychopathy can be measured reliably in juveniles (Forth et al., 2003), and that the construct is predictive of conduct problems in children (Frick et al., 1994), general criminal recidivism (Catchpole & Gretton, 2003), and sexual offense recidivism (Gretton et al., 2001) in juveniles. Interaction effects between a dispositional variable (PCL:YV Total Score; Psychopathy Total Score Variable) and an offense-related variable (Age and Gender of Victims; Victim Choice Variable) were obtained and analyzed. Most studies of juvenile sexual offenders are descriptive in nature. These investigations often fail to provide in-depth or comprehensive information on their sample, and treat juvenile sexual offenders as a homogenous group (Aljazereh, 1993). Rarely do these analyses feature a non-sexually offending comparison group, whether in the community, outpatient clinic, or detention center (Davis & Leitenberg, 1987). The present study featured relevant comparisons within the sample of sexual offenders, comparisons between groups of juvenile sexual offenders and a non-sexually offending juvenile delinquent sample, and also evaluated the impact of a promising measure of the psychopathic personality pattern in juveniles. Juvenile sexual offenders and non-sexually offending delinquents In comparing juvenile sexual offenders differentiated based on the age and gender of their victim(s) with juvenile delinquents without a documented history of sexual 104 offenses, a cohesive picture of distinctions emerged. Consistent with past research, juvenile delinquents without a documented history of sexual offenses were significantly older than the groups of juvenile sexual offenders. These youths, as in past studies, were likely older due to their longer history of delinquency prior to their most recent incarceration (Aljazireh, 1993; Vinogradov et al., 1988). Youths without a history of sexual offending featured a significantly lower than expected proportion of participants who had been physically abused, in comparison to juvenile sexual offenders who featured a history of having young victims. Non-sexual offenders, as in other studies, had significantly more school suspensions, juvenile delinquency commitments, past arrests, and fights than the groups of juvenile sexual offenders (Fagan & Wexler, 1988). Further, non-sexually offending youths featured greater exposure, in general, to varying types of violence across the home, school, and neighborhood contexts. These results, consistent with past research (Aljazireh, 1993), show that non-sexually offending juvenile delinquents tend to feature turbulent pasts characterized by violence exposure and family dysfunction (Caputo et al., 1999). They were generally less concerned with societal rules, more impulsive, and had more disruptive behavior tendencies than juvenile sexual offenders (Ford & Linney, 1995). They are older, and tend to feature a more entrenched pattern of delinquency (Fagan & Wexler, 1988; Hsu & Starzynski, 1990). For the non-sexually offending juveniles, it was their accumulation of offenses and escalation of offense severity that likely resulted in their incarceration. For many of the juvenile sexual offenders, it was the perceived heinousness of their sexual crimes, and not necessarily their risk for sexual offense or general criminal recidivism, that resulted in their incarceration. Juvenile non-sexual offenders were exposed to violence, on a near 105 continuous basis, where they lived, attended school, and spent time with friends. In comparison, the groups of juvenile sexual offenders were not as likely to experience this breadth and intensity of violence exposure or disruptive behavior problems (Caputo et al., 1999; Spaccarelli, 1997). As might be expected, delinquent youths without a history of sexual offending behavior featured fewer cognitive distortions surrounding their sexuality and sexual behavior, fewer justifications and excuses for past sexual behavior, and fewer cognitive and behavioral tendencies consistent with being or becoming a child molester or rapist. Unlike our groups of juvenile sexual offenders, these youths had not faced the social stigma associated with being charged with sexual offenses and, thus, likely did not possess the range of concerns, thoughts, and preoccupations associated with their sexual behaviors. When compared to juvenile sexual offenders who targeted at least one male sexual offense victim, non-sexual offenders were far less open in expressing their need for sexual education, guidance, or advice. In short, when it came to sexual matters, non- sexually offending juvenile delinquents were largely unconcerned compared to juvenile sexual offenders, in general. In general, delinquent youths without a history of sexual offenses, in comparison to all groups of sexual offenders except for the small sample of youths who sexually offended against peer-aged victims, at least one of whom was male, came from families where violence, arguments, substance abuse, and criminal activity were common (Awad & Saunders, 1991). They were more likely to abuse substances as well. Borduin et al. (1989) reported less family cohesion in non-sexually offending adolescents compared to a mixed group of juvenile sexual offenders. These results are entirely consistent with 106 these findings. In general, non-sexually offending juvenile delinquents could best be described as being the most insensitive to common social rules and conventions; which likely contributed significantly to their past involvement with criminal activity and delinquent peer groups (Blaske et al., 1989). They were not nearly as nervous and apprehensive about the future, and were more comfortable interacting with others. More specifically, compared to juvenile sexual offenders who targeted at least one victim younger than themselves by 4 years or more, the non-sexual offenders were more forceful, impulsive, and egotistic. They were not nearly as inhibited socially or submissive compared to juvenile sexual offenders who victimized females, at least one of whom was younger than themselves by 4 years or more. These findings are highly consistent with past studies that describe juvenile sexual offenders as being, in general, more submissive, inhibited, and less socially skilled than non-sexually offending juveniles (Awad & Saunders, 1991; Fehrenbach et al.,1986). However, as will be discussed, even this generalization needs to be tempered by the variability in the different groups of juvenile sexual offenders examined in this study. The Jesness Inventory was designed to measure behaviors and personality patterns consistent with multiple types of delinquent lifestyles. Non-sexually offending juvenile delinquents can best be described as being more distrustful of authority figures, less obedient, and less motivated to change their current life context or behavioral patterns. They reported being more comfortable around delinquent peers, and seemed to anticipate living a life characterized by varying degrees of antisocial behavior in the future. Juvenile sexual offenders who offended against at least one male victim and at 107 least one victim younger than themselves by 4 years or more, were more self-critical, ruminative, and uncomfortable than non-sexual offenders (Blaske et al., 1989). Compared to most of our sexual offender groups, non-sexually offending juvenile delinquents featured greater use of alcohol and illicit drugs. Their attitudes, in general, were more consistent with a lack of concern for future consequences and actions. They had observed their families using alcohol and drugs, tended to take similar risks, and many had chosen to abuse substances in their youth to a greater degree than the juvenile sexual offender groups. Within-Group Comparisons of Juvenile Sexual Offender Types The present investigation was consistent with past studies that examined characteristic differences amongst different types of juvenile sexual offenders, categorized in varying ways (Awad & Saunders, 1991; Benoit & Kennedy, 1992; Ford & Linney, 1995; Herkov et al., 1996; Hsu & Starzynski, 1990; Kaufman & Hilliker, 1996; Katz, 1990; Kavoussi et al., 1988; Smith, 1988; Worling, 1995). Distinguishing juvenile sexual offenders based on the age(s) and gender(s) of their sexual offense victims, if any, resulted in the observation of striking variability between groups. Compared to the other groups of juvenile sexual offenders, participants who committed sexual offenses against at least one victim younger than themselves by 4 years or more and at least one male featured a higher proportion than expected of Caucasians. Youths who sexually offended against female peers exclusively featured a higher proportion of African-Americans. The origin of these differences is difficult to interpret. Participants who committed sexual offenses against at least one male victim and at least one victim who was at least four years younger, had significantly higher Performance IQ 108 Scores on the Wechsler Intelligence Scales than those who sexually offended against female peers exclusively. The origin of this intellectual difference is difficult to trace developmentally; whereas racial differences across different groups of sexual offenders could stem, at least in part, from cultural norms associated with sexual activity. Information pertaining to racial, religious, and ethnic differences across juvenile sexual offender groups is needed (Aljazireh, 1993; Ryan et al., 1996), as cultural considerations have risen to the forefront in treatment studies for a wide range of mental health-related disorders characteristic of childhood and adolescence (Cichetti & Rogosch, 2002) Consistent with Worling (1995), juvenile sexual offenders who offended against at least one male victim and one victim younger than the offender by four years or more, featured a higher than expected proportion of youths who had a history of sexual abuse victimization. In addition, this group of juvenile sexual offenders, and the group who sexually offended against females, with at least one victim being younger than themselves by 4 years or more, had a higher than expected proportion of individuals who had been physically abused during the course of their life. Age and gender of victims is a variable that is often confounded in studies of adult and juvenile sexual offenders (Benoit & Kennedy, 1992; Worling, 1995). Juvenile sexual offenders who targeted at last one victim that was younger than themselves by four years or more, whether male or female, were most likely to have a history of abuse victimization (Seghorn et al., 1987; Worling, 1995). They were more likely to be raised in a culture in which deviant sexual behavior and/or interpersonal violence was experienced first hand, and at an early age. Consistent with the emerging developmental psychopathology perspective in mental health (Cichetti & Rogosch, 2002), the timing of these traumatic experiences, coupled with the sexual and 109 violent nature of the behavior they experienced, makes it more likely that these juveniles will develop deviant sexual behavior patterns, deviant sexual arousal patterns, and glaring interpersonal deficits surrounding how they view and treat others. Juvenile sexual offenders who targeted young victims, whether male or female, appear most likely to require specific treatment designed to address their abuse history and subsequent attachment disruptions. In addition to being more likely to have a sexual and/or physical abuse victimization history, juvenile sexual offenders who victimized at least one young child and one victim who was male, were also shown to be more likely to feature a history of exposure to interpersonal types of violence at school when compared to juvenile sexual offenders who targeted female peers exclusively. These sexual offenders? history of abusing young males, and their characteristic social awkwardness and isolation, seemed to place them at high risk for being labeled an outcast in their peer group. When left to their own devices during their formative years, they were more likely to be abused, and feel alienated, isolated, and awkward (Katz, 1990). This combination may place them at particularly high risk to develop a history of social deviance, which can and does sometimes take the form of sexual deviance (Murphy et al., 2001). Juvenile sexual offenders who target peer-aged females, as seen in other studies, appeared to feature sexual and physical abuse histories more similar to non-sexually offending juvenile delinquents. Juvenile sexual offenders who targeted at least one male victim and at least one victim younger than the offender by four years or more were much more likely to feature a range of deviant thoughts and actions associated with their sexuality (Murphy, et al., 2001). They had more persistent and extensive thoughts about sexual issues compared to youths who featured a sexual offending history against female 110 peers exclusively. In addition, this group featured a greater degree of exhibitionistic tendencies compared to youths who sexually offended against females, at least one of whom was younger than the offender by four years or more. They also placed more obscene phone calls during their lifetime and engaged in more voyeuristic behaviors than all other groups of sexual offenders, except for the small sample group of youths who sexually offended against peers, at least one of whom was male. Further, in comparison to juvenile sexual offenders who targeted peer-aged victims, whether they were male of female, juvenile sexual offenders who victimized at least one male and at least one individual younger than themselves by four years or more, had a more prominent history of exposure to sexually explicit material and experiences (i.e., sexual abuse victimization, pornography, sexually assaultive behavior). In short, this group of sexual offenders displayed characteristics that placed them at high risk for the development of sexual obsessions, and an entrenched pattern of sexual deviance that may last into adulthood. Despite the seemingly poor prognosis, youths who committed sexual offenses against at least one male victim and one victim younger than the offender by four years or more, also had attitudes more consistent with knowing that they had a problem that they needed help to address in comparison to those who sexually offended against female peers exclusively. Social mores and personal guilt associated with offending against younger victims and/or victims of the same sex, likely fuels these youths? desire to seek treatment for their problems, compared to juvenile sexual offenders who may believe that they were coaxed into offending against a similarly aged person of the opposite sex. Further, they are more aware of the persistent nature of their sexual thoughts, deviant sexual patterns, and history of exposure to confusing and traumatic experiences that were 111 often sexual in nature. It would seem that this group of juvenile sexual offenders, through their own reports and a review of their histories, possess a more entrenched personality structure, behavioral history, and history of abuse that places them at high risk for developing an embedded pattern of deviant sexual behavior. It is these youths that require a comprehensive treatment program to address their sexual deviancy, if their risk for future deviant sexual behavior is to be reduced. There was evidence to suggest that juvenile sexual offenders who victimized females, at least one of whom was younger than the offender by four years or more, featured more extensive and intense exposure to sexually explicit material and experiences when compared to juvenile sexual offenders who targeted peer-aged victims, whether they were male or female. This evidence seems to point to the fact that those who targeted younger female victims were also likely to feature some characteristics associated with modern conceptions of sexual deviance (i.e., exposure to sexual material at a young age, multiple sexual experiences, abuse history), however; not to the degree that those who targeted young victims, at least one of whom was male, were shown to possess. Participants who committed sexual offenses against female peers exclusively, and those who victimized at least one male and at least one victim younger than themselves by four years or more, had a lower than expected proportion of individuals who attempted to hide their sexual offenses or chide their victim into keeping their experiences a secret. This finding likely means different things for each group of youths. For the youths who targeted female peers exclusively, the impulsive and uninhibited quality of their offenses likely reflected a lack of insight into the nature of their behavior 112 and how it could affect other people. Further, it is possible that some youths in this category did not grasp the fact their behavior was illegal and unjust. Treatment focusing on educating these youths on appropriate sexual behavior, relevant sexual offense statutes, reviewing victim impact statements, and challenging maladaptive thought patterns might prove most beneficial in reducing their risk of sexual offense recidivism. For those who victimized at least one male and at least one victim younger than themselves by four years or more, their failure to attempt to hide their offenses likely reflects a level of impulsivity that stems from their more highly developed sexual preoccupation. Juvenile sexual offenders who targeted female victims, at least one of whom was younger than the offender by four years or more, had a higher proportion of youths who committed their sexual offenses in their own residence or the victim?s residence. Youths who offended against at least one male victim and at least one victim younger than themselves by four years or more, had a higher than expected proportion of youths who committed their sexual offenses at their own home, and a lower proportion than expected who committed their offenses outdoors or in a non-residential building. Youths who sexually offended against female peers exclusively, had a higher proportion than expected of individuals who committed their sexual offenses outdoors or in a non- residential building. These offense-related findings have obvious value for treatment providers working to establish relapse prevention plans for offenders and their caregivers. These initial findings suggest that individuals who target young victims are most likely to offend against others in their own homes, or the homes of their victim. Juvenile sexual offenders who targeted female peers are more impulsive and less calculated in their 113 sexual offending. They do not often believe that their behavior is wrong, and are more likely to act on their impulse suddenly and without planning. Youths who victimize young children may have frequent urges to offend sexually, yet because of the social mores associated with sexual contact with young children, they attempt to plan their offenses in order to avoid discovery and subsequent alarm from their family and the family of their victims. In examining personality-based differences between different types of sexual offenders, those juveniles who committed sexual offenses against females who were peer-aged exclusively, exhibited more efforts toward portraying themselves in a positive light, had a less extensive degree of childhood trauma, were less insecure about their peer relationships, and were more impulsive and dramatic than those who committed sexual offenses against at least one male and at least one victim younger than themselves by four years or more. They showed evidence of being more comfortable with their bodies than youths who committed sexual offenses against at least one male and one victim younger than themselves by four years or more. These youths were also shown to be less inhibited, less worried about their peer relationships, less sensitive to social mores, and more impulsive compared to youths who sexually offended against females, at least one of whom was younger than the offender by four years or more. In short, juvenile sexual offenders who targeted female peers exclusively were very similar to non-sexually offending juvenile delinquents in their personality structures. They tended to possess a range of risk factors for all types of violent crime, not just sexual offenses. In addition to sexual offense treatment, this group of youth would likely benefit from receiving treatment aimed specifically at reducing their impulsive behaviors (i.e., developing anger 114 management plans, developing coping strategies for dealing with other negative emotions). These juvenile sexual offenders could best be described as ?generalists?, whereas juvenile sexual offenders who victimized at least one victim younger than themselves by 4 years or more, are more ?specialized? in their sexually deviant behavior patters (Soothill et al., 2000). Psychpathy-Based Group Differences Youths judged to be high in total psychopathy scores for our sample had more half-siblings, and were more likely to feature a history of neglect than those judged to be low to moderate in psychopathy. In short, these findings are consistent with past literature investigating the construct of psychopathy with juveniles (Catchpole & Gretton, 2003; Forth et al., 2003; Frick et al., 1994; Gretton et al., 2991). Youths scoring high in psychopathy tend to possess a lack of empathy which contributes to them having poor social relationships in the future (Forth et al., 2003). They have witnessed their parents become involved, and re-involved, in multiple relationships. Further, high psychopathy youths also tended to have witnessed and experienced abuse and violence across their life contexts to a greater extent than those scoring low to moderate in psychopathy. Subsequently, it is not surprising that these youths featured dramatically more school suspensions, arrests, juvenile delinquency commitments, and fights than those judged to be low to moderate in psychopathic traits. These findings point to a group of individuals who feature inconsistent relationships with family, traumatic pasts, and disruptive behavior across multiple contexts for an extended period of time. In short, treatment with these individuals is likely to prove challenging, and may need to be of a long term duration (Forth et al., 2003). 115 Youths judged high in total psychopathy scores have witnessed and experienced considerably more types of violence across multiple life contexts in comparison to those judged low to moderate in psychopathy. They tended to possess more deviant sexual attitudes and maladaptive cognitions that do not support the change process advocated for in therapy. Further, they tend to feature lower intelligence quotients, disruptive behavior patterns, and greater consistency with characteristics shown to be predictive of criminal delinquency and future recidivism (Forth et al., 2003; Hare, 1991). In addition, those high in total psychopathy are most prone to substance abuse, as yet another example of their versatile and dangerous criminal lifestyle. . Interaction Effects In tandem, the Psychopathy Total Score and Victim Choice variables, were used to analyze for the presence of interaction effects that might further highlight specific ways to treat specific types of juvenile delinquents and juvenile sexual offenders. Juvenile sexual offenders who victimized at least one young victim and at last one victim who was male featured the most characteristics consistent with the development of a deviant sexual arousal pattern; a characteristic that has been shown in research with adults to place them at higher risk for sexual offense recidivism. They tended to feature a less extensive collection of factors that would place them at high risk for general criminal recidivism (Hare, 1991). However, it is important to note that juvenile sexual offenders who victimized at least one male and at least one person younger than themselves by 4 years or more, when also high in psychopathy, had a number of characteristics that would place them at high risk general criminal recidivism. This group had multiple school suspensions, an 116 extensive substance abuse history, and were more likely than expected to be forceful in their interactions with others. This suggests that when individuals featuring the highest levels of sexual deviance, and the highest levels of behaviors consistent with a versatile criminal lifestyle, their current and future prognosis is poor. These youth represent serious risks for future criminal and sexual offense recidivism due their impulsive and uninhibited style of interacting with others in their environment. Limitations This study represents an initial and tentative step toward defining important distinctions to be made among juvenile sexual offenders. Results are also likely to have utility for researchers working to refine their own research efforts pertaining to the assessment and treatment of juvenile sexual offenders. However, like in any study, limitations were present. First, our sample is restricted to youths living in the Southeastern United States. It is possible that our sample of juvenile sexual offenders may not be representative of other parts of the United States. Second, our categorized group of juvenile sexual offenders who targeted exclusively peer-aged victims, at least one of whom was male, was quite small in comparison to the other sub-samples of juvenile sexual offenders in this study, and, therefore, the power to detect differences was limited. Thus, even though mean scores on many of our dependent measures for this group of sexual offenders were different from the other groups, significant results could not often be obtained. This limitation is nothing new to research on juvenile sexual offenders, as descriptive studies featuring very large samples gathered throughout the United States, have reported that 117 juvenile sexual offenders who target male peers, or at least one male peer victim, are rare (Ryan, Miyoshi, Metzner, Krugman, & Fryer, 1996). Third, some of the measures utilized in this study are relatively new, and have not been used in work with juvenile sexual offenders in the past. For instance, the Screen for Adolescent Violence Exposure (SAVE) (Hastings & Kelley, 1997), although it represents a promising measure of useful and novel constructs, has never, to our knowledge, been used in a research study with juvenile sexual offenders. It is hoped that researchers will continue to incorporate a variety of new measures of novel constructs in order to increase our knowledge base pertaining to subtle differences among different types of juvenile sexual offenders. Finally, effect sizes (eta squared values) obtained for main effects and interactions were small. This suggests that the wide-ranging clinical significance of our results is in question. The value of the data examined in this study lies most in contributing to the knowledge base that future researchers will utilize to refining our treatment of specific types of juvenile sexual offenders. Researchers can take these exploratory findings, and through statistical means, combine information and results to arrive at a more advanced understanding of each individual offender. Future Directions The literature on the assessment and treatment of juvenile sexual offenders is in the beginning stages of development. Research efforts have intensified, considering the widespread nature of the problem (Davis & Leitenberg, 1987), and the enormous financial, emotional, and psychological costs to victims, offenders, families, and communities as a whole. Sexual offending behavior does not commonly begin in 118 adulthood (Groth, Longo, & McFadin, 1982), and unfortunately it does not end in adolescence (Abel, Becker, Mittelman, Cunningham-Rathner, Rouleau, & Murphy, 1987). There is a developmental progression in sexual offending behavior (Groth, 1977) that makes early and effective intervention imperative to our society. The developmental psychopathology perspective offers hope for assessment and intervention services aimed at interrupting the progression of sexual deviance into adulthood (Cichetti & Rogosch, 2002). However, sound empirical research is required to point to factors and programs most useful in reducing an individual sexual offender?s risk of sexual offense recidivism. This research study serves as an initial and tentative exploration into the differences between juvenile sexual offenders and other types of juvenile delinquents, and between different groups of juvenile sexual offenders. Conclusion The present study has served to highlight the importance of sub-categorizing juvenile delinquents and juvenile sexual offenders based on both historical/dispositional variables and offense-related characteristics. The importance of assessing each individual juvenile sexual offender in a comprehensive fashion cannot be overestimated. Although considerable evidence was obtained to support future research endeavors that utilize offense-related characteristics in differentiating among groups of juvenile sexual offenders. Many other promising avenues for categorizing juvenile sexual offenders exist. Evidence was provided that identifies juvenile sexual offenders who victimize young children, particularly those with at least one male victim, as possessing behavioral and personality-based tendencies consistent with sexual deviance. These youths were shown to be more frequently exposed to abuse, family instability, violence, sexually 119 explicit material, and sexual abuse victimization. They tended to express a desire to receive treatment to address their sexual preoccupations and extensive sexual abuse perpetration history. However, their levels of shyness, poor self-esteem, decreased self- confidence, and uncertainty calls for the creation of a therapeutic milieu that underlies the treatment modules they are expected to complete. Their past histories, coupled with the extensive and serious nature of their sexual offending behaviors, calls for the completion of a comprehensive treatment and aftercare monitoring program. This group of juvenile sexual offenders, if left untreated, represent a high risk for continuing their pattern of deviant sexual behavior; furthering the cycle of emotional and psychological pain experienced by sexual offense victims. Juvenile sexual offenders who victimized females who were peer-aged, were shown to be similar in many ways to non-sexually offending youths. They were, in general, a highly impulsive group of juvenile delinquents whose level of risk to society extends beyond sexual offense recidivism. These youths also tended to possess multiple risk factors for general criminal recidivism. In addition to sexual offense-specific treatment, these youth are likely to benefit from learning anger management and other emotion-coping lessons, designed to reduce their risk of both sexual and non-sexual criminal recidivism. Strong support was provided for the utility of the Psychopathy Checklist: Youth Version, in distinguishing youths based on their treatment needs and entrenched behavior patterns. Youths scoring high in psychopathy were shown to possess a range of historical and emotional tendencies that place them at high risk for future crimes, psychological deterioration, poor interpersonal relationships, and resistance to treatment. Treatment for 120 these youths must be of a long-term duration, and focused on altering the manner by which they view others, their society, and their futures across multiple life contexts. 121 REFERENCES Abel, G. G., Becker, J. V., Mittelman, M., Cunningham-Rathner, J., Rouleau, J. L., & Murphy, W. D. (1987). Self-reported sex crimes of nonincarcerated paraphiliacs. Journal of Interpersonal Violence, 2, 3-25. Aljazireh, L. (1993). Historical, environmental, and behavioral correlates of sexual offending by male adolescents: A critical review. Behavioral Sciences and the Law, 11, 423-440. Awad, G. A., & Saunders, E. B. (1991). Male adolescent sexual assaulters: Clinical observations. Journal of Interpersonal Violence, 6, 446-460. Becker, J. V. (1988). Adolescent sex offenders. Behavior Therapist, 11, 185-187. Becker, J. V., Cunningham-Rathner, J., & Kaplan, M. S. (1987). Adolescent sexual offenders: Demographics, criminal and sexual histories, and recommendations for reducing future offenses. Journal of Interpersonal Violence, 1, 431-445. Becker, J. V., Kaplan, M. S., Tenke, C. E., & Tartaglini, A. (1991). The incidence of Depressive symptomatology in juvenile sex offenders with a history of abuse. Child Abuse and Neglect, 15, 531-536. Becker, J. V., & Stein, M. (1991). Is sexual erotica associated with sexual deviance in adolescent males? International Journal of Law and Psychiatry, 14, 85-95. Benoit, J. L., & Kennedy, W. A. (1992). The abuse history of male adolescent sex offenders. Journal of Interpersonal Violence, 7, 543-548. 122 Blaske, D. M., Borduin, C. M., Henggeler, S., & Mann, B. (1989). Individual, family, and peer characteristics of adolescent sex offenders and assaultive offenders. Developmental Psychology, 25, 846-855. Brandt, J. R., Kennedy, W. A., Patrick, C. J., & Curtin, J. J. (1997). Assessment of psychopathy in a population of incarcerated adolescent offenders. Psychological Assessment, 9, 429-435. Brannon, J. M., Larson, B., & Doggett, M. (1989). The extent and origins of sexual molestation and abuse among incarcerated adolescent males. International Journal of Offender Therapy and Comparative Criminology, 33, 161-171, Butz, C., & Spaccarelli, S. (1999). Use of physical force as an offense characteristic in subtyping juvenile sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 11, 217-232. Caputo, A. A., Frick, P. J., & Brodsky, S. L. (1999). Family violence and juvenile sex offending: The potential mediating role of psychopathic traits and negative attitudes toward women. Criminal Justice and Behavior, 26, 338-356. Catchpole, R. E. H., & Gretton, H. M. (2003). The predictive validity of risk assessment with violent young offenders: A 1-year examination of criminal outcome. Criminal Justice and Behavior, 30, 688-708. Cicchetti, D., & Rogosch, F. A. (2002). A developmental psychopathology perspective on adolescence. Journal of Consulting and Clinical Psychology, 70, 6-20. Davis, G. E., & Leitenberg, H. (1987). Adolescent sex offenders. Psychological Bulletin, 10, 417-427. 123 Fagan J., & Wexler, S. (1988). Explanations of sexual assault among violent delinquents. Journal of Adolescent Research, 3, 363-385. Fehrenbach, P. A., Smith, W., Monastersky, C., & Deisher, R. W. (1986). Adolescent sexual offenders: Offender and offense characteristics. American Journal of Orthopsychiatry, 56, 225-233. Ford, M. E., & Linney, J. A. (1995). Comparative analysis of juvenile sex offenders, violent nonsexual offenders, and status offenders. Journal of Interpersonal Violence, 10, 56-70. Forth, A. E., Hart, S. D., & Hare, R. D. (1990). Assessment of psychopathy in male young offenders. Psychological Assessment: A Journal of Consulting and Clinical Psychology, 2, 342-344. Forth, A. E., Kosson, D. S., & Hare, R. D. (2003) Manual for the Hare psychopathy checklist: Juvenile version. Toronto, Canada: Multi-Health Systems, Inc. Frick, P. J., O?Brien, B. S., Wootton, J. M., & McBurnett, K. (1994). Psychopathy and conduct problems in children. Journal of Consulting and Clinical Psychology, 103(4), 700-707. Gretton, H. M., McBride, M., Hare, R. D., Shaughnessy, R., & Kumka, G. (2001). Psychopathy and recidivism in adolescent sex offenders. Criminal Justice and Behavior, 28, 427-449. Groth, N. A. (1977). The adolescent sexual offender and his prey. Journal of Offender Therapy and Comparative Criminology, 21, 249-254. Groth, N. A., Longo, R. E., & McFadin, J. B. (1982). Undetected recidivism among rapists and child molesters. Crime and Delinquency, 128, 450-458. 124 Hanson, R. K., & Slater, S. (1988). Sexual victimization in the history of child sexual abusers: A review. Annals of Sex Research, 1, 485-499. Hare, R. D. (1991). The Hare Psychopathy Checklist ? Revised version. Toronto, Canada: Multi-Health Systems. Hastings, T. L., & Kelley, M. L. (1997). Development and validation of The Screen for Adolescent Violence Exposure (SAVE). Journal of Abnormal Child Psychology, 25, 511-521. Herkov, M. J., Gynther, M. D., Thomas, S., & Myers, W. C. (1996). MMPI differences among adolescent inpatients, rapists, sodomists, and sexual abusers. Journal of Personality Assessment, 66, 81-90. Hildenbrand, M., de Ruiter, C., & de Vogel, V. (2004). Psychopathy and sexual deviance in treated rapists: Association with sexual and nonsexual recidivism. Sexual Abuse: A Journal of Research and Treatment, 16, 1-24. Hsu, L. K., & Starzynski, J. (1990). Adolescent rapists and adolescent child sexual assaulters. International Journal of Offender Therapy and Comparative Criminology, 34, 23-30. Jesness, C. F. (1962, 2002). The Jesness Inventory: Manual. Toronto, Canada: Multi -Health Systems, Inc. Kahn, T. J., & Chambers, H. J. (1991). Assessing reoffense risk with juvenile sexual offenders. Child Welfare, 70, 333-345. Katz, R. C. (1990). Psychosocial adjustment adolescent child molesters. Child Abuse and Neglect, 14, 567-575. 125 Kaufman, K. L., & Hilliker, D. R. (1996). Subgroup differences in the modus operandi of adolescent sexual offenders. Child Maltreatment, 1, 17-25. Kavoussi, R. J., Kaplan, M., & Becker, J. V. (1988). Psychiatric diagnoses in adolescent sex offenders. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 241-243. Kempton, T. & Forehand, R. (1992). Juvenile sex offenders: Similar to, or different from, other Incarcerated delinquent offenders? Behavior Research and Therapy, 30, 533-536. Miller, F. G., Renn, W. R., & Lazowski, L. E. (1990). The Adolescent Substance Abuse Subtle Screening Inventory (SASSI): User?s Guide. Baugh Enterprises, Incorporated: Bloomington, IN. Miller, F. G., Renn, W. R., & Lazowski, L. E. (2001). The Adolescent Substance Abuse Subtle Screening Inventory ? Second Edition (SASSI-A2): User?s Guide. Baugh Enterprises, Incorporated: Bloomington, IN. Millon, T. (1993). Millon Adolescent Clinical Inventory Manual. National Computer Systems, Incorporated: Minneapolis, MN. Murphy, W. D., DiLillo, D., Haynes, M. R., & Steere, E. (2001). An exploration of factors related to deviant sexual arousal among juvenile sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 13, 91-103. Nichols, H. R., & Molinder, I. (2001). Multiphasic Sex Inventory: Manual. Nichols and Molinder Assessments: Tacoma, WA. 126 Nisbet, I. A., Wilson, P. H., & Smallbone, S. W. (2004). A prospective longitudinal study of sexual recidivism among adolescent sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 16, 223-234. Oliver, L. L., Nagayama-Hall, G. C., & Neuhaus, S. M. (1993). A comparison of the personality and background characteristics of adolescent sex offenders and other adolescent offenders. Criminal Justice and Behavior, 20, 359-370. Otnow-Lewis, D., Shankok, S. S., & Pincus, J. H. (1979). Juvenile sexual assaulters. American Journal of Psychiatry, 136, 1194-1196. Rasmussen, L. A. (1999). Factors related to recidivism among juvenile sexual offenders. Sexual Abuse: A Journal of Research and Treatment, 11, 69-85. Reynolds, (1987). Reynolds Adolescent Depression Scale: Professional Manual. Psychological Assessment Resources, Incorporated: Odessa, FL. Ryan, G., Miyoshi, T. J., Metzner, J. L., Krugman, R. D., & Fryer, G. E. (1996). Trends in a national sample of sexually abusive youthss. Journal of the American Academy of Child and Adolescent Psychiatry, 35, 17-25. Seghorn, T. K., Prentky, R. A., & Boucher, R. J. (1987). Childhood sexual abuse in the lives of sexually aggressive offenders. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 262-267. Seto, M. C., & Barbaree, H. E. (1999). Psychopathy, treatment behavior, and sex offender recidivism. Journal of Interpersonal Violence, 14, 235-249. Smith, W. R. (1988). Delinquency and abuse among juvenile sexual offenders. Journal of Interpersonal Violence, 3, 400-413. 127 Smith, W. R., & Monastersky, C. (1986). Assessing juvenile sexual offenders? risk for Reoffending. Criminal Justice and Behavior, 13, 115-140. Soothill, K., Francis, B., Sanderson, B., & Ackerley, E. (2000). Sex Offenders: Specialists, generalists, or both. British Journal of Criminology, 40, 56-67. Spaccarelli, S., Bowden, B., Coatsworth, J. D., & Kim, S. (1997). Psychosocial correlates of male sexual aggression in a chronic delinquent sample. Criminal Justice and Behavior, 24, 71-95. Van Ness, S. R. (1984). Rape as instrumental violence: A study of youth offenders. Journal of Offender Counseling, Services, and Rehabilitation, 9, 161-170. Vinogradov, S., Dishotsky, N. I., Doty, A. K., & Tinklenberg, J. R. (1988). Patterns of Behavior in adolescent rape. American Journal of Orthopsychiatry, 58, 179 -187. Weinrott, M. R., & Saylor, M. (1991). Self-report of crimes committed by sex offenders. Journal of Interpersonal Violence, 6, 286-300. Worling, J. R. (1995). Sexual abuse histories of adolescent male sex offenders: Differences on the basis of the age and gender of their victims. Journal of Abnormal Psychology, 104, 610-613.