Types and Classifications of Male Adolescents Who Have Committed Sex Offenses

Although a variety of characteristics have been identified among juveniles who have sexually offended, few studies have attempted to classify these juveniles according to their similarities and differences. Weinrott (1996) noted that even though it is widely accepted that juveniles who have abused young children differ from those who have sexually assaulted peers, most studies of juvenile sex offenders combine these groups. Furthermore, in spite of alarming statistics pertaining to the prevalence of juvenile sex offending, sexual recidivism rates for juvenile sex offenders typically are very low (Weinrott, 1996). The apparently low recidivism rate suggests that there may be a significant subgroup of identified juvenile sex offenders who do not continue to commit sex offenses as adults (Knight and Prentky, 1993). Such a finding would be consistent with the literature on juvenile delinquents (Tolan and Gorman-Smith, 1998). Yet, studies investigating this probability are lacking. Research on juveniles who commit sex offenses that differentiates these juveniles according to their various behavior patterns, cognitive and emotional functioning, and other relevant factors is needed to determine and apply appropriate and effective treatment strategies.

Becker and Kaplan (as cited in Becker, 1998) proposed that an initial sex offense by a juvenile results from a combination of individual factors such as a lack of social skills, family factors such as familial relationships, and social-environmental factors such as social isolation. They suggested that three paths are possible after the initial offense: a dead end (no further crimes), a delinquency path, and a sex interest path involving continued sex offending and, frequently, the development of deviant sexual arousal patterns. Becker (1998) pointed out that this hypothesized model, like other classification models, has not been empirically validated.

In 1986, O'Brien and Bera (as cited in Weinrott, 1996) grouped juvenile sex offenders into the following seven categories:

• Naive experimenters.
• Undersocialized child exploiters.
• Sexual aggressives.
• Sexual compulsives.
• Disturbed impulsives.
• Group influenced.
• Pseudosocialized.

Although this classification scheme has been described as having much "face validity" (Weinrott, 1996) and has been recommended to facilitate interventions and treatment (e.g., Avalon Associates, 1986), systematic investigations of its reliability and validity are lacking. Some indirect support for the O'Brien and Bera (as cited in Weinrott, 1996) classification scheme comes through the work of Knight and Prentky (1993). These researchers reported that four of the O'Brien and Bera types overlap with factors supported by the empirical literature and by their own research with adult offenders who committed sex offenses as juveniles.

Knight and Prentky (1993) compared adult sex offenders who had official records of juvenile sex offending with those who did not. They also compared a third group of "hidden juvenile sex offenders"—individuals who reported they had committed sex offenses as juveniles but who did not have official records of such offenses. Their findings indicated that certain factors, such as low social competence and high rates of antisocial behavior and impulsivity, differentiated sex offenders who began offending as juveniles from those who did not. These factors also are significant in differentiating types of adult sex offenders. For example, of the nine rapist types, three—low social competence/opportunistic, low social competence/nonsadistic/sexual, and low social competence/vindictive—have low social competence as a defining characteristic. Combined, these findings suggest that the sex offender classification schemes validated by Knight and Prentky (1993) for adults may also be useful for differentiating juvenile sex offenders. The authors noted that social competence also is an important factor in the child molester typology, along with factors such as degree of sexual preoccupation and amount of contact with children.

Graves (as cited in Weinrott, 1996) conducted a meta-analysis of 140 samples involving 16,000 juvenile sex offenders. Results suggested three typologies: pedophilic, sexual assault, and undifferentiated. Pedophilic juveniles tended to lack social confidence and to be socially isolated, consistently molested much younger children (at least 3 years younger than themselves), and typically molested girls. The sexual assault group typically assaulted peers or older females. The undifferentiated group committed a variety of offenses, and the ages of their victims varied widely. This latter group engaged in hands-off offenses (e.g., exhibitionism) in addition to hands-on assaults. Compared with the other two groups, they began their abusive behavior when they were younger, had the most severe social and psychological problems, were more antisocial, and had more dysfunctional families.

Prentky et al. (2000) employed a rationally derived classification system to describe their sample of male juvenile sex offenders. They used the following six categories: child molesters (69 percent of the sample), rapists (12.5 percent), sexually reactive children (6.25 percent), fondlers (3 percent), paraphilic offenders (3 percent), and unclassifiable (6.25 percent). In the child molester category, all victims were under age 12 and offenders were at least 5 years older than victims. In the rapist category, all victims were age 12 or older and the age difference between offenders and their youngest victims was less than 5 years. Sexually reactive children were under age 11, as were their victims. In the fondler category (as in the rapist category), all victims were age 12 or older and the age difference between offenders and their youngest victims was less than 5 years; sexual acts in this category were limited to fondling, caressing, or frottage (i.e., touching or rubbing against a nonconsenting person for sexual arousal). Paraphilic offenders had no physical contact with their victims; acts included, for example, exhibitionism and obscene phone calls. Prentky et al. (2000) reported that these categories also were used by Becker and Kaplan, who found similar proportions of offenders in each category (with the exception of a somewhat higher proportion of rapists).

Weinrott (1998a) suggested four general types of juveniles who have sexually abused others. Three of these types are those who are juvenile delinquents in general, those who have deviant arousal, and those who are psychopathic offenders. The other type includes juveniles who fit none of these categories and may only require limited interventions, such as those that establish appropriate rules for future sexual behavior.

Malamuth's research with college students (as cited in Miner and Crimmins, 1995) suggested that sexual aggression resulted from the interaction of two pathways: hostile masculinity and sexual promiscuity. Hostile masculinity involves beliefs that to be male involves taking risks; being powerful, tough, dominant, competitive, and aggressive; and defending one's honor. The sexual promiscuity pathway reflects age at first intercourse and number of sexual partners since the age of 14. High scores on both pathways were associated with high rates of sexual aggression against women.

Sibling Incest

Few reports have specifically addressed issues pertaining to sibling incest. Araji (1997) noted that although sibling incest appears to be quite prevalent, often it is underreported and ignored. Various factors probably contribute to this apparent tendency to minimize the incidence of sibling abuse. For example, in contrast to extrafamilial sexual abuse, parents may be especially reluctant to report to authorities that one of their children has sexually abused another child in their home.

O'Brien (1991) emphasized the importance of "Taking Sibling Incest Seriously" with the title of his paper. He compared 170 juveniles who sexually offended against siblings (including stepsiblings, half siblings, and adoptive siblings) with extrafamilial offenders (those who offended against children other than their siblings), those who sexually victimized peers or adults, and those whose victims may have included a mix of sibling and extrafamilial children and/or peers and adults.⁴ As a group, sibling offenders perpetrated the greatest number of abusive acts (an average of 18 incidents, compared with 4.2 for extrafamilial offenders, 7.4 for peer/adult offenders, and 8.5 for the mixed group). The duration of sex offending was greatest for sibling offenders. Nearly 45 percent of the sibling offenders had been committing offenses for more than a year, whereas only 23 percent of the extrafamilial offenders and 24 percent of the adult/peer offenders had been offending for this long. In addition, the sibling offenders were more likely than the others to vaginally or anally penetrate their victims (46 percent, compared with 28 percent of the extrafamilial offenders and 13 percent of the adult/peer offenders). Sibling offenders also were more likely to have multiple victims. O'Brien hypothesized that specific issues, such as victim availability, the nature of the sibling relationship, and other factors, may have contributed to such serious offending histories. In spite of such abusive behaviors, only about one-third of the sibling offenders had court-ordered treatment, compared with three-quarters of the other offenders.

Family factors such as an increased rate of physical abuse were noted among the sibling offenders (61 percent, compared with 45 percent of the extrafamilial offenders and 37 percent of the adult/peer offenders). Sibling offenders also were sexually abused more frequently by their fathers than were other offenders, although only a small number of the sibling offenders had been sexually abused by any family member (including fathers). Interestingly, 36 percent of the sibling offenders' mothers and 10 percent of their fathers had been victims of sexual abuse as children, compared with 9.1 percent of the extrafamilial offenders' mothers and 5.5 percent of their fathers. Assessments also suggested that the rate of family dysfunction was higher for sibling offenders than for the other groups.

An early study of inner-city minority juveniles from low socioeconomic backgrounds (Becker et al., 1986) questioned the existing assumption that sibling offenders are significantly different from other juvenile sex offenders. Becker et al. noted that 9 of the 22 adolescents in their small sample also had evidenced nonsibling paraphilic behaviors. In view of the O'Brien (1991) study that found significant group differences between sibling and nonsibling sex offenders, it may be that, as in any group of sex offenders, juveniles who perpetrate sibling abuse are a heterogeneous mix. In fact, Becker and her colleagues noted that the juveniles in their sample included adolescents who engaged in consensual sexual behavior with a peer-aged relative, those whose sexual activity with a peer-aged relative began as consensual but became coercive when the relative withdrew consent, those who had developed deviant sexual interests, and those who engaged in both nondeviant and deviant sexual behavior.

In their discussion of sibling abuse, Bonner and Chaffin (1998) asserted that most interventions designed to address sibling sexual behavior assume a victim-perpetrator model. They noted that such a model may be appropriate when the sexual behavior has been abusive but cautioned that it is progressively less appropriate (and may be damaging) when sibling cases involve inappropriate mutual sexual behavior or, especially, age-appropriate sex play.

Girls Who Have Committed Sex Offenses

Incidence

Before 1986, references in the professional literature to female sex offenders were few and limited. Since then, some references have appeared, but research studies continue to be few, and studies of adolescent girls are relatively rare. Existing studies often are limited by small sample sizes and retrospective analysis of selected populations that may not be representative of the general population.

Lane and Lobanov-Rostovsky (1997) reviewed the literature on young sex offenders and found that adolescent female sex offenders represented between 2 and 3 percent of juveniles involved in two different treatment programs. These authors also cited the results of several statewide incidence studies conducted in the 1980's. The studies revealed that females represented 5 percent (19) of the juveniles arrested for sex offenses in Oregon in 1985, 8 percent (12) of the children identified as adolescent sex offenders by the Vermont Social Rehabilitation Services Department or Corrections Department in 1984, and 7 percent of the juveniles referred to juvenile court in Utah over a 5-year period. In a Maine study (Righthand, Hennings, and Wigley, 1989), females represented 11 percent (40) of the 348 juveniles identified as sex offenders by the Maine Departments of Human Services and Corrections during a 12-month period between 1988 and 1989.

In a more recent study by the Washington State Department of Social and Health Services, English and Ray (as cited in Araji, 1997) found that of 200 juveniles identified as sexually aggressive, 9.3 percent of those age 13 or older were female, compared with 19.1 percent of those age 12 and under. This relatively high rate of sex offending by young girls also was found by Johnson (as cited in Lane and Lobanov-Rostovsky, 1997). Girls who had been sexually abusive made up 21.6 percent of the children in her program for children ages 4 to 12 who had engaged in inappropriate sexual behavior. Gray et al. (1997) also found a relatively high rate of preadolescent girls who evidenced sexual behavior problems. In their sample, 35 percent of the children who evidenced such problems were girls.

Incidence reports on juvenile sex offenses may underestimate the extent of the problem for female offenders even more than for male offenders. Underestimates may occur because there is a general tendency to underreport sex crimes committed by females (Charles and McDonald, 1997). It has been hypothesized that this underreporting might result from a societal reluctance to acknowledge that girls are capable of committing criminal offenses, particularly sex offenses; even professionals may be reluctant to report female disclosure of sex offenses (Travin, Cullen, and Protter, 1990).

Characteristics of Female Offenders and Their Offenses

Ray and English (1995) compared girls and boys who were described as sexually aggressive and who were actively involved with their State's public social service agency. Findings indicated that the girls tended to be younger than the boys and were less likely to have perpetrated acts of rape. (Rape was defined as involving force or no consent and vaginal, oral, or anal penetration with a penis or object.) Approximately 94 percent of the girls in the sample had been victims of sexual abuse, compared with 85 percent of the boys. A greater percentage of girls than boys (94 percent versus 86 percent) had experienced multiple types of abuse, including sexual abuse, physical abuse, emotional abuse, and/or neglect.

All of the children in the Ray and English sample (1995) evidenced a wide range of behavior problems while under the State agency's supervision. Girls were significantly more likely than boys to steal and display temper tantrums. There also was evidence that girls were more likely to be truant. Girls appeared to have more adequate social skills and more empathy toward their victims, whereas boys tended to be more coercive and sophisticated in their sex offending. Use of sexual aggression appeared to be escalating more in boys than in girls. Another noteworthy difference was that although approximately one-third of all the juveniles studied were legally charged with an offense, only 2 girls (as contrasted with 93 boys) were charged. The study also found that girls were significantly more likely than boys to receive assessment and treatment for their experiences of being abused.

Fehrenbach and Monastersky (as cited in Bumby and Bumby, 1997) found that, in their sample, most adolescent girls who sexually victimized children age 12 or younger frequently did so while engaged in a childcare situation. In their sample, 53.6 percent of the adolescent girls committed some form of penetration (oral, anal, or vaginal intercourse or other forms of penetration); 46.4 percent engaged in fondling.

Hunter et al. (as cited in Bumby and Bumby, 1997) conducted a descriptive study of 10 girls who had sexually offended and who were in a residential care program for juveniles with emotional and behavioral problems. The study, although limited by the small sample size and lack of a comparison group, is informative. The girls had high rates of previous mental health services (80 percent). Many had a history of suicide attempts or ideation (60 percent), running away (60 percent), substance abuse (40 percent), enuresis (bed wetting, 40 percent), and/or learning disabilities (40 percent).

All of the girls reported a history of sexual victimization experiences, including the following:

• All had been sexually abused by more than one offender; the number of offenders ranged from two to seven.

• Experience of victimization began at early ages, ranging from 1 to 8 years; the median age was 4.5 years.

• All of the girls reported being sexually abused by a male; 60 percent also reported being abused by a female.

• Ninety percent of the girls reported actual or attempted vaginal penetration, 60 percent reported actual or attempted anal penetration, 70 percent reported having oral sex performed on them, and all reported being fondled.

• Ninety percent reported that force was used in their sexual victimization experiences, yet 80 percent reported that they experienced some sexual arousal during at least one of their experiences as a victim.

• They typically victimized younger children; victim ages ranged from 1 to 13 years, with a median age of 5.5 years.

• Their victims most frequently were strangers (39.4 percent); other victims were siblings (30.3 percent), other relatives (18.2 percent), and acquaintances (12.1 percent).

• They had fantasies about the deviant sexual behavior (in 89 percent of the cases).

• Their sexually offensive behavior included vaginal intercourse (70 percent), anal intercourse (10 percent), oral sex (70 percent), and fondling (100 percent).

Most of the girls in this sample also engaged in nonsexual delinquent behaviors such as stealing and physical aggression. Most had not been formally charged for these behaviors.

The findings of the above studies are fairly consistent with Bumby and Bumby's (1997) findings from their sample of 12 adolescent female offenders who were inpatients at a psychiatric facility for children and adolescents with emotional and behavioral disorders. Again, the girls in this sample tended to select young victims. Most often their victims were family members (75 percent). In contrast to the sample studied by Hunter et al. (as cited in Bumby and Bumby, 1997), none of these girls victimized strangers. Eleven of the twelve girls perpetrated their offenses when providing childcare.

A review of the characteristics of the girls in the Bumby and Bumby (1997) sample indicated that most (83 percent) experienced academic difficulties, although all but three fell within the average intellectual range; all but one had peer difficulties at school; and two-thirds had been suspended or expelled for physical aggression toward peers or teachers or for other causes. Behavior problems were common: 75 percent had abused alcohol, 58 percent had abused drugs, 58 percent had run away from home, 58 percent had been truant from school, and 33 percent had been arrested for stealing. Psychiatric diagnoses included conduct disorders, oppositional-defiant disorder, major depression, posttraumatic stress disorder, adjustment disorder, and chemical dependency. Most (83 percent) had received previous mental health services, 33 percent had histories of self-mutilation, and 58 percent had attempted suicide. Anger control problems were described as common (67 percent), as were low self-concepts (100 percent). Peer relations were very strained; 75 percent of the girls were described as significantly socially isolated, which possibly was related to their high rate of aggressive behavior toward peers (67 percent). A significant number (58 percent) were described as sexually promiscuous, having had many sexual relationships with older males. All of these girls had been sexually victimized themselves. They tended to have been sexually abused by more than one person. Seventy-five percent had been physically abused, and 42 percent had experienced emotional or physical neglect. In general, their families were described as dysfunctional and chaotic.

To provide additional information about their adolescent sample, Bumby and Bumby (1997) compared 18 female sex offenders to a group of female nonoffenders, male sex offenders, and male nonoffenders of similar age. All juveniles were inpatients at a psychiatric facility for children and adolescents with emotional and behavioral disorders. Psychological test results suggested that the adolescent female sex offenders experienced a number of psychological symptoms and difficulties. They had higher scores than the female nonoffenders on the psychopathic deviate, paranoia, and psychasthenia scales of the Minnesota Multiphasic Personality Index-Adolescent (MMPI-A). They did not, however, differ significantly from the male sex offenders and male nonoffenders.

The female sex offenders evidenced significantly more symptoms of anxiety and depression (including suicidal thoughts and behaviors) than the female nonoffenders but did not differ from the male sex offenders in this regard. The female sex offenders had higher rates of academic failure (having been retained for one grade in school) and truancy than the male sex offenders but did not differ from the female or male nonoffenders on these measures. Although delinquent behaviors, socially inappropriate behaviors, and status offenses were frequent among female sex offenders, the frequency did not differ significantly from that found in the other groups, with the exception that female sex offenders had higher rates of drug abuse and sexual promiscuity than the male sex offenders. In addition, although high rates of childhood sexual victimization occurred across all groups, the female sex offenders experienced significantly more sexual abuse than the other groups.

Again, these findings are informative but are limited by the small sample size. The authors point out, however, that their findings are consistent with the limited available information. The authors suggest that factors such as depression, suicidal ideation, anxiety, poor self-concept, and childhood sexual victimization are targets for assessment, and possibly treatment, in girls who commit sex offenses.

In perhaps the largest study to date, Mathews, Hunter, and Vuz (1997) compared 67 girls who were referred to community-based treatment or residential treatment subsequent to histories of sex offending with 70 boys who also had such histories. Because the samples did not meet scientific standards of comparability, statistical tests of differences between groups could not be used. A review of the findings, however, suggested some meaningful similarities and differences between the girls and the boys. The girls' offending behaviors were similar to the boys' in terms of offense types and style of victim selection. For example, both girls and boys committed the following types of offenses: fondling (77.6 percent girls, 75.4 percent boys), oral sex (47.8 percent girls, 29.7 percent boys), and vaginal or anal intercourse (26.9 percent girls, 54.5 percent boys). Also like the boys, the girls tended to victimize young children of the opposite gender. In contrast to the boys, and consistent with other studies, the girls typically had more severe victimization experiences themselves. These abusive experiences were characterized by a higher average number of perpetrators, younger age at the time of first victimization, and greater likelihood of having been a focus of their perpetrator's aggression. Girls also were three times more likely than boys to have been victimized by female perpetrators. Like boys, however, the girls' victimization by a perpetrator of the same gender seemed related to the girls' having sexual identity problems.

Other findings in Mathews, Hunter, and Vuz (1997) indicated that in addition to experiencing high rates of abuse and trauma, the girls in this study typically came from families evidencing high levels of dysfunction and an absence of parental support. Their family environments usually appeared detrimental for the development of healthy attachments and a positive sense of self. Although a small subgroup of the girls evidenced little psychopathology and limited offending behaviors, about one-third of the outpatient girls in the study evidenced mild to moderate levels of psychopathology, and about half of the entire sample appeared to have moderate to severe psychopathology. Problems included behaviors associated with conduct disorders, impulsivity, substance abuse, suicidal behaviors, and unprotected sex. A subgroup of the girls also evidenced deviant sexual arousal patterns, posttraumatic stress disorder, depression, and anxiety. In sum, the authors concluded:

Overall, the data from this study seem consistent with the authors' impression that biological and socialization factors create a higher threshold for the externalization of experienced developmental trauma in females than males. In this regard, it may be that females are generally less likely than males to manifest the effects of maltreatment in the form of interpersonal aggression or violence and that females who develop such patterns of behavior are generally those who have experienced remarkably high levels of such developmental trauma in the absence of environmental support for recovery and the presence of healthy female role models. (p. 194)

Young Children Who Have Committed Sex Offenses

Incidence

In the 1980's, after the problem of adolescent sex offending gained attention, similar behaviors in preadolescent and younger children also were recognized. Knopp (as cited in Araji, 1997) observed that the 1980 Uniform Crime Reports identified 208 children under the age of 12 who were arrested for rape. Thirty-seven of these children were age 10 or younger. Knopp found somewhat higher rates for 1979: in that year, 249 children under the age of 12 were arrested for rape; 66 of these children were age 10 or younger. The Uniform Crime Reports stopped reporting age ranges in 1980 (Araji, 1997).

Recent surveys of children with sexual behavior problems (including nonajudicated children) reveal substantially higher rates of sexually abusive behavior by preadolescent children than the rates cited in the Uniform Crime Reports. For example, English and Ray (as cited in Araji, 1997) reported that the Washington Department of Social and Health Services had 641 active cases of children under age 12 who had raped, molested, or engaged in noncontact sexual acts such as exposing, masturbating in public, or peeping. Gray and Pithers' Vermont studies (as cited in Araji, 1997) identified 200 children under age 10 who had sexually abused others between 1984 and 1989; even more striking, they identified 100 children who had sexually abused others in a single year, 1991. In addition, in a sample of 616 juveniles who had been referred for evaluation or treatment after the age of 12 for committing a sex offense, 25.9 percent had been sexually abusive prior to their 12th birthday (Ryan et al., 1996). This apparent increase in the rate of preadolescent children who evidence sexually abusive behaviors probably reflects a greater awareness of the problem.

In an extensive review of the published and unpublished literature pertaining to children who have been sexually aggressive, Araji (1997) stressed that research in this area is in its infancy and noted that many findings are simply clinical observations. Araji's point has relevance for the findings presented in this Report: because this area of research is so new, the findings presented must be considered preliminary and interpreted with caution.

Individual Characteristics

With this caution in mind, available studies (Araji, 1997) have reported that preadolescent children who have been sexually aggressive include children as young as 3 and 4, although the most common age of onset appears to be between 6 and 9. Contrary to findings regarding adolescent children who have committed sex offenses, girls were represented in much greater numbers among preadolescents who have sexually abused. Furthermore, these girls had often engaged in behaviors that were just as aggressive as the boys' actions. The number of reported victims for these preadolescent children ranged from one to nine; many had multiple victims. Victims tended to be quite vulnerable. They generally were young (averaging between ages 4 and 7); typically were siblings, friends, or acquaintances; and most often were female.

Studies generally have found high rates of sexual victimization histories among preadolescent children who have sexually abused: 50–75 percent of the boys and 100 percent of the girls in studies that provided this information by gender (Johnson, as cited in Araji, 1997; Friedrich and Luecke, as cited in Araji, 1997; Araji, Jache, Tyrrell, and Field, as cited in Araji, 1997; Araji, Jache, Pfeiffer, and Smith, as cited in Araji, 1997; Bonner, Walker, and Berliner, as cited in Araji, 1997; Pithers et al., 1998b). English and Ray (as cited in Araji, 1997) found that preadolescent children who have sexually abused have significantly higher rates of abuse and neglect victimization experiences than their adolescent counterparts. Furthermore, Friedrich and Luecke (as cited in Araji, 1997) also found severe sexual victimization experiences among sexually aggressive children when contrasted with two samples of children who were not sexually aggressive (one with a history of sexual victimization and one without). The children who were sexually aggressive experienced more severe types of sexual abuse that generally involved genital contact and penetration. Research by Friedrich and Luecke (as cited in Araji, 1997) and Pithers and Gray (as cited in Araji, 1997) also suggests that the children who engaged in sexually aggressive behaviors frequently experienced academic and learning difficulties and impaired peer relationships.

Family Characteristics

Studies described by Araji (1997) also suggest that the families of children who engaged in sexually aggressive behavior tended to be characterized as dysfunctional, evidencing high rates of parental separation, domestic violence, substance abuse, highly sexualized environments (e.g., exposing children to sexual activity, pornography, and both covert and overt sexual abuse), unsatisfactory role models, poor parent-child relationships, parental histories of childhood abuse, and so on. After reviewing the available research, Araji concluded, "The evidence . . . points to family interactions as a primary source of the problem" (p. 87).

The importance of family factors is supported by research conducted by Pithers et al. (1998a) concerning the caregivers of children with sexual behavior problems. These researchers used a structured interview and standardized measures to investigate the characteristics of these caregivers. Findings indicated that the caregivers and their families experienced much stress. Of the 72 children in the study (75 percent of whom resided with biological parents and 25 percent with foster parents), 38 percent resided in families whose income fell below the Federal poverty level (defined as a family of four or more with an annual income of less than $15,000). Comparisons between biological families and foster families revealed that 72 percent of the biological families and 28 percent of the foster families had incomes below the poverty level. Families also had a high rate of single parenting: approximately half of the parents (51.4 percent) were living with a partner.

The family environments of these children, particularly their biological families, were characterized as disorganized and as requiring much effort to meet the basic needs of the family. The families had a high rate of sexual abuse histories. Most families (72 percent) included at least one sexual abuse victim (other than the child being studied), and more than half of the extended families (62 percent) included at least one person (other than the child being studied) who had perpetrated sexual abuse. Sexual abuse victims of the children studied typically were relatives (94 percent). Very few of these children assumed responsibility for their sexually abusive behaviors (10.3 percent).

More than half of the children studied had witnessed domestic violence in the families with whom they were currently residing. Most witnessed violence between their biological parents (70.2 percent). Some observed partner violence in their foster homes (20 percent). In general, foster families seemed to provide more functional environments, experienced less conflict, and were more cohesive.

The individual functioning of the female caregivers⁵ was measured with the Brief Symptom Inventory (Derogatis, as cited in Pithers et al., 1998a; Derogatis and Spencer, as cited in Pithers et al., 1998a). The results suggested that, as a group, these women were significantly more psychologically distressed than most people in the general population. The biological parents evidenced significantly more distress than the foster parents. Parenting stress among female caregivers was measured by the Parenting Stress Inventory (Abidin, as cited in Pithers et al., 1998a). Both biological and foster parents appeared to experience significant parenting stress that warranted referrals for professional care. Again, the biological parents evidenced significantly more stress than the foster parents. Both groups cited their children as a major source of their parenting stress, and both groups evidenced impaired attachments to their children. In spite of these findings, parents typically denied having problems associated with parenting and appeared defensive about some personal difficulties.

Comparative Studies of Preadolescents and Adolescents Who Have Committed Sex Offenses

In one of the few existing comparative studies involving children who committed sex offenses, English and Ray (as cited in Araji, 1997) studied 271 juveniles who sexually offended by comparing the preadolescents (32.8 percent) with the adolescents (67.2 percent). Although the researchers found many similarities between the groups (e.g., previous aggressive behavior, psychiatric problems, and levels of intellectual functioning), the adolescents evidenced significantly higher rates of aggression and coercion and greater sophistication in committing their sex offenses. The older juveniles also were less empathic, were more likely to minimize the seriousness of their abusive behavior, and evidenced more escalating sexual violence. The adolescents also had higher rates of depressive symptoms and suicidal gestures. As Araji suggested, this latter difference may reflect developmental differences between the groups, as the older juveniles may have begun to internalize their difficulties in addition to expressing them outwardly.

English and Ray (as cited in Araji, 1997) also found that both groups had a moderate to moderately high number of risk factors that were considered by the authors to be associated with repeat offending. (The authors evaluated 32 risk factors in 3 categories: family and environment, juvenile characteristics, and victim characteristics.) The preadolescent children's families, however, evidenced significantly more family violence, anger management difficulties, blurred boundaries regarding the privacy of family members, family abuse histories, and parental problems coping with the child's alleged sexual misconduct. In addition, the younger group had significantly higher levels of social isolation and current life stresses.

Types and Classifications

Although, as noted below, some research studies have substantially advanced the body of knowledge about younger children who are sexually abusive and their difficulties, longitudinal studies following these children over time are lacking. Thus, it is not known whether childhood sexual behavior problems continue or, more accurately, which children persist in their sexual misconduct in adolescence or through adulthood.

Children who have sexual behavior problems are a heterogeneous group. Descriptions of these children typically differentiate normative sexual behavior exhibited by children from a continuum of progressively more excessive and abusive sexual behaviors (Araji, 1997; Johnson, 1991).

For example, Johnson (1991) described children referred for evaluation or consultation because of reported sexual acting-out behavior and identified four groups: normal sexual exploration, sexually reactive, extensive mutual sexual behaviors, and child perpetrators. Factors that distinguish these groups are as follows:

• Normal sexual exploration is an "information-gathering process" that involves children looking at and touching each others' bodies and trying out gender roles. The sex play is voluntary and typically involves same-age children. It usually is spontaneous and light hearted.

• Sexually reactive children have been sexually abused, have been exposed to pornography, and/or live in highly sexualized households. The behaviors of these children include exposing, touching the genitals of other children or adults, self-stimulating genitals or inserting objects, and so on. The emotions associated with these behaviors may reflect confusion and shame.

• The children in the extensive mutual sexual behavior group participate in extensive sexual behaviors on a continuous basis, including oral sex, vaginal intercourse, and anal intercourse. They do not appear to experience anxiety, guilt, shame, or confusion, and they evidence little desire to stop. The sexual activity is mutual; there is no offender or victim. Most of these children have previously been sexually abused. Sometimes their sexual behavior appears as a coping strategy in very chaotic, dysfunctional, and/or sexually abusive families. Some of these children have been placed in multiple foster homes and appear to cling to each other in this sexual way to assuage their feelings of fear and loneliness.

• The child perpetrator group includes children who engage in impulsive, compulsive, and aggressive sexual behavior. The sexual behaviors are not mutual and involve coercion, trickery, bribery, and force. The children in this group often associate feelings of anger and aggression (and sometimes rage) with sex. Other feelings associated with sex include fear, loneliness, or abandonment. These children typically have been exposed to high levels of sexual violence (including incest), promiscuity, pornography, and sexualized relationships.

In what appears to be the first attempt to develop empirically derived and clinically relevant classifications of children with sexual behavior problems, Pithers et al. (1998b) studied a sample of 127 children ages 6 to 12 who had evidenced sexual behavior problems. The authors defined "problematic" as sexual behaviors that were "(a) repetitive; (b) unresponsive to adult intervention and supervision; (c) equivalent to adult criminal violations; (d) pervasive, occurring across time and situations; or (e) highly diverse, consisting of a wide array of developmentally unexpected sexual acts" (p. 386).

Pithers et al. (1998b) found that children who evidenced sexual behavior problems varied significantly on several factors, including historical, demographic, behavioral, and diagnostic factors. They also varied according to number of victims, degree of aggression used during the sexual abuse, sexual penetration, psychiatric diagnoses, and internalizing and externalizing behaviors. Five subtypes were identified: sexually aggressive, nonsymptomatic, highly traumatized, abuse reactive, and rule breaker. Factors that distinguish these subtypes are as follows:

• The sexually aggressive children tended to have the highest rates of conduct disorder diagnoses. They were more likely to penetrate their victims and less often were victims of sexual or physical abuse themselves.

• The nonsymptomatic children were, as the classification name implies, within the normal range on most test measures. They typically did not have psychiatric diagnoses, evidenced low levels of aggression in their sexual behaviors, and had the fewest victims. These children were some of the most likely children to have in their extended family persons who had perpetrated sexual abuse.

• Both the highly traumatized children and the abuse reactive children typically were among the youngest and had the highest average number of victims. These two groups of children also had been victimized by the greatest number of sexual and physical abuse perpetrators.

• The highly traumatized children had the highest incidence of psychiatric diagnoses and posttraumatic stress disorders. Their parents were more likely than other parents to report feeling less attached to their children.

• The abuse reactive children had the shortest time between their own personal victimization experiences and the onset of their abuse against others. They experienced a high level of maltreatment and had a high number of sexual abuse perpetrators. This group had a high incidence of psychiatric diagnoses and the highest incidence of oppositional defiant disorders. They occasionally used aggression during their offenses.

• The rule-breaking group included a higher number of girls and had a greater time lag between their own victimization experiences and the onset of their abuse against others. These children had higher levels of sexualized and aggressive behaviors and also were more likely to act out in nonsexual ways. They had the highest number of sexual abusers within their extended families.

Across all five subtypes, certain factors were found to be associated with the number of victims abused by these children. The children who themselves had been abused by more perpetrators and the children who had impaired attachments with their parents had greater numbers of victims.

Juveniles With Developmental Disabilities and Mental Retardation Who Have Committed Sex Offenses

In one of the few studies focusing on adolescent sex offenders with mental retardation, Gilby, Wolf, and Goldberg (1989) compared sexual behavior problems in a sample of intellectually normal (defined by the authors as borderline intellectual functioning or higher) and mentally retarded (including mild and moderate mental retardation) adolescents. The sample included both outpatient and inpatient adolescents at an assessment and treatment center for children and adolescents. The authors found that the frequency of sexual behavior problems of the groups studied did not differ significantly according to their levels of intellectual functioning. They noted that, for both the "intellectually normal" and "mentally retarded" groups, the closer the adolescent was observed (e.g., within a residential setting), the greater the number of sexual behavior problems recorded. This finding was especially true for the mentally retarded inpatient group. The authors suggested that reports of a greater-than-expected number of sexual problems among persons with mental retardation may be related to the increased levels of supervision these individuals receive.

Gilby, Wolf, and Goldberg (1989) found increased levels of inappropriate, nonassaultive sexual behavior (e.g., exhibitionism and public masturbation) among the adolescents with mental retardation. Although the rate of sexual assault did not vary between the intellectually normal and mentally retarded groups, there were fewer "consented to" sexual activities among the mentally retarded outpatient group. The authors suggested that this difference could reflect a lack of opportunity. The authors also noted that sexual activity was frequent in both groups of adolescents once they were placed in residential settings. The adolescents with mental retardation, however, reportedly were more indiscriminant in their sexual activity: they were more likely to engage in both homosexual and heterosexual activities, whereas the adolescents with normal intellectual functioning were more likely to engage exclusively in either heterosexual or homosexual activity.

Analysis of offense patterns in the intellectually normal and mentally retarded groups revealed that both groups engaged in both consensual sexual behavior and assaultive and other inappropriate sexual behaviors. Adolescents with mental retardation, however, had a higher rate of sexual assaults against peers and were less likely to know their victims. Adolescents with normal intellectual functioning selected female victims more often, whereas those with mental retardation were equally likely to select male and female victims.

The Gilby, Wolf, and Goldberg (1989) study is informative. Additional research is needed, however, to determine whether the findings in this study can be generalized to other juveniles with mental retardation who have committed sex offenses.

Likelihood of being sexually victimized may be a special issue among juveniles with mental retardation and other developmental disabilities. Cowardin (as cited in Stermac and Sheridan, 1993) reported that developmentally disabled persons are four times more likely than nondisabled individuals to be sexually abused. Also, individuals with developmental disabilities usually are not encouraged to date and marry or to express their sexual needs (Brantlinger, as cited in Stermac and Sheridan, 1993) and typically are relatively uneducated about sexual matters (Edmondson, McCombs, and Wish, as cited in Stermac and Sheridan, 1993).

Juveniles Who Have Committed Sex Offenses Versus Other Types of Offenses

Given limited resources (funding and availability of treatment programs and providers) and reported similarities between juvenile sex offenders and other juvenile delinquents, the question arises as to how extensive the differences are between individual sex offenders and between sex offenders as a group and other juveniles who have been abused and traumatized and have had very difficult lives. Are juveniles who have committed sex offenses a distinct group in need of specialized intervention, or can their needs be best met through interventions that are effective with juveniles who have committed other types of offenses?

Again, research is limited. Available studies, however, suggest that juveniles who commit sex offenses and juveniles who commit other types of offenses share many characteristics (e.g., Miner and Crimmins, 1995).

Milloy (1994) conducted a comparative study of 59 juvenile sex offenders and 132 other juvenile offenders as part of a needs assessment survey. She found that although the juvenile sex offenders had some unique characteristics, they shared many more characteristics with juveniles whose offenses were nonsexual. In contrast to the juveniles whose offenses were nonsexual, the sex offenders were more likely to have been victims of sexual abuse, have major mental health problems, need health or dental hygiene education, lack appropriate peer relationships, and have problems with sexual identity. They also tended to have more adequate academic performance, fewer prior offenses and convictions, and less substance abuse. None of the sex offenders was convicted of a new sex offense. Their overall recidivism rate was lower than that of other offenders. When they did reoffend, their crimes tended to be nonsexual and nonviolent. By the end of a 3-year followup period, only 22 percent of the sex offenders had offense histories limited to sex offenses only. Only 15 percent had been adjudicated for multiple separate incidents of sex offenses. In contrast, 78 percent had been convicted of both sex offenses and other types of offenses. Milloy noted, "These findings suggest that when a longitudinal perspective is used, sex offending among juveniles appears to be but one piece of a pattern of generalized delinquency" (p. 9).

Miner and Crimmins (1995) compared juveniles in juvenile sex offender treatment programs with juveniles who self-reported committing other types of offenses and juveniles who reported no delinquent behaviors in a national survey of juveniles. Few differences were found in the delinquency-related attitudes of sex offenders and other offenders (e.g., whether it is okay to cheat on tests, be truant, use drugs, be violent, and commit theft). The sex offenders differed, however, from the other offenders in their overall negative attitude regarding most types of delinquent behavior. They also were more disengaged from family interactions. The authors proposed that it may be their social isolation from peers and family that allows juvenile sex offenders to violate a generally prosocial belief system and behave in antisocial ways toward others.

A more recent study by Spaccarelli et al. (1997) further supports findings suggesting that many juvenile sex offenders also commit other types of offenses and are difficult to distinguish from delinquents with no known history of sexual assault. Spaccarelli et al. examined a sample of 210 chronic delinquents, 24 of whom had been arrested for a sex offense and 26 of whom self-reported committing sex offenses for which they had never been arrested. There were no differences on any of the measured variables between the combined group of 50 juvenile sex offenders and a group of 106 juveniles who had been arrested for violent but nonsexual offenses.

⁵ The authors (Pithers et al., 1998a) noted that psychometric test results were reported for female caregivers only (94.5 percent of the sample), to facilitate the comparison of their scores with published norms.