FEMALE ADOLESCENT SEXUAL OFFENDERS IN RESIDENTIAL TREATMENT:
CHARACTERISTICS AND TREATMENT IMPLICATIONS

Germaine Lawrence Incorporated is a provider of residential treatment to emotionally disturbed adolescent girls, operating four programs spanning the continuum from short-term hospital diversion to long-term secure treatment. Since 1990 we have served increasing numbers of residents who were identified as having molested children. By mid-1994, we had served 20 adolescents with clear histories of aggressive sexual misbehavior.

In developing specialized programming and a protocol for this population, we were highly dependent on literature and treatment models concerning or designed for male sexual offenders. This was due to the paucity of literature that describes the characteristics or treatment needs of female adolescent child molesters (Johnson, 989). Moreover, the literature concerning females that does exist tends to be based on small sample sizes from out-patient programs (Fehrenbach & Monastersky, 1988; Ryan, 1989; Johnson, 1989). In fact, we could locate no studies describing:

• female adolescent sexual offenders requiring residential treatment, or,
• treatment protocols developed specifically for them

Our reliance on treatment methodologies developed for male juveniles may be problematic since it is not clear that female and male sexual offenders share the same characteristics and/or treatment needs (Mathews, 1987). Indeed some studies suggest that significant differences exist. Whereas the development of repetitive patterns of sexual offending behavior in males is well chronicled, in one of the few studies focusing on adolescent female sexual offenders, no subject exhibited a repetitive pattern of offending behavior (Fehrenbach & Monastersky, 1988). If female juvenile sex offenders (JSOs) do not develop repetitive patterns of sexual misbehavior, then they may not need specialized treatment protocols similar to those designed for males.

Another potential difference between male and female JSOs is their own histories of sexual abuse. Studies of male JSOs discussed in Knight and Prentky (1993) show a broad range of those sexually abuse children: 19%, 23%, 47%, and 81%. A study by Johnson (1988( of male child perpetrators showed 49% having been sexually abused. If almost all female JSOs have been sexually abused, it is possible that effective treatment might focus on sexual abuse survivor issues as well as on the sexual misbehavior.

A third area of possible difference is the context of the abuse, specifically the relationship between the perpetrator and the victim. Studies of females who have sexually misbehaved show that they tend to know the children with whom they acted out sexually In Johnson's (1989) sample (n=13) of latency age girls, all the subjects knew their victims, and for 10 girls their first victim was a family member, and for 8 of the subjects all sexual misbehavior was familial. If this tendency was supported by additional studies, it might have implications for understanding the behavior and designing treatment.

As a provider of residential treatment to this population, it is important to understand if female sexual offenders do differ significantly from their male counterparts, since, if they are different, it is possible that they require different treatment protocols than males. Hence this paper will focus on the three areas of potentially significant difference between male and female JSOs discussed above, in order to increase our understanding of female JSOs, and also to explore treatment implications of those possible differences.

Also, given the relative shortage of studies of female JSOs, it is important to contribute to the growing data base concerning this population. Hence, this article will present a descriptive study of the 20 female adolescent's served at Germaine Lawrence with histories of sexually offending.

Method

Subjects

This paper presents data describing all 20 female adolescents served at Germaine Lawrence programs between 1990 and 1994 who had sexually offended children. Girls were included in the sample if, at any time in their life, there was a documented incident of sexually aggressive behavior.

Two girls with histories of aggressive sexual misbehavior were referred to Germaine Lawrence programs but not admitted. Criteria for refusing to admit this type of child included denial of the sexual behavior, refusal to participate in a sexual offender evaluation, or a history of aggressive sexual misbehavior in a previous residential treatment center. These criteria are changing as we learn more about the risks associated with serving female JSOs, improve our treatment methods and protocol, and increase our capacity to ensure the safety of other residents.

It is problematic to generalize from our data since it is a sample of adolescents placed in residential treatment. Since only a subset of offenders are placed in residential treatment and such placement is, hopefully, not a random event but one based on an adolescents' behavior being dangerous to herself or others, our population seems not to represent a random sample of the offender population. Moreover, it is reasonable to believe that our sample is likely to represent either the more behaviorally disordered offenders or those who have been more chronic in their offending.

Data Collection

The data were collected through interviews with the subjects, their families, and their therapists, as well as through reviews of previous evaluations and documents from schools and previous placements. Additionally, each of the subjects was assessed by a senior clinician using a semi-structured sexual offender evaluation tool adapted by Germaine Lawrence from an instrument used by the Boston Court Clinic.

Given the age differential between the subjects and the victims (discussed below), and the ages of the victims, we have been dependent on our sample for much of the data concerning their sexual misbehavior. In addition, since sexual offending is socially unacceptable and punishable, offenders tend to deny the behavior. Thus it is possible that the nature and extent of these behaviors are understated.

Results

Child and Family Characteristics:

The girls ranged in age from 11 to 17 when they were admitted to Germaine Lawrence. However, their sexual misbehavior usually started significantly earlier. Only four (20%) girls in our sample were 13 or older when their first documented incident of offending occurred. The rest (80%) initiated their offending behavior before they turned 13.

18 (90%) of the girls had experienced sexual abuse as documented in their records. 14 (70%) had been the victim of familial abuse and 13 (65%) had been sexually abused by more than one perpetrator. In the two cases in which there was no documented sexual abuse, the mothers suspected the fathers of sexually abusing their daughters when they were pre-school age. 11 (55%) of the girls were victims of physical abuse within their home. All of the 20 subjects experienced either sexual abuse or physical abuse.

No subject came from an intact family: 13 girls' parents divorced, 4 never married, one were separated, and two sets of parents lost their parental rights. Most came from multi-problem families: in 14 (70%) at least one parent had a serious substance abuse problem, and in 9 (45%) at least one parent had been hospitalized due to mental illness (non-substance abuse) or a terminal disease.

95% of the subjects were the oldest girl in the household. Of the 20 subjects:

Twelve of the girls had experienced at least one hospitalization due to suicidal ideation or aggressive behaviors. Only four were adjudicated or court involved due to their sexual misbehavior.

Victim Characteristics

There were a total of 48 victims, all of whom were younger children, ranging in age from 2 to 12, with an average age of just under 6. The victims were evenly split between boys (24) and girls (24). In all cases of hands-on sexual offenses, the victims were known to the perpetrators. 28 were family members and an additional three were relatives living outside of the household. 12 victims were children in the neighborhood or on the school bus. 5 victims were children for whom the perpetrator was babysitting.

Relationship Between Perpetrator and Victim

As shown on graph x, 13 of our sample were under 12 at the time of their first known perpetration, with an average age of just under 11. Hence, there was an average age differential of over 5 years between the perpetrator and the victim.

All of the subjects with histories of hands-on offenses exclusively abused children they knew. 17 (85%) of the girls offended a sibling, and the only child offended cousins. Of the two remaining members of our sample, one was the youngest member of her family and abused four children in the neighborhood. The other was the oldest member of her family and abused children she was baby sitting. One subject committed hands-off offenses exclusively and offended only against strangers.

Sexual Behaviors

The girls in our sample practiced a wide variety of sexual behaviors with their victims. As shown in Graph x, the most common behaviors were genital contact without penetration (15 or 75%) and fondling (14 or 70%). Almost half of the sample (9 or 45%) practiced oral sex while 6 (30%) performed vaginal penetration with a finger (4) and/or and object (3). 2 girls had intercourse with their victims.

The majority (11 or 55%) of the girls in our sample were involved in a repetitive pattern of sexual offending behaviors, defined as behavior that continues for a period of over 6 months with a frequency of over 6 events. Among these 11 girls, 2 girls were involved in such a pattern for 8 to 10 months; 6 girls for 12 to 18 months; and 3 girls for over three years. One girl in this sub-group acted out sexually 6 reported times over 8 to 10 months. The rest of this sub-group acted out sexually at least twice monthly and as often as daily.

The subject who committed hands-off offenses made obscene telephone phone calls and propositioned a boy unknown to her. She committed these offenses with her boyfriend present.

There were nine girls in our sample who were not involved in a documented pattern of sexual offending behaviors. Each had two or three incidents of sexual offending documented by their own or their victim's disclosure. In three of the cases clinicians suspected that there were more, unreported incidents, and the existence of a repetitive pattern. In these cases under-reporting seemed possible since the victims were either pre-school age or mentally retarded.

Discussion

Repetitive Patterns

55% of the population that we studied have repetitive patterns of aggressive sexual misbehavior. In this regard, they are similar to their male counterparts. This finding supports the development and implementation f specialized treatment methods and protocols. In addition, treatment methods shown to be effective with males may prove to be useful models.

Why do our findings differ from those of Fehrenbach and Monastersky? Possibly because we studied significantly different populations. Their population was served on an our-patient basis, while our received residential treatment. Since adolescents are referred for residential treatment due the dangerous and chronic nature of their maladaptive behaviors, it seems reasonable to conclude that our sample population should be different behaviorally than one served on an out-patient basis.

Although the majority of our subjects did develop a documented pattern of behavior, some did not. In some cases it is possible that a pattern did exist but data has not been disclosed or documented. In other cases, it seems that the aggressive sexual misbehavior was isolated. Perhaps this suggests an empirically based typology that would guide treatment. Perpetrators with a history of only isolated incidents might be considered one type that requires one type of treatment. Those with a repetitive pattern of aggressive sexual misbehavior might be categorized differently and require qualitatively different forms of treatment to decrease their risk for future offending.

Victimization History

Almost all of the subjects in this study have a documented history of being sexually and/or physically abused. In this aspect, they are more similar to the other adolescents served at Germaine Lawrence than to male JSOs. The great majority of the adolescents served at Germaine Lawrence also have experienced severe trauma, most in the form of sexual abuse. As their counterparts with histories of sexual offending, they enter Germaine Lawrence because, in reaction to the trauma they experienced, they develop behaviors that are dangerous to themselves or others, and the behaviors are persistent over time, even after various community based interventions have been tried.

Although we have not collected data to support this argument, our experience with a residential treatment population suggests that trauma victims often develop "control seeking behaviors" (Ryan, 1989). These behaviors take a wide variety of forms. Some victims chose to internalize their control seeking behaviors and exhibit self-mutilation, self-destruction, substance abuse, or eating disorders. Some choose to externalize their control seeking behavior through assaultiveness, promiscuity, or sexual assault. From this perspective, sexual misbehavior is understood as one of several possible responses to severe trauma.

Relationship Between Perpetrator and Victim

This study's findings corroborate those of others that female perpetrators tend to know their victims well. In addition, these female perpetrators tended to have care giving responsibilities for their victims. Although only two subjects in the study molested children they were baby sitting, the great majority had are giving responsibilities in their own household for their victims. This suggests that female JSOs, in contrast to males, tend to offend within the context of an on-going relationship.

Given that the families have multiple problems and our subjects tend to be the oldest girl in the family, it is possible that a high percentage of these subjects were in parentified roles in their families. Additional data gained from interviews with the subjects show that many had extensive care giving responsibilities that fit the term parentified. This suggests that serving a parentified child role in the family may be a factor in sexually offending younger children. These girls not only have the opportunity to molest younger children, but other factors including their own abuse history, inadequate parenting, and resentment towards the children they must care for, may combine to form a witches' brew that can boil over to aggressive sexual misbehavior.

Typologies

As a treatment provider, a useful typology would help us differentiate:

Also, based on our findings we find a possible flaw with one typology that has been developed for girls who have sexually offended. Although Germaine Lawrence staff have used Johnson's category "sexually reactive" for several years, our study and others suggest this term is inaccurate and perhaps misdirects treatment.

Our point is not that Johnson's "Group II" is poorly described, but that labeling it with the term "sexually reactive" is a poor choice. We object to the use of the terms "sexually reactive" and "abuse reactive" to describe a sub-group of females who have sexually offended since it seems that almost all of them have been sexually abused, and that all of their offending behavior can be understood as abuse reactive.

Moreover, if only a sub-group is defined as sexually reactive, it suggests that only that sub-group might benefit from treatment directed at their abuse history. As discussed elsewhere, our findings suggest that almost all female adolescents who have sexually offended could benefit from the types of treatment designed for adolescents who have been sexually abused.

Conclusions and Implications

Treatment

These findings support Gil's (Gil & Johnson, 1993) concern with the tendency for service providers in sexual abuse treatment programs to develop rigid or polarized specializations, in which either the victim or offender issues are addressed in an exclusionary fashion. Female JSOs that we studied seem to share key characteristics with their peers in residential treatment, with the exception of their sexual misbehavior, and probably would benefit from similar treatment. Although sexual offenders need specialized treatment that focuses on their specific pattern of behavior, as well as special protocols to ensure the safety of those with whom they reside, the focus of therapy for female JSOs must also include key elements of effective treatment provided to non-sexual offending trauma victims:

System Response

As was found in other studies, very few of the girls in our sample were adjudicated. Our finding that most of the female JSOs at Germaine Lawrence do have repetitive patterns of sexual misbehavior suggests that there has been an inadequate assessment of the risk this population presents. We believe these adolescents should be adjudicated, and the juvenile justice system respond more actively, since such action provides increased leverage in treatment, and might decrease the risk of further victimization of children by these youth in the near term, and avoid victimization of their own children when they become mothers.

Future Research

This paper documents some similarities and differences between male and female JSOs. Clearly, this is just the start of such an investigation: it seems imprudent to assume that the factors leading to aggressive sexual behaviors, and the motivations that fuel its continuation, are identical. To better understand female juveniles who have sexually offended, future studies could focus on:

This paper also suggests that female JSOs, or at least those in residential treatment, may share many characteristics and treatment needs with their non-sexual offending peers in residential treatment. A careful comparison of their victimization histories, family structures, psychopathology, and social and academic skills might help us understand the behavioral choices they have made, and develop better treatment interventions.