The Precondition Model as a Method for Developing Understanding of Female Contact and Non-Contact Sex Offending: A Single Case Study

Abstract This research evaluates the use of an established model, typically used for understanding male sex offenders, to understand the behavior of a female sex offender. The Finkelhor (1984) Precondition Model of offending is used to provide a rare opportunity to explore the process of offending for a female contact and non-contact offender, whose offenses were against children. It reviews the efficacy of utilizing this model in the rehabilitation and collaborative risk management of a female sex offender. The results suggest that this approach can be applied to Internet and contact sex offenses to develop understanding of the progression of offending, including issues such as sexual arousal and the impact of a male co-perpetrator. In this case, the results indicate a post-intervention improvement in areas such as affect control, ability to maintain positive relationships, self-support, and reduced dissociation and dysfunctional sexual behavior. This project provides support for the development of a treatment approach that explores the individual nuances of female sex offending.


Background
Our developing understanding of female sex offenders While research has begun to identify themes and trajectories within female sex offender characteristics (Gannon et al., 2014;Johansson-Love & Fremouw, 2009), the reporting of these cases is still relatively rare in comparison to the male sex offender literature. Gannon, Rose, and Ward (2010) have described a number of pathways to female offending, the Descriptive Model of Female Sexual Offending, based on the accounts of 22 individuals, but as the researchers identify, this project is in its infancy and these pathways may not accurately reflect the specific offending process and treatment needs of a particular individual. Research exploring the female non-contact (Internet) offender is rarer still (Elliot & Ashfield, 2011). However, some authors such as Elliot and Ashfield (2011) have extrapolated from minimal data to clinical practice noting, for example, that female Internet sex offenders often have difficulties with interpersonal relationships offline. They recommended that researchers focus exploration in three key areas: how female offenders use the Internet to fulfil a desire to connect with others, how exposure to abusive material may manufacture a desire to view more abusive material, and how the exposure to abusive material may impact attitudes towards children and sex. Researchers have also recognized the importance of increasing awareness and understanding of the female sex offender and the characteristics of her crimes, and reducing the gender bias through which female sex offenders are perceived (Cortoni & Gannon, 2011;Johansson-Love & Fremouw, 2006). However, further research is necessary to draw conclusive profiles and comparisons between male and female sex offending behavior (Gannon & Cortoni, 2010).
In their comparison of five case studies of Italian female sex offenders, Grattagliano et al. (2012) found that the cases did not reveal a consistent or typical pattern. Furthermore, the existing typologies used to describe these women may be insufficient in adequately representing the full spectrum of female-perpetrated sexual offending against children. The authors concluded that female sex offenders abuse for a variety of reasons. However, these authors were unable to undertake clinical interviews with the offenders studied, and although explanations regarding the motivation for offending were logical, the authors highlighted the need for more qualitative data in this area.  also highlighted that many studies of female child sex offenders have focused on developing profiles of the offender or offense characteristics and have not concentrated on the treatment needs of these women. The authors presented a detailed review of female sex offending characteristics and typologies, but expressed concern that this method of understanding female child sex offenders has limited usefulness when working with an individual case. For example, female offenders tend to be younger than their male counterparts (Faller, 1995), often experience financial issues (Allen, 1991), have experienced frequent and severe abuse themselves (Miccio-Fonseca, 2000), and show high levels of emotional dependency (Hunter & Mathews, 1997), amongst other traits. They also tend to use less physical violence and more persuasion and coercion than males (Grayston & De Luca, 1999). It is not yet clear how these factors may be related to offending behavior nor what are the core issues to be addressed to reduce recidivism. However, there are many features that are shared between male and female offenders, but a recent review (Tsopelas, Spyridoula, & Athanasios, 2011) concludes that we are yet to reach an agreement on how to accurately conceptualize female offenders.
In their review of female sex offending literature  suggested that while women commit fewer sexual offenses than males, the statistics are unlikely to accurately reflect the reality of female sex offending. They argue that this is due to widespread denial of femaleperpetrated abuse, societal perception that this abuse is less harmful than male-perpetrated abuse, and significant underreporting of these crimes. Elliott, Eldridge, Ashfield, and Beech (2010) suggested that a relatively small number of convictions results in a lack of data concerning the psychological profiles of female sex offenders, the factors linked to re-offending, and what should be targeted in treatment. Despite these concerns, it is estimated that approximately 5% of all sexual offenses are committed by females (see Cortoni, Hanson, & Coache, 2009). Using this figure further analysis has estimated that female offenders are responsible for 1.4% of all child victims (Pereda, Guilera, Forns, & G omez-Benito, 2009).
Addressing female sex offenders' treatment needs Ford (2010) suggests that although female sex offender treatment needs appear similar to those of male sex offenders it is crucial that treatment providers recognize gender-specific nuances in relation to those treatment needs. For example, in contrast to their male counterparts, female sex offenders tend to demonstrate an absence of beliefs associated with an entitlement to sexually abuse children and are often impacted by the negative environment created by a male co-perpetrator (Beech, Parrett, Ward & Fisher, 2009). Prior to Ford (2010),  suggested it may be beneficial for professionals to familiarize themselves with individuals' vulnerability factors highlighting the potential need for female sex offender treatment to explore sexual interests, empathy, intimacy deficits, and emotional regulation. In addition, Gannon and Alleyne (2012) suggested that exploration of offense-supportive cognitions represents one of the many treatment needs of female sex offenders. Finally, Elliott et al. (2010) recommend that, due to the relative infancy of female sex offender assessment and treatment, examining the grounding of theoretical findings in clinical experience can be of great benefit to the development of practice in this area.
The precondition model as a suggested treatment for female sex offenders The Finkelhor (1984) Precondition Model is described as one of the most promising etiological theories for use in the rehabilitation of sexual offenders (Ward & Beech, 2006), it has had an important role in both research and practice (Ward & Hudson, 2001), and has been used as a framework for understanding aspects of male offending (e.g., Duff & Willis, 2006). It was developed to meet the need for a broad integrative theoretical framework with the aim of elucidating the aetiology of child sexual abuse (Howells, 1994). However, it was originally developed to support the exploration and understanding of factors that lead to male sexual offending against children.
The model is based on research suggesting that in order for sexual offenses to occur the offender must pass through four planning stages: motivation to offend, overcoming internal inhibitions, overcoming external barriers, and overcoming victim resistance. These stages highlight both the intrapersonal (within perpetrator) and external factors relevant to the offending behavior. According to Finkelhor (1984), the motivation to offend is determined by three underlying factors: emotional congruence (the fit between the adult's emotional needs and the characteristics of a child), sexual arousal, and blockage (a failure to meet sexual and emotional needs in a prosocial way). Overcoming internal inhibitions relates to the perpetrator's disinhibition of beliefs and attitudes that usually act to control deviant sexual desires, for example, due to impulse disorder, alcohol consumption, or the presence of severe stress. The remaining preconditions, overcoming external barriers and overcoming victim resistance, relate to the external factors (rather than causal factors) that allowed the offense to progress. Ward and Hudson's (2001) critique of the model highlighted a number of weaknesses, including a lack of attention to developmental factors, as well as a lack of detail concerning the way the psychological vulnerability factors are specifically linked to the perpetration of sexually abusive behaviours. However, a significant strength of the model is the way it relates a broad range of causal factors to the offense process, providing a useful framework for therapists (Ward & Hudson, 2001). Indeed, it was selected for the present study based on this merit, as the model explores specific details of the offending behavior that can inform risk management strategies for the individual. For example, if the individual identifies work stressors as a disinhibiting factor steps can be taken to manage this risk in future, such as additional support/ monitoring at work. The application of the model also aims to increase the individual's awareness of the offending behavior in detail, establish acceptance of the premeditated nature of sexual offending, reduce denial, and increase empathy for the victim. The Precondition Model was also selected over other theoretical models of sexual offending as, according to Thakker and Ward (2012), it benefits from focusing specifically on child sexual offending, unlike Ward and Beech's Integrated Theory (2008), and can adequately account for nonaggressive sexual offending and offenders who commit their first offense at a later life stage, unlike Marshall and Barbaree's Integrated Theory (1990). Furthermore, it prioritizes the impact of external and environmental factors in the understanding of the process of offending, unlike the Quadripartite Model (Hall & Hirschman, 1991).
The Precondition Model is recommended by Forensic Psychology Practice Ltd. guidelines (1999), is widely used in the rehabilitation of sex offenders in both UK community forensic services and Probation Services (e.g., the Ministry of Justice's NSOG programme; see Harkins, Flak, Beech, & Woodhams, 2012) and forms the basis upon which many subsequent models of child sexual abuse have been based (Whittle, Hamilton-Giachritsis, Beech, & Collings, 2013). The model is commended for being a broad, coherent, explanatory framework (Howells, 1994). Furthermore, according to Howells (1994), the model has face validity and directs the clinician's attention to important assessment and therapeutic targets, a feature that was considered to be of particular relevance within the current study. However, it has rarely been systematically critically examined or reviewed (Ward & Hudson, 2001) and when the model was developed there was little awareness of female sex offenders and the Internet had not been invented.

Project aims
This case study was completed as an audit for the service in which it was completed, regarding the use of the Precondition Model in individual sex offender rehabilitation. However, the service-user and psychologists involved were keen to share understanding of this female perspective of contact and noncontact (Internet) sex offending against children and to support the development of service provision that is suited to the individual needs of female sex offenders.
The study aims were to increase awareness of a female perpetrator's perspective of child sex offending and provide evidence regarding the relevance and efficacy of using the Precondition Model in the understanding, rehabilitation, and risk management of female sex offenders. There have been no previously published single case studies of female sex offenders using the Precondition Model.
The aims were to answer the following questions: Method Design A single case design was utilized. The qualitative data gathered via the therapeutic intervention outlined in this report were analyzed retrospectively, along with quantitative outcome measures that had been administered pre-and post-intervention.

Participant
The service-user was a female, aged 40-50 of White ethnicity. She had completed a custodial sentence imposed for offenses of sexual assault against a child under 13 years old, making indecent photographs of a child, and possessing indecent photographs of children. The victim of the contact offenses was a female relative aged 5 years. The victim had been involved in multiple sexually abusive contacts with the service-user and her male partner (and co-defendant), which included her involvement in sexual activity, exposure to pornography and abusive images, and the creation of sexually abusive photographs. The service-user was a registered sex offender with a Sexual Offences Prevention Order (SOPO). She had no previous convictions. She was unemployed at the time of the intervention. Prior to the therapeutic work reported within this project, the service-user had completed a 15-session intervention with another service that had commenced during her period of custody and continued following her reinstatement into the community. This initial work had focused on establishing engagement with a therapeutic relationship over a 15-month period, also identifying links between core thinking, emotional processing, and risky behavior. This work had highlighted a history of complex childhood trauma including experiences of victimization in the context of sexual abuse. No mental health issues had been identified during the previous intervention. The service-user was referred to a community Forensic Psychology Service based in the United Kingdom to engage with the intervention presented in this study. Specifically, for further psychological input to address her understanding of the internal and external risk factors for her offending behavior, to develop her ability to protect her relatives from abuse, to manage her risk of sexually offending against children and to manage her own experience as a victim of abuse as a child.

Procedure
Following an assessment interview, in which the serviceuser's motivation to engage with a further psychological intervention was established, the service-user engaged with 22 individual 50-minute sessions. Sessions took place at a community Forensic Psychology Service with expertise in the psychological treatment of male and female sex offenders and in the therapeutic use of the Precondition Model. Sessions occurred weekly over an 8-month period, with allowance made for planned leave from therapy.
Of these 22 sessions, one session was used to collaboratively agree on the goals of the intervention. Next, each of the four planning stages (preconditions) were written down on a sheet of paper to form a Precondition Model framework. Each stage of the model was considered in order starting with motivation to offend, followed by overcoming internal inhibitions, overcoming external barriers, and overcoming victim resistance. Ten sessions were taken to complete the two Finkelhor Precondition Models, with five sessions allocated to the Precondition Model of the noncontact offending behavior and five sessions to the Precondition Model of the contact offending behavior. Noncontact offending had occurred prior to contact offending, therefore the Precondition Model for the noncontact offending was completed first. The therapist supported the service-user to generate relevant information by using explorative questions about the intrapersonal and external factors relating to the offense process. During this process 'meaning units' (direct quotes from the service-user) were recorded in writing by the therapist beneath the relevant precondition heading on the Precondition Model framework and then reviewed with the service-user in situ to ensure that they encapsulated the intended meaning of her statements. The beginning and end of each session were used to review the work completed so far. This created a working document that could be amended at any time and encouraged a reflective process. Once the service-user was satisfied with the completed Precondition Model of the noncontact offending, this method was then repeated to produce the Precondition Model of the contact offending behavior.
Of the remaining 11 sessions, nine were used to review and challenge the completed Precondition Models and generate 'safe-living guidelines.' These guidelines addressed the major themes within the Precondition Models, encouraging the service-user to recognize the many factors, both internal and external, that had 'allowed' the behavior to occur. The aim of this aspect of the therapy was to encourage the service-user to recognize her emotional and sexual needs and to establish more appropriate ways of meeting these, while empowering the service-user to set her own boundaries in terms of risk management. Finally, one session was used to go through the psychological report of therapeutic outcome, and one session was used to exchange therapeutic 'goodbye' letters.

Measures
Quantitative outcome measures were completed by the service-user pre-and post-intervention, following the first session and the penultimate session. These included: the Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983), the Trauma Symptom Inventory (TSI; Briere, 1995), and the Inventory of Altered Self-Capacities (IASC; Briere, 2000). Reliability was assessed using Cronbach's alpha values, in which .70-.95 is considered to provide a good degree of reliability (Tavakol & Dennick, 2011).
The HADS is one of the most widely used self-report measures of anxiety and depression. It is recommended that on both subscales scores between 8 and 10 identify mild cases, 11-15 moderate cases, and 16 or above severe cases (Snaith & Zigmond, 1994). Bejelland, Dahl, Haug, and Neckelmann (2002) report very good reliability, with mean Cronbach's alpha values of .83 for the anxiety scale and .82 for the depression scale. This outcome measure was selected due to the service-user's report, provided on her presentation to the community Forensic Psychology service in which the presented intervention took place, of low mood as a potential factor in her offending and self-reported anxiety since her recent release from prison.
The TSI is a self-report measure of post-traumatic stress and other psychological sequelae of traumatic events. It is designed for the evaluation of acute and chronic symptomology as well as the lasting effects of childhood or early traumatic events. The TSI contains 10 clinical subscales. On all subscales a T score of 65 or over falls within the clinical range. Snyder, Elhai, North, and Heaney (2008) report very good reliability with Cronbach's alpha values ranging from .73 to .91 for the subscales. This measure was selected due to the service-user's report of childhood trauma and the possible association between the ongoing psychological impact of these experiences and maladaptive coping strategies that may have been factors in her offending.
The IASC is a self-report measure of psychological functioning based on three key constructs: the ability to maintain a sense of personal identity and self-awareness, the ability to control and tolerate strong affect, and the ability to form and maintain meaningful relationships. The IASC contains 11 clinical subscales. On all subscales a T score over 70 falls within the clinical range. Briere and Runtz (2002) report impressive reliability, with Cronbach's alpha values ranging from .86 to .96 within a clinical sample. This measure was selected due to the potential association between the three key constructs and the service-user's vulnerability factors for offending, for example, patterns of abusive relationships, susceptibility to influence of dominating male partners, difficulty managing mood states appropriately, and limited self-awareness.

Ethical considerations
Consultation with the University of Hertfordshire Research Team and the supervising qualified psychologist confirmed that this project was considered a retrospective service provision evaluation. Therefore, ethical approval was not considered necessary. Consent to engage with the therapeutic intervention was sought during the assessment interview, prior to treatment. The potential for the case to form a research project was initially discussed with the service-user at the end of the therapeutic intervention. Twelve months' post-intervention, when the opportunity to write up the project arose, formal consent was attained in writing pertaining to important information about the study, issues of confidentiality, and informed consent. Protecting the anonymity of the service-user and victim has been a priority. Therefore, specific participant information and the full list of meaning units from the completed Precondition Models have not been included in this study.

Analysis
Quantitative content analysis (Berg & Lune, 2013) was used to analyze the service-user's responses to each aspect of the Precondition Model. These had been summarized into meaning units and collaboratively agreed with the service-user. This method supported a collaborative process that complemented the therapeutic intervention. The meaning units were then analyzed for themes post-intervention. Previous research was considered throughout the coding process, as with a deductive approach. For example, Finkelhor's (1984) suggestions regarding the factors contributing to each of the planning stages of the Precondition Model, and  suggestions regarding common factors associated with the planning and completion of sexual offenses by females, such as issues of dependency on males, deviant sexual interests, and offense supportive cognitions. However, an inductive analysis of the data, whereby the analysis involved an immersion in the data to extrapolate themes, formed the primary approach. Therefore, a mixed approach to coding the data was applied. The content analysis provided a final coding scheme, including themes endorsed by quantifiable meaning units, which was applied to the data on two occasions over a 2-week time period (Joffe & Yardley, 2004).
The primary coder was the therapist. To address the risk of bias, the final coding scheme was presented to a qualified psychologist for verification and the opportunity to make alternative suggestions to the coding scheme. Interrater reliability was also estimated by using an independent coder who was not involved in the therapy or study, to code a random selection of 25% of the qualitative data taken from both the Precondition Models. The agreement between the coders was greater than 90%, suggesting a good interrater reliability.
Where appropriate, analysis of the quantitative outcome measures involved calculation of a Reliable Change Index (RCI). Jacobson and Truax (1991) suggested that to demonstrate significant improvement change must be reliable; it must be over and above the fluctuations of an imprecise measuring instrument. Secondly, change must be clinically significant; the service-user must belong to the functional or normal population post-intervention (within two standard deviations of the normal mean). Jacobson and Truax (1991) did assert that operationalizing clinical significance in terms of recovery or return to normal functioning may not be appropriate for all disorders treated by psychotherapy. However, within the present study, this cut-off was deemed to provide a useful point of reference regarding the degree of change observed, particularly as it is currently unclear whether a return to normal functioning is an appropriate expectation for the female sex offender population. Therefore, this definition of reliable and significant change has been used within the current study and, in line with Cahill et al. (2003), the results can be said to have achieved reliable and clinically significant change if the pre-intervention measures indicate that the individual is outside normal limits but they move to within normal limits post-intervention, having changed by at least the reliable change index.

Motivation for Internet offending
This indicates that the service-user's responses in relation to motivation for Internet offending were summarized into six meaning units. These were coded into four main themes. The most frequently endorsed comments were: the service-user's curiosity about aspects of child abuse, such as the impact on the child; and how the abuse is completed. Also, the service-user's relationship with her partner, such as how the abusive images provided a way of communicating and meeting sexual and emotional needs for both partners.

Overcoming internal inhibitions for Internet offending
The service-user's responses to the overcoming internal inhibitions aspect of the Precondition Model for the Internet offending behavior could be summarized into four meaning units. These were coded into three main themes. The service-user's projection of enjoyment into the children depicted in the images was the most endorsed theme.

Overcoming external barriers for Internet offending
The service-user's responses to the overcoming external barriers aspect of the Precondition Model for the Internet offending behavior could be summarized into 11 meaning units. These were coded into three main themes. The most frequently endorsed comment was the service-user's use of secrecy to overcome external barriers, such as the use of passwords on computers and keeping materials unlabelled and locked away.

Overcoming victim resistance for Internet offending
The service-user's responses to the overcoming victim resistance aspect of the Precondition Model for the Internet offending behavior could be summarized into five meaning units. These were coded into two main themes. The most frequently endorsed comment was the serviceuser's emotional detachment from the abuse depicted in the images, the service-user described not thinking about the victim's resistance at the time of viewing the images, but did acknowledge that she would delete images in which the child looked upset.

Motivation for contact offending
The service-user's responses in relation to motivation for the contact offending behavior could be summarized into seven meaning units. These were coded into three main themes. The most frequently endorsed comments were related to the service-user's relationship with her partner, such as using the abuse to please her partner's interests and as a shared interest that maintained the relationship.

Overcoming internal inhibitions for contact offending
The service-user's responses to the overcoming internal inhibitions aspect of the Precondition Model for the contact offending behavior could be summarized into 11 meaning units. These were coded into four main themes. The most frequently endorsed comment was the abdication of responsibility, for example, putting responsibility onto the victim for not asking for the abuse to stop or onto the partner who had a more active role in the physical abuse. This was followed in frequency by the minimizing of the harm of the sexual abuse.

Overcoming external barriers for contact offending
The service-user's responses to the overcoming external barriers aspect of the Precondition Model for the contact offending behavior could be summarized into seven meaning units. These were coded into three main themes. The most frequently endorsed comment was the service-user's removal of potential witnesses to the abuse, for example, by encouraging other family members to stay away from home overnight and giving alcohol to the victim's young sibling.

Overcoming victim resistance for contact offending
The service-user's responses to the overcoming victim resistance aspect of the Precondition Model for the contact offending behavior could be summarized into eight meaning units. These were coded into four main themes. The most frequently endorsed comment was the serviceuser's normalizing of the abuse to the victim, acknowledging a slow and gradual grooming process.

Quantitative outcome measures
Q3: Is the Finkelhor Precondition Model efficacious in the rehabilitation/risk management of a female contact and noncontact child sex offender? Table 3 indicates the service-user's scores on the Hospital Anxiety and Depression Scale (HADS; Zigmond & Snaith, 1983) pre-and post-intervention. This table highlights the reduction in symptoms on both subscales post-treatment, with anxiety remaining within the severe range and depression lowered from the moderate range to within the mild range. The total change from 34 to 26 post-intervention exceeds the 95% RCI of 6.20, suggesting a clinically reliable change overall. However, the post-intervention total score is not within two standard deviations of the normal population mean, so according to the parameters outlined within this study the improvement is not clinically significant.  Table 4 indicates the service-user's scores on the Trauma Symptom Inventory (TSI; Briere, 1995) pre-and postintervention. This table shows the reduction in symptoms on all subscales post-treatment, except for 'tension reduction behavior' which increased but remained within the nonclinical range. Subscales for 'dysfunctional sexual behavior' and 'impaired self-reference' reduced significantly, from within the clinical range to below clinical cut-off post-intervention. Table 4 also shows the 95% RCI values for each subtest of the TSI. This indicates that the change for 'dissociation,' 'dysfunctional sexual behavior,' and 'impaired self-reference' was reliable. Improvement in these three areas was also clinically significant, falling to within two standard deviations of the normal mean. Table 5 indicates the service-user's scores on the Inventory of Altered Self-Capacities (IASC; Briere, 2000) pre-and post-intervention. This table shows the improvement of self-capacities on nine subscales post-treatment. The improvement demonstrated by the 'interpersonal conflicts,' 'affect skill deficits,' and 'tension reduction activities' subscales are of significance, with scores reducing from the clinical range to below the clinical cut-off. The score for 'susceptibility to influence' increased from below the clinical cut-off to within the clinical range post-intervention. The score for 'affect instability' also increased post-intervention, remaining within the clinical range. Table 5 also shows the 95% RCI values for each subtest of the IASC. This indicates that the change on nine of the subtests was reliable. However, further analysis of clinically significant change demonstrated that only the improvements in 'interpersonal conflicts,' 'affect skill deficits,' and 'tension reduction activities' were clinically significant, falling to within two standard deviations of the normal mean.

Summary of results
The results demonstrate that the Precondition Model can be applied with a female contact and noncontact sex offender, providing a beneficial framework for the exploration of the offending behavior. The service-user in this case example was able to engage with the model and it provided a platform for empowering the service-user to identify risk factors and develop risk management strategies. The results are summarized below, demonstrating how the findings correspond with exploration of the outlined project aims. The results suggest that the Finkelhor (1984) Precondition Model can be utilized to explore the process of offending for noncontact (Internet) offending behavior. Delmonico and Griffin (2008) highlighted the lack of  Internet-focused interventions for sex offenders. These authors also recommend the modification of already existing cognitive-behavioral techniques, such as the Precondition Model, which although originally developed for contact offenders may be suitable for use with Internet offenders.
In this case, the use of the Precondition Model illuminated several key themes in the development of this offending behavior. The service-user identified her sense of curiosity regarding the sexual abuse of children, which she linked to her own experience of early sexualization and the abuse of her own children by a male partner. This is in line with research that suggests that the majority of female sex offenders have experienced a great deal of developmental adversity, including poor parental relationships and considerable emotional, physical, and sexual abuse (Elliott et al., 2010), although, it is important to note that most victims of childhood trauma do not go on to commit sexual offenses. The authors link these early abusive experiences with similarly chaotic adult lives that are characterized by feelings of inadequacy and a tolerance of abusive relationships.
The service-user acknowledged how her curiosity and initial shock had developed into sexual arousal when viewing or thinking about abusive images. These motivations are commonly reported by male offenders (Quayle & Taylor, 2001). In addition, of interest in this case was the element of progression identified by the service-user. This included how the use of abusive images during sexual contact with her adult partner generated further interest in seeking and using abusive images. Also, how it encouraged discussion about how they could commit contact abuse themselves. This is in contrast to research into the progression from noncontact to contact offending of males, which largely suggests that the risk of progression is low (Seto, Hanson, & Babchishin, 2011).
However, research does acknowledge the use of abusive images in the grooming of children in contact offenses (Quayle & Taylor, 2001), without establishing the direction of causality.
The use of the Precondition Model also highlighted practical aspects of the Internet offending behavior, such as the use of passwords, secrecy, and taking advantage of other's lack of knowledge. This created the opportunity to develop collaborative risk management strategies and encourage engagement with Probation-enforced rules around computer and Internet use. According to Delmonico and Griffin (2008) these basic Internet management techniques can appear superficial but are often overlooked and underutilized in interventions for Internet sex offenders. The results of this single case study suggest that the Finkelhor (1984) Precondition Model can be used to develop further understanding of the nature of female contact sex offending behavior and can be a useful tool in the exploration of an individual's perspective. Elliott et al. (2010) suggested that female sex offenders demonstrate many similarities to their male counterparts, for example, their beliefs about children as sexual beings and the minimized perception of harm caused by sexual abuse, and therefore also endorse approaches towards the assessment of risk and treatment targets through the adaptation of frameworks that have been developed for male sex offenders.
The quantitative content analysis has highlighted several key themes that mirror recent research into victim, crime, and female sex offender characteristics. Of particular interest in this case is the involvement of a male co- offender. Early research suggested that female sex offenders were typically coerced by dominant male partners (Saradjian, 1996). More recently this view has developed further to include male-accompanied female offenders who offend without coercion (Nathan & Ward, 2002). The Precondition Model highlighted the complicated nature of the task of unpicking responsibility when more than one perpetrator is involved. It proved useful in generating discussion about responsibility and identifying dominance in relationships. The service-user identified both with the role of dominance and as someone who had at times been coerced. This highlighted an aspect of the model that may benefit from development specific to its application with female offenders. However, the model did facilitate the awareness and reduction of dependency on males, as recommended by Matthews (1993). By reflecting on the service-user's own experience of victimization it also promoted feelings of empathy that may have been suppressed by cognitive mechanisms designed to cope with the trauma, as suggested by Elliott et al. (2010), such as cognitive minimizations of the impact of abuse. The Precondition Model highlighted elements of the offending behavior in line with recent research into the victims of female sex offenders. For example, the victim was a young female (Grayston & DeLuca, 1999), and victim and offender were biologically related (Johansson-Love & Fremouw, 2009). Quantitative content analysis also illuminated the role of cognitive distortion in the selection, grooming, and abuse of the victim. For example, the abdication of responsibility, minimization and justification at the time of offending. Also highlighted was the impaired emotional regulation and interpersonal problems that motivated the offending (Beech & Ward, 2004). This created opportunity to explore and challenge these risk factors, while focusing on the service-user's strengths and her innate capabilities to overcome these difficulties. This approach is recommended by Ward and Stewart (2003), who advocate focus on positive states of mind, personal characteristics, and experiences that provide a viable alternative to the offending behavior.
In line with Grattagliano et al.'s (2012) research, the present case does not fit neatly into a single typology. For example, the quantitative content analysis suggests that the offender was at times both passive and actively involved in the abuse (Grayston & De Luca, 1999). She also reported assuming the teacher/lover role, could be described as a predisposed molester, and male-coerced molester (Mathews, Matthews, & Speltz, 1989). According to Gannon, Rose, and Ward's (2010) Descriptive Model of Female Sexual Offending, she would also fit within the Directed-Avoidant pathway, for women who tend to offend in order to achieve intimacy with their male co-perpetrator, and the Explicit-Approach pathway, for women who seemed to plan their offenses in pursuit of specific goals, such as sexual gratification and intimacy with the victim, and experienced positive affect, such as excitement and satisfaction. This further supports the movement towards individual formulation of sex offending for both females and males, and a movement away from attempts to categorize sexual offenders in research. Q3: Is the Finkelhor Precondition Model efficacious in the rehabilitation/risk management of a female contact and noncontact child sex offender?
The quantitative outcome measures used did highlight some significant and reliable change for the individual post-intervention. In terms of risk management, the improvement in areas such as affect control, ability to maintain positive relationships, the reduction of externalizing behaviors in reaction to painful internal states, reduced dissociation, improved self-identity and selfsupport and reduced dysfunctional sexual behavior is highly relevant. The increased endorsement of affect instability post-intervention may indicate the destabilizing effect of challenging maladaptive coping strategies throughout this intervention. This may highlight the potential for increased risk of self-harm or relapse of historic maladaptive coping strategies for some serviceusers. However, in this case the service-user reported recognizing that she would previously have used sexual behaviors to tolerate these feelings and described motivation to use more positive strategies and avoid previous maladaptive patterns of coping. These warning signs and alternative coping strategies were then incorporated in to her 'safe living guidelines.' The high endorsement of the Identity subscale of the IASC both pre-and post-intervention may highlight a significant limitation of this model. These subscales are related to difficulties with self-awareness, self-assertion, and satisfying one's own interpersonal needs. For a female sex offender that has co-offended and reported experiencing being dominated and influenced by her partners a lack of change in this area may indicate ongoing risk within her relationships. This was discussed and addressed as part of her 'safe living guidelines,' however, future replications of this study may benefit from additional attention to this area.
The outcome measures used, while valid and relevant to the rehabilitation of the service-user in terms of affect, did not target changes in cognition or empathy. Johansson- Love and Fremouw (2006) also refer to the lack of standardized measures available in relation to female sex offenders. This issue continues to impact on the interpretability of research into female sex offending.
These authors suggest that further research is required to develop the cognitive distortion measures that are frequently cited in male sex offender literature, for use with female sex offenders.
In this case, service-user report and clinical judgment were also used to interpret efficacy. The service-user reported positive change in her understanding, motivation, and ability to protect herself and others from the risk of sexual abuse. She was enthusiastic about the individualized approach used to gain understanding of her offending and had made strong links between her developmental experiences and her offending behavior. Johansson-Love and Fremouw (2006) also advocate the use of an individualized approach to the rehabilitation of female sex offenders due to the heterogeneity of the female sex offender population.
The service-user also accepted responsibility for the offending and appeared more empathic toward victims, as well as demonstrating increased compassion for herself as a victim of sexual abuse as a child. Following this intervention, she expressed a desire to continue to increase her understanding of the impact of destructive early experiences, including her experience as a victim of sexual abuse as a child, and was referred for further psychological input within a Community Mental Health Team.

Service/clinical implications and recommendations
Due to increasing recognition of the perpetration of sexual abuse by females , there is a need for services to provide suitable psychological interventions for the rehabilitation and risk management of sex offenders, both male and female. The Finkelhor (1984) Precondition Model is a well-established framework for exploring the process of offending and developing factors such as responsibility and empathy that reduce the risk of future offending behavior in males. This project highlights the applicability of this model to the rehabilitation and risk management of female sex offenders. Furthermore, it contributes to the movement away from the exploration of typologies of female sex offenders towards the development of theory and treatment-driven models that focus on relapse prevention, as prioritized by Gannon, Rose, and Ward (2010). The present study also demonstrates the suitability of the Precondition Model to an inductive approach to exploring offenders' own accounts, as recommended by , to explain the sequence of behavioral, contextual, cognitive and affective factors that facilitate and maintain female sexual offending against children.
It is widely recognized that group interventions add a beneficial dynamic in the rehabilitation of male offenders (Perkins, Hammond, Coles, & Bishopp, 1998). Some research suggests the possible benefits of mixed-gender group intervention for non-sexual offenders (Burrowes & Day, 2011). However, Blanchette and Taylor (2010) state that wider research is not supportive of mixed-gender group interventions for sex offenders. It was considered in this case that the inclusion of the female offender in a group of male sex offenders would prove detrimental to the female service-user, and possibly to the group of male offenders. However, the number of female sex offenders referred for psychological intervention in relation to their offenses is much reduced in comparison to males. Therefore, it will not always be possible, as in this case, for the group approach to be facilitated. It is therefore essential for services to consider the individual needs of female sex offenders, many of whom are likely to have experienced sexual abuse themselves (Johansson-Love & Fremouw, 2009). In this case the gender of the psychologist facilitating the intervention was carefully considered and collaboratively agreed with the service-user.
This project supports research such as Gannon and Rose (2009), which suggests that female sex offenders appear to display psychological deficits in similar domains to male sex offenders. However, it also highlights the benefit of an intervention that is individualized and tailored to the needs of the sex offender. It provides further evidence to support Johansson-Love and Fremouw (2006) who argued that differential treatment still needs to be scientifically studied and evaluated, and that the scientific bar needs to be raised to better support the conclusions that are made about this population.
In their review of treatment initiatives for female sex offenders, Blanchette and Taylor (2010) also highlighted the dearth of knowledge regarding the female sexual offender and the challenges this presents to the development of gender-appropriate interventions. They recommended the following considerations for the provision of gender informed services for this population: gender should be central to guiding women out of sexual offending (Gannon, Rose, & Ward, 2008); female-perpetrated sex offenses are more likely to occur in the context of a caregiving situation, and with a male co-perpetrator (Grayston & De Luca, 1999); these females generally present with an interrelated set of needs, for example victimization, traumatic history, and mental health needs (Bloom, Owen, & Covington, 2003); interventions should target deficits in interpersonal, self-regulation, and distress-tolerance skills and should assist females to establish and maintain pro-social, supportive, and equitable relationships (Blanchette & Taylor, 2010). The present study provides support for the use of the Precondition Model as a framework for exploration of these issues, particularly when reinforced with a relapse prevention approach, such as the development of individualized 'safe living guidelines,' as in this case. However, this study also highlights limitations in the model's application with a female sex offender. These limitations and methodological limitations of the present study are discussed below.

Methodological limitations
Due to the minimal number of female child sex offenders referred for therapeutic intervention and further reduced number of female contact and noncontact offenders, female sex offender research is significantly impacted by natural limitations to sample size. In this project, the generalizability of results is compromised by the single case study. Therefore, caution must be taken when considering the external validity of the findings. However, the exploration of this female perspective of contact and noncontact sex offending is an important issue that warrants further research, and the present study may inform theory development and further conceptual understanding.
As described above, the outcome measures used were also not sensitive to nor directly linked to changes in thinking about deviant sex behavior or victim empathy. Therefore, assessing the efficacy of the intervention based on these measures is far from ideal. Each measure was selected due to its potential to capture factors relevant to the service-user's offending behavior. However, retrospectively there are clear disparities between the factors targeted by the therapeutic model and those measured. While this is a potential limitation of the current study, it does successfully illuminate a number of areas that the Precondition Model does not appear to address. Considerations for future research are discussed below.
The retrospective quantitative content analysis of the Precondition Models, while informative, may not provide the most useful account of intervention efficacy and outcome. It may also be subject to bias on behalf of both the clinician and service-user who were motivated to report improvement. As described by Johansson-Love and Fremouw (2006) the validity of self-report is a concern given that this population may also have secondary motives for reporting victimization, dysfunction, and emotional difficulties. In addition, Saunders (1991) suggested that the self-report responses to quantitative outcome measures completed by individuals who perpetrate interpersonal violence, including sex offenders, are likely to be affected by Social Desirability Response Bias (SDRB). SDRB includes both 'faking good' to make a good impression and 'faking bad' to highlight difficulties, care-seek, and potentially justify undesirable behavior. Therefore, it may have been beneficial to include collateral reports about changes in affect from sources that have prolonged contact with the service-user, such as the service-user's Probation worker, in addition to administering a separate scale of SDRB that can be used to statistically remove SDRB from the scores of self-report measures (Saunders, 1991).
Finally, a major limitation of the Precondition Model is that it does not directly address the moral emotion of shame. Klein, Joseph, and Zambrana (2012) discussed the role of shame in risk of recidivism for both male and female sex offenders. They highlighted a gap in the research in this field and recommended further exploration of gender differences in sex offenders' experience of shame.

Conclusion and recommendations for future research
This case study provides a unique insight into the perspective of a female contact and noncontact sex offender in relation to her offending behavior. It also highlights the benefits of using the Precondition Model with a female to address both types of offending, including elaboration on underresearched issues such as progression from noncontact to contact offending and motivation. Therefore, it would be of substantial benefit for the study to be replicated with additional female sex offenders. This would provide further insight into this underresearched group and allow for meaningful comparisons of the cognitive aspect of male and female sex offending. Future applications of this model would do well to be supported by collateral reports from other professionals in close contact with the service-user. In addition, replications of this study would reduce bias further by having coding schemes checked by completely independent reviewers.
This study illuminates a number of limitations of the model itself, in particular the absence of focus on factors such as shame and identity. Therefore, a priority for future research must be the development of a framework for exploring the individual nuances of sexual offending that addresses both the cognitive and emotional factors of the offending. Furthermore, the development of suitable outcome measures that have been validated for female sex offenders will greatly support understanding in this field.