JustResearch no. 12

Research in Profile (cont'd)

Family Involvement in the Treatment of Sexually Abusive Youth

By Kelly E. Morton Bourgon[22][23] & Guy Bourgon 1[24]


Research on general delinquency has shown that youth who have family involvement in treatment show more positive results (Howell & Hawkins, 1998; Latimer, Dowden & Morton-Bourgon, 2004). For example Multi-Systemic Therapy, which has shown positive results in the United States, includes the entire family system (as well as other systems) in the treatment process (Henggeler, Mihalic, Rone, Thomas, & Timmons-Mitchell, 1998). However, not all families choose to be involved in a youth's treatment process. Identifying the characteristics of a family that participates in treatment can assist service providers in pre-screening which families are more or less likely to become involved in treatment. This could allow clinicians to focus on the pertinent factors that may impede family involvement. In addition, studies have shown that parenting style affects a delinquent youth's future criminality (Hawkins, Herrenkohl, Farrington, Brewser, Catalano, & Harachi, 1998). However, little is known about parenting behaviours and sexually abusive youth. The identification of influential parental behaviours may assist clinicians in identifying the specific behaviours that require attention during “family work” in the treatment of sexually abusive youth.

This study investigated the following three questions:

  1. What type of family gets involved in a youth's sexual offender-specific treatment for youth?
  2. Do youth with family involvement demonstrate more improvement on outcome measures than youth without family involvement?
  3. Are parental behaviours related to treatment benefits for youth with family involvement?


  1. Estimate of Risk of Adolescent Sexual Offense Recidivism (ERASOR: Worling & Curwen, 2001) is a 25-item empirically guided risk instrument designed to assess risk for sexual recidivism in 12 to 18 year old youth who have previously committed a sexual assault.
  2. Youth Level of Service/Case Management Inventory (YLS/CMI: Hoge & Andrews, 1999) is a structured interview designed to assess the criminogenic risk and needs of youthful (< 19 years) offenders.
  3. Bronfenbrenner Family Questionnaire (BFQ: Bronfenbrenner, 1965) is a self-report instrument completed by the youth that attempts to measure the frequency of various parental behaviours as rated by the adolescent.
  4. The Treatment Outcomes for Adolescent Sexual Offenders (TOASO: Rich & Sauer, 1997) is a therapist rating scale designed to evaluate the adolescent's demonstration, or understanding, of certain behaviours and/or concepts generally covered in sex offender specific treatment

The participants in this study were part of a larger Canadian study on specialized services for sexually abusive adolescents with data provided by 15 agencies across Canada. Prior to data collection, each youth (and parent/guardian, where necessary) gave consent to participate in the project. Data was collected using a specific battery of tests administered to the youth and another battery administered to the youth's primary therapist. The method requested that each youth be assessed pre-treatment, halfway through the treatment program, and again at the end of treatment. Data on 127 youth had been collected at the time of this analysis: 126 males and 1 female. Ninety-seven youth were from Ontario, 22 from Manitoba and the remaining 8 were from Saskatchewan, the Yukon or Newfoundland. The mean age at the start of treatment was 15.4 years old ( SD = 1.8). One-third of the youth reported being sexually abused and 48% reported being physically abused. In 51% of the cases, there was no history of abuse reported. The family was involved in treatment in 72% of the cases.

Data on family composition were available for two-thirds of the youth ( n =84). Fifty-five percent of youth came from dual parent (intact) families; the remainder came from lone-parent or blended families. At the time of assessment, 59% were living with their family, 20% were living in a temporary placement situation, and 21% were permanently separated from their families.


What type of family gets involved in the treatment of sexually abusive youth?

Two strategies were employed to examine the characteristics of youth and their families that were associated with family involvement. The first strategy involved calculating correlations of specific demographic factors with family involvement. These factors were: age of the youth, family composition, closest parent to the youth, residing with potential victims, history of sexual abuse, previous agency involvement, having siblings, the victim's gender, type of offence, worst intrusion, number of victims, whether or not the youth was convicted, the age difference across victims, the source of the referral and the type of treatment.

Six variables were found to be significantly related to family involvement in treatment (see Table 1).

Table 1: Family more likely to be involved in youth's treatment if…
Characteristic r
Youth is younger .329
Closest parent is biological mother or father .328
Youth not convicted of the offence .307
Youth in less restrictive treatment (community, probation, residential) .304
Family has no/lower incidence of involvement with other agencies .191
The family is intact (dual-parent) .168

The second strategy involved comparing the risk level of youth with family involvement (FI) to youth with no family involvement (FNI). FI youth had significantly lower risk levels for both sexual and non-sexual recidivism, when compared to FNI youth (see Figure 1). In fact, FNI youth were assessed at a higher risk on 12 of the 13 subscales of the ERASOR and the YLS/CMI compared to FI youth.

Figure 1. Risk level of youth at intake.

Figure 1, vertical bar graph illustrating a comparison of risk levels for recidivism among young offenders whose family was involved in treatment versus those whose family was not, using two criminogenic risk scales.



It was found that a family was more likely to be involved in treatment when the youth was younger, their closest parental figure was a biological parent, the youth had not been convicted of their offence, the youth participated in less restrictive treatment, the youth/family had less prior agency involvement, and the family was considered biologically intact. This suggests that less problematic families, where the sexual offending is addressed at a younger age are more likely to be involved in treatment. Whereas the more problematic families, those with multiple experiences with the system (criminal justice or mental health), as well as other potential stressors (e.g., divorce, legal proceedings), were less likely to be involved in the youth's treatment. The significant differences found on both risk measures between FI youth and FNI youth provided additional support to this observation ­ that those youth with higher levels of risk and families with more needs are less likely to be involved in treatment. It is precisely in these higher risk cases that both youth and their family require services to prevent the probability of their criminal behaviour continuing into adulthood.

These results suggest that two different approaches, for the two different types of families, may be useful during the initial contact phase to enhance the possibility of involving families in the treatment process and improving the quality of the involvement. In the case of low risk youth, clinicians can adjust their approach to focus on the issues that are more likely to be salient to them, such as understanding the stress and/or the shame associated with a youth's sexual offending. In the case of the higher risk youth, a more effective approach may involve spending more energy persuading the family of the importance of their participation in the treatment process and how they can assist the youth in initiating and maintaining change. As such, motivational interviewing techniques may be of value. Even if the family remains reluctant, revisiting the possibility of their involvement at a later date should be encouraged. Additionally, it may be important to focus specifically on parental issues that may be impeding their involvement (e.g., divorce, tension within the family, stress from legal proceedings).

Do youth with family involvement demonstrate more improvement on outcome measures than youth without family involvement?

This question was investigated using separate Analysis of Variance examining pre ­ post changes on the ERASOR, YLS/CMI, BFQ and the TOASO. Surprisingly, only one significant difference was found, on negative parental behaviours. FNI youth rated their parents higher on negative parental behaviours after treatment compared to their pre-treatment levels. Contrary, FI youth rated their parents as significantly reducing their level of negative parental behaviours following treatment (See Figure 2).

Figure 2. Negative Parental Behaviours Scale

Figure 2. Negative Parental Behaviours Scale


Figure 3. ERASOR Psychosocial Functioning Scale

Figure 3. ERASOR Psychosocial Functioning Scale


Interestingly, we noted a trend, although not significant, across a number of the scales: treatment appeared to result in a greater degree of positive changes for the FNI youth in comparison to FI youth. Typically, FI youth were at lower risk, or demonstrating less of a problem in the area assessed, than the FNI youth. By the end of treatment, the two groups were about equivalent. The FI youth appeared to remain at the same level from pre to post treatment, whereas FNI youth appeared to show a decrease in risk level or problem. For example, the Psychosocial Functioning Scale of the ERASOR (Figure 3), the YLS/CMI total score (Figure 4) and the Social Skills Scale of the TOASO (Figure 5) illustrate this trend.

Figure 4. YLS/CMI Total Score

Figure 4. YLS/CMI Total Score


Figure 5. Social Skills Scale (TOASO)

Figure 5. Social Skills Scale (TOASO)



It was surprising that we did not find support for our hypothesis that FI youth would exhibit more positive changes as a result of treatment when compared to FNI youth. The data appeared to indicate that it was the FNI youth that were exhibiting a better response to treatment. A plausible explanation for this contra-intuitive finding can be found in the risk principle of Andrews and Bonta (2003). They state that level of services should be matched to the offender's risk level in order for the services to be most efficacious. That is to say that intensive treatment is most effective when provided to higher risk individuals and that treatment for lower risk individuals is most effective when those services are minimal and/or less intense. As described earlier, FNI youth were significantly higher risk pre-treatment than the FI youth. These higher risk youth have more needs and their problems are likely more complex and severe than the lower risk youth. According to the risk principle, it is these higher risk youth that demonstrate a greater degree of amenability to treatment.

Of course, this is not to say that youth who are of lower risk can not benefit from treatment for sexual offending. However, the question remains as to what level of intensity and duration of such services is needed in order to change the offending behaviour of lower risk youth. This is an important question as the average length of sex offender specific treatment in this sample was 18 months; it is possible that a shorter duration of treatment for lower risk youth may be sufficient. Future investigations into the relationship between treatment duration and re-offending would be helpful to determine the necessary and sufficient "dosage" level of treatment for youth of differing levels of risk in order to reduce the likelihood of future sexual and nonsexual criminal behaviour.

Are parental behaviours related to treatment benefits for youth with family involvement?

For this question, we examined data on only FI youth ( N =89). Parental behaviour was characterized on two different dimensions, positive and negative parental behaviours. For each dimension, parental behaviour was categorized as high or low using a mean split on each of the two parental behaviour scales of the BFQ at intake. In this fashion, a youth was considered to have parents scoring either high or low on positive parental behaviours and either high or low on negative parental behaviours. The high and low groups on each dimension were then compared on their post treatment BFQ scores, as well as on their pre-and post-risk scores. In this manner, we hoped to identify if the level of positive parental behaviours or negative parental behaviours were related to treatment changes.

Youth with high positive parental behaviours showed no significant changes across treatment compared to youth with low positive parental behaviour. Parents who started treatment with many positive behaviours, continued to maintain a high level of positive behaviours. The same was true for parents with a low level of positive parental behaviours ­ they maintained low levels. Additionally, there were no significant differences between high and low positive parental behaviour groups on the risk measures across treatment. This suggested that positive parental behaviours demonstrated no predictive ability to changes in the adolescent's behaviour over the course of treatment. In other words, for youth with parents involved in treatment, those with parents demonstrating many positive behaviours (high) did as well in treatment as those adolescents with parents demonstrating few positive behaviours (low).

Similar to positive parental behaviour, negative parental behaviour also showed little change over the course of treatment as approximately 50% of the parents rated as high on negative behaviours remained high after one year of treatment. However, negative parental behaviours appeared to play a mediating role in the youth's risk level over the course of treatment. Specifically, youth in the high negative behaviour group showed deterioration on 8 of the 12 risk subscales (the criminal history scale of the YLS/CMI was not included, as it is static). In other words, these youth were rated to be at higher risk at post-treatment compared to pre-treatment. On the other hand, youth in the low negative parental behaviour group were found to reduce their level of risk on 11 of the 12 subscales.


These results indicated that parenting style did not appear to change over the course of treatment, indicating its resiliency. Nonetheless, it appears that negative parental behaviours have an influence on the youth's responsiveness to treatment as seen by the changes in risk levels over treatment. Youth who reported their parents exhibited high levels of negative parental behaviour were found to increase their risk level over the course of treatment. Whereas youth who reported their parents exhibited few negative parental behaviours were found to reduce their level of risk over the course of treatment. This result, coupled with the resiliency of parenting style, would suggest that it may be beneficial for therapists to specifically address negative interactions within the family, with an emphasis on teaching and reinforcing positive and encouraging interactions between parent and child.


It has been illustrated that there are certain characteristics of families of sexually offending youth that are related to the family's participation in sex-offence specific treatment. Recognizing that families of higher risk youth are less likely to participate in the treatment process, treatment providers can take a more cautious approach to engaging these families in the treatment of the youth. It may be beneficial to utilize motivational interviewing techniques with the parents and/or provide assessment/counselling services for marital and family problems that are sometimes viewed as distal to the sexual offending behaviours of the youth. On the other hand, families of lower risk youth are more likely to be involved in the treatment process. Being cognizant of this can permit clinicians to expedite the engagement of the family and parents into treatment and may start to deal with key treatment issues earlier on in the treatment.

Nonetheless, the results of the present investigation did not support the general view that a critical factor to a youth's responsiveness to treatment is the involvement of his/her family in the treatment process. Our results indicated the opposite; youth responded better to treatment if their family was not involved in treatment, in comparison to those youth with their families involved. However, upon closer examination of the data, we found that youth without their families participating in treatment were at higher risk to re-offend than those youth that had their families participating in treatment. In light of this difference, we propose that these results reflect Andrews and Bonta's risk principle: that treatment intensity and duration should match the level of risk. This principle indicates that, when treatment is rather intense and lengthy, as it typically is for sexually aggressive youth, it is these higher risk cases that would demonstrate the most benefit.

Finally, for those youth whose families were involved in treatment, it was found that negative parental behaviours, not positive parental behaviours, played a significant mediating role to changes in risk for sexually abusive youth in treatment. These results suggest that it may be very important to include specific interventions (such as ways to replace negative behaviours with more positive and encouraging ways of interacting and disciplining the youth) in the family/parental components of a treatment program. This may be crucial, as parenting behaviours appeared to be rather resilient to change.

In conclusion, the precise role of the family in the treatment of youth, particularly with sexually abusive youth, remains to be elucidated. There are still few published studies examining the overall effectiveness of sexual offence specific treatment with youth (Worling & Curwen, 2000). Further research is necessary to identify the nature of a youth's responsiveness and progress, and how this is associated with the involvement of his family in the treatment process.