Efficacy of Treatment of Juvenile Sexual Offenders
A Meta-analytic study
Master Thesis Forensic Child and Youth Care Sciences
University of Amsterdam
L.M.Th. Blok
Student number 10013598
Under supervision of: Drs. E. Ter Beek, Prof. Dr. J. Hendriks, and Prof. dr. G.J.J.M. Stams Amsterdam, January, 2014
Abstract
The aim of this study is to examine the efficacy of treatment of juvenile sexual offenders in
the reduction of (sexual) delinquent behavior. For this study all reports from 1985 until 2013
addressing the efficacy of treatment of juvenile sexual offenders were included, with a total of
8 studies comprising 1224 assigned juveniles. The overall mean effect size for the
effectiveness of treatment was not significant (d = 0.177, Z = 0.720), which indicates that
treatment does not (yet) contribute to reduction in recidivism. However, the current
meta-analysis showed that treatment results of juvenile sexual offenders on recidivism are affected
by important moderators. Treatment of juvenile sexual offenders has more effect on reducing
recidivism when it is briefly (outpatient) offered. Further research is important, and should
examine how juvenile sexual offenders can be offered effective evidenced based treatment.
Keywords: juvenile sexual offenders, effectiveness, treatment, multilevel meta-analysis
Introduction
Sexual offenses usually trigger a sense of horror and an unscrupulous societal
response, especially when children are the victims. These sexual offenses are not only
committed by adults, but also by adolescents. Of all sexual offenses in the Netherlands one in
five are committed by adolescents (Brouwers & Smit, 2005). From 2008 until 2012, 496
adolescents were suspected or convicted of a sexual offense in the Netherlands (Valstar &
Afman, 2013). After treatment, adjudicated juvenile sexual offenders return to society.
Although research shows that the probability of adolescent sexual reoffending is low (e.g.,
Caldwell, 2010; Hendriks & Bijleveld, 2005; Reitzel & Carbonel, 2006; Wijk, Bullens, &
Eshof, 2007), it is still important to provide treatment. Acts of sexual aggression against
children can never be left without response, since society and victims are both harmed.
Furthermore, proper treatment is thought to be effective in reducing (even the smallest
amount) of recidivism (e.g., Hendriks & Bijleveld, 2005; Worling & Curwen, 2000).
The past years a growing body of research has been conducted to examine the efficacy
of treatment of sexual offenders. Every study still stresses the importance of further research
on this topic (e.g., Fanniff & Becker, 2006; Reitzel & Carbonell, 2006; Worling, Litteljohn, &
Bookalam, 2010; Worling & Curwen, 2000). Research of Worling and Curwen (2000) shows
that in the year 2000 ten studies were published that examined specialized adolescent sexual
offender treatment and subsequent recidivism rates.
The treatment of juvenile sexual offenders became an issue since the publication of
inter alia Abel, Mittelman, and Becker (1985). Abel et al. (1985) suggested that experiencing
of deviant arousal for a lot of adult sexual offenders started when they were adolescents. This
is in line with research by Abel, Osborn and Twigg (1993), showing that a lot of adult sexual
delinquents fantasized in their youth about sexual abuse or started with sexual abuse.
Studies examining differences between adult and adolescent sexual offenders
(Caldwell, 2010; Letourneau & Borduin, 2008) show that juvenile sexual offenders do not
often become adult sexual offenders (e.g., Hendriks & Bijleveld, 2012; Wijk, Bullens, &
Eshof, 2007). Also, the study of Zimring, Jennings, Piquero and Hays (2009) showed that
juvenile sexual offenders not automatically become adult sexual offenders. Only one in ten
juvenile sexual offenders becomes an adult sexual offender. However, in recent years several
meta-analyses on treatment effectiveness were conducted, in which adolescent and adult
sexual offender treatment programs were evaluated together (e.g., Losel & Schmucker, 2005).
This despite the fact that it is not entirely clear whether the factors that influence recidivism in
adulthood also affect recidivism among juvenile sexual offenders (Caldwell, 2010; Hendriks
& Bijleveld, 2005). Moreover, is it not possible to generalize to the population of adolescents
when adults are included.
To our knowledge, there is only one meta-analysis of juvenile sexual offenders
focusing on efficacy of treatment. Reitzel and Carbonell (2006) integrated nine studies of
adolescent sexual offender treatment. Their meta-analysis, however, has a number of
drawbacks. Firstly, Reitzel (2006) used a fixed effects models, which limits the
generalizability of the results. In the second place, Reitzel examined only few moderators,
including maximum follow-up time, recidivism adequacy (appropriateness of follow-up for
identifying sexual recidivism), study setting, average duration of treatment, treatment format
of primary treatment, treatment modality of primary treatment and recidivism criteria (e.g.
suspected, arrested, conviction). Finally, Reitzel made use of both published and
unpublished studies. By the lack of peer review these unpublished studies may be restricted in
quality. By conducting peer review the inadequate studies are removed and a minimum of
study quality will be preserved (Jefferson, Alderson, Wager, & Davidoff, 2002).
The present study makes use of multilevel analytic techniques, which makes it
possible to investigate more moderators. Also, more statistical power can be generated
because more effects per study can be included. Finally, new studies have been published the
last six years, which will be included in the present meta-analysis.
Interventions that are offered to juvenile sexual offenders are mostly compulsory
because of the forensic context. Opposed to other treatment settings, the information that is
discussed during juvenile offender treatment is often (partly) available to other stakeholders,
such as the Council for Child Protection, the court, juvenile probation officers and the lawyer
of the victim. The goals of the treatment for juvenile sexual offenders are primarily to prevent
deviant sexual or non- sexual behavior against others by learning skills and techniques that
contribute to a social function in society and raise the offenders sense of responsibility for
their own behavior (Center for Sex Offender Management, 2006).
Timmer, Workel and Dijk (2001) divided the various interventions that are offered to
juvenile delinquents in the Netherlands roughly into four categories; 0-variant, Community
service sexuality,Outpatient treatment and Residential treatment. When the 0-variant is
indicated the juvenile sexual delinquents can, for example, be placed in a juvenile justice
institution. When the chance of recurrence is minimal and the juvenile sexual delinquents do
not have any form of psychopathology, then the 0-variant will be imposed. The Community
Service Sexuality will be imposed for first offenders and consists of ten meetings in which
attention is given to taking responsibility and education. These first offenders do not have
mental disorders and have not committed a serious sexual offense. When juvenile sexual
delinquents have a sexual disorder and committed one or more serious sexual offenses than
outpatient treatment is indicated. Conditions for this form of treatment are that
parent/caregivers can be involved in the treatment and that sufficient protective factors be
present. Outpatient treatment can be given individually or in groups and lasts on average 20
months. Finally, residential treatment will be indicated for juvenile sexual delinquents who
have a disorder in sexual matters, committed serious sexual offenses and do not have
protective environments (Timmer, Workel, & Dijck, 2001).
Since 2010, however, the landscape of treatment opportunities changed. A separation
has been made between juveniles who are place on a civil justice title and a criminal justice
title. Based on this title juveniles can be placed in a youth welfare institution or a juvenile
correctional institution. In both cases they receive forced treatment. In case of juveniles
having many problems (e.g., aggression, addiction) and consequently treatment in regular
care is not possible (anymore), they will be placed in youth welfare institutions. Juveniles
who have committed offense(s) can be placed in a juvenile correctional institution on the
basis of custody, detention or placement in a correctional institution (Van Dam, Nijhof,
Scholte, & Veerman, 2010). As a results, classification of the various forms of treatment is
nowadays more difficult.
All forms of current treatment in the Netherlands use cognitive behavioral techniques
and relapse prevention strategies. All are based on the Risk-Need-Responsivity (RNR)/ What
Works principles (Andrews & Bonta, 2010). So all treatment forms can be described as
established treatment. Differences are found mostly in the way that treatment is administered
(to an individual, to a family system or to groups) and the intensity of the treatment (from a
few hours per week to admission in a 24 hour setting).
Biological forms of treatment (surgical castration and hormonal medication) also exist
and have been shown to be more effective than psychosocial interventions in terms of
recidivism reduction (Losel & Schmucker, 2005). For adolescents, however, these forms of
treatment are not preferred because of the serious damage that is done by them to an
individual growing up (ethical issues) and the fact that the medication used cannot be taken
continuously through life without severe side effects (e.g., Bruinsma, 1996). If such treatment
is considered in severe cases, it is administered on a voluntary basis or through consent by an
expert committee (Wille & Beier, 1989). Studies focusing on hormonal and surgical castration
are not included in the current meta-analysis because the effects of the biological forms of
treatment are not sufficiently comparable to the effects of psychosocial forms of treatment.
The aim of the current meta-analysis is to examine the efficacy of treatment in juvenile
sexual offenders in the reduction of (sexual) delinquent behavior and to conduct moderator
analyses by means of multilevel meta-analytic techniques.
Method
Inclusion criteria
The inclusion criteria were drawn up allowing population samples that are
representative of the juvenile offenders who have committed a sex crime. Studies included in
the present study had to meet the following criteria: 1).the participants in the study had to be
arrested, convicted of and/or acknowledged a sexual offense against a minor, 2).the
intervention should focus on the effectiveness of treatment of deviant sexual behavior, 3).the
study had to be conducted in the period of 1985 until 2013, 4).the study had to provide an
experimental and control group receiving treatment “as usual”, another kind of (established)
treatment or no treatment, 5).the total sample size of the study had to contain at least a sample
of 15 persons, 6).the study had to provide sufficient data for effect size calculation, 7).the
study should be reported in English or Dutch, 8).the study had to be published in a peer
reviewed journal.
Literature search
Different databases were consulted (e.g., Scienes Direct, Google Scholar, Springer, the
library of the University of Amsterdam, PsycINFO) between February 2013 en May 2013
using the search string “Sexual assault”, “Incest”, “Recidivism”, “Sex offenders”, “Sex(ual)
offender treatment”, “Residential treatment”, “Secure setting”, “Treatment efficacy”,
“Treatment effectiveness”, Juvenile sex(ual) offenders”, “Adolescent sex(ual) offenders”,
“Rapist”, “Child molest”, “Treatment outcome”, “Program” and “Intervention”. The search
was for books, articles, reviews, chapters and paper presentations. Also reference lists of
articles in this area were consulted (e.g., Caldwell, 2010; Reitzel & Carbonell, 2006). This
resulted in a total of ten studies (N= 1447 participants) that met the inclusion criteria, ending
in eight independent reports (N= 1299) that could be included. The included studies are listed
in Table 1.
Two reports could not be included (Seabloom, Seabloom, Seabloom, Barron, &
Hendricksons, 2003; Van Outsem, 2009). In the study of Seabloom et al. (2003) the issues of
the participants who received treatment were widespread and ranged from medical issues to
transgender issues and rape. So it is a contaminated group, considering several “problems”
that are combined and called sexual deviancy. Moreover, no data on recidivism was reported.
Finally, no established treatment was offered. In the study of Van Outsem (2009) the control
group was recruited at secondary schools, among non-offending adolescents. Considering the
fact that the current meta-analysis focusses on the effectiveness of treatment of juvenile sex
offenders this control group was not relevant.
Coding the studies
Study characteristics, participants characteristics, treatment characteristics and study
quality characteristics were coded. The following study characteristics were coded: year of
publication, authors, authors’ independence, study design (Randomized Controlled Trail or
Quasi- Experimental), year of publication, country where the study was conducted and the
year of data collection. Coded participants characteristics were gender, mean age, percentage
living with parents, percentage immigrants, percentage previously arrested, percentage type of
offense (hands-on, child molest, mixed). Treatment characteristics were coded for the
experimental group and the control group. The coded characteristics were treatment
completers, mean duration of treatment, total of participants, duration follow up, recidivism
and control for differences between the control and the experimental group. It was also
reported whether the control condition consisted of multiple treatment, single treatment or
non- treatment.
Analysis
For all studies Cohen’s d was computed by using the formulas fromLipsey en Wilson
(2001) and Mullen (1989). An effect of d= .80 was considered large, an effect of d= .50
medium and an effect of d= .20 small (Cohen, 1988). The program MLwiN was used for
conducting multilevel analyses (Hox, 2002). By making use of this program it is possible,
through the multilevel random effects model to analyze the hierarchical structure of the data.
Also, combined effect sizes and moderator- analyses can be calculated, whereby the lowest
level (effects sizes or study result) are embedded in the highest level (studies) (Hox, 2002;
Van den Noortgate, & Onghena, 2003).
First an empty (without moderators) random intercept only model was fitted to the
data. Subsequently, to investigate significant variation in effect sizes between the studies a
test for heterogeneity was carried out. Finally, the continuous and discrete moderators were
added to the model.
Results
For this thesis all reports from 1985 until 2013 addressing the efficacy of treatment at
juvenile sexual offenders were included, with a total of eight studies (presented in Table 1).
These eight studies comprised 1224 assigned juveniles. In total, 97,40 percent of all
participant were boys. All boys and girls were arrested, convicted for and/or suspected of
sexual offending. The mean age of al juveniles was M = 14.92, sd = .96. In total 556
juveniles received experimental treatment and 668 constituted the control group. The mean
duration in months of treatment in the experimental group was M = 17,11, sd = 10,61. The
mean duration in months of treatment in the control group was M = 10,42, sd = 3.87.
The overall mean effect size for the effectiveness of treatment was not significant (d =
0.177, Z = 0.720), which indicates that treatment does not (yet) contribute to reduction in
recidivism. The overall effect size proved to be heterogeneous (see Table 2). Therefore, we
conducted moderator analyses.
Participants characteristics
The effectiveness of treatment was moderated by the living situation of participants
Mixed samples, which consisted of juveniles living at home or in a residential setting, yielded
smaller effect sizes than studies consisting of juveniles living in a residential setting. Also, a
moderator effect was found for average previous arrests, indicating that studies consisting of
juveniles with a higher average previous arrest rate had larger effect sizes. Moderator tests showed also that an effect towards significance was found for percentage mixed offenses, indicating that studies consisting of juveniles with mixed index offenses (hands-on & child
molest) showed smaller effects sizes. Finally, a trend was found for the average age of the
participant, indicating that studies with higher average ages had smaller effect sizes.
Moderator test showed that the effect of treatment was not moderated by the status of
juveniles (M1), the type of recidivism (M3), percentage immigrants (M6), percentage living
with parents (M7), percentage hands-on offenses and percentage child molest offenses.
Study/ program characteristics
A moderating effect was found for the year in which the treatment was carried out ,
indicating that older studies showed smaller effects sizes. Also, an effect was found for study
design, indicating that studies with randomized controlled designs yield larger effects than
studies with quasi- experimental study designs. How larger the differences in months between
the experimental treatment and the control treatment how smaller the effects sizes, indicating
that longer treatment had less effect. Also, longer duration in months in both the experimental
and control group yielded smaller effects sizes. A moderating effect was found towards
percentage treatment completers in the control group, indicating that higher percentages of
treatment completers in the control group were associated with smaller effect sizes. Finally,
intention to treat analysis showed positive effects, whereas no intention to treat analysis
showed negative effects.
Discussion Introduction
The aim of the current meta-analysis was to examine the efficacy of treatment on
juvenile sexual offenders, in the reduction of (sexual) delinquent behavior and conduct
moderator analyses with a different methodology, multilevel meta-analytic techniques, in
comparison with the meta-analysis of Reitzel (2006). In contrast to Reitzel (2006) in the
current meta-analysis no significant overall effectiveness of treatment is found, so treatment
of adolescents who sexually offend does not (yet) contribute to reduction on recidivism. Also,
there is no found difference in the current meta-analysis between the type of re-offenses.
However, in agreement with Reitzel (2006) larger effect sizes were found in higher quality
studies. In the current meta- analysis it showed that the randomized controlled design yielded
more positive effects. This is also in line with research by Shadish, Matt, Navarro, and
Phillips (2000), who showed in a meta-analysis that selective allocation in nonrandomized
studies causing smaller effects.
The current meta-analysis showed that effects of treatment of juvenile sexual offenders
on recidivism are moderated by several of the included participants, study and program
characteristics. This suggests that the reduction of recidivism after treatment is influenced by
certain participants/program characteristics and that treatment types are not equally effective
and appropriate for each juvenile sex offender. These findings are in line with what Hendriks
(2011) pleads for; the need of the individual juvenile should be central during the treatment
rather than the treatment itself.
Outpatient treatment and residential treatment
When juvenile sexual offenders live at home it positively influences the effectiveness
of treatment in the reduction of (sexual) delinquent behavior. This in in line with previous
research (e.g., Hendriks & Bijleveld, 2005) that showed that the living situation is an
important factor that influences inter alia the effect of treatment on recidivism. Hendriks and
Bijleveld (2005) showed that there is a difference between juvenile sexual offenders who
received treatment in a residential setting and juvenile sexual delinquents who received
outpatient treatment. Juvenile sexual offenders who received outpatient treatment reoffended
relatively less than juvenile sexual offenders who received residential treatment. Also, the
study of DiGiorgio-Miller (2007) indicates that juvenile sexual offenders who received
residential treatment showed more abnormal sexual fantasies than juvenile sexual offenders
who received outpatient treatment. This is in line with research of Zakireh, Ronis and Knight
(2008), who claimed that juvenile sexual offenders receiving treatment in a residential setting
reported a more negative aggressive and sexual mentality than juvenile sexual offenders who
received outpatient treatment. This raises the question whether juvenile sexual delinquents
who are treated residentially are in any case more at risk of recidivism. If so, the positive
effects of living at home would be interpreted in this light rather than the effectiveness of
treatment type. More research into the differences between groups will have to be conducted
to explain this.
Risk- Need- Responsivity model
The current meta-analysis shows also that treatment is more effective when juvenile
sexual offenders have been arrested more often. This could be explained by research
conducted by Andrews and Bonta (2010). Andrews and Bonta (2010) emphasized in their
research how important it is to offer treatment whereby a difference is made between the risk
of the offenders to reoffend, the criminogenic needs of offenders and to reduce delinquent
behavior by cognitive social learning adapted to the learning capacity and style of offenders.
This is known as the Risk- Need- Responsivity (RNR) model. Offenders with higher risks of
reoffending (i.e., sex offenders with higher rates of previous arrest) need a more intensive
treatment than offenders with low or moderate risks of reoffending. Also, research by
Andrews and Dowden (2006) showed that when treatment is offered to low- risk offenders
this gives negative results on recidivism in some cases.
In the current research this result is also supported by the duration in months in the
experimental and the control treatment and the differences in treatment duration. When the
duration of treatment in the experimental group and the control group is longer, the effect on
reducing recidivism is less. Also, if the duration in treatment between the experimental and
control group is longer, the effect of treatment in reducing recidivism is less. The current
research also supports that offenders who have committed different offenses (hands-on &
child molest) benefit less from treatment, also older adolescents benefit less. These results
stress again the importance of the RNR model from Andrews and Bonta (2010):the type of
treatment should be attuned on the individual.
It can be derived from the results that the treatment of juvenile sexual offenders has
more effect on reducing recidivism when it is relatively brief, preferably offered in outpatient
settings. Thereby, treatment for this heterogenic group should be sufficiently intensive, but
also brief, for the juveniles who have a higher risk of reoffending. The juvenile sexual
offenders who have a lower risk of reoffending need less intensive and also brief outpatient
treatment.
Besides the risk of reoffending, it is also important to look at the factors that influence
the chances to become a juvenile sexual offender. Various studies (e.g., Hendriks, 2006;
Hissel, Bijleveld, Hedriks, Jansen, Collot, & d’Escury- Koenings, 2006) have shown that
juvenile sexual offenders are a heterogenic group, and distinction can be made between the
juvenile sexual offenses, their re-offenses and criminogenic factors.
Generalist and specialist juvenile sexual offenders
In research of Wijk (2005) there is made a distinction between generalist and specialist
juvenile sexual offenders on the basis of criminogenic factors. ‘Generalists’ are defined as
juveniles whose sexual offense is just one of many different type of offenses, mainly the
antisocial (criminal) behavior is central. ‘Specialists’ are defined as juveniles who (almost)
only commit sexual offenses. This is in line with findings of Butler and Seto (2001) who
suggest that adolescent sex only offenders (specialists) reoffend less than adolescent sex-plus
(generalist) offenders. Hissel et al. (2006) suggest that it would be logical in treatment to put
more focus at specialist offenders on the sexual aspect and with generalist offenders on the
antisocial aspect.
In response to the difference between specialist en generalist offenders, described
above, is it a remarkable result that in the current meta-analysis no difference is found in the
type of offending. No difference was found for effectiveness of treatment between sex
offenses, any offenses, and non- sex offenses. Reitzel (2006) found a difference in recidivism
on sexual offenses, non-sexual offenses and non-violent offenses. Recidivism on sexual
offenses proved to be less than recidivism on non-sexual and non-violent offenses. As
previously discussed, the probability of juvenile sexual reoffending is low (e.g., Caldwell,
2010; Hendriks & Bijleveld, 2005; Reitzel & Carbonel, 2006; Wijk, Bullens, & Eshof, 2007).
This results in the expectation that the percentage of sex offenses would significantly
differentiate negatively from any offenses and non-sex offenses. As a difference can be made
between juvenile sexual offenders and their background, the expectation is that the influence
of treatment on re-offenses could be different. Further research into the differences between
different groups sexual offenders will have to be conducted in different treatment settings.
Treatment efficacy
The period in which the juvenile sexual offenders have received treatment also
effected the efficacy of treatment (a trend). Research conducted after the nineties yielded
more positive effect sizes compared to less positive effects sizes before the eighties and
negative effects sizes between de eighties and nineties. This indicates that more recent
programs show better results. This is in line with the meta-analysis of Hanson, Bourgon,
Helmus and Hodgens (2009).
Most of the studies included in the current meta-analysis are cognitive behaviorally
based. Previous meta-analysis showed that cognitive behaviorally based treatment yielded
larger effect sizes than non-cognitive behaviorally based treatment (e.g., Hanson et al., 2002;
Lösel & Schmucker, 2005).The current meta-analysis showed that cognitive based treatment
after the nineties yielded better results than cognitive behaviorally treatment before the
nineties. This indicates that more recent cognitive behaviorally based treatment shows better
results in the reduction of recidivism.
Considering the relative short period of specific treatment developed for juvenile
sexual delinquents, and in addition the limited availability of efficacy studies, it could be
argued that treatment for juvenile sexual delinquents is still in its early stages. However, in
recent years there has been a lot of focus on the use of Risk Need and Responsivity and the
What Works principles of effective judicial interventions (Andrews and Bonta, 2010). It can
be postulated that these current additions to the use of cognitive behavioral techniques have
improved treatment and increasingly meet the need of the individual juvenile.
Limitations
The strong points of the current meta-analysis are that it included the most recent
information about the effectiveness of treatment for juvenile sexual offenders. It is the first
multilevel meta-analytic study conducted on the effectiveness of treatment targeting
recidivism of juvenile sexual offenders. However, this study also has limitations. Firstly,
there is a limited number of studies available on the effectiveness of treatment offered to
juvenile sexual offenders. Only eight published studies could be included. Therefore, the
results of the analyses should be interpreted with caution.
A second limitation is the exclusion of unpublished studies. Several studies (e.g.,
Hopewell, Mc Donald, Clarke, & Egger, 2007) show that unpublished studies, on average,
showed smaller effects on treatment than published studies. Exclusion of unpublished studies
could influence the actual effects size and could lead to overestimation of the effects. Also,
Littell (2008) stresses the importance of more scientific studies whereby non-significant
results will not be ignored. A combination of published and unpublished studies could lead to
more objective results, which might improve treatment.
Thirdly, in the current meta-analysis the outcome measure is recidivism. This while
treatment focuses on dynamic factors that theoretically help prevent the chance of recidivism,
not the recidivism itself. Therefore, treatment effectiveness might not be accurately
measurable by recidivism only. For further research not only recidivism must be examined as
an outcome measure, but also dynamic criminogenic factors, such as psychopathology,
cognitive biases, family functioning and social functioning. This makes it possible to focus on
the factors that are more directly the focus of the interventions.
Further research
The current meta-analysis stresses the importance of further research on the
effectiveness of treatment for juvenile sexual offenders. It is known that there is a dark
number of sexual offenses (e.g., Hissel, Bijleveld, Hedriks, Jansen, Collot, &
d’Escury-Koenings, 2006), which makes it hard to gain full insight in the offense and re-offenses
committed. Considering that acts of sexual aggression against children can never be left
without response, since society and victims are both harmed, it is important to provide
treatment that is effective in reducing (even the smallest amount) of recidivism. Further
research should ensure that juvenile sexual offenders will be offered effective, evidenced
based, treatment.
References
Abel, G.G., Mittelman, M.S., & Becker, J.V. (1985). Sexual offenders: Results of assessment
and recommendations for treatment. In Ben-Aron, M.H., Hucker, S.J., & Webster,
C.D. (Eds). Clinical Criminology: The assessment and treatment of criminal behavior
(191-205) Toronto: University of Toronto.
Abel, G.G., Osborn, C.A., & Twigg, D.A. (1993) Sexual assault through life span: Adult
offenders with juvenile histories. In H.E. Barbaree, W.L. Marshall, & S.M. Hundosn
(Eds.), The juvenile sex offender (104-117). New York: The Guildford Press.
Andrews, D., & Bonta, J. (2010). Rehabilitating criminal justice policy and practice.
Psychology, Public Policy, and Law, 16, 39-55. doi: 10.1037/a0018362
Andrews, D. A., & Dowden, C. (2006). Risk principle of case classification in correctional treatment. International Journal of Offender Therapy and Comparative Criminology, 50, 88–100.
Borduin, C.M., Henggeler, S.W., Blaske, D.M., & Stein, R.L. (1990). Mutlisystemic
treatment of adolescent sexual offenders. International Journal of Offender Therapy
and Comparative Criminology, 996(3), 105-113.
Butler, S.M., & Seto, M.C. (2001). Distinguishing two types of adolescent sex offenders.
American Academy of Child & Adolescent Psychiatry, 41, 83-90.
Caldwell, M.F. (2010). Study characteristics and recidivism base rates in juvenile sex
offender recidivism. International Journal of Offender Therapy and Comparative
Criminology, 54(2), 197-212.
Center for Seks Offender Management. (2006). Understanding treatment for adult and
juveniles who have committed sex offenses. U.S. Department of Justice, Office of
Justice Programs.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ:
Lawrence Erlbaum.
Dam, C. van, Nijhof, K., Scholte, R., & Veerman, J.W. (2010). Evaluatie Nieuw Zorgaanbod:
Gesloten jeugdzorg voor jongeren met ernstige gedragsproblemen. Eindrapport.
Nijmegen: Radboud Universiteit & Praktikon.
DiGiorgio-Miller. J. (2007). Emotional variables and deviant sexual fantasies in adolescent
sexual offenders. The Journal of Psychiatry & Law,35, 109-124.
Fanniff, A.M., & Becker, J.V. (2006). Specialized assessment and treatment of adolescent sex
offenders. Aggression and Violent Behavior, 11, 265-282.
Hanson, R.K., Bourgon, G., Helmus, L., & Hodgens, S. (2009). A meta-analysis of the
effectiveness of treatment for sexual offenders: Risk, Need, and Responsivity 2009-01.
Ottawa, Ontario, Canada: Public Safety Canada.
Hanson, R. K., Gordon, A., Harris, A. J. R, Marques, J. K., Murphy, W. D., Quinsey,
V. L., & Seto, M.C. (2002). First report of the collaborative outcome data project on
the effectiveness of psychological treatment for sex offenders. Sexual Abuse:
Journal of Research and Treatment 14, 169-194.
Hendriks, J. & Bijleveld, C. (2005) Recidive van jeugdige zedendelinquenten na residentiele
behandeling. Tijdschrift voor Seksuologie, 29, xxx-xxx
Hendriks, J. (2011). Wie Joost weet, mag het zeggen. Den Haag: Boom Lemma uitgevers
Henggeler, S.W., Schoenwald, S.K., Borduin, C.M., & Rowland, M.D. (2009). Multisystemic
therapy for antisocial behavior in children and adolescents (2nd. ed). New York: The
Guilford Press.
Hissel, S., Bijleveld, C., Hendriks. J., Jansen, B., & d’Escury-Koenings, A.C. (2006).
Jeugdige zedendelinquenten: specialisten, generalisten en ‘first offenders’. Tijdschrift
voor Seksuologie, 30, 215-225.
Hopewell, S., McDonald, S., Clarke, M., & Egger, M. (2007). Grey literature in
meta-analyses of randomized trials of health care interventions. Cochrane Database of
Systematic Reviews 2007, (2), Art. No.: MR000010. doi: DOI:
10.1002/14651858.MR000010.pub3.
Hox, J. (2002). Multilevel analysis: techniques and applications. Mahwah: Lawrence
Erlbaum Associates Publishers.
Jefferson,T., Alderson, P., Wager, E., & Davidoff, F. (2002). Effects of editorial peer review.
JAMA, 287,2784-2786.
Letourneau, E.J., & Borduin, C.M. (2008). The effective treatment of juveniles who sexually
offend: An ethical imperative. Ethics & Behavior, 18, 286-306.
Lipsey, M. W., & Wilson, D. B. (2001). In Bickman L., Rog D. J. (Eds.), Practical
meta-analysis. Thousand Oaks, CA: SAGE Publications.
Lösel, F., & Schmucker, M. (2005). The effectiveness of treatment for sexual offenders: A
comprehensive meta-analysis. Journal of Experimental Criminology, 1, 117-146.
Mullen, B. (1989). Advanced BASIC meta-analysis. Hillsdale, New Jersey: Lawrence
Erlbaum Associates, Publishers.
Seabloom, W., Seabloom, M.E., Seabloom E., Barron, R., & Hendrickson, S. (2003). A 14-to
24 year longitudinal study of a comprehensive sexual health model treatment program
for adolescent sex offenders: Predictors of successful completion and subsequent
criminal recidivism. International Journal of Offender Therapy and Comparative
Criminology, 47(4), 468-481. DOI: 10.1177/0306624X03253849.
Shadish, W. R., Matt, G. E., Navarro, A. M., & Phillips, G. (2000). The effects of
psychological therapies under clinically representative conditions: A
meta-analysis. Psychological Bulletin, 126(4), 512-529.
Reitzel, L. & Carbonell, J.L. (2006) The effectiveness of sexual offender treatment for
juveniles as measured by recidivism: A Meta-analysis. Sex Abuse, 18, 401-421.
Thomas, B. H., Ciliska, D., Dobbins, M., & Micucci, S. (2004). A process for systematically
reviewing the literature: Providing the research evidence for public health nursing
interventions. Worldviews on Evidence-Based Nursing, 1(3), 176-184. doi:
10.1111/j.1524-475X.2004.04006.x.
Timmer, S., Workel, M., & Dijck, H.M. (2001). Opvang en behandeling van jeugdige
zedendelinquenten in Nederland. Een verkennend onderzoek. Den Haag: Ministerie of
justitie.
Van den Noortgate, W., & Onghena, P. (2003). Multilevel meta-analysis: A comparison with
traditional meta-analytical procedures. Educational and Psychological Measurement,
63(5), 765-790. doi: 10.1177/0013164402251027.
Van Outsem, R.E. (2009). Exploring psychological characteristics of sexually abusive
juveniles (Doctoral dissertation). Vrije University Amsterdam.
Valstar. H. & Afman, T. (2013). JJI in getal 2008-2012. Dienst Justitiële Inrichtingen.
Ministerie van Veiligheid en Justitie.
Van Wijk, A. Ph. (2005). Juvenile sex offenders and non- sex offenders. A comparative
study (Doctoral Dissertation). Wageningen: Ponsen & Looijen.
Van Wijk, A.Ph., Bullens, R.A.R., & Eshof, P. van den. (2007). Facetten van
zedencriminaliteit. ’s-Gravenhage: Reed Business Information bv.
Worling, J.R., Curwen, T. (2000). Adolescent sexual offender recidivism: Succes of
specialized treatment and implications for risk prediction. Child Abuse & Neglect, 24,
965-982.
Worling, J.R., Litteljohn, A., & Bookalam, D. (2010). 20-Year Prospective follow-up study
of specialized treatment for adolescents who offended sexually. Behavioral Sciences
and the Law, 28, 46-57.
Zakireh, B, Ronis, S.T., & Knight, R.A. (2008). Individual beliefs, attitudes, and victimization
histories of male juvenile sexual offenders. Sexual Abuse: A journal of Research and
Treatment, 20, 232-351. DOI: 10.1177/1079063208322424.
Zimring, E.F., Jennings, W.G., Piquero, A.R., & Hays, S. (2009). Investigating the continuity
of sex offending: Evidence from the second Philadelphia Birth Cohert. Justice
Quarterly, 26(1), 58-76.
Table 1. Summary of the studies that were included in this meta-analytic study
Study Year Background Participants
Total N Man (%) Migrant(%) Study quality
Borduin, Henggeler, Blaske, & Stein 1990 Arrested 16 100 38 Weak Lab, Shields, & Schondel 1993 Convicted 155 98 43 Moderate Worling & Curwen 2000 Convicted
and/or admitted
148 94 - Moderate
Hendriks & Bijleveld 2005 Convicted 325 100 32 Moderate Waite, Keller, McGarvey,
Wieckowski Pinkerton, & Brown
2005 Convicted 261 100 57 Moderate
Henggeler, Letourneau, Chapman, Borduin, Schewe, & McCart
2009 Suspected 127 98 54 Strong
Borduin, Schaeffer, & Heiblum 2009 Arrested 48 96 29 Strong Worling, Litteljohn, & Bookalam 2010 Convicted
and/or admitted
148 94 - Moderate
Table 2
Results for the overall mean effect size and discrete moderators for recidivism (bivariate models).
Moderator variables # Studies # ES Β1 ZB1 Mean d Zd Hetero X2 Overall 8 20 0.177 0.720 1.762* || Participant characteristics M1 Status 1.77* 0.57 Convicted (RC) 3 15 0.066 0.237 Suspect 2 5 0.411 0.765 0.477 1.035 M2 Living 1.532+ 9.709* Residential (RC) 1 3 0.385 1.390 + Living at home 4 8 0.222 0.659 0.607 3.178* Mixed 2 6 -1.065 -3.096* -0.68 3.350* Unknown 1 3 -0.113 -0.289 0.272 0.986 M3 : Type of recidivism 1.764* 2.611 Sex offenses (RC) 7 7 ||- 0.133 0.528 Any offenses 7 7 0.127 1.411 + 0.260 1.032 Non-sex offenses 6 6 -0.002 -0.021 0.131 0.516 Study/program characteristics M11: Region of collected data 1.757* 0.03 USA (RC) 7 17 ||0.272 0.429 Europe 1 3 -0.112 -0.163 0.160 0.602 M12:Year Published 1.77* 0.121 Before 2004 (RC) 3 8 0.075 0.199 After 2004 5 12 0.173 0.352 0.248 0.782 M13: Year Treatment 1.683* 5.60+ >90 3 6 0.705 2.304* 90-80 3 9 -1.060 -2.704* -0.355 -1.455 + <80 2 5 -0.275 -0.611 0.430 1.307 + M14: Authors 1.76* 0.08 Independent (RC) 1 2 0.009 0.014 Dependent 7 18 0.196 0.283 0.205 0.774 M15: Study design 1.66* 6.81* Quasi-experimental (RC) 5 14 ||- -0.132 -0.733 Randomized controlled 3 6 |1.194 ||3.210* 1.062 3.258* M16: Experimental group 1.76* 0.08 Established treatment (RC) 7 18 0.205 0.774 Treatment as usual 1 2 -0,196 -0.283 0.009 0.014 M17: Control group 1.767* 0.713 Treatment as usual (RC) 3 5 0.475 1.044 Established treatment 4 12 -0,463 -0.843 0.012 0.039 Non treatment 1 3 -0.203 -0.267 0.272 0.448 M18: Control group 1.749* 2.11 Single treatment (RC) 3 3 1.046 1.663* Multiple treatment 4 14 -1.037 -1.532 + 0.009 0.040 Non treatment 1 3 0.774 0.926 1.820* 0.494 M19: Intention to treat 1.528+ 9.18** Yes (RC) 6 14 0.478 3.319* No 2 6 -1.157 -4.484* -0.679 -3.158*
RC = reference category; # Studies = number of independent studies; # ES = number of effect sizes; Z = difference in mean d with reference category; mean d = mean effect size (d); heterogeneity = within class heterogeneity (Z) +p<0.1; *p<0.05 (tested one-sided).
Table 3
Results for the continuous moderators for recidivism (bivariate models).
Moderator variables # Studies # ES β0 (SD) β1 (SD) Z Heterogeneity
Participant characteristics
M4: Average age of participants 8 20 0.176 (0.219) -0.326 (0.228)
-1.430 + 1.753
M5: Average previous arrest 3 8 0.501 (0.120) 0.230 (0.079)
2.911 * 0.8123
M6: Percentage immigrants 6 14 0.507 (0.166) -0.017 (0.016)
-1.063 1.329
M7: Percentage living with parents 5 20 0.180 (0.244) -0.004 (0.011)
-0.364 1.759
M8: Percentage handson offenses 6 20 0.173 (0.244) -0.002 (0.008)
-0.25 1.754
M.9: Percentage childmolest offenses 6 17 -0.003 (0.218) -0.011
(0.013)
-0.846 1.611
M10: Percentage mixed offenses 4 12 -0.171 (0.156) -0.122
(0.047) -2.596* 1.304 Study/program characteristics M21: Sample size 8 20 0.185 (0.239) -0.002 (0.002) -1.000 1.769 M22: Year of publication 8 20 0.176 (0.243) -0.017 (0.035) -0.486 1.765
M23: Different treatment duration 4 9 0.853 (0.109) -0.035 (0.009)
-3.889* 0.000
M24: Mean duration months exp 7 18 0.233 (0.218) -0.051 (0.023)
-2.217* 1.585
M25: Mean duration months ctrl 4 9 0.853 (0.109) -0.124 (0.031)
-4.000* 0.000
M26: Mean duration follow up recidivism
6 18 0.210 (0.288) -0.001 (0.014)
-0.071 1.647
M27: Percentage treatment completers Experimental group
8 20 0.173 (0.236) 0.008 (0.012)
0.667 1.766
M28: Percentage treatment completers Control group
6 15 0.231 (0.265) -0.038 (0.020)
-1.9* 1.457
RC = reference category; # Studies = number of independent studies; # ES = number of effect sizes; Z = difference in mean d with reference category; mean d = mean effect size (d); heterogeneity = within class heterogeneity (Z), = difference with model without moderators (χ2). +p<0.1; *p<0.05 (tested one-sided).
