The relationship between the ‘Central Eight’ risk
factors and self-reported recidivism after discharge
from youth psychiatric residential care
Master thesis Forensic Orthopedagogy Esther de Vries [11408219] Graduate School of Child and Education, University of Amsterdam (UvA) Research Group Forensic Mental Health, GGzE First supervisor: Dr. Titia van Zuijen (UvA) Second supervisor: Drs. Cisem Gürel (UvA) External supervisors: Dr. Ilja Bongers & Drs. Lisette Janssen – de Ruijter (GGzE)
2 Abstract
Andrews and Bonta’s Central Eight risk factors best predict recidivism among delinquent adolescents. In this study, the question which of Andrews and Bonta’s Central Eight risk factors best predicted general and violent recidivism in a group of patients who were discharged from youth psychiatric residential care was answered. The predictive validity of the Central Eight risk factors was examined, using self-report recidivism data. Hereby, the risk factors that were present at the moment of discharge from youth residential care were used. The sample included 46 former patients of a clinic for youth forensic psychiatry and orthopsychiatry in the Netherlands. Recidivism during the 1 to 6 year follow-up period was measured using self-report data. Concerning general recidivism, a model consisting of one Central Eight risk factor, antisocial personality pattern, was found to be best predictive (Nagelkerke R² = .27, OR = 13.85, 95% CI = 1.56-122.97). Regarding violent recidivism, a model with antisocial personality pattern (model Nagelkerke R² = .45, OR = 15.45, 95% CI = 2.60-91.87) and low interest or commitment to school and/or work (OR = 7.31, 95% CI = .93-57.21) was best predictive in our study. This study suggests that aftercare programs for delinquent adolescents should focus on maladaptive, antisocial personality characteristics. Furthermore, interventions that focus on the adolescent at school and/or work should be applied in order to reduce violent recidivism.
Keywords: Central Eight risk factors, delinquent juveniles, general recidivism, violent
3 Introduction
Recidivism rates of adolescents who were discharged from youth judicial institutions are high (Wartna, Tollenaar, Verweij, Alberda & Essers, 2016). Official data show that more than half (57.6%) of the adolescents who were discharged in 2012 recidivated in violent or non-violent offenses within two years after discharge (Wartna, Tollenaar, Verweij, Alberda & Essers, 2016). Moreover, in the years between 2002 and 2012, recidivism rates in ex-detained adolescents showed a smaller decrease (-5.3%) than in ex-detained adults (-14.8%) (Wartna, Tollenaar, Verweij, Alberda & Essers, 2016). Despite the small decrease in recidivism rates, recidivism among delinquent adolescents is still a serious problem which threatens the safety of society and causes financial burdens (Rijksoverheid, 2015).
In the past years, much research has been conducted on risk factors for recidivism in youth. In a well-known study, Andrews and Bonta (2010) found eight risk factors predicting recidivism among delinquent juveniles. These so-called ‘Central Eight’ risk factors are subdivided in two categories: the ‘Big Four’ risk factors (history of antisocial behavior, antisocial personality pattern, antisocial cognition, and antisocial associates) with a high impact on recidivism, and the ‘Moderate Four’ risk factors (family, school, leisure/recreation, and substance abuse) with a smaller impact on recidivism. The Big Four are crucial risk factors that need to be addressed in treatment directly, according to Andrews and Bonta (2010). Additionally, the Moderate Four are environmental risk factors which have both a direct effect (by causing opportunities for criminal behavior) as well as an indirect effect (by interacting with the Big Four) on criminal behavior, and these factors need to be worked on in treatment as well (Andrews & Bonta, 2010).
The predictive validity of all Central Eight risk factors for recidivism was confirmed in a study of Grieger and Hosser (2013). In their study, Grieger and Hosser (2013) also investigated whether there were individual Central Eight risk factors that better predicted
4
general and violent recidivism than the other Central Eight factors. Concerning general recidivism, they found that four factors are best predictive, i.e. history of antisocial behavior, antisocial associates, school, and substance abuse. Other studies, in which not all Central Eight factors were included, confirm the impact of history of antisocial behavior (Donker & Bakker, 2012; Olver, Stockdale & Wong, 2012), antisocial associates (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2014) and school (low interest or commitment to school) (Olver, Stockdale & Wong, 2012; Van der Put et al., 2014). In addition, regarding violent recidivism, Grieger and Hosser (2013) found that the best predictive risk factors are partly different from the risk factors for general recidivism. History of antisocial behavior and school were predictors for both general and violent recidivism, whereas antisocial cognition and leisure/recreation were strong predictors for violent recidivism but not for general recidivism. Additional evidence exists for the impact of history of antisocial behavior and school on violent recidivism (Olver, Stockdale & Wong, 2012). Although the impact of some of the risk factors is confirmed, this is not the case for all Central Eight risk factors. Hence, the current knowledge of the factors that best predict general compared to violent recidivism is inconsistent.
Knowledge of risk factors for general and violent reoffending is essential for health care professionals to address relevant risk factors in treatment (Krug, Dahlberg, Mercy, Zwi & Lozano, 2002). In order to lower the chance that these adolescents will recidivate, treatment in youth judicial institutions in the Netherlands is aimed at decreasing the presence of risk factors in adolescents who engaged in criminal behavior (Dienst Justitiële Inrichtingen, 2016). However, whereas risk factors are generally less present at the moment of discharge from residential care, several risk factors (e.g. substance abuse, low interest or commitment to school) could still be present (Megens & Day, 2007). Therefore, after residential care adolescents often move to a less secured setting or are provided with outpatient aftercare from
5
home (Dienst Justitiële Inrichtingen, 2016) instead of going home directly without help. However, in a less secured setting the adolescent has more freedom, which could increase the chance that he or she will reoffend (Kempes, 2012). A study of Wartna and his colleagues (2005) shows that most juvenile recidivists recidivate within one year after discharge from residential care. After one year after discharge, former patients were suspect in an average of two new criminal cases (Wartna et al., 2005). Therefore, in order to decrease recidivism among delinquent adolescents, relevant risk factors that are still present at the time of discharge from residential care need to be addressed in aftercare programs. To our knowledge, at this moment there are no studies which specifically focused on risk factors at the time of discharge from youth psychiatric residential care, in order to be able to improve aftercare programs.
In most studies on risk factors for recidivism, recidivism rates were measured using official data from police registrations and judicial documentation (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2012; Van der Put et al., 2014). A disadvantage of using official data is that many offenses are unknown by the police (Loeber, Slot & Sergeant, 2001; Van Domburgh, Geluk, Jansen, Vermeiren & Doreleijers, 2016; Weijters & Van der Laan, 2016). Due to the incompleteness of official data, self-report data have been used in an increasing amount of studies (Piquero, Schubert & Brame, 2014). A benefit of the use of self-report data is that it provides a more valid picture of the number of offences compared to official measures of offending (Joliffe, 2013). Furthermore, self-report data have been found to be reliable and useful (Coleman & Moynihan, 1996; Loeber, Slot & Sergeant, 2001). Therefore, using self-report data provides added value when collecting information about criminal behavior (Thornberry & Krohn, 2000; Weijters & Van der Laan, 2016). We believe that there are no studies which investigated all Central Eight risk factors using self-report recidivism data.
6
The aim of this study is to expand the existing research investigating risk factors for general and violent recidivism. This study examined the relationship between all Central Eight risk factors and general as well as violent self-reported recidivism in a sample of adolescents who were discharged from youth psychiatric residential care. The research questions are: 1) Which of the Central Eight risk factors best predict self-reported general recidivism in a group of patients who were discharged from youth psychiatric residential care?, and 2) Which of the Central Eight risk factors best predict self-reported violent recidivism in a group of patients who were discharged from youth psychiatric residential care? This study may help to expand the current knowledge about targeted aftercare programs for delinquent adolescents in order to reduce recidivism risk. Therefore, this study focused on risk factors which are present at the moment of discharge from youth residential care. The risk factors history of antisocial behavior, antisocial associates and school (low interest or commitment to school) were expected to best predict (self-reported) general recidivism (Donker & Bakker, 2012; Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2014), whereas the risk factors history of antisocial behavior and school were expected to best predict (self-reported) violent recidivism after discharge from youth psychiatric residential care (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012).
Methods Setting
The Catamaran is a hospital for youth forensic psychiatry and orthopsychiatry in Eindhoven, the Netherlands. At this hospital, residential treatment is offered to adolescents between 14 and 23 years who exhibit psychiatric and behavioral problems. Adolescents who have been admitted to the Catamaran have been sentenced under the Dutch juvenile criminal law, the Dutch juvenile civil law or on a voluntary basis. The Dutch juvenile criminal law
7
comprises treatment and rehabilitation of adolescents who have committed serious offenses. The Dutch juvenile civil law is applied to adolescents whose development is at risk and whose parents or caregivers are not able to provide the needed help. The aim of admission to the Catamaran is to help patients to control problem behavior and to work on a positive future perspective (GGzE, 2017).
Measures
Central Eight risk factors
The Structured Assessment of Violence Risk in Youth (SAVRY) was used to operationalize the Central Eight risk factors. The SAVRY (Lodewijks, Doreleijers, De Ruiter & De Wit-Grouls, 2006) is a risk assessment tool which is developed for adolescents between 12 and 18 years. However, when carefully used, the SAVRY can be used for young adults as well (Lodewijks, Doreleijers, De Ruiter & De Wit-Grouls, 2006). The SAVRY consists of 24 risk items (historical risk factors, social/contextual and individual risk factors) and six protective items. The presence of the risk items is coded as low, moderate or high; the protective factors are coded as present or absent. The inter-rater reliability of the SAVRY has been found to be good; the predictive validity for physical violent behavior against persons has been found to be excellent (Lodewijks, Doreleijers, De Ruiter & Borum, 2008).
In this study, the SAVRY items that best fit the definitions of the Central Eight risk factors were used. See Table 1 for an overview of the operationalization of the Central Eight risk factors using items of the SAVRY. A ‘high’ score on one or more of the SAVRY items belonging to the Central Eight risk factor was recoded into ‘present’ on the Central Eight risk factor, whereas ‘low’ and ‘moderate’ scores on the SAVRY item(s) were recoded into ‘absent’ on the Central Eight risk factor. Missing scores on SAVRY items because of insufficient information were coded as absent.
8 Table 1
Overview of the Operationalization of the Central Eight Risk Factors
Central Eight risk factor SAVRY items
History of antisocial behavior Item 1: History of violence
Item 2: History of nonviolent offending Antisocial personality pattern Item 18: Risk taking/impulsivity
Item 20: Anger management problems Item 21: Low empathy/remorse
Antisocial cognition Item 17: Negative attitudes Antisocial associates Item 11: Peer delinquency
Family/marital circumstances Item 15: Poor parental management
School/work Item 24: Low interest/commitment to school
or work
Leisure/recreation Item P1: Prosocial involvement (reversed) Substance abuse Item 19: Substance use difficulties
Self-reported recidivism
Self-reported recidivism data were collected by using a semi-structured interview. This semi-structured interview contains questions concerning different areas of the lives of former patients, for instance about treatment and jobs they had after discharge from the Catamaran. Regarding delinquent behavior after discharge, participants were asked whether they were convicted of one or more offense(s) after discharge and if yes, for which type(s) of offense(s) and how many times. Furthermore, they were asked whether they committed one or more offense(s) after discharge for which they were not convicted and if yes, for which type(s) of offense(s) and how many times. Finally, participants were asked whether they committed one or more offense(s) after discharge which are unknown by the police and if yes, which type(s) of offense(s) and how many times.
9
Both general recidivism and violent recidivism were used as dependent variables. General recidivism was defined as “any new custodial sentence after discharge” (Grieger & Hosser, 2013). Violent recidivism was defined using the definition of the SAVRY: “Violence is a deed of abuse or physical violence sufficient to cause an injury to one or more persons (for instance, cuts, bruises, bone fractures, dead, et cetera), no matter whether this injury really occurred or not; every form of sexual assault; or threat with a weapon. In general, these deeds need to be sufficiently serious to (could) have led to prosecution for criminality” (Lodewijks, Doreleijers, De Ruiter & De Wit-Grouls, 2006). First, a researcher and a trainee independently decided whether every offense either fell in the category “general recidivism” or in both categories “general recidivism” and “violent recidivism. After that, consensus was reached (in case the researcher and trainee did not agree).
Procedure
In order to score the SAVRY items, which were used to operationalize the Central Eight risk factors, the SAVRY was scored at the moment of discharge from the Catamaran. Scoring of the SAVRY items was based on file information concerning the last six months of admission, for example treatment plans and reports written by therapists. An exception was the scoring of the SAVRY items history of violence and history of non-violent offending, for which the entire patient files including information before admission were used. Scoring of the SAVRY was done by trained and certified researchers and trainees under supervision. The SAVRY was completed by consensus scoring until an inter rater reliability of at least 80 percent was achieved. After reaching an inter rater reliability of at least 80 percent, the certified researchers scored individually. Every SAVRY that was completed by a trainee was checked by a trained researcher.
10
Self-reported recidivism data were collected in a follow-up study. The researcher tried to find current contact information of the former patients to be able to ask them to participate in the follow-up study. In case no contact information could be retrieved, a thorough ‘internet search’ was conducted to find contact information, for instance at social media. In case no address but only a telephone number could be retrieved, a researcher called the former patient and shortly explained the study to him. Afterwards, the researcher asked for the address of the former patient and for his permission to send an information letter. If the former patient immediately said that he did not want to participate, he was not contacted again. In case only the address (and not the telephone number) could be retrieved, the information letter (together with a reply card and envelope) was sent to the most recent known address(es) of the former patient. On the reply card, the former patient could fill in whether he wanted to engage in the study or not, and his telephone number (in case he wanted to participate). A telephone number and e-mail address of the researcher were stated in the letter as well, in order to allow the former patient to contact the researcher via telephone, WhatsApp or e-mail instead. In case both an address and a telephone number could be retrieved, an information letter containing information about the study was send to the last known address of the former patient. A week after sending the letter, the researcher called the former patient in order to inform him about the research project orally and to ask him to engage in the study.
Interviews were conducted at a public location, the participant’s home or a(n) (judicial) institution. Before starting the interview, every participant was informed about the study once again (verbally). After that, he was asked to sign an informed consent. All participants decided to sign an informed consent. In total, completion of the questionnaires and the interview took about 1,5 hours. Participation in the study was voluntary.
11
Ethical approval
The proposal of this study was submitted to the science committee of GGzE (Geestelijke Gezondheidszorg Eindhoven en De Kempen). They concluded that this study was in accordance with the prevailing medical ethics in the Netherlands. They declared that this study is a non-WMO study, which means that no medical health care examination was needed for this study.
Sample
Former patients who were discharged from the Catamaran between April 2009 and October 2013 were approached to ask them to participate in the study. Inclusion criteria were: a) being male (99% of the patients who are admitted to the Catamaran are male); b) admission period of at least three months, c) minimum age of 18 years at the moment of follow-up, and d) discharged from the Catamaran a minimum of one year before the follow-up study. In total, 144 former patients met the inclusion criteria of this study. Nineteen former patients could not be reached despite extensive searches or because they died. Of the 125 former patients who were requested to participate in the study, 79 (63%) were active exclusions, since they refused to participate, for example because of a lack of time. The remaining 46 (37%) former patients were included in the follow-up study.
The sample comprised 46 male former patients of the Catamaran with a mean age of 21.85 years (range 18-27 years) at the time of the follow-up interview. The majority of the participants (95.7%) were convicted for at least one offense before being admitted to the Catamaran, two participants did not commit an offense before admission. The most common offenses were moderate violent offenses (37.0%), property offenses without violence (34.8%) and vandalism (property) (30.4%). The average time between discharge from the Catamaran
12
and the follow-up interview was 2.80 years (range 1-6 years). See Appendix I for more sample characteristics.
In order to measure the potential impact of attrition, an analysis on the differences on study variables between former patients who were included and former patients who were excluded from this study was done. The study variables that were included in the attrition analysis were age at discharge in years, duration of stay at the Catamaran, judicial measure and previous delinquent behavior. No significant differences between both groups have been found on these variables. A significance level of p ≤ .05 was used. See Appendix II for an overview of the attrition analysis.
Statistical Analyses
The analyses were conducted using Statistical Package for the Social Sciences (SPSS), version 19. Descriptive statistics were used to analyze the prevalence of the different Central Eight risk factors in our sample. Thereafter, two separate logistic regression analyses were conducted with general recidivism and violent recidivism as dependent variables. In both analyses, the eight Central Eight factors were independent variables. In the first block of both logistic regression analyses, the confounders (length of stay and length until follow-up) were entered. In the second block of both analyses, all Central Eight risk factors were added and analyzed using forward stepwise logistic regression analyses. Nagelkerke R², odds ratios and 95% confidence intervals were used to determine whether the variables added predictive validity to the models.
13 Results
Descriptive statistics
As can be seen in Table 1, the risk factor leisure/recreation had the highest prevalence in our sample: 80.4% of the participants had problems with leisure/recreation. Furthermore, more than half of the participants (54.3%) had a history of antisocial behavior. The risk factor antisocial associates had the lowest prevalence: only 8.7% of the participants had antisocial associates at the moment of discharge from residential care. For an overview of the prevalence of all Central Eight factors, see Table 2.
General recidivism
Twenty-nine participants (63.0%) recidivated in one or more (general) offenses after discharge from the Catamaran. The confounders were not significantly related to the dependent variables. Forward stepwise logistic regression analyses showed that a model with the factor antisocial personality pattern (Nagelkerke R² = .27, OR = 13.85, 95% CI = 1.56-122.97) is best predictive for general recidivism. An overview of the results can be found in Appendix III.
Violent recidivism
Fifteen participants (32.6%) recidivated in at least one violent offense after discharge from the Catamaran. The confounders were not significantly related to the dependent variables. Forward stepwise logistic regression analyses showed that a model with the factors antisocial personality pattern (model Nagelkerke R² = .45, OR = 15.45, 95% CI = 2.60-91.87) and school/work (OR = 7.31, 95% CI = .93-57.21) is best predictive for violent recidivism. Although the factor school/work was not significantly related to violent recidivism, the factors antisocial personality pattern and school/work together better predicted violent recidivism
14
than antisocial personality pattern alone. An overview of the results can be found in Appendix III.
Table 2
Prevalence of the Central Eight risk factors in the sample (N=46)
Item N % present
Big Four risk factors
History of antisocial behaviour 25 54.3
Antisocial personality pattern 14 30.4
Antisocial cognition 6 13.0
Antisocial associates 4 8.7
Moderate Four risk factors
Family/marital circumstances 10 21.7
School/work 8 17.4
Leisure/recreation 37 80.4
15 Discussion
The aim of this study was to expand existing research investigating the Central Eight risk factors for general and violent recidivism. More specifically, we were interested which of the Central Eight risk factors (Andrews & Bonta, 2010) best predicted self-reported general and violent recidivism in a group of patients who were discharged from a clinic for youth forensic psychiatry and orthopsychiatry. In this study, we focused on the Central Eight risk factors that were present among the patients at the moment of discharge from the clinic. The results show that the risk factor antisocial personality pattern best predicts general recidivism, whereas the risk factors antisocial personality pattern and low interest or commitment to school and/or work best predict violent recidivism.
Thus, antisocial personality pattern appeared to be an important factor in predicting both general and violent recidivism. This is not in line with earlier studies (e.g. Grieger & Hosser, 2013; Olver, Stockdale & Wong), in which antisocial personality pattern was not one of the most predictive factors for recidivism. The substantial differences in time at risk could explain the large influence of antisocial personality pattern in this study. The time at risk in this study was between one and six years after discharge, whereas in other studies usually a shorter, fixed time at risk was applied, for instance two years (e.g. Van der Put et al., 2012). Traits like an antisocial personality pattern could have a larger impact on recidivism a longer time after discharge (Walton et al., 2016) than other risk factors, of which several are states. Traits are more rigid than states, which makes traits harder to change (Eurelings-Bontekoe, Verheul & Snellen, 2017; Walton et al., 2016) and are more likely to have a longer-term effect on recidivism.
Apart from antisocial personality pattern, in line with our expectations, low interest or commitment to school and/or work was important in predicting violent recidivism (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2014). Adolescents who
16
have a low interest in school and/or work generally have less future perspective and less financial resources. Individuals who are unemployed are more likely to experience hopelessness, which is associated with delinquent behavior (Morselli, 2016). Furthermore, a lack of social sources of capital at school has been found to be related to delinquent behavior (Dufur, Hoffmann, Braudt, Parcel & Spence, 2015). However, in our study a low interest or commitment to school and/or work was predictive for violent recidivism, but not for general recidivism. An explanation could be that the SAVRY risk factor school/work, that was applied in this study, does not only include the factor school, but also the factor work. In earlier studies, only the risk factor school was included (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2014). Problems at school could be better predictive for general recidivism than problems at work (Agnew, 2005; Ngo & Paternoster, 2014).
Several risk factors did not predict recidivism in our study, whereas we expected these factors to be predictive. For instance, the risk factor history of antisocial behavior did not predict general recidivism, neither did it predict violent recidivism. This is a notable result, as history of antisocial behavior was one of the best predictors for recidivism in multiple earlier studies (Donker & Bakker, 2012; Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012). According to Moffitt (Moffitt, 1993), there is an ‘adolescence-limited’ group of offenders, who only show antisocial behavior before the age of eighteen. In contrast to several other studies (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012), all participants in our study were 18 years or older at the time of follow-up. Possibly, the adolescence-limited group is highly represented in the sample of this study. Therefore, the people in this group did not reoffend at adult age, even though they have a history of antisocial behavior. This could have led to the relatively small impact of history of antisocial behavior on recidivism in this study.
17
Next to history of antisocial behavior, the risk factor antisocial associates did not appear to be an important predictor for general recidivism, whereas this was an important risk factor in several other studies (Grieger & Hosser, 2013; Olver, Stockdale & Wong, 2012; Van der Put et al., 2014). The very small prevalence of participants with antisocial associates in this study could be an explanation for the lack of effect that was found in this study. This small prevalence could be explained by the high prevalence of participants with a pervasive development disorder in the sample of this study. Adolescents with a pervasive development disorder generally have less contact with peers than adolescents without this disorder (Volkmar, Rogers, Paul & Pelphrey, 2014). Therefore, the behavior of the adolescents is also less influenced by the negative behavior of antisocial associates in the sample of this study.
An important strength of this study is the use of self-report data, which provides added value to other studies on the Central Eight risk factors in which official data from police registrations and judicial documentation were applied. Due to the incompleteness of official data, self-report data provides added value when collecting information about criminal behavior (Thornberry & Krohn, 2000; Weijters & Van der Laan, 2016). Next to advantages, applying self-report data has a disadvantage as well; it can lead to socially desirable answers (Loeber, Slot & Sergeant, 2001). However, there is no reason to assume that socially desirable answers played a crucial role in this study, as the found recidivism rates were slightly higher than the recidivism rates that were found in a study using official data (Wartna et al., 2016). Nonetheless, this study has limitations. The first limitation is the small sample size (N=46), which could have biased the results. A study with a small sample size has a reduced chance of detecting a true effect. Furthermore, a small sample size reduces the accuracy of the results (Button et al., 2013). The second limitation is the use of the SAVRY to operationalize the Central Eight risk factors. Therefore, some of the risk factors were not exactly the same as for instance in the study of Andrews and Bonta (2010).
18
Despite these limitations, the results of this study have clinical implications. The importance of the risk factors antisocial personality pattern and low interest in school and/or work in predicting recidivism can be used in aftercare treatment programs for delinquent adolescents. In these programs, the risk factor antisocial personality pattern (which in our study consisted of risk taking/impulsivity, anger management problems and low empathy/remorse) needs to be addressed, in order to decrease general as well as violent recidivism. Interventions should specifically focus on maladaptive, antisocial personality characteristics. For instance, schema therapy is an intervention that has been found to be effective in treating personality disorders in adults (e.g. Van Genderen, Jacob & Seebauer, 2012) and recently its effectivity has been proven for adolescents as well (Van Wijk-Herbrink, Broers, Roelofs & Bernstein, 2017). Furthermore, the risk factor school/work needs to be addressed additionally. It is important to stimulate adolescents to go to school or work after discharge from residential care (Grieger & Hosser, 2013). Interventions that focus on the adolescent at school, for instance Multisystemic Therapy (MST), can be effective in reducing the presence of this risk factor (Johnides, Borduin, Wagner & Dopp, 2017; Van der Stouwe, Asscher, Stams, Deković & Van der Laan, 2014). The importance of school in aftercare programs was underlined by research on youth and parent perceptions of aftercare support after residential care. A study of Trout and her colleagues (2014) shows that both adolescents and parents indicate supports in education as most important subject in aftercare.
The results of this study show that antisocial personality pattern is an important predictive factor for general recidivism. In addition, an antisocial personality pattern combined with a low interest or commitment to school and/or work best predicts violent recidivism. Therefore, these factors need to be particularly addressed in aftercare programs. This study has added value as it was a first study to measure the impact of the Central Eight risk factors using self-report recidivism data. This study shows different results than earlier
19
studies on risk factors for recidivism, in which official recidivism data were used. Additional research on the impact of the Central Eight risk factors on self-reported recidivism, using larger samples, is needed.
20 References
Agnew, R. (2005). Why do criminals offend? A general theory of crime and delinquency. Los Angeles, CA: Roxbury Publishing Co.
Andrews, D. A., & Bonta, J. (2010). The psychology of criminal conduct (5th ed.). New Providence, NJ: LexisNexis.
Button, K. S., Ioannidis, J. P. A., Mokrysz, C., Nosek, B. A., Flint, J., Robinson, E. S. J., & Munafò, M. R. (2013). Power failure: why small sample size undermines the reliability of neuroscience. Nature Reviews Neuroscience, 14, 365-376.
Coleman, C., & Moynihan, J. (1996). Understanding crime data: haunted by the dark figure. Criminal Behaviour & Mental Health, 8(3), 241-242.
Dienst Justitiële Inrichtingen (2016). Ten uitvoerlegging van een straf of maatregel in een
Justitiële Jeugdinrichting. Retrieved from https://www.dji.nl/binaries/DEF%20
informatieblad%20JJI's%20oktober%202016_tcm41-120764.pdf
Donker, A., & De Bakker, W. (2011). Vrij na een PIJ: Voorspellende factoren van
acceptatie en vrijwillige nazorg en recidive na een PIJ-maatregel. Retrieved from
https://www.wodc.nl/onderzoeksdatabase/succesfactoren-pij-maatregel.aspx Dufur, M. J., Hoffmann, J. P., Braudt, D. B., Parcel, T. L., & Spence, K. R. (2015).
Examining the effects of family and school social capital on delinquent behaviour.
Deviant Behavior, 36(7), 511-526.
Eurelings-Bontekoe, E. H. M., Verheul, R., & Snellen, W. M. (2017). Handboek
persoonlijkheidspathologie. Houten: Bohn Stafleu van Loghum.
GGzE (2017). De Catamaran. Retrieved from https://www.ggze.nl/organisatieonderdelen/de- catamaran
21
Andrews and Bonta’s “Central Eight’ risk factors for recidivism in German correctional facility inmates. Criminal Justice and Behavior, 41(5), 613-634. Johnides, B. D., Borduin, C. M., Wagner, D. V., & Dopp, A. R. (2017). Effects of
multisystemic therapy on caregivers of serious juvenile offenders: A 20-year follow-up to a randomized clinical trial. Journal of Consulting and Clinical
Psychology, 3, 492-499.
Joliffe, D. (2013). Exploring the relationship between the Five-Factor Model of personality, social factors and self-reported delinquency. Personality and Individual Differences,
55, 47-52.
Kempes, M. (2012). Overzicht van onderzoek naar de PIJ-maatregel tussen 2006-2011. ’s-Gravenhage: WODC.
Krug, E. G., Dahlberg, L. L., Mercy, J. A., Zwi, A. B., & Lozano, R. (2002). World report on
violence and health. Geneva: World Health Organization.
Lodewijks, H. P. B., Doreleijers, T. A. H., De Ruiter, C., & Borum, R. (2008). Predictive validity of the Structured Assessment of Violence Risk in Youth (SAVRY) during residential treatment. International Journal of Law and Psychiatry, 31, 263-271. Lodewijks, H. P. B., Doreleijers, T. A. H., De Ruiter, C., & De Wit-Grouls, H. F. (2006).
SAVRY. Handleiding voor de gestructureerde risicotaxatie van gewelddadig gedrag
bij jongeren [SAVRY. Manual for the structured assessment of violence risk in youth].
Zutphen: Rentray.
Loeber, R., Slot, W., & Sergeant, J. A. (2001). Ernstige en gewelddadige jeugddelinquentie:
Omvang, oorzaken en interventies. Houten/Diegem: Bohn Stafleu van Loghum.
Megens, K., & Day, S. (2007). Delinquentie van meisjes: over de relatie tussen risicofactoren en delinquent gedrag na behandeling in een JJI. Tijdschrift voor Criminologie,
22
Moffitt, T. E. (1993). Adolescence-limited and life-course persistent antisocial behaviour: a developmental taxonomy. Psychological Review, 100, 674-701.
Morselli, D. (2016). Contextual determinants of hopelessness: Investigating socioeconomic factors and emotional climates. Social Indicators Research, 133(1), 373-393.
Ngo, F. T., & Paternoster, R. (2014). Contemporaneous and lagged effects of life domains and substance use: A test of Agnew’s general theory of crime and delinquency.
Journal of Criminology, 32, 111-131.
Olver, M. E., Stockdale, K. C., & Wong, S. C. P. (2012). Short and long-term prediction of recidivism using the youth level of service/case management inventory in a sample of serious young offenders. Law and Human Behavior, 36(4), 331-344.
Piquero, A. R., Schubert, C. A., & Brame, R. (2014). Comparing official and self-report records of offending across gender and race/ethnicity in a longitudinal study of serious youthful offenders. Journal of Research in Crime and Delinquency, 51(4), 526-556.
Rijksoverheid (2015). Rijksbegroting Artikel 35 Jeugd. Retrieved from
http://www.rijksbegroting.nl/2015/voorbereiding/begroting,kst199430_10.html Thornberry, T. P., & Krohn, M. D. (2000). The self-report method for measuring delinquency
and crime. Criminal Justice, 4, 33-83.
Trout, A. L., Hoffman, S., Huscroft-D’Angelo, J., Epstein, M. H., Duppong Hurley, K., & Stevens, A. L. (2014). Youth and parent perceptions of aftercare supports at discharge from residential care.Child and Family Social Work, 19, 304-311.
Van Domburgh, L., Geluk, C., Jansen, L., Vermeiren, R., & Dorelijers, T. (2016). Antisocial behavior and victimization over 2-year follow-up in subgroups of childhood arrestees.
Journal of Interpersonal Violence, 1, 1-27.
23
Langewouters, F. E. M. (2014). Risk assessment of girls: Are there any sex differences in risk factors for re-offending and in risk profiles? Crime & Delinquency, 60(7), 1033-1056.
Van der Put, C. E., Stams, G. J. J. M., Hoeve, M., Deković, M., Spanjaard, H. J. M., Van der Laan, P. H., & Barnoski, R. P. (2012). Changes in the relative importance of dynamic risk factors for recidivism during adolescence. International Journal of Offender
Therapy and Comparative Criminology, 56(2), 296-316
Van der Stouwe, T., Asscher, J. J., Stams, G. J. J. M., Deković, M., & Van der Laan, P. H. (2014). The effectiveness of multisystemic therapy (MST): A meta-analysis.
Clinical Psychology Review, 34(6), 468-481.
Van Genderen, H., Jacob, G., & Seebauer, L. (2012). Patronen doorbreken, Negatieve
gevoelens en gewoonten herkennen en veranderen. Amsterdam: Uitgeverij
Nieuwezijds.
Van Wijk-Herbrink, M. F., Broers, N. J., Roelofs, J., & Bernstein, D. P. (2017). Schema therapy in adolescents with disruptive behavior disorders. International Journal of
Forensic Mental Health, 16(3), 261-279.
Volkmar, F. R., Roger, S. J., Paul, R., & Pelphrey, K. A. (2014). The Handbook of Autism and
Pervasive Developmental Disorders. Hoboken, NJ: John Wiley & Sons.
Walton, K. E., Krueger, R. F., Elkins, I., D’Accordo, C., McGue, M., & Iacono, W. G. (2016). Personality traits predict the developmental course of externalizing: A four-wave longitudinal study spanning age 17 to age 29. Journal of Personality, 85(3), 364-375. Wartna, B. S. J., Harbachi, S. L., & Van der Laan, A. M. (2005). Jong vast: een cijfermatig
overzicht van de strafrechtelijke recidive van ex-pupillen van justitiële jeugdinrichtingen. ’s-Gravenhage: WODC.
24
Recivebericht 2015: Landelijke cijfers over terugval van justitiabelen bestraft in de periode 2002-2012. Retrieved from
https://www.wodc.nl/binaries/fs-2016-1-nw_tcm28-74220.pdf
Weijters, G., & Van der Laan, A. M. (2016). De overeenstemming tussen zelfgerapporteerde
jeugdcriminaliteit en bij de politie bekende jeugdige verdachten. Retrieved from
25 Appendix I
Sample Characteristics (N = 46)
M (SD) Range
Age at discharge (in years) 18.6 (1.9) 16-23
Duration of stay at the Catamaran (in months) Age at follow-up (in years)
Duration until follow-up interview (in years)
20.2 (11.8) 21.8 (2.4) 2.8 (1.4) 3-50 18-27 1-6 n % Judicial measure Criminal law 21 45.7 Civil law 21 45.7 Voluntary 4 8.7
Previous delinquent behaviour
No previous delinquent behaviour 2 4.3
Drug offense 2 4.3
Vandalism (property) 14 30.4
Property offense without violence 16 34.8
Moderate violent offense 17 37.0
Violent property offense 6 13.0
Serious violent offense 2 4.3
Sex offense 10 21.7
Manslaughter 3 6.5
Arson 0 0.0
26 DSM-IV-TR Axis I diagnosis
Pervasive development disorder 21 45.7
Attention deficit/hyperactivity disorder 19 41.3
Disruptive behaviour disorder 19 41.3
Substance disorder 10 21.7
Reactive attachment disorder 7 15.2
Other disorder usually first diagnosed in infancy, childhood or adolescence
6 13.0
Schizophrenia or another psychotic disorder 5 10.9
Mood disorder 5 10.9
Impulse regulation disorder, not classified elsewhere 4 8.7 DSM-IV-TR Axis II diagnosis
Personality disorder 5 10.9
27 Appendix II Attrition Analysis Inclusion (N=46) M (SD) Exclusion (N=98) M (SD) p-value
Age at discharge in years 18.6 (1.9) 18.7 (2.2) .69
Duration of stay at the Catamaran in months 20.2 (11.8) 19.6 (13.2) .77 Age at (estimated) follow-up in years 21.8 (2.4) 23.4 (1.5) .17
% % p-value
Judicial measure .61
Criminal law 45.7 54.1
Civil law 45.7 39.8
Voluntary 8.7 6.1
Previous delinquent behaviour .84
No previous delinquent behaviour 4.3 5.1
One or more previous offense(s) 95.7 94.9
28 Appendix III
Logistic Regression Analyses General and Violent Recidivism
Omnibus Chi-square
Nagelkerke R²
Exp(B) Exp(B) 95% C.I.
General recidivism
Antisocial personality pattern 10.25 0.27 13.85 1.56-122.97 Violent recidivism
Antisocial personality pattern 13.50 0.36 15.45 2.60-91.87
