Childhood traumatic experiences and mental health problems in sexually offending and non-sexually offending juveniles

The sample included 44 male juveniles who sexually offended (i.e., JSOs) and 44 propensity score matched male juvenile non-sexual offenders (i.e., non-JSOs) incarcerated in two juvenile detention centers in the Netherlands between May 2008 and March 2014. Youth were classified as JSO if their official judicial record showed at least one conviction for a sexual offense (n = 17), if at least one index offense was a sexual offense (n = 26), or if they reported during the assessment that they ever engaged in sexual behavior against someone else’s will (n = 6). Non-JSOs were suspected or convicted of violent (e.g., manslaughter, armed robbery) and/or non-violent (e.g., theft, drug dealing) crimes, but did not have a history of sexual offense perpetration. JSOs and non-JSOs were propensity score matched on age and ethnicity. The age range of the total sample was between 13 and 24 years (33% of the offenders were 18 years or older, and 18% were 19 years or older). The mean age of JSOs and non-JSOs was similar [JSOs: 17.0 (SD = 2.0), non-JSOs: 17.7 (SD = 1.8); t = 1.8; p = .97], as was the proportion of participants who were native Dutch (JSOs: 40.9%, non-JSOs: 22.7%; χ² = 3.4, p = .11).

Assessment was part of a standardized self-report mental health screening procedure in the juvenile detention centers used for clinical purposes. Master students and test assistants with a Master’s degree trained by clinically experienced researchers performed the comprehensive assessments. Juveniles and their parents were informed that all information was also used for scientific research after encryption. The relevant institutional review and scientific boards of the juvenile detention centers approved the study and procedure (for more details, see: [16]).

Data were analyzed using International Business Machines Corporation Statistical Package for Social Sciences, version 19 (IBM SPSS 19). The level of significance was set at .01 in order to account for Type I error inflation due to multiple testing. First, differences in childhood trauma scores and mental health scores between JSOs and non-JSOs were examined using t-tests. Second, as our data were not normally distributed, we used Spearman Rho Correlations to examine the relation between childhood abuse and mental health problems in JSOs and general offending juveniles. Third, we compared the strength of the relationship between childhood abuse and mental health problems in JSOs and non-JSOs by calculating the difference between the two independent correlation coefficients using software available from http://​quantpsy.​org [33]. Although the Fischer r-to-z transformation is a method usually applied to Pearson correlation coefficients, Myers and Sirois [34] showed that this approach performed best in terms of control of Type I error when compared to other strategies. To interpret the magnitude of the correlation coefficients, we followed Cohen’s [35] benchmark of small (r = .10), medium (r = .30) and large (r = .50).

Findings of this study must be interpreted in the context of some limitations. First, previous research showed that JSOs constitute a heterogeneous group with differences in childhood abuse and mental health problems [43‐45]. Especially JSOs with child victims, when compared to JSOs with adolescent/adult victims, show more childhood abuse, especially sexual abuse, and more mental health problems. We did not examine subgroups given that our sample of JSOs constituted only 44 offenders. Second, we did not assess the extent, frequency and duration of childhood abuse, which also might have influenced our results. Third, the juvenile detention centers in the current study only admitted males. Therefore, our results cannot be generalized to female offender populations. The fourth limitation refers to the reliability of the results. The CTQ-SF and the MAYSI-2 are both self-report instruments. Therefore, our results may have been biased due to social desirability (e.g., on one hand it is conceivable that a history of maltreatment is kept a secret because of shame or loyalty to the perpetrator, but on the other hand it can be suggested that a history of maltreatment is over-reported to gain justification and/or compassion for one’s behavior). Furthermore, as youths were told that their answers would be used for clinical purposes and for evaluation of their interventions, the (lack of) confidentiality could have affected our results. Moreover, retrospective recall bias [46] also may have played a role in the over- or under-reporting of perceived maltreatment; it has been suggested that more recent maltreatment is more accurately recalled than more distal maltreatment. In addition, amplification of the negativity of the maltreatment (e.g., recall of own abuse history when charged with perpetration of a sexual offense) also could lead to over-reporting of maltreatment [47]. Fifth, the internal consistency of the MAYSI-2 scales Thought Disturbance and Traumatic Experiences have been found to be poor to acceptable [22, 23, 25, 27]. Although lower consistency may be explained by the broadness of the constructs measured, this should be taken into account when interpreting the results. Sixth, by lowering the level of statistical significance to .01 we reduced the probability of making a Type I error. As a result, however, the probability of making Type II errors increased (and power reduced), which should also be taken into account. Seventh, the current study was cross-sectional and, therefore, causal relationships between childhood abuse and mental health problems could not be established. Longitudinal studies are needed to establish this relationship. Finally, beyond mental health problems, other variables could have influenced the relationship between childhood (sexual) abuse sexual offending behavior, such as genetic predisposition, various family factors, and peer influences.

Our results suggest that if a youth with a history of perpetrating a sexual offence reports mental health symptoms, especially internalizing mental health problems such as suicidal ideation and thought disturbance or externalizing mental health problems such as angry-irritable problems, there is stronger reason to suspect these symptoms are related to childhood abuse or neglect, especially sexual abuse, than if a youth without a history of sexual offending reports similar symptoms. As internalizing mental health problems are harder to detect than externalizing mental health problems, it is of great importance to assess both internalizing and externalizing mental health problems in JSOs at entry to juvenile detention centers. Furthermore, as we found a stronger relationship between childhood sexual abuse and both internalizing and externalizing mental health problems in JSOs than in non-JSOs, our results suggest the need for a different focus for treatment of JSOs and non-JSOs. For JSOs, perhaps the treatment needs to focus on dealing with the childhood sexual abuse (e.g., trauma-based therapy) if this is determined to be a key risk factor for future offending for that particular youth. Finally, as there is evidence that the relationship between sexual abuse and sexual offending behavior could be caused indirectly through mental health problems, one aspect of sexual violence risk management among juveniles who have experienced sexual abuse could comprise treatment with a focus on healthy development and behaviors in order to prevent sexual offending behavior.