Behavioral Monitoring of Sexual Offenders Against Children in Virtual Risk Situations: A Feasibility Study

Abstract

The decision about unsupervised privileges for sexual offenders against children (SOC) is one of the most difficult decisions for practitioners in forensic high-security hospitals. Facing the possible consequences of the decision for the society, a valid and reliable risk management of SOCs is essential. Some risk management approaches provide frameworks for the construction of relevant future risk situations. Due to ethical reasons, it is not possible to evaluate the validity of constructed risk situations in reality. The aim of the study was to test if behavioral monitoring of SOCs in high-immersive virtual risk situations provides additional information for risk management. Six SOCs and seven non-offender controls (NOC) walked through three virtual risk situations, confronting the participant with a virtual child character. The participant had to choose between predefined answers representing approach or avoidance behavior. Frequency of chosen answers were analyzed in regards to knowledge of the participants about coping skills and coping skills focused during therapy. SOCs and NOCs behavior differed only in one risk scenario. Furthermore, SOCs showed in 89% of all cases a behavior not corresponding to their own belief about adequate behavior in comparable risk situations. In 62% of all cases, SOCs behaved not corresponding to coping skills they stated that therapists focused on during therapy. In 50% of all cases, SOCs behaved in correspondence to coping skills therapists stated that they focused on during therapy. Therapists predicted the behavior of SOCs in virtual risk situations incorrect in 25% of all cases. Thus, virtual risk scenarios provide the possibility for practitioners to monitor the behavior of SOCs and to test their decisions on unsupervised privileges without endangering the community. This may provide additional information for therapy progress. Further studies are necessary to evaluate the predictive and ecological validity of behavioral monitoring in virtual risk situations for real life situations.

Keywords: child abuser; pedophilia; pedophilic disorder; risk assessment; risk management; virtual environment; virtual reality.

Figure 1

The virtual supermarket. (I) Bird's eye view of the virtual supermarket. The supermarket comprised an entrance area (A) and a product area (B). (II) Entrance area of the supermarket with the info screen. (III) Product area of the supermarket with the shopping list Head-up display. Red crosses symbolize products not yet in the basket; green check marks symbolize products that are already in the basket.

Figure 2

Experiment procedure. The experiment started with the initial rating. Goal of the initial rating was to identify the individual most unattractive virtual adult and most attractive virtual child character. In the main experiment phase, the participants learned the controlling of the virtual supermarket and the task in the tutorial and training. Afterwards, the participant was exposed to the most attractive virtual child character consecutively in three risk scenarios. The risk scenarios differed from each other with regard to their difficulty to avoid the contact to the virtual child character.