Simulation in the assessment and treatment of violent offenders

Abstract

With the aim of ultimately reducing the burden of violent crime on society, research has identified a number of areas in which conventional methods of assessment and treatment of violent offenders may be improved. The areas examined in this thesis are gender, mental illness, and intimate partner violence, and they have been integrated with research in the field of case and computer simulation for assessment, training, and teaching. Firstly, paper-andpencil case vignettes were used to investigate possible gender bias in forensic psychiatric assessment (study I). Secondly, interactive computer simulation was applied in two contexts; augmentation of the assessment of mentally disordered violent offenders within forensic psychiatry (study II) and the assessment and treatment of intimate partner violence (IPV) offenders within the Prison and Probation Service (studies III and IV).

In study I, the case vignette method did not reveal a significant association between the gender of the simulated offender and the judgement, by forensic psychiatric assessors, that the offender suffered from a severe mental disorder and was in need of compulsory, inpatient forensic psychiatric care. However, a simulated offender depicted as having mental retardation was more likely to be assessed as being at high risk of criminal recidivism if portrayed as female, regardless of the sex, place of work, or level of experience of the assessor.

In the studies of the interactive computer simulation program Reactions on Display (RoD), study II revealed significant differences between mentally disordered offenders (MDOs) and controls in the area of interpretation of the simulated characters’ emotions and physical reactions, and the MDOs chose more violent actions on behalf of the main character compared with controls. There were also incongruent or deviant responses made by individual MDOs. This indicates that the Reactions on Display/Forensic Psychiatry (RoD/FP) has the ability to differentiate between MDOs and controls in several ways.

In study III, Reactions on Display/Intimate Partner Violence (RoD/IPV) was well received and understood, both by intimate partner violence (IPV) offenders and healthy, non-offender controls. Significant differences between the groups emerged in the area of interpretation of the simulated characters’ emotions. Offenders who had not previously undergone treatment with the Integrated Domestic Abuse Programme (IDAP) showed a trend towards making more violent choices than other IPV offenders and controls.

In the healthy, non-offender controls who took part in study IV, the violent film sequences in the RoD/IPV elicited a subjective feeling of increased arousal and decreased valence on the Self-Assessment Manikin (SAM) scale. It did not, however, elicit a physiologically detectable stress response, i.e. a negative mood. This may allay fears that the film content is disturbing to a degree that it interferes with assessment/learning or pleasurable to a degree that it will be sought after for entertainment purposes.

While largely showing a consensus between Swedish forensic psychiatric assessors, the results provided by study I call for further research into the conformity of Swedish and international forensic psychiatric assessment practices. They also add to the body of research advocating the development of improved methods for risk assessment of offenders, which is part of the rationale behind the computer simulation systems presented in studies II-IV.

The interactive computer simulation tools, studied in a pilot fashion in this thesis work (studies II-IV), have been found to be acceptable to all study groups. Significant differences between the offender and control groups have emerged, particularly with regard to their interpretations of the simulated characters’ emotions and also differences in the level of violence chosen. This indicates that the offenders’ interactions with the RoD program may provide additional information to traditional violence risk assessment methods (e.g. checklists) by showcasing dynamic risk factors (such as difficulties in emotional processing). The above properties may also be seen to support the future use of such programs for offender treatment, in which they can provide an engaging, safe, and controlled environment in which the user can observe and practice interactions and receive feedback from which he/she can learn desirable non-violent behaviours, and, in the specific case of the RoD/IPV, with minimal risk of negative side effects such as being excessively disturbed by its contents or using it as entertainment.