WED, Oct 6 \ 17.00-19.00 \ Risk Assessment, II

010
EVALUATING THE PREDICTIVE ACCURACY OF SEX OFFENDER RISK ASSESSMENT MEASURES ON UK SAMPLES: A CROSS-VALIDATION OF THE RISK MATRIX 2000 SCALES

Leam A Craig
Principal Forensic Psychologist, Forensic Psychology Practice, Birmingham, UK & University of Birmingham, UK
Anthony Beech
Senior Lecturer in Forensic Psychology, University of Birmingham, UK
Kevin D. Browne
Director of Postgraduate Course in Forensic and Family Psychology, University of Birmingham, UK, & The World Health Organisation, Denmark

The predictive accuracy of the newly developed risk measures Risk Matrix 2000 Sexual/Violence (RMS, RMV) were cross validated and compared with four risk scales (RRASOR, SACJ-Min, SVR-20, and Static-99) in a sample of sexual (n = 85), violent (n = 46), and general (n = 22) offenders. The sexual offence reconviction rate for the sex offender group was 18% at 10 years follow-up, compared with 2% for the violent offenders. Survival analyses revealed the violent offenders were reconvicted at twice the rate of any other group. Reconviction data were analyzed using the area under the curve (AUC) of the Receiver Operating Characteristic (ROC). The RMV significantly predicted violent recidivism in the sex and combined sex/violent offender groups. Although the RMS obtained marginal accuracy in predicting sexual reconviction, none of the scales significantly predicted sexual reconviction. An item analysis revealed four factors not included in the risk scales significantly correlated with sexual and violent reconviction. Combining these factors with Static-99, RMV and RMS increased the accuracy in predicting sexual reconviction.

011
COMMUNITY LIVING SKILLS IN OFFENDER GROUPS (COSMSKILLS) – A EUROPEAN NETWORK STUDY

Thomas Ross,
Dipl.- Psych., Forensic Psychotherapy, University of Ulm, Ulm, Germany
ross@sip.medizin.uni-ulm.de

A large scale international project involving parallel, complementary interventions in three EU countries and an associate country (Germany, The Netherlands, United Kingdom, and Norway) is presented. The aim is to develop a unified approach to assessment of social risk and related behaviours in (sexual) offender groups using the Behavioural Status Index (BEST-Index) and a set of cross-validating instruments (HCR-20; PCL-R, SCL-90-R; BDHI-D). Important elements of the project are improvement of carer assessments; development of offender profiles; related care planning and delivery; studies of the change process; and theoretical implications of what works and what doesn´t to prevent further violence.
The focus is on the BEST-Index, a wide-spectrum classification instrument aimed at reducing the probability of social risk when psychiatric patients are returned from hospital to community care and improving all-round social functioning and integration in such groups. The theoretical model underlying the BEST-Index posits that the social risk presented by an offender varies inversely with his/her degree of personal insight and capacity to perform well in key areas (communication and social skills; empathy, self and family care; work and recreation). These repertoires provide critical focal areas for the treatment planning and delivery in forensic psychiatric care. The study runs in cooperation with clinics in Germany, the Netherlands, Norway, and the UK, involving cohort samples of mentally disordered sexual offenders and violent offenders (N=210). Three measures are taken over the course of three years, using the complete set of research instruments applied.
Results are reported and implications discussed.

012
RISK ASSESSMENT SHOW IMPROVED RELIABILITY AND VALIDITY BUT SHOULD WE QUESTION THE ETHICS OF BASING DECISIONS ON THEM?

Aaron W. B. Konopasky,
Tulane University, New Orleans, LA, USA
Robert Konopasky, Saint Mary's University, Halifax, NS, Canada
rkonopasky@hfx.andara.com

Sexual assault causes incalculable injury and harm. As prevention is a by-product of incarceration, it makes sense that length of sentence and percentage of imposed sentence that is actually served be determined in part by the risk of a sexual offender recidivating.
Psychologists know that past behaviour, sexual offender attributes and even circumstantial variables can be quantified and used to predict recidivism. These predictions are relied on to set sentences, even to incarcerate indefinitely, to make decisions about release to the community, and to set conditions of parole.
As the decisions are grave, it is reasonable to question and consider the models underlying the predictions, the steps needed to standardize the administration of the tests, the samples on which the tests are based, the ways in which the results are reported, and the ways in which the results are used by decision makers. These routine considerations are put into the context of various models of ethics to sharpen the practitioner's sensitivity to the balance of science and the needs of the various constituencies in risk assessments. Making ethical issues explicit will not change the reliability and validity or risk assessments but it suggests guidelines for application.

013
THERAPISTS ASSESS DENIAL AND IMPROVEMENT THROUGH IMPLICIT LINGUISTIC ANALYSIS OF SEMANTIC ROLES AND MODALITY: EXPLICIT LINGUISTIC ANALYSIS CLARIFIES TREATMENT MILESTONES FOR THERAPISTS AND CLIENTS

Abigail Konopasky,
University of New Orleans, New Orleans, LA. USA
Aaron W. B. Konopasky,
Tulane University, New Orleans, LA. USA
Robert Konopasky,
Saint Mary's University, Halifax, NS, Canada
rkonopasky@hfx.andara.com

Believing that acceptance of responsibility for offensive behavior is generally positive, sexual offender therapists routinely assess type and level of denial during treatment. Less formal and less thorough than risk assessments that are done both before and, sometimes, after treatment, this intra-therapy judgment of denial often relies on an unarticulated analysis of the content of the patient's verbal acts, e.g., “She came on to me” means that he thinks it wasn't his fault. We argue that this assessment is actually dependent upon the linguistic resources of both the clinician and clients and how they use those resources to construct physical, emotional, and verbal processes. A formal linguistic framework to explicitly analyze these resources in order to reveal clear links between real-world states, client verbalizations, and therapist verbalizations is suggested. A formal linguistic assessment identifies 4 key participant types: a.) the client (and his/her actions), b.) the therapist, c.) the victim(s) (real and potential), and d.) psychopathology or other mental states. A quantitative analysis of these participant types reveals (a) their relative semantic status as Agent of action versus Patient of action, (b) their relative involvement in physical versus emotional versus verbal actions and (c) their relative involvement in indicative versus subjunctive constructions. A qualitative analysis of word co-occurrence patterns reveals each speaker's idealized linguistic model for each participant type. Together, these analyses can illuminate underlying assumptions of denial and treatment progress and offer more explicit ways of measuring change.

014
EFFECTIVE COMMUNICATION OF RISK ESTIMATES WHEN USING A COVERGENT APPROACH TO RISK ASSESSMENT

Douglas P. Boer, Ph.D.
Senior Psychologist, Correctional Service of Canada, Canada
boerinc@telus.net

As discussed in my first paper, there are academic proponents of actuarial approaches and there are opposing academic proponents of structured clinical methods for the risk assessment of sexual offenders. To date, neither type of approach has convincingly proven superior to the other in terms of assessing the likelihood of future sexual violence. However, these methods do have their unique strengths and weaknesses, and fortunately, the strengths and weaknesses of each type are largely opposite in nature. As a result, a convergent approach in which both types of instruments are used in a combined manner may serve both the offender and public interest best. A series of common-sense rules to help communicate risk estimates will be provided that may aid clinicians when faced with explaining findings from several instruments of different methodological natures.