Evaluating Risk Assessment Instruments for Intimate Partner Stalking and Intimate Partner Violence

Abstract

Research indicates that most stalking victims are harassed by a prior intimate partner and that this group of victims is at a higher risk for violence than other stalking victims. Furthermore, many researchers found a relationship between prior intimate partner stalking (IPS) and intimate partner violence (IPV). The present study focused on identification of overlapping and unique risk factors among IPS and IPV offenders. This was done by examining the accuracy and utility of abridged versions of the Brief Spousal Assault Form for the Evaluation of Risk (B-SAFER) and the Guidelines for Stalking Assessment and Management (SAM) in a sample of 158 low-risk offenders charged with crimes targeting an intimate partner. Follow-up ranged from one to 8.5 years; outcome variables were renewed (or continued) stalking and violent reoffending. Receiver operating characteristic (ROC) curves indicated that both instruments had weak, nonsignificant predictive validity for both violent or stalking reoffending when considered in isolation. However, each instrument added significantly to the logistic regression model when predicting stalking recidivism when entered after the PCL:SV and relevant covariates, but neither improved predictive accuracy when added after the other was already in the model. Neither instrument added to logistic regression models predicting violent reoffending. Analysis of individual items revealed that items measuring distress and violations of supervision (SAM) and violations of court orders (B-SAFER) were the most consistent predictors of renewed stalking but not violence. The findings of this study question the utility of these tools for predicting violent reoffending but provide qualified support for the prediction of stalking.