Practical Implications of Current Domestic Violence Research: For Law Enforcement, Prosecutors and Judges
Published June 2009
Chapter 3. Offender Characteristics
Section 18 — Do the widely used risk instruments accurately predict reabuse?
All of the common risk instruments in use are insufficient. The best instruments have been found to falsely predict 40 to 43 percent of abusers in both directions. [24, 190] For example, a study of a risk instrument used by police in Berkeley, Calif., found that those abusers classified at highest risk for reoffending did have the highest rate of reoffending but also that the instrument generated 43 percent false positives for predicting reabuse. Those abusers gauged as having the lowest risk of reoffending had 2 percent false negatives. 
Implications for Law Enforcement and Prosecutors
Given high base rates of reabusing, the default presumption should be that the defendant is likely to reoffend until proven otherwise. Risk instruments do not significantly improve upon victim perception and basic actuarial data. (Research basis: Not only is there wide agreement among multiple studies but it is also agreed that the same instrument may have different results in different jurisdictions.)