Reference: Dunford, F. W. (1992). The measurement of recidivism in cases of spouse assault. Journal of Criminal Law and Criminology, 83, 120-136.

Strategy: Reactive arrest for domestic violence-specific deterrence

Matrix Dimensions:
X-axis: Individuals
Y-axis: Focused
Z-axis: Mostly Reactive

Results: Nonsignificant
Arrest increased offense frequency at 12 months

Methodological Rigor: Highly rigorous- randomized experiment

Abstract: (from NCJRS) This article presents findings associated with six additional months of outcome (for a total of 12 months) used to assess recidivism in the Omaha replication of the Minneapolis experiment in the effectiveness of various police responses to domestic violence; the full Omaha data set is used to examine the amount of time required to assess recidivism accurately in cases of spouse assault reported to the police. Of the six replications of the Minneapolis study, four were funded on the basis of 6-month followup period. Atlanta used an 8-month followup, and Omaha used a 12-month followup. This analysis found that repeat offending in the second 6 months was more likely to involve perpetrators who did not recidivate in the first 6 months than it was to involve the same perpetrators recidivating in both time periods. Even though more abusers recidivated in the first 6 months than in the second 6 months, the prevalence of offending in the second period was from 70 to 80 percent of that of the first period. The frequency of repeat offending among offenders was relatively stable across the two time periods. When the proportion of those at risk of reoffending for the first time was calculated for both halves of a 1- year followup period, the recidivism rate as measured by arrests was almost identical for the two halves. The study concludes that limiting to 6 months the period for measuring spouse-abuse recidivism has resulted in underestimates of the extent to which such abuse continues and has thus fostered spurious or misleading policy recommendations.