The On-Scene Warrantless Arrest Experiment involved 198 suspects arrested at the scene of a domestic violence incident, while subjects in the Victim-Initiated Complaint Experiment included 480 suspects identified by victims' affidavits. Prosecutors tracked cases to one of three outcomes¦-pretrial diversion to rehabilitative counseling, adjudicated guilt with counseling as a condition of probation, or other sentencing, i.e., fines, probation, or incarceration. Following court settlement, each case was followed for 6 months to determine if any of the alternative criminal justice policies had reduced the prevalence, severity, or frequency of violence in each relationship, or the length of time before the next episode of violence erupted. Results showed that neither policy alternative to traditional sentencing was more effective in protecting victims. However, when victim- initiated defendants were arrested under a warrant, and their victims dropped charges, women were significantly more likely to be safe from continuing violence. 28 tables, 7 figures, and 54 references