The study defined domestic violence as "the actual or threatened physical or sexual violence directed toward a spouse, ex-spouse, current or former boyfriend or girlfriend, or current or former dating partner." The dependent variable was the probability of arrest in a domestic assault case. Independent variables encompassed situational variables (victim's gender, whether the crime was interracial, the extent of victim injury, whether a weapon was involved, whether the suspect was intoxicated, and the location of the crime). Other characteristics included the race, gender, and age of the offender and the age of the victim. The key independent organizational variable was whether the police had a mandatory arrest policy for domestic violence cases. The study involved 57,000 domestic violence cases across 115 police departments. The findings show that mandatory arrest policies were linked to an increased arrest risk in domestic violence cases. Still, the average arrest risk in jurisdictions with a mandatory arrest policy was only 50 percent, suggesting considerable police discretion regarding arrest in domestic violence cases. Race was a less significant factor in arrests if a jurisdiction had a mandatory arrest policy in domestic violence cases. These findings thus suggest that police agencies' formal mandatory arrest policies will reduce the likelihood of racial bias in police arrest decisions. 3 tables, 11 notes, and 118 references