Purpose – The purpose of this paper is to examine domestic violence calls for service data in one Florida county for a two-year period from July 2004 through July 2006. Design/methodology/approach – Data for this study include information gathered on domestic violence calls for service during a two-year period (n=3,200). This secondary data were analyzed by logistic regression to determine statistically significant predictor variables. Findings – This study found that severity of crime, presence of children, presence of an injunction, and victim injury increased the likelihood of an arrest. Victim race, location of call, victim alcohol use, and length of relationship did not affect likelihood of arrest. Research limitations/implications – Use of secondary data precluded examination of additional relevant variable information. Practical implications – The research shows clear law violations and seriousness of the acts correlate to an increased likelihood of an arrest. Arrest research should inform police training and policy. Social implications – The research is consistent with other research that shows that law enforcement officers continue to play a significant role in responding to domestic violence crime and suggest that researchers should continue to study arrest practices. In this study, arrest was more likely when factors existed that may have indicated a more serious crime. Originality/value – There is an ongoing need to examine agency-level response to domestic violence. This paper adds to the literature on the law enforcement response to domestic violence. The paper suggests areas for future research.