Subjects were 111 police officers from 3 city and 7 small- town police departments in Wisconsin. Approximately three-fourths of the officers had the rank of patrol officer, and nearly all of them had recent patrol experience. Six of the officers were women; 97 percent were Caucasian; 79 percent were married, 16 percent were separated, and 4 percent were divorced. The dependent measure consisted of responses to two vignettes of domestic violence. Ten response options followed each vignette. The primary independent variables, which were measured with a number of instruments, were attitudes toward women, negative evaluations of battered women, a belief that marital violence is justified, beliefs about why domestic-assault victims stay in the relationship, comfort with victims, and correction for response bias. Several other variables were used for post hoc analyses, including demographic and background variables and orientation to policing. Twenty-eight percent of the officers reported that they might arrest the victim. In response to one of the vignettes, 15 percent said there was a good chance (50 percent or greater likelihood) that they would arrest the victim. Officers with an inclination to arrest victims believed that domestic violence is justified in some situations and that women stay in violent relationships for psychological reasons. These officers also reported less comfort in talking with victims. General sex-role beliefs were unrelated to their responses. Implications of these findings for officer training in responses to domestic violence are discussed. 3 tables, suggestions for further research, and 31 references