This study addressed this issue with data from 72 midwestern male police officers. The officers were asked to respond to vignette scenarios in which husbands had assaulted their wives. Three variables previously related to spouse abuse (sex role egalitarianism, approval of marital violence, and marital stress) and the officer's use of violence in his own marriage were entered into a path model to predict three different police responses (i.e., mediating response, arrest, and antivictim response). The model predicted significant variance in the officer's hostile response to victims of domestic violence but not to his likelihood of arresting abusers or mediating between the abusive couple. Implications for understanding domestic violence and police response are presented. 2 tables, 2 figures, 33 references. (Author abstract)