The police agency had a written pro-arrest departmental policy prior to Florida's pro-arrest legislation that took effect in January 1992. The department had also taken action designed to let all police officers know that an arrest was the preferred outcome when dealing with a domestic call that had become or had the potential for becoming violent. Data were collected in the summer of 1992 by means of a two-page anonymous survey questionnaires to all police officers attending roll call on all shifts. Responses came from 297 of the 412 officers; the 115 officers were either on vacation or did not regularly attend roll call. The analysis focused on five factors and their impact on self-reported likelihood to arrest: (1) officers' demographics and attitudes toward women, (2) the amount of domestic assault training received, (3) understanding of the departmental policy on domestic assault, (4) the degree to which they regarded police involvement as beneficial in responding to calls, and (5) their use of violence in their own homes. Results indicated that the belief in the usefulness of police involvement, knowledge of the agency policy, and attitudes toward women were all significant in accounting for the variation in the likelihood of making an arrest. Tables, note, and 58 references (Author abstract modified)