Prior research has speculated about but has not provided systematic empirical data on how police officers use their prior knowledge to interpret wife assault situations and how these interpretations shape their responses. In the current study, 128 police officers in North Georgia were classified into several groups based on their answers to questions about experience and training (untrained novice, trained novice, and trained experienced). Respondents were randomly assigned to read one of eight scripts to determine how they made decisions about domestic violence situations. By manipulating whether a wife exhibited abnormal behavior, the study found that experienced police officers do not focus on whether wives can control their "provoking" actions and are to blame. Instead, police officers consider the husband's relative credibility and dangerousness. Prior experience with handling wife assault cases thus shifts the focus of decisionmaking from normative considerations such as blameworthiness to efficiency considerations such as substantiating claims for successful prosecution. Both novice and experienced police officers, however, base their arrest decisions on prior beliefs about whether wives provoke their husbands when wives have alcohol problems. Further research is recommended to examine how prior knowledge shapes interpretations to gain a better understanding of police decisionmaking. Pitfalls of preferred and mandatory arrest policies are noted, and a study script based on a normal, low-income couple is appended. 86 references, 26 footnotes, and 7 tables