A review of the relevant literature on violence against women and police responses found no clear evidence regarding the leniency hypothesis. The data used to test the hypothesis were obtained from a study of police behavior in Dade County, Fla., during 1985 and 1986. Trained civilian observers rode 877 8-hour shifts with officers of the Metro-Dade Police Department who had been randomly selected to participate in the study; observers recorded selected aspects of police-citizen encounters in observation schedules and summary narrative reports. At the time of this study, neither Florida law nor police departmental policy mandated arrest in cases of suspected spousal violence. Even though the current data were collected in the mid-1980's, they afford the opportunity to examine the leniency hypothesis under the statutory and policy conditions that characterized most U.S. jurisdictions at the time. Over the study period, 245 disputes that involved police intervention were observed. Eighty disputes involved evidence of physical violence. Twenty-eight of the violent disputes involved a male and a female who were either dating, cohabiting, married, divorced, estranged, or former lovers. The woman was the assailant in three of these cases. A comparison of arrest rates for spousal violence and other violence provided an initial empirical assessment of the spousal violence leniency hypothesis. Although the arrest rate was slightly lower in the spousal assault cases compared to the other cases, the difference was not statistically significant. Possible explanations for these findings are discussed. 3 tables, 11 notes, and 38 references