The scientific literature on alcohol and sexual risk behavior is marked by multiple theoretical perspectives and inconsistent findings from global-frequency and event-level studies. Multilevel measures of alcohol use and multiple sexual risk outcomes can be used to evaluate these perspectives and resolve these inconsistencies. Among women recently involved in the criminal justice system in Portland, Oregon, daily alcohol use and sexual behavior were measured during four 30-day intervals over one year. In mixed effects models, person-level, month-level, and day-level alcohol use were significantly associated with the occurrence of intercourse but not with the use of condoms during intercourse. Findings are also reported for main, casual, and exchange partners. The relationships between alcohol use and sexual risk behavior are complex: No single theoretical perspective is sufficient to account for the study findings, and increased risk may be mediated through changes in intercourse rather than through changes in condom use.