Background: Few population-based national studies include complete measures of alcohol consumption, alcohol-related problems, and sexual orientation. When measures of sexual orientation are included in alcohol surveys, typically only one measure is included. The purpose of this paper is to compare two ways of measuring sexual orientation and to explore the relationship of each measure with alcohol use and alcohol-related problems. Methods: Data were used from the 2000 National Alcohol Survey, a telephone national alcohol probability survey representative of U.S. adults 18 and over (N = 7,612). Sexual orientation was measured by two self-reported measures: sexual orientation identity and sexual behavior. Alcohol use and problems were assessed by number of drinks in the last year, number of days drank 5 or more drinks in the last year, social consequences, and DSM-IV alcohol dependence. Results: There is no consistent pattern of alcohol use or alcohol-related problems based on how sexual orientation is measured for men. For women, bisexual identity or bisexual behavior was consistently associated with greater alcohol use and problems as compared to heterosexual identity or heterosexual behavior. Use of behavior measures alone appears to provide lower estimates of alcohol-related consequences and dependence among lesbian respondents. Conclusions: These results underscore the utility of using both identity and behavior items to measure sexual orientation particularly for women when assessing alcohol use and alcohol-related problems.