This study examined the operating characteristics of the Alcohol Use Disorders Identification Test (AUDIT) as a screen for “at-risk” drinking in a multi-ethnic sample of primary care patients, from a family practice center located in the southwestern United States. A probability sample of 1333 family medicine patients, stratified by gender and racial/ethnic background (white, African-American and Mexican-American) completed the AUDIT, followed by the Alcohol Use Disorders and Associated Disabilities Interview Schedule (AUDADIS) to determine ICD-10 diagnoses. Indicators of hazardous alcohol use and alcohol-related problems were included as measures of “at-risk” drinking. Despite differences in the spectrum of alcohol problems across patient subgroups, there was no evidence of gender or racial/ethnic bias in the AUDIT as indicated by Receiver Operating Characteristic Curve analysis. Excluding abstainers from the analysis had little impact on screening efficacy. In this patient population, the A UDIT appears to be an unbiased measure of “at-risk” drinking.