Retrospective self-reports of behavior are widely used in alcohol and drug research. However, assessments of the reliability and validity of such data among certain populations are nonexistent. This study examines the ability of the Addiction Severity Index (ASI), a widely used clinical and research instrument, to provide valid and reliable data within a homeless population of drug misusers. The results support the usefulness of the ASI in producing quality data among homeless substance misusers seeking treatment. Qualitative data gathered from field interviewers are used to highlight strategies for enhancing the quality of ASI data in the future.