Public service utilization data are often used as key outcomes in studies on homelessness. Although self-report data on these outcomes are accessible and cost-effective, various factors may affect retrospective recall in homeless populations. It is therefore necessary to establish validity of self-report to ensure the integrity of studies involving such populations. Participants (N=134) were chronically homeless individuals with severe alcohol problems who participated in a housing first effectiveness trial (Larimer et al., 2009). The authors compared 30-day and 3-year retrospective self-report data on sobering center, jail, and hospital use with archival records corresponding to the same timeframes. Analyses indicated good category-specific agreement for 30-day self-report and archival data on sobering center (82%; [kappa]=0.58) and jail use (89%; [kappa]=0.60). Hospital use, however, was self-reported significantly more frequently than indicated by archival data (78%; [kappa]=0.30). Three-year data showed inadequate agreement across all three variables. (Contains 2 figures, 3 tables, and 1 footnote.)