Researchers have used both self-reports and official records to measure the prevalence and frequency of crime and delinquency. Few studies have compared longitudinally the validity of these two measures across gender and race/ethnicity in order to assess concordance. Using data from the Pathways to Desistance, a longitudinal study of 1,354 serious youthful offenders, we compare official records of arrest and self-reports of arrest over seven years. Findings show moderate agreement between self-reports and official arrests, which is fairly stable over time and quite similar across both gender and race/ethnicity. We do not find any race differences in the prevalence of official arrests, but do observe a gender difference in official arrests that is not accounted for by self-reported arrests. Further work on issues on the validity and reliability of different forms of offending data across demographic groups is needed. Abstract published by arrangement with Sage.