The authors investigated the reliability of self-reported age of onset (AO) for alcohol, tobacco (cigarette), and illicit drug involvement. Participants were 410 young adults taking part in an 11-year longitudinal study. A moderate degree of reliability was found for the 3 substances. Despite this level of stability, results illustrate a tendency for reported AOs to increase over time. The trend is more salient for participants who reported younger AOs at the initial assessment. Findings also indicate that, for alcohol and tobacco, more individuals were classified as early onset based on Year 1 compared with Year 11 reports. Despite these systematic changes, at least for alcohol and illicit drugs, age at which onset was assessed did not moderate the association between AO and substance-related outcomes.