Can psychopathy be identified as accurately during adolescence as adulthood? To address this developmental question, this study compared the stability of scores on the leading measure of psychopathy, the Psychopathy Checklist (PCL), among 202 adolescent (M = 15.8 years, SD = 0.89) and 134 adult (M = 27.5, SD = 1.08) offenders. Over a 2-year period, adolescents’ total scores on the PCL (r = .33) were less stable than those of adults (r = .71). Adolescents’ baseline PCL scores also weakly predicted psychopathy classifications 2 years later (AUC = .62), particularly compared with those of adults (AUC = .85). Finally, increases in psychosocial maturity over time predicted decreases in PCL scores for adolescents, but not adults. These results raise questions about reliance upon psychopathy measures to inform decisions about youthful offenders that will have long-term consequences.