Ratings by different informants such as parents, teachers, and youths are important to accurately assess adolescent problem behavior. Agreement among informant ratings on adolescent problem behavior, however, is typically only low to moderate. Rather than dismiss these low levels of agreement between informants' ratings as being unreliable, low agreement is considered to reflect each informant's unique view on problem behavior. The overall aim of this study is to examine how much repeated parent, teacher, and self-ratings of internalizing and externalizing problems in adolescents add to single informant ratings in the prediction of internalizing and externalizing DSM disorders in adulthood. Parent, teacher and self-ratings were obtained in 588 adolescents (49.1% boys) aged 11-14 years (at baseline) at 3 time points, spanning 4 years. Twenty years after the first assessment, DSM diagnoses were obtained through structured psychiatric interviews when individuals were 31-34 years. We used structural equation models to investigate whether discrepancies and changes in discrepancies over time between parent, teacher, and self-reports of problem behavior contributed to the prediction of DSM diagnoses. We found that higher levels of internalizing problems in adolescence correlated, r =.14, SE =.06, p =.035 with more internalizing disorders in adulthood and that higher levels of externalizing problems in adolescence correlated, r =.23, SE =.07, p =.001 with more externalizing disorders in adulthood. Increasing discrepancies across 2 time periods between teacher and self-reports of internalizing problems correlated r =.14 (SE =.06, p =.033) and r =.13 (SE =.05, p =.016), respectively, with fewer DSM internalizing disorders in adulthood. Further, higher baseline discrepancies between parent and selfreports of externalizing problems correlated, r =.11, SE =.05, p =.018 with fewer DSM externalizing disorders, whereas increases of discrepancies over time correlated, r =.13, SE =.06, p =.036 with more DSM externalizing disorders in adulthood. We found no association of problem scores in one domain with adult disorders of another domain, for example, no association between internalizing symptoms and externalizing disorders. Discrepancies and changes in discrepancies over time of parent and self-reports and of teacher and self-reports of internalizing and externalizing problems contributed to the prediction of adult internalizing and externalizing DSM disorders.