Most studies rely on cross-sectional retrospective reports from adult samples to collect information about adverse childhood experiences (ACEs) to examine relationships with adult outcomes. The problems associated with these reports have long been debated, with only a few studies determining their reliability and validity and fewer still reaching consensus on the matter. This paper uses repeat prospective and retrospective reports of adverse childhood experiences from two respondent sources in the South African Birth to Twenty Plus (Bt20+) cohort to explore agreement and concordance in the prospective reporting of ACEs by caregivers and respective children as adolescents and then as young adults. The findings demonstrate little overall agreement between prospective and retrospective accounts of childhood experiences, with 80% of kappa values below the moderate agreement cutoff (k = .41). The highest levels of agreement were found between prospective and retrospective reporting on parental and household death (kappas ranging from .519 to .944). Comparisons between prospective caregiver reports and retrospective young adult reports yielded high concordance rates on sexual and physical abuse and exposure to intimate partner violence (91.0%, 87.7% and 80.2%, respectively). The prevalence of reported ACEs varied with the age of the respondent, with adolescents reporting much higher rates of exposure to violence, physical and sexual abuse than are reported retrospectively or by caregivers. This variation may partly reflect actual changes in circumstances with maturation, but may be influenced by developmental stage and issues of memory, cognition and emotional state more than has been considered in previous analyses. More research, across disciplines, is needed to understand these processes and their effect on recall. Long-term prospective studies are critical for this purpose. In conclusion, methodological research that uses a range of information sources to establish the reliability and validity of both retrospective and prospective reports ‒ recognizing that the two approaches may fundamentally answer different questions ‒ should be encouraged.