In this secondary systematic review of single-case and controlled group design intervention studies conducted with transition-age autistic youth, we examined features of 48 studies with 273 participants that measured at least one "problem behavior" outcome (Prospero registration number: 231764). We searched 11 databases for relevant studies, and the final search date was November 2022. Our primary aims were to determine how problem behaviors were defined and selected for reduction, how functions were determined, and the interventions used to address them. Studies were coded and codes were tabulated and converted to percentages to answer each research question. Thirty-eight percent of studies defined problem behavior, and 88% of studies implemented behavioral strategies to reduce problem behaviors. Behaviors with low potential for harm constituted the majority of the 67 outcome variables (61%), while behaviors with high potential for harm were a minority (39%). The most common intervention target was stereotypic behavior. Fewer than half of studies: reported procedures for selecting behaviors, reported procedures to determine behavior function, or ascribed functions to behaviors. We were unable to report on some demographic features of participants (e.g. race/ethnicity) because they were rarely reported in primary studies. We conclude that problem behavior is poorly conceptualized in this research.