Objective: Almost no studies identify mediators of psychosocial interventions for attention deficit/hyperactivity disorder (ADHD)—largely due to design limitations. Understanding mediators can promote streamlined interventions in usual care (UC) settings. When individual studies are insufficient to pursue complex questions, integrative data analysis (IDA) allows researchers to pool raw data from multiple studies to produce cumulative scientific knowledge. Method: We leveraged IDA to pool and harmonize data from four randomized controlled trials of ADHD psychosocial treatment (N = 854) with three time points. Linear growth curve analyses examined the impact of four psychosocial treatment conditions on ADHD symptom outcomes and five candidate mediators (compared to no treatment). To test mediation, we examined whether treatment condition predicted linear growth in the mediator at posttreatment, and if the mediator predicted linear growth in the outcome at follow-up. Results: Compared to no treatment, engagement-focused parent–teen treatment (d = .43–.72; Supporting Teens’ Autonomy Daily [STAND]) and community-based usual care (d = .54–.99) led to greatest reductions in parent-rated ADHD symptoms, followed by the Summer Treatment Program–Adolescent (d = .29–.30; STP-A) and standard behavioral parent training + organization skills training (d = .26–.31; BPT/OST). Improvements in organization, time management, and planning skills mediated outcome for all treatments. BPT/OST and STP-A prevented deterioration of social skills, in turn mitigating escalation of ADHD symptoms. Improvements in parent–teen communication skills mediated outcome for STAND, BPT/OST, and the STP-A. Parent contingency management and disruptive classroom behavior were not treatment mediators. Conclusions: Psychosocial treatments for adolescent ADHD primarily improve ADHD symptoms through development of teen organization, time management (OTP), and parent–teen communication skills, as well as slowing deterioration of social skills. (PsycInfo Database Record (c) 2022 APA, all rights reserved)