To elucidate processes underlying therapeutic change in a large-scale randomized clinical trial, we examined whether alterations in self-reported parenting practices were associated with the effects of behavioral, medication, or combination treatments on teacher-reported outcomes (disruptive behavior, social skills, internalizing symptoms) in children with attention-deficit hyperactivity disorder (ADHD). Participants were 579 children with Combined-type ADHD, aged 7-9.9 years, in the Multimodal Treatment Study of Children with ADHD (MTA). We uncovered 2 second-order factors of parenting practices, entitled Positive Involvement and Negative/Ineffective Discipline. Although Positive Involvement was not associated with amelioration of the school-based outcome measures, reductions in Negative/Ineffective Discipline mediated improvement in children's social skills at school. For families showing the greatest reductions in Negative/Ineffective Discipline, effects of combined medication plus behavioral treatment were pronounced in relation to regular community care. Furthermore, only in combination treatment (and not in behavioral treatment alone) was decreased Negative/Ineffective Discipline associated with reduction in children's disruptive behavior at school. Here, children in families receiving combination treatment who showed the greatest reductions in Negative/Ineffective Discipline had teacher-reported disruptive behavior that was essentially normalized. Overall, the success of combination treatment for important school-related outcomes appears related to reductions in negative and ineffective parenting practices at home; we discuss problems in interpreting the temporal sequencing of such process-outcome linkages and the means by which multimodal treatment may be mediated by psychosocial processes related to parenting.