Objective. The primary goal of this paper was to evaluate independent effects of stepfather parenting behaviors within the context of a parent training efficacy trial designed for recently married couples with children exhibiting behavior problems. A secondary goal was to examine measurement properties of a multiple-method, multiple-source construct of effective stepfathering including direct observation. Stepfather hypotheses were derived from a social interaction learning model of child adjustment and specifically evaluated the Oregon model of Parent Management Training (PMTO) intervention. Design. In a randomized control trial. 110 recently married families consisting of an early-elementary-school-aged focal child, biological mother, and stepfather were assessed over 2 years. Assessment included direct observation of stepfather-stepchild interactions. Analyses first tested intervention effects on change in stepfathering and second tested independent effects of stepfathering on change in children’s depression and noncompliance at follow ups. Results. The intervention produced medium effect sizes at 6 and 12 months for improved stepfathering with parenting effects diminishing at 24 months. Hierarchical regression models showed that intervention group improvements in stepfathering predicted greater reductions in children’s depression and noncompliance at 2 years relative to controls, controlling for change in mothering. Conclusions. These findings underscore the preventive utility of the PMTO intervention for stepfathers. Implications for research, translation, timing of intervention, and implementation are discussed.