This multimethod study of mothers and toddlers (a) examined the effectiveness of a play-based intervention (child-oriented play vs. play-as-usual) on children's cooperation with their mothers and socioemotional competence; (b) introduced a robust new measure of maternal engagement in the intervention, reflected in the dose of child-oriented play the mother delivered to the child; and (c) examined ecological factors that predicted maternal engagement, and the effect of engagement on the outcomes. Low-income mothers (N=186, 11% Latino, 27% minority) were randomized into child-oriented play group or play-as-usual group, participated in 8 play sessions, and played daily with their children for 10 weeks. Microscopic coding of mothers' behavior in play sessions assessed the dose of child-oriented play delivered to children; mothers' diaries assessed time in daily play. Children's cooperation with maternal control, observed in the laboratory, and mother-rated competence were measured before randomization (Pretest), after play sessions (Posttest 1), and 6 months later (Posttest 2). Children in both groups made significant gains in both outcomes. The gains in cooperation appeared longer lasting in child-oriented play group. Both groups made significantly greater gains than a historical community control group, an unrelated longitudinal study without any intervention. Structural equation analyses revealed that married mothers and those with fewer children delivered higher doses of child-oriented play, and those doses predicted children's higher cooperation and competence, with the effects of earlier scores covaried. The dose of time spent in daily play had no effect. Child-oriented play may be a promising, effective, and inexpensive means of promoting toddlers' positive development.