Objective: This study presents outcomes from a randomized trial of a novel Parent-Child Interaction Therapy (PCIT) model for foster families. Differential effects of two intervention doses on child externalizing and internalizing symptoms are examined. Method: A sample of 102 foster children was assigned to one of three conditions--brief PCIT, extended PCIT, or wait-list control. The brief and extended groups received 2 days of PCIT training and 8 weeks of telephone consultation. The extended PCIT group received an additional booster training plus 6 more weeks of consultation. Wait-list controls received services as usual. Tests of change over time were estimated using mixed-model repeated measures analysis of covariance. Results: Compared to controls, children in both PCIT groups exhibited a greater reduction in externalizing and internalizing scores over time. Pairwise contrasts of the two PCIT conditions yielded mixed results. Conclusion: Results indicate that PCIT can be tailored efficaciously for foster families using alternative treatment modalities.