Relation of the alliance with outcomes in youth psychotherapy: A meta-analysis

The goal of this meta-analytic review was to provide a reliable estimate of the alliance–outcome relation in youth psychotherapy. Previous meta-analyses focused upon the alliance–outcome association in youth and adult psychotherapy have produced effect size (ES) estimates above r = .20. In the current study, meta-analytic methods were applied to the largest study sample collected (N = 38) to date in the youth psychotherapy field and the mean weighted ES estimate was r = .14, which is smaller than previous estimates. The child- and parent-therapist alliances were not differentially associated with outcomes. However, the alliance–outcome association did vary across theoretical (i.e., child age, problem type, referral source, and mode of treatment) and methodological (i.e., source and timing of alliance assessment; domain, technology, and source of outcome assessment; single vs. multiple informants) variables. Existing client-, therapist-, and observer-report alliance measures evidenced adequate reliability; however, substantial variability exists in how the alliance is conceptualized and measured. Though the magnitude of the ES estimate raises questions about the role that the alliance may play in youth psychotherapy, the findings also suggest that the extant literature represents a heterogeneous group of studies whose effects vary according to theoretical and methodological factors. Addressing existing knowledge and measurement gaps in the field may therefore lead to a more robust estimate of the alliance–outcome association in youth psychotherapy.

Research Highlights
► The mean weighted effect size estimate for the alliance–outcome association was r = .14.
► The mean weighted effect size estimate is smaller than previous estimates.
► Theoretical and methodological variables moderated the alliance-outcome association.