Abstract
Objective: Better alliance is known to predict better psychotherapy outcomes, but the interdependent and interactive effects of both therapist- and patient-reported alliance levels have yet to be systematically investigated. Method: Using actor–partner interdependence model analysis the authors estimated actor, partner, and 2 types of interactive effects of alliance on session outcome in a sample of 241 patient–therapist dyads across 30 sessions of cognitive–behavioral and alliance-focused therapy. Results: Findings suggest that the most robust predictors of session outcome are within-treatment changes in patient reports of the alliance, which predict both patient and therapist report on outcome. Within-treatment changes in therapist reports of the alliance, as well as differences between patients and between therapists in their average ratings of alliance levels across treatment, predict outcome as reported by the specific individual. Although alliance was found to be a significant predictor of outcome in both treatments, for therapist-reported alliance and outcome it had a stronger effect in alliance-focused therapy than in cognitive–behavioral therapy. Additionally, dyads with the highest pooled level of alliance from both partners fared best on session outcome. Conclusions: The results are consistent with a 2-person perspective on psychotherapy, demonstrating the importance of considering the interdependent and interactive nature of both patient and therapist alliance levels on session outcome.