A growing body of research has consistently demonstrated a relationship between alliance and treatment outcomes in youth psychotherapy. However, previous research often suffered methodological issues that prevented detailed investigation of temporal relationships between alliance and symptomatology. The current study explored the directions of effect between alliance and outcome by examining the associations between early alliance and subsequent outcome while controlling for patients' baseline severity and prior symptom change. It also examined potential moderators of this association. Data were drawn from the IMPACT study, a randomized controlled trial comparing cognitive-behavioral therapy and short-term psychoanalytic psychotherapy versus a brief psychosocial intervention in the treatment of adolescent depression. Adolescents (N = 224) and therapist (n = 139) rated the alliance 6 weeks after randomization. Depression severity and overall psychopathology were assessed at baseline, 6, 12, 36, 52, and 86 weeks after randomization. Patients' age, gender, baseline depression severity, conduct disorder symptoms, and treatment type were examined as potential moderators of the alliance-outcome association. Data were analyzed using multilevel models. Findings suggested that higher early alliance ratings were associated with subsequent symptom reduction, even after controlling for prior symptom change and baseline severity. There was some evidence that the strength of this association was strongest in cognitive-behavioral therapy, weaker in short-term psychoanalytic psychotherapy, and statistically indistinguishable from zero in brief psychosocial intervention. These findings suggest that early therapeutic alliance with adolescents may influence subsequent outcome independent of prior symptom change and initial severity but that the effect of the alliance on outcome might vary across treatment types. (PsycInfo Database Record (c) 2021 APA, all rights reserved).