Objective: The Working Alliance Inventory short form (WAI-S) is one of the most commonly used alliance measures with adolescents. Yet, its factor structure has received minimal attention in the youth alliance literature. This study investigated the factor structure of the WAI-S in psychotherapy for adolescent depression and explored its measurement invariance across time, therapeutic approaches and patients’ and therapists’ perspectives. The existence of method effects associated with the negatively worded items of the scale was also assessed. Method: The setting of this study is the IMPACT trial, a randomized controlled trial assessing the effects of three therapeutic interventions in the treatment of adolescent depression. The WAI-S was completed at 6, 12 and 36 weeks after randomization by 338 adolescents and 159 therapists. Data were analysed using confirmatory factor analysis. Results: The hypothesized Bond-Task-Goal alliance structure was not supported and a general, one-factor model was found to be more psychometrically valid. The existence of a method effect and measurement invariance across time and treatment arms were also found. Conclusions: While the distinction between the specific alliance dimensions is conceptually and clinically interesting, at an empirical level the alliance features of the WAI-S in youth psychotherapy remain strongly intercorrelated.