This study explores the influence of adult attachment patterns on working alliance, presumed to share with the alliance construct several important conceptual elements such as bond. Although attachment in adults has been shown to have a strong relationship to working alliance, there is not complete convergence between the two concepts (Reis & Grenyer, 2004). It was expected that client perceptions of the therapy process itself may have a strong influence on working alliance as well, and may explain elements of the working alliance not accounted for by attachment. The relative weight of adult attachment and perceptions of therapy process were explored in regard to therapy outcome. In particular, supportive and confrontational behaviors by therapists are presumed to have relevance to both adult attachment patterns, measured using the Adult Attachment Questionnaire (ASQ; Feeney, Noller, & Hanrahan, 1994) and working alliance, measured using the Working Alliance Inventory-Short Form (WAI-S; Tracey & Kokotovic, 1989). To measure such a relationship, a scale of therapist support and confrontation was developed using principal components analysis, the Therapist Behavior Items (TBI), based on the therapist intentions work of Hill and O'Grady (1985). An established measure of session quality, the Session Evaluation Questionnaire (SEQ; Stiles, 1980; Stiles & Snow, 1984) was used for convergent validation of the concepts in the new scale. Hierarchical regression analyses showed that therapist behaviors of support were significant predictors of working alliance when controlling for demographics and other therapy variables, and that client reports of adult attachment security also predicted working alliance.