Few studies have examined the role of the therapeutic alliance scores in active versus control psychotherapies. Using data from a randomized clinical trial of psychotherapy and pharmacotherapy for cocaine dependence, it was found that therapeutic alliance scores were rated as significantly more positive in cognitive-behavioral treatment than clinical management, a psychotherapy control condition. However, level of the therapeutic alliance was associated with outcome for the control but not the active psychotherapy. These data suggest that control conditions, which are intended to control for common factors of psychotherapies such as the therapeutic alliance, may exert their effect on outcomes primarily through the level of those common factors.