In a randomized trial for generalized anxiety disorder, cognitive-behavioral therapy (CBT) and CBT integrated with motivational interviewing (MI) promoted comparable worry reduction at posttreatment, whereas MI-CBT outperformed CBT over 12-month follow-up (Westra, Constantino, & Antony, 2016). Secondary analyses revealed competing mediators of the long-term treatment effect: MI-CBT related to lower patient resistance to the treatment, which promoted lower follow-up worry, whereas CBT related to greater increases in patient friendly submissiveness (FS), or compliance, which also promoted lower follow-up worry (that suppressed an even greater long-term advantage of MI-CBT). In this study, we tested these competing, though theoretically consistent, variables as mediators of the nonsignificant treatment effect on posttreatment worry, as there could also be treatment-specific means to arriving to these comparable ends. Eighty-five patients received 15 sessions of MI-CBT or CBT. Therapists rated patient FS through treatment, observers rated resistance at midtreatment, and patients rated worry at posttreatment. Bootstrap analyses indicated that MI-CBT patients exhibited less resistance, which promoted lower posttreatment worry, whereas CBT patients had greater increases in FS, which also promoted lower worry. CBT and MI-CBT achieved comparable posttreatment outcomes through separate indirect paths that each conferred an advantage for one treatment over the other (and had canceled out a direct treatment effect immediately after therapy). The composite trial findings have significance for understanding different pathways to personal change in both the short- and long terms, and for the importance of testing indirect effects both when direct treatment effects do and do not emerge. (PsycINFO Database Record (c) 2018 APA, all rights reserved).