Objective: The authors investigate a model on how clients’ differential involvement in therapeutic methods mediates the effect of therapist support in psychological treatment—in this case, a cognitive behaviorally based bibliotherapy for insomnia, administered with or without supportive telephone calls. Method: Eighty-nine participants, who fulfilled diagnostic criteria for insomnia, had a mean age of 49.1 years (range, 18–73 years) and were predominantly female (77%), fairly well educated, and mainly Caucasian. Participants were randomized between a bibliotherapeutic self-help treatment and the same treatment with the addition of therapist support. Primary outcome measure was the Insomnia Severity Index. Data on involvement in different methods and aspects of the treatment were estimated by clients at posttreatment and validated against therapist ratings of client involvement during treatment. Structural equation modeling was used to test if the effect of therapeutic support on outcome was mediated by involvement in treatment. Results: Carrying out the treatment with therapist support significantly boosted the therapeutic effects. A mediational analysis with involvement in the three key treatment methods (sleep restriction, sleep compression, and stimulus control) as the mediator fully mediated the differential effect between the two conditions (Sobel test; r = .31; z = 2.173; p < .05) and explained 68.4% of the total effect. Conclusions: Therapeutic support improved outcome via higher patient involvement rather than having a direct effect on outcome. Thus, relationship and methods could be regarded as interactional, and patient involvement should be considered. These factors could be further studied in treatments where specific ingredients within the therapeutic contact can be experimentally manipulated.