Abstract An adaptation of the Collaborative Study Psychotherapy Rating Scale (CSPRS) was used to rate therapist adherence to three psychotherapies for anorexia nervosa. One audiotaped psychotherapy session from each of 30 patients was rated independently by two raters. Analysis of the psychometric properties of the instrument revealed good interrater agreement and high internal consistency. For all three therapies, therapists were rated as exhibiting significantly more behaviors appropriate to the therapy to which patients were randomized than the other two therapies, indicating very satisfactory adherence to therapy. The three therapies were clearly distinguishable by raters unaware of therapy condition using the modified CSPRS.