Used an observation system developed to measure client resistance to code videotaped therapy sessions at the beginning, middle, and end phases of treatment for 27 families referred for child management problems. Only 18 families completed treatment. Resistance scores covaried with phase of treatment, probability of completing treatment, source of referral (i.e., agency vs self), and therapist ratings of treatment outcome. The utility of a molecular coding system for quantifying client resistance in treatment and the construct validity of the observation system received moderate support.