Proposes a classification scheme that integrates previous outcome studies in psychotherapy. Studies are classified into 1 of 3 frameworks according to the question addressed, the assumptions made, and the experimental designs used. Strengths and weaknesses of the 3 frameworks are examined within the context of specific studies of outcome. Evidence from the 3 frameworks suggests an alternative formulation of the question about outcome and a new experimental design. Specifically, the design involves evaluating the effectiveness of psychotherapy with respect to a population of patients deemed appropriate for psychotherapy. This design avoids random assignment of patient to psychotherapist and permits the investigator to address the question of whether psychotherapists are effective with patients whom they regard as appropriate for their particular type of therapeutic intervention.