Therapist characteristics have long been hypothesized to impact directly on treatment outcome. However, the empirical evidence relative to this hypothesis is contradictory and inconclusive. Using a new model of the therapeutic process in which four main sources of variance are identified, this present study investigated the impact of therapist characteristics on client outcome in a clinical setting. Client and therapist characteristics, programme and intermediate outcome measures were assessed using scales derived from the Family Service Assessment Project. The therapist measures included demographic, training and experience variables as well as preferred styles of interpersonal interaction and model of therapy. The sample for the study was comprised of thirty-four therapists treating 394 clients associated with seven Canadian Family Service Agencies, Pre and post assessments of clients were made by the therapist, and after termination, client feedback questionnaires were completed. This feedback included self-reported assessments of satisfaction with the treatment and with the therapist. At closure, reasons for termination were also recorded by the therapist. Analysis of variance was employed to determine if there were differential effects in treatment outcome due to therapist characteristics, and to explore the possibility of there being interactions amongst worker characteristics and between worker and client characteristics. It was found that there were differential effects in outcome of therapy as a function of therapists' interpersonal style and client factors. Other types of demographic and training measures contributed minimal variance to the clients' progress in therapy. Stepwise heirarchical regression was employed to determine to what extent therapist variables were predictive of outcome. It was found that the interpersonal style of the therapist was a significant predictor of outcome accounting for the largest share of variance.