The reliability and validity of the Diagnostic and Statistical Manual of Mental Disorders-IV Global Assessment of Relational Functioning (GARF) was evaluated in consenting participants presenting to a family therapy training clinic. The purpose of this study was to (a) assess whether the GARF could be administered quickly in a marriage and family therapy training clinic, (b) assess the inter-rater and internal reliability of GARF ratings of intake sessions, (c) examine the relationship of GARF ratings to established measures of relationship functioning, and (d) compare GARF ratings to a measure of depressive symptoms. Study participants completed the General Functioning Subscale of the Family Assessment Device (GFS/FAD), the Quality of Marriage Index (QMI), and the Center for Epidemiologic Studies-Depression Scale (CES-D). After intake sessions were conducted with study participants, GARF ratings were made by an American Association for Marriage and Family Therapy Approved Supervisor or Supervisor-in-Training, a therapist trainee, and two observing trainees who had observed the session from behind a one-way mirror. Cronbach's coefficient alphas and the average intraclass correlation coefficients were both .82 when ratings of the supervisor and two observers were compared and .78 and .79, respectively, when ratings of the supervisor, two observers, and therapist were evaluated. There was significant agreement between GARF ratings made by the supervisor and therapist (but not observers) and scores on the GFS/FAD, QMI, and CES-D. Results suggest that GARF ratings can be made quickly and reliably, especially among raters with greater clinical experience, and are related to measures of relational functioning and depression. Implications for research and marriage and family therapy training are discussed.