A PROGRAM OF SERVICES MUST SEEK ITS BASIC DEFINITION AND OBJECTIVES IN TERMS OF CONSUMER IMPACT AND GOALS. TWO GROUPS OF CONSUMERS ARE IDENTIFIED: THE ACTUAL RECIPIENTS OF SERVICES AND THE COMMUNITY AT LARGE. WHILE PROGRAM IMPACT ON SERVICE RECIPIENTS CAN BE MEASURED USING GLOBAL LEVEL OF FUNCTIONING SCALES, SUCH SCALES ARE SUSCEPTIBLE TO BIAS. TWO CONDITIONS, HOWEVER, APPEAR TO INCREASE THE RELIABILITY OF GLOBAL SCALES: IF THE CLINICAL STAFF IS ACTIVELY INVOLVED IN THE ADAPTATION OF THE SCALE'S VOCABULARY TO LOCAL MEANS AND IF THERE IS REGULAR FEEDBACK CONCERNING THE CLINICIANS' COLLECTIVE USE OF THE SCALE. A STUDY IS DESCRIBED IN WHICH THE RESULTS OF A CLINICAL JUDGMENT TASK ARE ANALYZED USING ANDERSON'S STIMULUS INTEGRATION MODEL. THE SAME 20 CASE HISTORY ABSTRACTS WERE RATED BY 65 CLINICIANS IN 6 MENTAL HEALTH PROGRAMS FOR GLOBAL LEVEL OF FUNCTIONING USING THEIR OWN LEVEL OF FUNCTIONING SCALES. ANALYSIS OF THE 65 CLINICIANS' GLOBAL LEVEL OF FUNCTIONING JUDGMENTS SHOWED THAT A CLIENT'S PRIOR PSYCHIATRIC HISTORY DID AFFECT DIFFERENTIALLY CLINICIANS WITH DOCTORAL-LEVEL TRAINING RELATIVE TO THOSE WITH MASTER- AND BACHELOR-LEVEL TRAINING. THERE WERE A NUMBER OF METHODOLOGICAL PROBLEMS THAT INTERFERED WITH THE ASSESSMENT OF THE EFFECTS OF CLASS AND GENDER OF CLIENTS ON CLINICAL JUDGMENT. IT WAS CONCLUDED, HOWEVER, THAT THE GENERAL METHOD OF CLINICAL JUDGMENT STUDIES PRESENTS A VIABLE MEANS FOR TESTING WHETHER DIFFERENCES IN CLIENT OUTCOMES AND RELATED SERVICE COSTS ARE RELATED TO DIFFERENCES IN CLINICIANS' JUDGMENTS. FEEDBACK CAN BE PROVIDED FOR STAFF DEVELOPMENT IN A PROGRAM'S SERVICE PLANNING, DELIVERY, QUALITY ASSURANCE, AND EVALUATION PROCEDURES. REFERENCES ARE PROVIDED. (AUTHOR ABSTRACT MODIFIED--KJM)