There is an increasing need for practical instruments that can rapidly and accurately assess the effectiveness of treatments for mental illness in clinical settings. Symptom rating scales used in clinical research are too complex and time-consuming to be useful in these settings. In contrast, single-item global measures of severity such as the Clinical Global Impression-severity scale (CGI) and the Global Assessment of Function scale (GAF) are brief and easy to complete, but little is known about their relationship with the specific symptoms of severe mental illnesses. In this study, we examine the extent to which CGI and GAF scores reflect the severity and the change in severity of positive, negative, depressive, and agitation symptoms in a sample of 102 schizophrenia inpatients at the University of Michigan Medical Center. At admission, positive symptoms were the strongest correlates of both CGI and GAF scores, followed by negative symptoms, and agitation. Depressive symptoms did not correlate significantly with either global measure. The three symptom scores together explained 58% of the variation in CGI and 39% of the variation in GAF. A similar pattern of association was found for the scores measured at discharge and for the relationships between the change in global measures and change in specific symptom clusters. Thus, by demonstrating that single-item global measures, particularly the CGI, can be reasonably good indicators of psychopathology, this study suggests that these measures may be practical tools for routine monitoring of the effectiveness of treatments for severe mental illness in community settings.