OBJECTIVE: The demand to measure the clinical outcomes of persons with serious mental illness in the community is growing; however, there is no consensus about how to do this task. This paper identifies challenges in measuring the outcomes of persons with serious mental illness and reviews selected instruments that measure the community functioning of this population. METHODS: Papers in peer-reviewed psychiatric journals for the years 1986 to 1996 were reviewed to select instruments that measure two or more domains of community functioning and for which data on reliability and validity have been published. Selected instruments were evaluated, focusing on their format, content, item scoring, length, and original sample population. RESULTS AND CONCLUSIONS: Challenges to measuring the community functioning of persons with serious mental illness include the multiplicity of domains that must be measured, the conflicting interests of various stakeholders involved in care, the limitations of self-report data, and other methodological problems. Nine instruments that met the study criteria were selected from the literature. Three are self-report instruments, and six are based on the report of an informant or independent rater. The instruments vary in length and in their original sample population. The content areas most consistently represented are self-care and social relationships. Life satisfaction, health status, psychiatric symptoms, and work skills are not consistently addressed. Individual instruments have additional limitations, including the absence of behavioral anchors for scale items and the lack of specificity to persons with serious mental illness. Effort should be directed toward sharing data across settings, measuring the effects of treatment interventions, and demonstrating the predictive validity of outcome data.