The need to portray accurately the level of functioning and severity of psychiatric symptoms among mentally ill offenders (MIOs) is paramount from several perspectives. The prison environment may cast aspersions on the reliability and validity of commonly used functional assessment tools. In addition, these tools do not capture environment-specific areas that may be of interest to the courts, clinicians, community mental health centers, and other correctional facilities. Male MIOs (n=61) who had been treated for at least 3 months in a male Washington State prison mental health program were evaluated by using clinical assessment tools, data abstraction from medical records, and structured assessments from correctional officers. Clinical assessments occurred at their current site of incarceration. The semistructured clinical assessment had high construct validity and correlation for psychiatric symptoms and diagnosis. The ability of evaluators to determine accurately relative treatment compliance within the prison was low compared with the reports from correctional staff, particularly with respect to attendance at programs. In general, the officers did not recognize lack of program participation and reclusive behavior as potential signs of mental illness. Despite a significant history of psychiatric symptoms severe enough to warrant inpatient treatment, 70 percent of the MIO individuals were functioning reasonably well in a general population. A fully informed functional assessment of MIO's likely requires input from both clinicians and correctional officers. 4 tables, 1 figure, and 19 references