Previous research has shown the performance of the CJDATS Co-Occurring Disorders Screening Instruments (CODSI-MD and SMD)—six- and three-item instruments to screen for any mental disorder (CODSI-MD) and for severe mental disorders (CODSI-SMD), respectively—to be comparable or superior to other, longer instruments. This study tested the stability of the performance of the CODSI-MD and SMD across three racial/ethnic groups of offenders entering prison substance abuse treatment programs (n = 353), consisting of 96 African American, 120 Latino, and 137 White admissions. The Structured Clinical Interview (SCID) was used to obtain DSM-IV Axis I and II diagnoses; a lifetime SCID diagnosis of a mental disorder or a severe mental disorder was the criterion against which the CODSI-MD and SMD were validated. Results showed no statistical differences in sensitivity or specificity for either the CODSI-MD or SMD across the African American, Latino, and White prisoner groups. The value of the CODSI-MD and SMD as brief screens for mental disorders among offenders with diverse racial/ethnic backgrounds is discussed. Copyright © 2008 John Wiley & Sons, Ltd.