In 1991, Colorado instituted a standardized assessment and treatment process designed to match offenders to treatment services. The assessment battery measures criminal risk and substance abuse needs. The scores from these assessments are used to assign offenders to one of the seven classifications in the treatment system. The sample consisted of transitional community corrections and parole offenders. The research considered the utility of the LSI in predicting recidivism rather than identifying and treating substance abusing offenders. The findings are intended to assist the Colorado Department of Corrections in deciding whether to fully implement the LSI as a classification system with community-based offenders. The sample included 172 parolees and 85 transitional community corrections residents. Offenders were assessed upon release to the community and after 6 months of community supervision. To establish the predictive validity of the LSI, two outcome variables were collected. The first included a 10-point rating of the offenders' compliance with their supervision, completed by the parole officer or ComCor case manager at 6-month and 12-month follow-up periods. The second outcome variable was whether or not offenders were reincarcerated within a year of release to the community. The findings show that the LSI was predictive of outcome with parolees but not community corrections offenders. Time 2 LSI reassessment generally did not improve prediction of 12-month outcomes over the initial assessments. Also, the LSI was more predictive of recidivism than the Wisconsin classification system for parolees. Implications of the findings are discussed, and recommendations are offered. 5 tables, 10 figures, and 30 references