Hamilton focuses on a population of offenders released from prison to halfway houses. Based on assessments administered prior to admission, intervention groups were established using Latent Class Analysis, which identified several classes or types of substance abusing offenders. The resulting classification was utilized to assess responsivity among four program orientations: 12-Step, Cognitive-Behavioral, Rehabilitation and Therapeutic Communities. A matching strategy was then created from class-program interactions. This study is first to examine typology construction as a method for improving treatment effectiveness. Policy implications describe extensions of the study methodology - refining target populations and improving resource allocations for reentry.