CONSTRUCT REFINEMENT AND CONFIRMATION FOR THE ADDICTION SEVERITY INDEX

A series of analyses reconstruct and confirm the validity and reliabil ity of hypothesized scales for the Addiction Severity Index. Using a d iverse sample (N = 990 ) of methadone maintenance substance abuse pati ents, a multistage scaling strategy was applied to identify 7 psychome trically integral addiction problem scales. Exploratory item analyses, confirmatory oblique item clustering, and 2nd-order factor analysis v erified that the scales comprised relatively little common variance an d that each retained a substantial amount of unique and reliable varia nce. Concurrent and predictive validity were supported with a sample o f 244 methadone patients assessed throughout treatment. The advantages of the new scales are discussed, including provision of computer code for calculating normalized standard scores for use in clinical practi ce and treatment outcome research.