The Addiction Severity Index: a field study of internal consistency and validity

Abstract

This study investigated whether the use of the Addiction Severity Index (ASI) in a network of inner-city alcohol and drug abuse clinics under nonideal conditions would yield internally consistent and valid data. A sample of 8,984 ASI scores was collected over a 34-month period. Construct validity was examined by computing the internal consistency of all subscales. Convergent and divergent validity of composite scores and of severity ratings were evaluated using correlation matrices. Findings demonstrated that ASI scores were internally consistent and valid, even though the recommended administration protocol may not always have been followed as faithfully as might be desirable. This robustness bodes well for the use of the ASI in on-line clinical environments. Results should be viewed with caution until the reliability of ASI administration is tested under similar nonideal conditions and until permissible deviations from standard protocol can be quantified.