Relationship of Latent Profiles from the RIA Self Inventory with Various Outcomes

Recently Robert Mann and his colleagues, using a Canadian sample, performed a factor analysis on the RIA Self Inventory (RIASI), a screening instrument developed for use with DUI offenders. Results showed differential relationships of the eight identified factors with various outcomes. The current study was designed to confirm the factor structure of the RIASI in a sample from the United States and then to assess whether specific profiles could be identified that would help in development of intervention strategies. Subjects were referred to the Research Institute on Addictions (RIA) for clinical evaluation. As part of that process, the DUI offenders were extended an offer to participate in this research project. Of the 765 individuals referred to the RIA from various courts in the Western New York area, 549 agreed to participate in the study, with 520 having valid data. The assessment included alcohol and other drug use and problems, abstinence self-efficacy, psychiatric distress, hostility, family history, readiness to change, and the RIASI. An 18-month follow-up was also conducted, with driver records obtained. Confirmatory factor analysis on the RIASI showed a relatively good fit: Root Mean Square Error of Approximation (RMSEA) = .018; Comparative Fit Index (CFI) = .95; Tucker-Lewis Index (TLI) = .95. Indications for the latent class analyses based on the subdimensions of the RIASI indicated the most optimum solution was for 4 classes. There were significant associations of the profiles with alcohol problems, drug problems, alcohol expectancies, abstinence self-efficacy, psychiatric distress, hostility, and treatment entry (all p’s < .05) but not for recidivism. The results indicate that the RIASI has reliable underlying dimensions that can be used to identify subgroups of offenders. Differences in the subgroups can lead to more effective intervention development and use.