Bridging Disciplinary Boundaries (January 11 - 14, 2007)

Purpose: Identification of factors that are associated with problem drinking, problem drug use, and Driving under the influence (DUI) recidivism are important because they can lead to early recognition of potential high-risk individuals. Many models explaining risk factors of DUI offenders mainly rely on simplified information such as demographics or driving characteristics (reviewed in Nochajski & Stasiewicz, 2006). The RIA Self Inventory (RIASI), a 49 item screening instrument, was developed to screen for alcohol and/or drug problems, as well as risk for DUI recidivism in the New York State Drinking Driver Programs (DDP). However, there has been little work on the psychometric properties of the RIASI. The current study was designed to evaluate the underlying structure of the RIASI.

Method: Data were drawn from 11,397 convicted drinking-drivers who completed the RIASI between 1996 and 2005. Participants were older than 18 and participated in a voluntary seven-week, 16 hour educational program (DDP). The mean age of the sample was 35.5 (SD=11.96) with the majority being male (79%). The RIASI is scored using dichotomous items, reflecting risk and no risk. For this paper, the sample was randomly split in half, with one half used for construction and the second for validation purposes. The exploratory factor analysis used a maximum likelihood method of extraction with a varimax rotation procedure. Confirmatory Factor Analyses (Mplus version 4, 2006) were used to test and validate the factor model.

Results Six factors were extracted from the exploratory factor analysis, accounting for 32.5% of the variance. The final model included 41 items that significantly contributed to the final solution. The factors represented are: Psychiatric Distress (14 items, 14% of variance); Drinking Practices (6 items, 5% of variance); Preoccupation with Alcohol Use (8 items, 4% of variance); Sensation Seeking (6 items, 3.5% of variance); Deviance (5 items, 3% of variance); and Family History of Alcohol or Drug Problems (2 items, 3% of variance). Correlations among factors varied from .312 to .752. The fit for the initial model was adequate according to the CFI (.955), TLI (.951) and RMSEA (.028) but not the Chi-Square (4245.871, df=804). Replication in the second half of the sample also provided an adequate fit according to the CFI (.956), TLI (.953) and RMSEA (.027) but not the Chi-Square (4007.495, df=804).

Implications: Initial development of the RIASI involved domains pertaining to psychiatric distress, aggression/hostility, interpersonal incompetence, drinking practices, preoccupation with alcohol, family history for alcohol or drug problems, deviance, and sensation seeking. The results from this study show that many of the initial domains remain independent but others (aggression, interpersonal competence) appear to be encompassed by more general constructs (psychiatric distress). Implications for practice stem from the underlying constructs represented by the factors. High scores on the various factors may indicate potential areas where interventions need to be targeted. By channeling the individuals into appropriate interventions, alcohol-related crashes and fatality rates that may come from the future DUI should decrease. Future research needs to consider the relationship of these factors to relevant outcomes.