Establishing the construct validity of rapid assessment instrument's (RAI's) is essential to their effective clinical utilization. A sample of multiple offender druk drivers were administered the Severity of Alcohol Dependency Questionnaire (SAD-Q, n=132), Alcohol Use Disorders Identification Test (AUDIT, n=256) and the Michigan Alcoholism Screening Test (MAST, n=257). 187 were then evaluated using a semi-structured clinical interview designed to establish a DSM-IV diagnosis: the Alcohol Use Disorders and Associated Disabilities Interview Schedule (AUDADIS). Using the interview's items as criteria, the validity of each instrument's theoretical and conseptual constructs were evaluated by examining statistical convergence between test and criterion using Pearson's r. Moreover, each instrument was factor analyzed to examine the factorial validity of its constructs. Receiver Operating Characteristics (ROC) was applied to determine sensitivity, specificity, and optimal cut points for each instrument. The MAST exhibited slightly better convergence with DSM-IV (r=.628; p<.01) than either the SAD-Q (r=.568; p<.01 or the AUDIT (r=.548; p<.01). Overall, the MAST exhibited better construct validity than the other two instruments and a short version of this test is recommended for continued use by the program. AUDADIS interviewing is suggested as the only effective manner of assessing physiological dependence. Social workers are advised to develop an educated contextual understanding of the theory, conceptual constructs, and scoring properties of the RAI's they use in clinical practice. (A)
Abstract