Construct validity of the Substance Abuse Subtle Screening Inventory (SASSI) as a screening instrument for alcoholism

The purpose of the study was to evaluate the construct validity of the Substance Abuse Subtle Screening Inventory (SASSI) as a screening test for alcoholism. The study sample consisted of 238 participants. These included 50 alcoholics, 50 normals, 50 psychiatric outpatients, 50 co-dependent family members and 38 drug addicts, so classified by clinician diagnosis.All participants were administered the SASSI and also the Michigan Alcoholism Screening Test (MAST) and the MacAndrew Alcoholism-Screening Scale (AMAC). Correct classification rates for all three tests were determined using both a five-group criterion of classifying as an alcoholic, normal, psychiatric outpatient, co-dependent or drug addict, and a simpler two-group criterion of classifying as a substance abuser or non-abuser. Five null hypotheses were tested using Chi-square (alpha = .01) tests for equal proportions of classification accuracy. A sixth null hypothesis was tested using linear discriminant function analyses.Results Five-group criterion:1. The three tests differed in correctly classifying alcoholics, with the MAST statistically significantly superior to the AMAC and the SASSI.2. The three tests differed in correctly classifying normals, with the MAST and the AMAC statistically significantly superior to the SASSI.3. The three tests differed in correctly classifying psychiatric outpatients, with the AMAC statistically significantly superior to the SASSI.4. The three tests differed in correctly classifying drug addicts, with the AMAC and the SASSI statistically significantly superior to the MAST.5. The three tests did not differ to a statistically significant degree in classifying codependents.Two-group criterion (abuser/non-abuser):1. The three tests differed in correctly classifying alcoholics, with the MAST statitically significantly superior to the AMAC and the SASSI.2. The three tests differed in correctly classifying normals, with the SASSI statistically significantly superior to the AMAC and the MAST.3. The three tests differed in correctly classifying psychiatric outpatients, with the SASSI and the AMAC statistically significantly superior to the MAST.4. The three tests differed in correctly classifying drug addicts, with the MAST statistically significantly superior to the AMAC and the SASSI.5. The three tests differed in correctly classifying co-dependents, with the SASSI statistically significantly superior to the AMAC and the MAST.Employing the two-group criterion, all tests vastly improved their performance with the SASSI significantly superior to the MAST and the AMAC at author-recommended cutting scores. Increasing the cutting score for the MAST improved classification accuracy even further. Employing the linear discriminant function, the three tests differed significantly, with the MAST statistically significantly superior to all other scales.ConclusionsNone of the three scales performed adequately with the five-group criterion. The simpler two-group criterion produced an 87% classification accuracy rate for the SASSI with the study sample at author-recommended cutting scores, which was statistically significantly superior to the MAST and the AMAC. The highest classification accuracy rates of 90.3% and 91.2% respectively were produced by the MAST at elevated cutting scores of 10 and 12 in the two-group criterion. These parallelled linear discriminant function results for the MAST.Department of Counseling Psychology and Guidance ServicesThesis (Ph. D.