Abstract
Investigated the reliability, validity and dimensionality of scores obtained by acute psychiatric inpatients (N = 120) on the Michigan Alcoholism Screening Test adapted to assess lifetime and current symptomatology. Under most conditions, test scores and classifications were reasonably consistent over brief intervals and the underlying structure of test content was unidimensional. The test showed high sensitivity (88%), insufficient specificity (50%), and modest classification accuracy overall (58%) and thus underscored the problem of excessive false positives and/or an inadequate diagnosis as criterion.