The Children of Alcoholics Screening Test (CAST): test-retest reliability and concordance validity.

Abstract

The stability of the Children of Alcoholics Screening Test (CAST) has been called into question by several authors (Hart, 1989; Maxwell, 1985; Schinke, 1989; Stacey 1985). In a bid to verify these reservations, 376 college students were each evaluated twice by means of the CAST. Results confirm the high reliability and validity of this instrument in identifying grown-up children raised in homes with at least one alcoholic parent. Reliability was demonstrated by good internal consistency, high item-total scale correlations, and near perfect test-retest agreement (k = .83). Furthermore, concordance between high CAST scores and satisfaction of recognized criteria for alcohol dependence, as measured with the Structured Clinical Interview for the DSM-III-R (SCID), attests to the CAST's external validity (k = .78). Moreover, factor analysis shows the instrument to have a unidimensional structure. Finally, with a cutoff score of 6 or more, the CAST demonstrates excellent sensitivity and specificity.