Problem drinking: A construct and its measurement

Abstract

A problem drinking screening test for general hospital patients was developed in five experiments as a measure of the construct of problem drinking. In Experiment I, test questions were selected from a group of 59 psychological and medical items on the basis of responses from 1613 hospital patients. The resulting Problem Drinking Screening Test (PDST) has a 23 item self-report section and an optional physician's section of four items. Experiments II to V are concerned with the validity and reliability of the PDST.

Viewing problem drinking as a construct is justified theoretically, and its relationship to alcoholism, to locus of control, to labelling theory and to idiographic and nomethetic measurement is given. A method for measuring the construct is examined from the perspective of George Kelly's 1955 Personal Construct Theory, and a mathematical explanation is presented.

The experiments are introduced by a comprehensive review of the available problem drinking instruments. This review is divided into psychological measures, biomedical measures, combined psychological and biomedical measures, and others. Finally, a brief review is provided of the statistical considerations in developing a test.