Validation of the Alcohol Expectancy Questionnaire (AEQ-A) for Peruvian University students

Abstract

The paper describes the adaptation of the Adolescent version of the Alcohol Expectancy Questionnaire (AEQ-A) for university students in Peru. 672 male and female university students enrolled in the first years at a private university in Peru completed the AEQ-A translated into Spanish. All 90 items were scored on 5-point rating scales. A principal component analysis with orthogonal Procrustes rotation towards the a priori 7 component structure yielded congruence measures (Tucker's phi) of .82 or less, implying that the original 7 dimensions of the AEQ-A could not be reproduced. A second, explorative principal component analysis produced a three-component solution explaining 27.8% of the variance: two unipolar components measuring positive and negative personal and social expectancies, and a bipolar component measuring expectancies regarding cognitive and motor capacities. A cross-validation using a random-split half of the sample and performing VARIMAX rotations on each subsample confirmed this three-component solution, which was further confirmed by an orthogonal Procrustes rotation of the first subsample towards the second revealed. After the elimination of 27 items due to overlapping factor loadings, scales were produced with Cronbach alpha of .94, .78 and .70. It is argued that this three-component structure corresponds with the "minimal" structure of the original AEQ.