Alcohol Use Disorder Identification Test (AUDIT): Exploring its psychometric parameters

Abstract

Objective: This study aimed to know evidences of validity and reliability of the Alcohol Use Disorder Identification Test (AUDIT). Methods: For this purpose, we considered 547 college students from Fortaleza (CE) with a mean age of 21.6 years (sd = 4.86, varying from 18 to 53), mostly male (51.5%), single (91.4%) and catholic (62.5%). The participants answered the AUDIT and demographic questions. Looking to know AUDIT's factorial structure, we carried out a principal component analysis. Additionally, in order to evaluate its reliability, we calculated Cronbach's alpha (internal consistency), Pearson correlation r and intraclass coefficient correlation - ICC (test-retest). Also, descriptive statistics were evaluated. Results: According to the principal component analysis using oblimin rotation, the bifactorial structure of the AUDIT proved to be consistent, once all items showed satisfactory saturations exceeding |0.40|, with the first factor explaining 47.5% of the total variance with alpha of 0.84 and Factor 2 explaining 11.6% of the total variance with alpha of 0.69. The results of the test-retest showed strong correlation between data obtained in the first (t 1) and in the second (t 2) applications (r tt = 0.94, p < 0.01), with no significant difference between the means in both applications (m t1 = 0.37, sd = 0.49, = 0.34 m t2, dp 2 = 0.47, p > 0.05) with satisfactory ICC (0.96). Conclusions: The findings support the psychometric adequacy of the AUDIT, with the exploratory factor analysis pointing as more satisfactory the structure with two factors, and also showing its good temporal stability.