An Examination of the Factor Structure of the Alcohol Use Disorders Identification Test in Two High-Risk Samples

Abstract

The Alcohol Use Disorders Identification Test (AUDIT) was examined by employing confirmatory factor analytic techniques to data from two samples collected 1998–1999: college students (n = 465) and court-referred, substance use treatment outpatients (clinical sample; n = 135). Despite the fact that the AUDIT was originally designed as a three-factor measure (consumption, dependence, and consequences), previous studies have lent support to one- and two-factor models. The results of this study support a two-factor model (alcohol consumption and dependence/consequences) in both samples. As further evidence that the two-factor model is appropriate, a psychometric evaluation suggested that the AUDIT generated reliable scores in both groups when used as either a one- or two-factor measure, but not when three scores are derived in the student sample.

• Alcohol
• Screening
• AUDIT

Resumen

La prueba Alcohol Use Disorders Identification Test (AUDIT) fue examinado empleando técnicas analíticas de factores confirmatorios en información obtenida de dos muestras: Estudiantes universitarios (n = 465) pacientes externos referidos por el juzgado para tratamiento por uso de sustancias (muestra clinica; n = 135). A pesar de que el AUDIT fue diseñado originalmente como una medida de tres factores (consumo, dependencia, y consecuencias), estudios previos han apoyado los modelos de uno y dos factores. Los resultados del presente estudio apoyan el modelo de dos factores (consumo de alcohol y dependencia/consecuencias) en ambas muestras. Como mayor evidencia de que el modelo de dos factores es apropiado, una evaluación psicométrica sugiere que el AUDIT genera cifras confiables en ambos grupos cuando se usa como medida tanto como modelo de uno como de dos factores, pero no cuando las tres cifras como son derivadas de la muestra de estudiantes.

Résumé

Le Test d’Indentification d’Abus d’Alcool et de la Dépendance Alcoolique (l’AUDIT) a été examiné en appliquant des techniques d’analyse factorielle confirmatoire à des données prises de deux échantillons: des étudiants en université (n = 465); et des patients renvoyés par le système judiciaire en consultation externe à un service de traitement pour la toxicomanie (échantillon clinique; n = 135). Malgré le fait que l’AUDIT ait été conçu à l’origine comme une mesure à trois facteurs (la consommation, la dépendance, et ses conséquences), les recherches jusqu’à présent ont soutenu des modèles à un ou à deux facteurs. Les résultats de la recherche actuelle soutiennent un modèle à deux facteurs (la consommation de l’alcool, et la dépendance ainsi que ses conséquences) dans les deux échantillons. En tant que preuve supplémentaire que le modèle à deux facteurs est convenable, il existe une mesure psychométrique qui a suggéré que l’AUDIT génère des résultats fiables dans les deux groupes lorsqu’il est employé comme une mesure à un ou deux facteurs, mais non pas lorsque les résultats sont divisés en trois au sein de l’échantillon des étudiants.

Notes

^aMaisto et al. (2000b) used an interpretive technique requiring them to drop items that loaded significantly (e.g., loadings with absolute value greater than 0.40) on both factors (Stevens, 1996). This caused them to drop item 10 from the final solution. The same rule applied to Karno et al. (2000) would have led to dropping item 4 from their MLF two-factor solution.

Additional information

Notes on contributors

Alan L. Shields

Alan Shields received a B.A. in psychology from the University of Dayton. Alan is currently a doctoral student in clinical psychology at the University of Montana and is on clinical internship as a Harvard Fellow at McLean Hospital in Belmont, MA. He will be awarded a Ph.D. in July 2003, following completion of his internship. As a graduate student, Alan has been an active member of the Addictive Behaviors Research Laboratory and Psychological Statistics and Measurement Center, both housed at the University of Montana. This exposure led him to develop research interests in the phenomenon and treatment of alcohol use disorders. Alan is primarily interested in exploring novel data analytic and data gathering methods, and understanding the conceptual and methodological processes affecting self-reports of drinking and other health-related behaviors. Clinically, Alan has almost 8 years of experience working in outpatient and inpatient mental health and substance user treatment settings.

Katarina Guttmannova

Katarina Guttmannova, M.A., came to the United States from her home country of Slovak Republic almost 8 years ago to pursue a higher education in psychology. Since then, she has received a Master's degree and is currently working on her dissertation for a Ph.D. in Developmental Psychology at the University of Montana. Katarina also works full-time as a Research Associate for the National Rural Bioethics Project and is an active member of the Psychological Statistics and Measurement Center at the University of Montana. Her professional passions include, but are not limited to, psychometrics, statistics, metacognition, and child development. Katarina currently resides in Seattle with her husband and their talkative cat.

John C. Caruso

John C. Caruso received his Ph.D. in Quantitative Psychology from the University of Southern California in 1997. Prior to this, he received an M.S. in Clinical Psychology from Mississippi State University in 1993 and a B.S. in Psychology from Michigan State University in 1990. During his doctoral studies, John worked as a statistical consultant at a Los Angeles-based market research firm, Horizon Research Corporation. John is currently an Assistant Professor where he teaches both undergraduate and graduate statistics and methods classes, mentors graduate students, and directs the Psychological Statistics and Measurement Center. Dr. Caruso is also the Director of Services for Dissertation Statistical Services. John's primary research interests include ordinal statistics, multivariate statistics, and psychometrics. Most recently, his research has examined a new statistical technique called reliable component analysis. The procedure is similar to other forms of data reduction (or factor analysis), except that the composites that result have maximum reliability, a very desirable property. In addition to examining statistical techniques, John has substantive interests in the areas of autobiographical memory, personality theory, and intelligence.