Factorial Validity of the Alcohol Use Disorders Identification Test (AUDIT) with Drunk Drivers, Emergency room patients, and Colleges Students

Timothy Conley
School of Social Work
Boston College
Commonwealth Ave.
Boston MA 02467
978-465-7999
FAX: 781-594-8509
ttconley@mediaone.net
 
Kathleen Carty
School of Social Work
Boston College
Commonwealth Ave.
Boston MA 02467
401-456-2526
kacarty@earthlink.net
 
Richard Longabaugh
Center for Alcohol and Addiction Studies
Brown University
Box G-BH
Providence RI 02906
401-444-1836
 
Robert Woolard
Rhode Island Hospital
393 Eddy St.
Providence RI 02903
401-444-4000
 
Allison Minugh
DATACORP
1 Richmond Square
Providence RI 02906
401-331-1500
 
Thomas O'Hare
School of Social Work
Boston College
Commonwealth Ave.
Boston MA 02467
617-552-4060
The AUDIT is one of the most widely used rapid assessment instruments for detecting the presence of an alcohol use disorder.  Its ten items are presumed to measure three distinct conceptual domains:  hazardous drinking, dependency and harmful drinking.  This study uses exploratory factor analysis to test for the emergence of these three domains as latent variables in three populations.
 
Secondary analysis was undertaken on the data from three previous studies involving the AUDIT:  256 multiple offender drunk drivers, 530 emergency room patients and 505 college drinkers.  An extensive series of  factor analysis' were run to examine the factorial validity of the hypothesized constructs.  Populations were examined separately and in combination.  Principal components and varimax extraction methods were employed as well as both orthogonal and non-orthogonal rotation procedures. Factor matrices and scree plot output was reviewed.
 
Factor analysis of the AUDIT failed to validate the existence of its three hypothesized constructs  and, regardless of sample population, a review of eigenvalues indicates that there is a two factor structure.  These factors are labeled problem drinking and hazardous drinking.   Review of scree plots indicates evidence supporting the argument for only a single factor: problem drinking.
 
The AUDIT is a good and effective rapid assessment instrument for detecting the presence of an alcohol use disorder in diverse populations.  Its individual item groups should not however be viewed as representing conceptually different aspects of an alcohol use disorder in drunk drivers emergency room patients or college drinkers.  Positive test results should be followed up aggressively with diagnostic interviewing to determine the exact nature and makeup of an individual's drinking problem.