Background: The Alcohol Use Disorders Identification Test (AUDIT) was developed as a screening tool for problematic alcohol use and an intervention framework to aid those who drink excessively. While the AUDIT is widely used with at-risk populations, such as military veterans, major gaps exist in the research literature regarding the construct validity of the AUDIT in military samples. Objectives: This study assessed the factor structure and measurement invariance of the AUDIT in a large sample of Canadian military veterans (N = 1669; 94.94% male). Methods: Exploratory factor analysis (EFA) was conducted using a random subsample (n = 825) to assess the underlying factor structure of the AUDIT. Confirmatory factor analysis (CFA), using the second subsample (n = 844), was used to cross-validate the factor structure revealed by EFA and compare it to other model variants. Finally, multigroup CFAs were conducted using the whole sample to further cross-validate the factor structure and examine measurement invariance in military veterans with and without clinical elevations in posttraumatic stress disorder (PTSD) symptoms. Results: Factor analyses revealed that a modified two-factor model provided a statistically better fit to the data compared to all other model variants; yet, the results did not confirm measurement invariance across military veterans with and without clinically significant symptoms of PTSD. Conclusions/Importance: The findings are in line with increasing evidence suggesting that two subscale scores should be calculated for the AUDIT. Results further suggest that care should be taken in interpreting AUDIT scores when PTSD symptoms are present for military veterans.
Keywords: AUDIT; Military; alcohol; confirmatory factor analysis; exploratory factor analysis; measurement invariance.