Problematic anger is prevalent and associated with adjustment difficulties in military populations. To facilitate measurement of problematic anger, a very brief valid measure is needed. To reduce the Dimensions of Anger Reactions 5-item (DAR-5) scale to a very brief measure. This cross-sectional study used survey data collected between 2014 and 2016 in the Australian Transition and Well-Being Research Programme and US Millennium Cohort Study. Participants were service members who were actively serving or had transitioned out of the military (separated). Statistical analyses were performed from September 2021 to June 2023. The DAR-5 was reduced to the 3 experiential items: frequency, intensity, and duration (the DAR-3). Psychometrics for the DAR-3 and DAR-5 were compared in terms of standardized Cronbach α, positive screening result, mean, and SD. Analyses were stratified by Australian and US military service status cohorts (active duty and separated). A total of 71 010 participants were included from Australia and the US. Of 10 900 Australian participants (8145 active duty participants [74.7%]; 2755 separated participants [25.3%]), 5893 (55.2%) were aged 40 years or older and 8774 (80.5%) were male; of 60 110 US participants (24 706 active duty participants [41.1%]; 35 404 separated participants [58.9%]), 28 804 (47.9%) were aged 30 to 39 years and 43 475 (72.3%) were male. The DAR-3 demonstrated good internal consistency in the active duty (Australia: mean [SD] score, 4.97 [2.5]; α = 0.90; US: mean [SD] score, 5.04 [2.6]; α = 0.87) and separated (Australia: mean [SD] score, 6.53 [3.4]; α = 0.92; US: mean [SD] score, 6.05 [3.2]; α = 0.91) samples. The cutoff score of 8 or greater on the DAR-3 had optimal sensitivity and specificity across all samples. DAR-3 and DAR-5 were associated with posttraumatic stress disorder (PTSD), depression, aggression, and relationship conflict. While the scales did not significantly differ in their associations with PTSD, depression, and relationship conflict, the magnitude of association for aggression was significantly lower in US samples using the DAR-3 (eg, US active duty sample: DAR-5 OR, 9.96; 95% CI, 9.01-11.00; DAR-3 OR, 8.36; 95% CI, 7.58-9.22). In this cross-sectional study of a very brief measure of anger, each item contributed to the overall strength of the measure without losing psychometric strength compared with the DAR-5. The consistency of these findings across military and veteran samples in Australian and US populations demonstrated the psychometric robustness of the DAR-3.
Read full abstract