Objective: The aim of this study was to examine the abbreviated Desires for Alcohol Questionnaire (DAQ) with respect to component structure and concurrent validity. Method: The DAQ was administered to 2,960 adults participating in the Collaborative Studies on the Genetics of Alcohol. Rotated principal components analysis was conducted on 1,500 subjects with an alcohol-use disorder (AUD) and on 1,460 non-AUD subjects. Total DAQ scores were compared for these two subsamples. In addition, correlations were computed between DAQ scores and the following: (1) a sum of alcohol symptoms, and (2) endorsement of a single interview craving question. Results: Similar solutions emerged in the AUD and non-AUD subsamples, with dimensions characterized by (1) strong desires/intentions to drink, (2) negative reinforcement, and (3) positive reinforcement + ability to control drinking. Each component was significantly correlated with the alcohol symptom scale in both subsamples (rs = .25-.64 and .31-.40, respectively, p < .0001) and with the interview craving item in the AUD subsample (rs = .22-.55, p < .0001). Total DAQ score was significantly higher for AUD subjects (40.5) than for non-AUD subjects (23.1, p < .0001) and exhibited significant correlations with the alcohol symptom scale in the AUD and non-AUD subsamples (rs = .61 and .39, respectively, p < .0001) and with the interview craving item in the AUD subsample (rs = .51, p < .0001). Conclusions: The DAQ is an appropriate measure of alcohol craving, as demonstrated by similar component structures across two samples as well as its concurrent validity.