Objective: To confirm the construct validity of the Depression Anxiety Stress Scales-21 (DASS-21) by investigating the fit of published factor structures in a sample of adults with moderate to severe traumatic brain injury (posttraumatic amnesia > 24 hours). Participants: Archival data from 504 patient records at the Brain Injury Rehabilitation Unit at Liverpool Hospital, Australia. Participants were aged between 16 and 71 years and were engaged in a specialist rehabilitation program. Main Measure: The DASS-21. Results: Two of the 6 models had adequate fit using structural equation modeling. The data best fit Henry and Crawford's quadripartite model, which comprised a Depression, Anxiety and Stress factor, as well as a General Distress factor. The data also adequately fit Lovibond and Lovibond's original 3-factor model, and the internal consistencies of each factor were very good (alpha = 0.82-0.90). Conclusion: This study confirms the structure and construct validity of the DASS-21 and provides support for its use as a screening tool in traumatic brain injury rehabilitation.