The structure of posttraumatic stress is of both theoretical and clinical interest. In the present study, seven models of posttraumatic stress were compared using confirmatory factor analysis. A sample of 528 Western New York undergraduate students was assessed 1 and 3 months after the September 11th, 2001 terrorist attacks. At the Month 1 assessment, the current three-factor Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; American Psychiatric Association, 1994) model, which consists of Intrusions, Avoidance/Numbing, and Hyperarousal, did not provide a good fit to the data; however, a four-factor model consisting of factors labeled Intrusions, Avoidance, Dysphoria, and Hyperarousal did fit the data well and provided better fit than the three-factor model and other competing models. Importantly, Dysphoria spans symptoms from the traditional DSM Numbing and Hyperarousal clusters. The four-factor model continued to fit the data well at Month 3. These findings parallel the results of earlier studies which suggest that a four-factor model better reflects the nature of posttraumatic stress than do simpler models, including the DSM. The present work is consistent with a dimensional model of stress responses and calls for further longitudinal work in this area.