Background:: The Harvard Trauma Questionnaire (HTQ) is the predominant questionnaire for assessing PTSD in trauma-affected refugees. Although the scale is increasingly used for measuring treatment outcomes, it has never been specifically validated for such use. The current study does so by testing the HTQ with the Rasch model. Methods:: The analysis is based on 641 Arabic and Persian speaking refugees, diagnosed with PTSD and undergoing psychiatric treatment in Denmark. The responses were tested against the assumptions of the Rasch model, including unidimensionality, local independence and the absence of differential item function across subgroups. Results:: Results reveal two subscales that, when accounting for local dependence and differential item functioning, meet criteria for the Rasch model in the included samples: An arousal/intrusion subscale and an avoidance/numbing subscale. Limitations:: The included sample was highly chronic and suffered from major depressive disorder. Results may not be fully representative of less chronic populations, e.g. recently arrived refugees. Conclusions:: We recommend the use of the arousal/intrusion subscale and the avoidance/numbing subscale as an optimized way of summarizing responses to the HTQ, which remains simple to administer while effectively summarizing all the information available in the scores.