Background
The Hopkins Symptom Checklist-25 (HSCL-25) is a commonly used questionnaire to assess anxiety and depression in trauma-affected refugees. Despite numerous studies evaluating criterion validity and reliability, few studies have assessed the construct validity of the scale.

Methods
In the current study, a Rasch analysis was carried out based on responses from 635 Arabic and Persian speaking refugees. Results supported the use of an eight- or nine-item anxiety subscale score.

Results
The anxiety subscale fit the Rasch model with limited modifications, including the exclusion of purely somatic items. Results for the depression subscale revealed a range of issues. While a shortened scale which accounted for gender differences allowed for a fit with the Persian speaking sample, no fit could be obtained for the Arabic speaking sample. We discuss whether more flexible psychometric models may be necessary to derive valid estimates of depression in some cultures.

Limitations
Both constructs studied, anxiety and depression, were comorbid to PTSD in the included sample. Most respondents were chronically affected and often troubled by extensive post-settlement stress. Results may not generalize to other populations, e.g. non-refugees or more recently arrived refugees.

Conclusions
Purely somatic items have no clear place in the HSCL-25 anxiety subscale, while the anxiety items show good potential to fit a Rasch model. The HSCL-25 depression subscale may not be suitable for summating a single total score, reflecting the need for either a simpler scale or a more complex approach to assessing responses.