Reliability and validity of a Norwegian version of the awareness of social inference test.

Abstract

To establish reliability and construct validity of a Norwegian version of the social cognition test, the Awareness of Social Inference Test (N-TASIT). Participants with traumatic brain injury (TBI; n = 101) and 50 matched healthy controls performed either a virtual reality (VR) or 2D version of N-TASIT at baseline and 16 weeks later. Reliability measures were test-retest reliability and internal consistency. Intraclass correlation (ICC) and Cronbach's alpha (α) were calculated for the overall sample and both groups separately. Construct validity was tested with known groups validity, convergent and discriminant validity. Known groups analysis was conducted for both versions separately and combined. Convergent and discriminant validity were determined by associations between N-TASIT and established measures of social cognition and with measures of cognition, emotional distress and fatigue. ICC for the total sample was 0.63 (95% CI 0.49-0.73), and α was 0.88, when presentation modes (VR and 2D) were combined. The healthy control group outperformed the TBI group in both presentation modes. Medium-to-large associations were found between N-TASIT performance and social cognitive measures, and mostly weak or no significant correlations with non-social domains. N-TASIT appears psychometrically sound and comparable to the original. The influence of presentation mode on performance remains unclear.