Cross-Cultural Measurement Invariance of the Harvard Trauma Questionnaire Across Nine Adolescent Samples.

Abstract

Trauma researchers often make claims about the severity of posttraumatic stress disorder (PTSD) across populations, and yet cross-cultural measurement invariance (MI) is rarely assessed. Nine youth samples with Harvard Trauma Questionnaire (HTQ) responses were grouped based on sampling strategy used into two sets: representative (Denmark, the Faroe Islands, Iceland, and Lithuania, n = 1,457), and convenience (Greenland, India, Kenya, Malaysia, and Uganda, n = 2,036). Confirmatory factor analysis (CFA) was used to gauge whether configural, metric, scalar, and residual invariance of different models held between national samples within the two sets. Configural invariance held for most PTSD models in convenience samples, not in representative samples. Metric invariance was less common, and scalar and residual in general did not hold. Cultural similarity between samples seemed to be associated with invariance. Findings suggest that although PTSD symptoms may cluster similarly across culturally distal groups, comparisons of the severity of symptoms using the HTQ across adolescent samples are not likely valid.