The World Health Organization five-item Well-Being Index has been widely used to assess adolescents' well-being. However, its structural properties and measurement invariance at both between- and within-country levels have been underexamined in this population. The present study aimed to fill this gap. Relying on a nationally representative sample of 9,007 school attendees in Luxembourg (MAGE = 14.7; SDAGE = 2.4; 48.7% female) and on both classical test theory and item response theory, I assessed the scalability, reliability, monotonicity, factorial structure, convergent validity, and sociodemographic measurement invariance of the WHO-5. Measurement invariance was investigated across groups differing in terms of academic performance, country of birth, language, sex, socioeconomic status, and type of school. The WHO-5 exhibited strong scalability (Loevinger's H = .550) as well as satisfactory reliability (McDonald's ω = .84), monotonicity, factorial validity, and convergent validity. Measurement invariance held across all the examined subgroups. This study suggests that the WHO-5 is a sound, unidimensional measure of well-being that can be used for comparisons across various sociodemographic adolescent groups.
Romain Brisson (Tue,) studied this question.