Towards Cross-Cultural Research on Personal Light Exposure: Evaluating Translation Fidelity and Language Effects

Abstract Background Light exposure is a key regulator of human circadian physiology and health, yet real-world evidence remains limited, particularly across diverse cultural and geographic contexts. Large-scale, cross-cultural studies require harmonised and linguistically valid tools to assess personal light exposure and its behavioural correlates. Methods Within the MeLiDos multicentre cohort study, we translated and culturally adapted a comprehensive set of English sleep-, circadian-, and light-related questionnaires into five languages (Dutch, German, Spanish, Swedish, and Turkish) using the TRAPD framework, followed by cognitive pretesting. Quantitative analyses were conducted on complete data from 147 participants across five countries (Costa Rica, Ghana, Germany, Spain, and Türkiye) to evaluate potential language effects using one-way ANOVAs, linear mixed-effects models, and principal component analysis (PCA). In addition, language-specific Pearson correlation matrices were computed for momentary mood and alertness data (N = 4,025). Results Questionnaires assessing sleep, light exposure behaviour, light sensitivity, momentary affect and alertness, as well as real-world data collection logs related to device wear and participant experience, were successfully translated and harmonized. The translation and pretesting process identified challenges related to sensitive topics and culturally unfamiliar concepts such as chronotype and light sensitivity , which were addressed through clear, standardized explanations. It was identified that instruments require adaptation to accommodate local norms, technological accessibility, and seasonal interpretation to enhance sociocultural applicability. Quantitatively, robust country effects were observed across multiple scales, whereas language effects were small and inconsistent. Mixed-effects modelling revealed no significant fixed effect of language, with most variance attributable to individual- and country-level factors. PCA showed that multivariate response patterns clustered by country rather than language. Correlation structures of momentary affect and alertness were highly preserved across languages. Conclusions Harmonized multilingual questionnaires can be reliably used in cross-cultural light-exposure research without introducing substantial language-related bias. Observed differences are likely driven by geographic and cultural factors rather than language, supporting the global applicability of subjective light-exposure assessments while underscoring the necessity of culturally diverse, multicentre data collection. This work provides a strong methodological foundation for advancing global research on the interplay between culture, light exposure, and circadian health, with implications for understanding environmental influences on health.