Variation in sleep duration across latitudes and countries: a Bayesian hierarchical analysis of wearable data.

Seasonal effects on sleep have been documented but remain poorly characterised at a global scale. Using wearable-derived data from 697 individuals contributing 185,143 nights across 49 countries, we examined how sleep duration varies with season, day length, latitude, and country using Bayesian hierarchical models. Across the dataset, sleep duration decreased with increasing day length, with a reduction of roughly 4.4 minutes per additional hour of daylight. In contrast, categorical season indicators contributed little explanatory power once individual- and country-level differences were modelled, suggesting that coarse season labels do not capture consistent structure in globally distributed samples. We found limited evidence that photoperiod sensitivity scales with latitude, despite large differences in seasonal light amplitude. At the same time, country-level variability emerged in both baseline sleep duration and estimated photoperiod responses. These findings indicate that global variation in sleep duration is shaped more strongly by individual and national differences than by calendar season.