Digital Traces of Everyday Smartphone Usage Predict Fluid Intelligence

Scalable objective judgements of individuals’ everyday cognitive functioning may promote favorable outcomes throughout the human lifespan in educational, occupational, and healthcare contexts.Taking a theory-informed machine learning approach, we examine how digital traces of everyday smartphone usage predict fluid intelligence, the most central component of human intellect. Using a German quota sample of 367 volunteers providing a total of 1,269,491 phone usages, cross-validated results demonstrate that everyday patterns in smartphone usage consistently predict fluid intelligence (r_Md = .43, r_IQR[.35, .51), and that predictions relate to socio-economic and behavioral outcomes coherent to psychometric tests. High-dimensional combinations of behavioral features (particularly short-term usages related to basic task completion and dealing with complexity) were predictive for individuals’ fluid intelligence, highlighting the pervasiveness and idiosyncratic diversity of manifestations. The findings suggest that smartphone-based cognitive inference could complement traditional assessments, enhancing early identification of cognitive fit and personalized support.