StudentSense: Primary outcomes of a smartphone digital phenotyping study of university students in Australia

Abstract

University students face increasing mental health challenges, yet traditional assessment methods often lack timeliness and contextual relevance. This study presents StudentSense, a digital phenotyping investigation examining associations between smartphone-sensed behavior and psychological wellbeing over a 17-week academic semester. A total of 151 students from the University of Melbourne participated, providing continuous passive smartphone data (e.g., screen activity, app usage, location, keyboard, and on-screen text) and self-reported assessment data, including the Depression Anxiety Stress Scales (DASS-21), Mental Health Continuum-Short Form (MHC-SF), College Student Subjective Wellbeing Questionnaire (CSSWQ), and UCLA Loneliness Scale (ULS-8). We used Monte Carlo subsampling correlation analysis to explore relationships between behavioral and psychometric variables. Findings showed that levels of depression, anxiety, stress, and loneliness declined following the examination periods. Screen usage was positively correlated with the reported level of a loneliness item, while time spent at significant locations (e.g., university, home) was negatively correlated with depression. Students reporting higher wellbeing and extroversion reported more positive university experiences.