Smartphone Use and Mental Health: Accounting for idiosyncrasy with person-specific models

Contrary to concerns about media use harming mental health, research on this relationship is ambiguous - jeopardizing theory, interventions and policy. Small or null results are often blamed on cross-sectional studies that have imprecise measures of media use and mental health symptoms. To address these critiques, we captured moment-by-moment recordings of smartphone use for one entire year, accompanied by fortnightly clinical surveys about symptoms of depression, anxiety, and adult ADHD, and about psychological well-being. Screens were recorded from smartphones every five seconds they were in use, resulting in approximately one million screens collected per person. Using canonical correlation analysis of media and mental health metrics, we found strong relationships for some participants (r from .78 to .92) between variations in smartphone use and variations in mental health measures over one year. The person-specific models captured heterogeneity that might better guide theory development, and precision screening, diagnosis and interventions.