User-Donated Screenshots Analysis: Feasibility of a New Approach to Collect Objective Social Media App Usage in Adolescents

Objective data on social media use is now urgently needed for understanding its impact on adolescent well-being. Traditional objective social media data collection methods, such as data donation and passive sensing, face challenges including intrusiveness, privacy concerns, and limitations in adolescent—a critical demographic in this research area. In our study, we introduced a novel, less intrusive method using user-donated screenshots within an ecological momentary assessment (EMA) framework. We recruited 374 adolescents from Switzerland, who were instructed to capture and share three daily screenshots detailing their total and app-specific usage across screentime, activations, and notifications. From this, we collected 6,819 screenshots, with 25% of participants failing to submit any screenshots, 14% submitted incorrect or incomplete ones, while 64% provided complete data for more than five days. To process this data, we developed an image-to-text pipeline using Tesseract OCR that achieved a 96% average accuracy rate. This user-donated screenshot method proved to be less burdensome than traditional data donation, capable of capturing detailed app-specific usage across smartphone operating systems, and applicable among adolescents. Nonetheless, success of the user-donated screenshot approach hinges on user compliance. We analyze attrition sources and suggest six strategies to enhance future research, such as incentivizing participation, implementing pre-upload image checks, and improving participant onboarding and education.