Problem: The increasing impact of mobile phones on adolescents has raised concerns among educators and parents regarding their potentially negative effects on learning. Schools and parents have responded with restrictive policies and screen time limits, but these approaches fail to recognize the diverse nature of phone usage, which includes both harmful and productive activities. To address this, it is essential to establish a scientific basis for distinguishing harmful from productive phone use in adolescents, allowing us to mitigate phone-related distractions while harnessing their potential to enhance learning. Research suggests that the harm associated with phone use is linked to habitual behaviors, where individuals repeatedly engage with their phones without conscious intent. The proposed project aims to investigate whether habitual and non-habitual phone users can be differentiated based on their patterns of mobile-touch interactions immediately after unlocking their phones, referred to as 'digital phenotypes'.
Research Questions: (1) Can we use passive smartphone data (touch patterns and Apps used) to distinguish between habitual and non habitual phone use in adolescents? (2) Is there a relationship between habitual phone use and mental health?
Keywords: Mobile Touch Interactions, Habits, Digital Phenotypes.
iTACO: We developed a real-time mobile Android app called 'iTACO': Interactive Touch Application for Conductual Observation, to capture user gestures (e.g., tap, swipe, scroll, and drag), track their interactions with various apps, and record the components within these apps. The collected data from this app is anonymized and securely hosted on an institutional server, protected by username and password access.
Mexico data collection: During our first-year progress, we recruited 30 adolescents aged between 18-22 years (mean = 20.26 years ; sd = 1.01 years; women = 16) and installed iTACO on their phones for a week. As total we have 5,377,032 touch based interactions, with a mean of 35,609 daily touch based interactions per participant (sd = 30664.07).
Our next steps involve starting data analysis and initiating data collection on the Cambridge side.