Phones have become part of young people's daily lives in ways that could affect their sleep and mood. Yet most of the evidence rests on surveys asking young people to recall how much time they spent on their devices. A 2021 meta-analysis found that self-reported media use corresponds only moderately with logged use.

A new study in the Journal of Research on Adolescence by TIP Center postdoctoral scholar Anthony Chen, with Monika Lind and Candice Odgers of the University of California, Irvine and Nilam Ram of Stanford, addresses this problem by logging adolescents' phone use directly. Using these logged data, the study finds that the causal arrow may run backward.

Negative affect predicts phone use (not the other way around)

Rather than relying on recall, the researchers installed a research keyboard on adolescents' own phones through the Effortless Assessment Research System (EARS). Over 90 days, the app timestamped keystrokes and logged time spent in each application. Keyboard activity marked active use (messaging, posting, commenting). Time in an app without typing marked passive use (scrolling, watching, browsing). Each morning, participants reported their mood and sleep quality.

The design yielded roughly 3,200 participant-days from 40 adolescents aged 15 to 20. The number of repeated observations allowed each adolescent to serve as their own control. Instead of comparing a light user to a heavy one, two young people who likely differ in many other ways, the analysis asked whether a given teenager behaved differently on their own worse-than-usual days.

What do they find? On mornings when adolescents reported more negative affect than their own average, they spent more time in passive use that day, about eight minutes for each point on a five-point scale. Morning positive affect predicted nothing. The reverse pathway did not appear. Neither active nor passive use predicted self-reported sleep quality, passively sensor-measured sleep duration, or next-morning mood. 

Takeaway: Phone use as symptom rather than source?

The pattern fits a mood-management account: passive use takes little effort and is always at hand, which makes it an easy way to cope with a bad morning. On that reading, scrolling may be better understood as a marker of how an adolescent already feels than as the thing making them feel that way.

The study is limited by a small sample of Android users spanning late adolescence and emerging adulthood. Still, its findings are consistent with a 2025 study of 250,000 days of logged use among 10,099 adults, which also found weak or null links between phone use and mood.

Taken together, the work suggests that the stronger predictors of adolescent wellbeing may lie elsewhere (e.g., family dynamics, peer relationships, etc.), and that treating screen time as the main lever may direct attention toward the explanation with the least support.