In a direct challenge to screen-time panic, Orben & Przybylski (2019) analyzed data from over 350,000 adolescents across three large datasets and found that digital technology use accounts for only 0.4% of the variation in adolescent wellbeing.
Key findings:
- The negative association between technology use and wellbeing was smaller than the negative association between wellbeing and regularly eating potatoes, wearing glasses, or skipping breakfast
- Using Specification Curve Analysis (testing all 20,000+ possible analytical combinations), the authors showed that previous alarming results were driven by selective analysis choices
- The relationship was not linear: moderate technology use was associated with slightly higher wellbeing
Implications
This study reframed the technology-wellbeing debate by showing that alarmist conclusions often stem from cherry-picking analytical approaches. However, it predates the LLM revolution — AI chatbots represent a qualitatively different form of digital interaction than passive screen time.
Source
Orben, A., & Przybylski, A. K. (2019). Nature Human Behaviour, 3(2), 173-182.