Is screen time at age 11 associated with depressive symptoms in adolescence. A monozygotic twin differences study.

Screen-based media use in childhood and early adolescence has been widely linked to poorer mental health outcomes, but the current literature remains contested about the size of these associations and if there is a causal link. Large population studies have reported associations between higher daily screen time and lower psychological well-being, but these findings are often cross-sectional and therefore cannot account for directionality or confounding (Twenge & Campbell, 2018). Evidence suggests screen time can predict later depressive symptoms, with meta-analytic work revealing elevated risks in cohort studies. This strengthens the evidence for temporal ordering, but these estimates may still reflect unmeasured family-level and genetic confounding (Li et al., 2022). When screen-time data are analysed in many different ways, the average association with wellbeing is very small and often depends on modelling choice (Orben & Przybylski, 2019). This reveals the fragility of correlational designs and helps explain why designs like the monozygotic (MZ) twin differences design are useful. Theoretical accounts also make competing predictions about why screen time might relate to wellbeing. For example, displacement-based explanations propose that screen use may replace physical activity and face-to-face interaction, whereas social compensation frameworks suggest that psychologically vulnerable young people may use online interaction to cope or seek connection. This implies potential selection effects and bidirectionality (Valkenburg & Peter, 2011). One key limitation across much of this literature is that typical observational analyses cannot separate environmental effects from stable predispositions (e.g genetically influenced traits or family environments that can increase screen time and depression risk). Behavioural genetics highlights that behavioural traits are heritable, that shared family environment often explains less variance than genetic factors, and that substantial variance is attributable to the nonshared environment (Turkheimer, 2000). MZ twin comparisons provide a quasi-causal approach by controlling for genetic factors and shared environment: if the twin who used more screen media later shows higher depressive symptoms than their co-twin, this pattern is more consistent with an environmentally mediated effect than with genetic/shared-family confounding. This study will test whether total screen time at age 11 predicts depressive symptoms at age 15 using data from the German TwinLife longitudinal twin cohort, focusing on Cohort 2 and restricting analyses to MZ twin pairs. We plan to extend beyond population-level cross-sectional associations by applying an MZ within-pair framework to see if there is an association between screen time and depressive symptoms, even when we control for genetic factors and shared environment, helping clarify whether observed links reflect causal influence from the environment or confounding.