"[Longitudinal effect of early social media use on standardized learning outcomes during school career](https://www.nature.com/articles/s41562-026-02522-4)" Gui, M., Respi, C., Abbiati, G. _et al._ _Nat Hum Behav_ (2026). https://doi.org/10.1038/s41562-026-02522-4 *Abstract:* This study uses difference-in-differences models to evaluate the impact of early social media use on academic performance among children aged 7–16 years. A dataset comprising 5,227 students from northern Italy was constructed by merging data from a questionnaire on lifetime digital media habits with longitudinal standardized test scores in Italian, English and mathematics. Participants who created a personal social media account in 6th, 7th or 8th grade were separately matched with late adopters who initiated their first social media account in the 9th grade or later. The findings provide compelling evidence that early engagement with social media adversely affects learning outcomes. Additional findings on the effects of early social media use on other academic outcomes for which we do not have longitudinal data are in line with our main findings. We also find supporting evidence that a substantial part of this negative effect is mediated by students’ smartphone pervasiveness in key moments of the day. **Figure 1** ![[2026_Gui_SocialMedia_Fig1.png]] Notes: _N_ (students) = 5,227. The centre line indicates the median; box bounds indicate the 25th and 75th percentiles; whiskers extend to the lower and upper adjacent values, defined as the most extreme observations within 1.5 times the interquartile range from the lower and upper quartiles, respectively. The horizontal red dashed line represents the timing of the INVALSI competence tests, which take place at the end of grades 2, 5, 8 and 10. The original questions were worded differently according to each device or account type. Students were asked to indicate the grade in which they first had (i) their own internet-connected smartphone, (ii) access to video game consoles, (iii) independent access to tablet and computer, and (iv) a personal profile for messaging apps and social media accounts. **Figure 2** ![[2026_Gui_SocialMedia_Fig2.png]] Note: results are presented as regression coefficient estimates, with error bars indicating 95% confidence intervals from two-sided tests. The coefficients display the interaction parameters between treatment status and grade from student fixed-effects models.