How much screen time is too much for kids?

The guidance. WHO (2019) says screen time is not recommended for infants or 1-year-olds, and should be no more than an hour for ages 2 to 4, less being better; reading and storytelling with a caregiver are encouraged¹. The American Academy of Pediatrics' 2026 policy statement says children's media use cannot be judged through screen limits alone. Digital products built around children's development can support learning, while those built for engagement and commercial gain encourage long sessions that displace sleep and movement¹⁰.

Young children. In Calgary, among 2,441 mother-child pairs, more screen time at 24 and 36 months predicted lower developmental screening scores at 36 and 60 months, with small effects⁷. In Japan, among 7,097 pairs, 1-year-olds with 4 or more hours a day had about 4.8 times the odds of communication delay at age 2 compared with those under an hour¹¹. A meta-analysis of 42 studies tied more screen time to weaker language (r = -0.14), and co-viewing and educational programs to stronger language³. A 2024 review of 100 studies added that background television, age-inappropriate content and caregiver phone use during routines went with poorer outcomes⁴.

Not all screens are equal. Toddlers learned new verbs from live and live video-chat interaction, but not from non-interactive video².

Sleep. In children aged 6 to 19, bedtime device use went with inadequate sleep (odds ratio 2.17), and so did having a device in the bedroom without using it⁵. Of 67 studies, 90% found screen time linked to shorter or later sleep⁶.

Attention and mood. Media use showed a small link with ADHD-related behaviors (r = 0.12)¹². For teen well-being, one analysis found at most 0.4% of the variation explained⁹, while a 2026 meta-analysis found consistent associations with risk, especially for social media⁸.

Practical rules. Use screens together, favor sleep, play and reading, keep devices out of bedrooms overnight, and turn them off before bed. See blue light and evening screens and digital detox.

Early childhood development

Madigan's cohort used a three-wave cross-lagged model in 2,441 Calgary mother-child pairs, with screen hours and Ages and Stages Questionnaire scores reported at 24, 36 and 60 months. Higher screen time at 24 and 36 months predicted poorer screening performance at 36 months (beta -0.06, 95% CI -0.10 to -0.01) and 60 months (beta -0.08, 95% CI -0.13 to -0.02), after controlling for stable differences between children⁷. Takahashi's Japanese cohort (7,097 pairs) grouped 1-year-olds as under 1, 1 to under 2, 2 to under 4, or 4 or more hours a day. Compared with under 1 hour, communication delay at age 2 had odds of 1.61, 2.04 and 4.78 (95% CI 3.24 to 7.06) across the higher groups, and the 4-hour group still showed more communication delay at age 4 (OR 2.68, 95% CI 1.68 to 4.27)¹¹. Both rely on parent-reported development and cannot rule out confounding by family circumstances.

Language and context

Madigan's meta-analysis (42 studies, 18,905 children) found screen time correlated with lower language (r = -0.14, 95% CI -0.18 to -0.10) and background television even more (r = -0.19), while educational programs (r = 0.13) and co-viewing (r = 0.16, 95% CI 0.07 to 0.24) went with stronger language, as did later onset (r = 0.17)³. Mallawaarachchi pooled 64 observational studies from 100 (176,742 participants). Program viewing was linked to poorer cognitive outcomes (r = -0.16), background television too (r = -0.10), age-inappropriate content and caregiver screen use during routines to poorer psychosocial outcomes (both r = -0.11), and co-use to better cognitive outcomes (r = 0.14). The authors called effects small to moderate and urged attention to context beyond time limits⁴. In Roseberry's experiment, 36 toddlers aged 24 to 30 months learned verbs in live and video-chat conditions but not from yoked video². It was a small, short laboratory study.

Sleep

Carter's meta-analysis of 20 cross-sectional studies (125,198 children aged 6 to 19) found bedtime device use linked to inadequate sleep (OR 2.17, 95% CI 1.42 to 3.32), poor sleep quality (OR 1.46) and daytime sleepiness (OR 2.72); mere access without use also predicted inadequate sleep (OR 1.79) and sleepiness (OR 2.27). Heterogeneity was high⁵. Hale and Guan's review of 67 studies found adverse sleep associations, mostly shorter duration and later timing, in 90%, but noted unconfirmed causation, measurement error and little data on content⁶.

Attention and well-being: conflicting results

Nikkelen's meta-analysis of 45 studies found a small relationship between media use and ADHD-related behaviors (r = 0.12)¹². Orben and Przybylski analyzed three datasets (n = 355,358) with specification-curve analysis; digital technology use was negatively linked to adolescent well-being but explained at most 0.4% of its variation, which the authors judged too small to warrant policy change⁹. Teague's 2026 meta-analysis of 153 longitudinal studies (ages 2 to 19, mean 12.8) found social media associated with more depression, internalizing and externalizing behavior (r = 0.09 to 0.21) and lower academic achievement, video gaming with aggression (r = 0.16), and concluded digital media use was consistently associated with risks, particularly social media⁸. Both small effects and consistent direction can be true at once, and none of these designs prove cause.

Guidance

WHO sets numeric limits for under-5s¹. The 2026 AAP statement moves toward family, platform and policy recommendations rather than hour counts alone¹⁰.

Bottom line (confidence 2 of 5)

Associations are consistent but small, mostly observational and open to reverse causation or confounding. The sturdier practical points are no screens at bedtime, live interaction and co-viewing for young children, and limits for under-5s. Beyond that, no single number is proven.