Physical activity and screen time of children and adolescents before and during the COVID-19 lockdown in Germany: a natural experiment

Physical activity (PA) and low levels of sedentary screen time (ST) are linked to better physical, psychosocial, and mental health in youth. WHO recommends at least 60 minutes of moderate-to-vigorous PA daily for children and adolescents, alongside low recreational screen time. Prior to COVID-19, many countries supported youth PA through organized (e.g., school sports, clubs) and nonorganized opportunities. In Germany, the first COVID-19 lockdown (March–May 2020) closed schools, kindergartens, sports clubs, gyms, and leisure institutions, shutting down organized sports while still allowing nonorganized outdoor activities (e.g., jogging, walking) with distancing. Given evidence that structured days (e.g., school days) foster healthier movement behaviors than unstructured days, there were concerns that lockdowns would exacerbate inactivity and sedentariness. The study investigates, in a natural experiment, how organized sports activity (SA), habitual physical activity (HA), and recreational screen time changed among German children and adolescents during the strictest lockdown, using longitudinal data from the MoMo cohort with measurements immediately before and during the lockdown.

Emerging international studies during COVID-19 generally reported decreased PA and increased screen time among youth, though methodologies and representativeness varied. Shanghai data (n=2426, ages 6–17) showed weekly PA reduced from 540 to 105 minutes and ST increased from 170 to 450 minutes; guideline adherence for PA dropped from 60% to 17.7%. Croatian adolescents (~16 years) showed PA declines, greater in urban settings. Canadian children with congenital heart disease reduced steps by 21–24%. Italian children with obesity reduced PA by ~2.5 h/week. In Canada, during COVID-19, 18% adhered to PA guidelines and 11.2% to ST recommendations, compared to pre-pandemic report card ranges of 35% and 8–64% respectively. Differences across countries likely reflect heterogeneous policy measures (e.g., some disallowed outdoor exercise, as in parts of China, while others, such as Belgium, had milder restrictions). Prior work indicates non-structured days promote less healthy movement behaviors than structured days. Overall, the literature suggests risk of increased inactivity and sedentariness during lockdowns but comparability is limited.

Design: Natural experiment using longitudinal, self-reported measures from the German Motorik-Modul (MoMo) cohort study, Wave 3. Two measurement points: immediately before lockdown (pre) and during the strictest lockdown period (peri). Setting and timeline: Germany imposed closures of schools, kindergartens, sports clubs, gyms, and other leisure institutions between March 16–18, 2020, with contact restrictions; nonorganized outdoor activities were allowed with distancing. Peri data were collected 04/20/2020–05/01/2020; pre data were collected onsite as part of Wave 3 up to March 2020. Participants: Nationwide, representative longitudinal sample of children and adolescents aged 4–17 years. Of 2722 Wave 3 participants invited for the peri survey, 1711 provided complete longitudinal data (boys 50.2%, girls 49.8%). Age (mean±SD): pre 10.36±4.04 years; peri 11.34±4.06 years. Sampling used a multi-stage, stratified approach (167 primary sampling units across Germany; age-stratified random sampling from resident registers). Ethics: Conducted per Declaration of Helsinki with approvals from Charité Universitätsmedizin Berlin, University of Konstanz, and Karlsruhe Institute of Technology. Written informed consent obtained; data protection approved by the Federal Commissioner for Data Protection and Freedom of Information. Measures: - Physical Activity: MoMo Physical Activity Questionnaire (MoMo-PAQ, 28 items; test-retest ICC≥0.68) assessed organized sports activity (school curricular/extracurricular classes; sports club activities) and nonorganized leisure-time sports, as well as habitual physical activities (playing outside, gardening, housework) and active travel (walking, cycling). Organized sports minutes were adjusted for seasonality/vacations; non-energy-expending activities (e.g., esports) were excluded. WHO guideline indicator: days per week with ≥60 minutes of moderate-to-vigorous PA. - Recreational Screen Time: Self-reported time for television watching, gaming (any device), and recreational internet use for weekdays/weekends via 7-point scale (almost never up to 4 h/day); aggregated to daily averages (5:2 weekday:weekend ratio). Guideline adherence: ≤120 minutes/day of recreational ST. Data collection mode: Pre via onsite laptops; peri via online survey (organized sports questions omitted peri since activities were shut down). Statistical analysis: Repeated-measures ANOVAs tested lockdown (time) effects, and interactions with age group (4–5, 6–10, 11–13, 14–17 years) and sex. Reported means, SDs, differences (peri–pre) with 95% CIs, partial eta-squared, and Cohen's d. Analyses performed in IBM SPSS 25. Seasonal sensitivity analysis restricted to participants measured in April 2019 was also conducted.

Overall PA and guideline adherence: - Active days (≥60 min/day) increased by 0.44 days/week (F1,1686=72.9, p<0.01, ηp2=0.041); larger increases in younger children (4–5 years +0.76 days) vs. adolescents (14–17 years +0.26 days). - WHO PA guideline adherence increased by 11.1% overall (pre 19.1% to peri 30.2%), with larger gains in younger children (4–5 years +14.7%) than older adolescents (14–17 years +4.8%). - Recreational ST guideline adherence (≤120 min/day) decreased overall by 23.3% (pre 60.9% to peri 37.6%). Sports activity (SA): - Organized sports dropped to 0 during lockdown by policy. - Nonorganized sports increased by 17.7 min/day (F1,1601=626.3, p<0.01, ηp2=0.281), with age interaction: +17.0 min/day (14–17 years) vs. +11.8 min/day (4–5 years). - Total sports (organized + nonorganized) decreased by 10.8 min/day (F1,1644=122.3, p<0.01, ηp2=0.069); the decline was greater in adolescents (14–17 years −15.6 min/day) vs. younger children (4–5 years −2.2 min/day). Habitual physical activity (HA): - Total HA increased by 36.2 min/day (F1,1673=181.4, p<0.01, ηp2=0.098). - Components: playing outside +21.4 min/day (F1,1665=12.6), gardening +6.7 min/day (F1,1638=80.7), housework +4.0 min/day (F1,1645=62.2); walking/cycling showed no main effect of lockdown. Several age interactions indicated larger increases in younger age groups for playing outside and differential changes by age for gardening and housework. Recreational screen time (ST): - Total recreational ST increased by 61.2 min/day (F1,1676=494.7, p<0.01, ηp2=0.228). - Components: TV +21.2 min/day (F1,1667=248.8), gaming +21.5 min/day (F1,1654=209.5), recreational internet +18.5 min/day (F1,1646=149.9). - Age interactions for internet, gaming, and total ST; sex interaction evident for gaming (boys increased more). Increases were present across all age groups and both sexes. Net effect on total daily PA: - Despite the decline in organized sports, increases in nonorganized sports and especially HA resulted in higher total daily PA (from 142.2 to 167.8 min/day reported in discussion), with consistent direction across sexes and age groups (magnitude varied).

Contrary to many international reports of overall PA declines, German children and adolescents increased total PA and WHO guideline adherence during the strictest lockdown. This was driven by a substantial rise in habitual activities (playing outside, gardening, housework) and an increase in nonorganized sports that partially compensated for the shutdown of organized sports. Adolescents experienced a larger loss in total sports and smaller gains in HA than younger children, indicating developmental differences in how opportunities translate into activity. Recreational screen time increased markedly across all subgroups, and guideline adherence for ST decreased, yet PA and ST rose concurrently, supporting evidence that they are not functional opposites and share only a weak negative association. Cross-country differences likely reflect contextual factors, notably policy restrictions on outdoor activity; Germany allowed nonorganized outdoor activities, which appears crucial for maintaining PA. Inequality analyses suggested more equal distributions in sports activity and screen time (lower Gini coefficients) during lockdown, possibly due to removal of entry barriers to organized sports and heightened health focus. Seasonal/weather effects and socioeconomic factors were examined and did not alter the main conclusions. Overall, the findings underscore the context-driven nature of youth PA and the importance of enabling autonomous, accessible, nonorganized opportunities for movement during disruptions.

In this natural experiment using a representative German cohort, the first COVID-19 lockdown led to reduced organized sports, increased nonorganized sports and habitual activities, and net higher total physical activity alongside increased recreational screen time. PA increases were more pronounced in younger children, while adolescents showed larger declines in total sports. PA and ST changed in the same direction, reinforcing that they are distinct behaviors requiring separate intervention strategies. Policy implications include ensuring access to safe outdoor spaces and nonorganized PA opportunities during lockdowns, and leveraging digital supports (e.g., online club offerings, emerging virtual/augmented reality solutions) to reach adolescents and lower entry barriers. Future research should incorporate device-based measures where feasible, explore causality and mediators, assess long-term behavioral trajectories post-lockdown, and evaluate interventions that simultaneously address PA promotion and ST reduction in diverse contexts.

All outcomes were self-reported, introducing possible recall and reporting biases; device-based validation was not implemented peri-lockdown due to ethical/logistical constraints. As a natural experiment, no control group was available, limiting causal inference. The peri measurement occurred about one year after the pre measurement on average, potentially biasing estimates given age-related declines in some PA domains; this would likely underestimate the observed PA increases. Weather during April 2020 was unusually warm and sunny in Germany, though sensitivity analyses restricting to April-tested participants showed stable effects. Representativeness of the longitudinal sample was slightly limited due to the pandemic interrupting pre-sample collection, but main effects were robust across socioeconomic and environmental strata.