*4.3. Limitations*

There are limitations regarding sample and selection bias since we investigated an ad hoc sample with a cross-sectional cohort design of children from one primary school in a rural community of the German federal state Baden-Württemberg. The sample is not representative and its results show a selection effect concerning a higher physical activity and fitness compared with the whole of Germany. However, the long study period and the number of cohort measurements before and during the COVID-19 pandemic ensure that children's physical fitness is considered globally and not just a one-point statement based on a one-point measurement.

Moreover, in this study were some confounding factors, e.g., socio-economic status, educational level of parents, and level of testosterone, which may affect the results, but were not controlled. However, we can use some physical activity data to classify. For future investigations, the methodology should be improved and possible cofounding data collected.
