Lastly, the above analyses involve numerous statistical tests, and consequent risk of committing a familywise error; as such it is recommended that results falling below highly significant ( p < 0.001) are viewed with caution.
Perhaps surprisingly, although there was a mild trend of lower levels of physical activity with rising weight status, more evident in the larger 2015 sample, these differences were comparatively mild compared to what might be expected if weight status and physical activity levels were strongly associated.