Abstract
Studying within-person variability in children's behavior is frequently hindered by challenges collecting repeated observations. This study used wearable accelerometers to collect an intensive time series (2.7 million observations) of young children's movement at school (N = 62, M(age) = 4.5 years, 54% male, 74% Non-Hispanic White) in 2021. Machine learning analyses indicated that children's typical forward acceleration was strongly correlated with lower teacher-reported inhibitory control and attention (r = -.69). Using forward movement intensity as a proxy for impulsivity, we partitioned the intensive time series and found that (1) children modulated their behavior across periods of the school day, (2) children's impulsivity increased across the school week, and (3) children with greater impulsivity showed greater variability in behavior across days.