Children's physical activity (PA) patterns change from day to day. This intra-individual variability affects precision when measuring key physical activity and sedentary behaviour variables. This paper discusses strategies used to reduce the random error associated with intra-individual variability and demonstrates the implications for assessing PA when varying number of days are sampled. Self-reported data collected on two hundred and ninety eight 13-14-year-olds were used to compare estimates of PA and sedentary behaviour derived from between 1 and 7 days of recall data. Large intra-individual coefficients of variation were calculated for physical activity level (14.5%), moderate-to-vigorous physical activity (83.4%), screen time (60.8%) and sleep (12.2%). While the magnitude of error associated with estimating means decreased as more days were sampled, the paper notes that depending on the nature of the research question being asked, sampling fewer days may yield sufficiently precise estimates. Therefore, researchers should conduct power analyses based on estimated inter- and intra-individual variability and sample size to determine how many days to sample when assessing children's PA patterns.