To assess the association between the amount of recess provision and children's accelerometer-measured physical activity (PA) levels. Parents/guardians of 6- to 11-year-olds (n = 451) in the 2012 National Youth Fitness Survey reported recess provision, categorized as low (10-15min; 31.9%), medium (16-30min; 48.0%), or high (>30min; 20.1%). Children wore a wrist-worn accelerometer for 7days to estimate time spent sedentary, in light PA, and in moderate to vigorous PA using 2 different cut points for either activity counts or raw acceleration. Outcomes were compared between levels of recess provision while adjusting for covariates and the survey's multistage, probability sampling design. Children with high recess provision spent less time sedentary, irrespective of type of day (week vs weekend) and engaged in more light or moderate to vigorous PA on weekdays than those with low recess provision. The magnitude and statistical significance of effects differed based on the cut points used to classify PA (eg,4.7 vs 11.9 additional min·d-1 of moderate to vigorous PA). Providing children with >30minutes of daily recess, which exceeds current recommendations of ≥20minutes, is associated with more favorable PA levels and not just on school days. Identifying the optimal method for analyzing wrist-worn accelerometer data could clarify the magnitude of this effect.