This study sought to assess criterion validity of the Actical monitor step-count function in children via ankle and waist placement, compared with observed video recordings. Children attending a summer program (12 boys, 7 girls, mean age = 9.6yrs, range 7-11yrs) wore two synchronized Acticals, attached at the ankle (AA) and waist (AW). Children performed treadmill walking at varying speeds, and two research assistants counted steps using observed video recordings (OVR). Results showed high correlations for AW-OVR (r = .927, p < .001) and AA-OVR (r = .854, p < .001), but AW and AA were significantly lower than OVR (t > 11.2, p < .001). AW provided better step estimates than AA for step rates above 130 steps per minute. In contrast, AA was superior to AW for slow walking, and measured more steps during the (nontreadmill) program time. Overall, the Actical monitor showed good evidence of validity as a measure of steps in children for population-based studies.