The aim of this study of 56 children aged 6-9 years was to identify measures of physical activity that could be used in either clinical or population studies. Comparisons were made between four measures of physical activity: a three day parent-reported activity diary, a parent-reported physical activity questionnaire, the Tritrac-R3D™ accelerometer (worn three days) and physical activity energy expenditure calculated over 10 days by the doubly labeled water (DLW) technique. The strongest correlation between methods was for the diary and Tritrac-R3D™ during the two hour after-school period (1530-1730 hours) (r = 0.75, P < 0.0001). Activity level in this after-school period was positively correlated with average activity level over three days for both Tritrac-R3D™ (r = 0.53, P < 0.01) and diary (r = 0.54, P < 0.0001). No associations were found between measures of activity from DLW and activity measures from the Tritrac-R3D™, diary or questionnaire. These results suggest that the two hour after-school period is of high interest for future population studies of physical activity in school-age children.