Thirty (10 male, 20 female) children, ages 41 to 71 months, were observed simultaneously in a freeplay setting by two observers using a validated assessment of physical activity in children (i.e. the FATS). One observer utilized a whole interval time sampling method of data recording and the second observer used a hand-held computer and continuously entered observations whenever a child's activity level changed, while the computer simultaneously calculated elapsed time (in tenths of seconds) spent in each behavioral category. Overall activity scores generated by the two approaches correlated 0.90 although the time sampling method produced lower estimates of physical activity. It is concluded that whole interval time sampling is a reliable cost-effective method of assessing physical activity levels in children and may be preferable to the continuous computer-assisted approach that was utilized.