To express intensity of physical activity, energy expenditure is often divided by either body weight, resting metabolic rate, or fat-free mass. These calculations are used widely as the physical activity index. However, it is unclear how body size influences the valid estimation of intensity of various kinds of activities. In the present study, we investigated whether these indices are able to adjust for body size when calculating energy expenditure in various kinds of activities. In addition, we examined to what extent the error of index is introduced by differences in body size. Resting metabolic rates and energy expenditure during sitting light work, 4 lifestyle and 7 ambulant activities were measured in the postabsorptive state using indirect calorimetry in 71 healthy Japanese adults. We regarded an index as an inappropriate adjustment for body size when there was a significant correlation between it and body weight. Energy expenditure normalized by body weight correlated with body weight in all sedentary states; when normalized by lying resting metabolic rate it correlated with body weight in 3 ambulant activities; when normalized by sitting resting metabolic rate it correlated with body weight in 2 lifestyle and 5 ambulant activities; and when normalized by fat-free mass it correlated with only 1 ambulant activity. The indices caused errors in estimates of activity intensity of less than +/-10% when body weight was more than 10 kg above average. In conclusion, the body weight-normalized index was inappropriate for sedentary activities and the other three indices were inappropriate for ambulant activities. However, the use of any of these indices introduces an error in the estimate of total energy expenditure of considerably less than +/-10% for body weights within the normal range.