Estimation of energy expenditure in Western Samoa, American Samoa, and Honolulu by recall interviews and direct observation.

Reliable and accurate estimates of energy expenditure are a fundamental requirement for research on energy balance, energy flow, and other biocultural phenomena. Cross-validation of estimates derived from more than one time-allocation method can detect sources of bias and may lead to estimates which are more credible than those from activity recalls used alone. This paper compares 24-hour activity recalls and direct observations as methods of estimating the energy expenditure of 145 18-37-year-old Samoans residing in rural Western Samoa, American Samoa, and Honolulu. Daily energy expenditure was estimated on 2 workdays by activity recalls (n = 214) or by focal-individual observations (n = 91). Samoan men expend 2,900-3,400 kcal/day and women expend 2,250-2,650 kcal/day. Estimates of energy expenditure in the Western Samoa sample are comparable to other subsistence horticultural societies, while the American Samoan and Honolulu samples are comparable to other sedentary working populations. Group estimates from recall interviews are 100-400 kcal/day higher than from observations, although this is only a difference of 5-10%. Overestimation of energy expenditure is more evident at higher work intensities than for sedentary activities. The apparent exaggeration of strenuous activity in this, and other, studies indicates that without cross-validation, self-reported activity data may often be confounded by an unknown amount of subjective bias.