Cover Image
close this bookResearch Methods in Nutritional Anthropology (UNU, 1989, 201 pages)
close this folder3. Methodological procedures for analysing energy expenditure
View the document(introductory text...)
View the documentIntroduction
View the documentSurvey of habitual activities
View the documentDetermination of critical activities
View the documentDetermination of key participants
View the documentMeasuring energy expenditure rates
View the documentTime-motion analysis
View the documentEstimation of energy expenditure rates from time-allocation data
View the documentAssessment of endurance capacity
View the documentSummary
View the documentReferences

Estimation of energy expenditure rates from time-allocation data

As previously discussed, although indirect calorimetry or heart-rate monitoring can provide relatively precise determination of energy-expenditure rates, they are high cost, time-intensive techniques. Hence, it becomes difficult to use these techniques to examine all activities or large segments of a population. At the other extreme, the self-assessment method of estimating energy costs allows for only very gross categorization of the cost of activities (e.g. "light," "moderate," and "heavy") with respect to the effort involved in their performance.

Table 9. A comparison of energy expended in herding between a 12-year-old Nuñoa boy and a man

Activity

Energy expenditure rate (kcal/min)

Time expenditure (mini/day)

Energy cost (kcal/day)

12-year-old boy (30 kg)  
Lying

0.9

2

1.8

Sitting

1.0

343

343.0

Standing

1.0

20

20.0

Squatting with arm motion

1.5

2

3.0

Walking slowly

2.3

59

135.7

Walking moderately

2.7

11

29.7

Walking with light load

3.3

18

59.4

Walking up and down hills

3.5

20

70.0

Running

4.5

5

22.5

Total  

480

685.1

Man (54 kg)  
Lying

1.2

2

2.4

Sitting

1.3

343

445.9

Standing

1.5

20

30.0

Squatting with arm motion

1.9

2

3.8

Walking slowly

3.3

59

194.7

Walking moderately

4.5

11

49.5

Walking with light load

5.5

18

99.0

Walking up and down hills

6.0

20

120.0

Running

7.5

5

37.5

Total  

480

982.2

Source: Thomas, 1973a, p 141.

Fortunately, there are two techniques with which one can build on the limited data obtainable through indirect calorimetry while avoiding the imprecision of self-assessment methods. One is the use of standardized published values of similar activities, and the other is the interpolation of estimates from activities for which energy cost has been derived through indirect calorimetry (or heart-rate monitoring) on the same group.

The use of published values is a common technique for estimating the energy cost of habitual activities in energy flow studies (Rappaport, 1968; Lee, 1969; Kemp, 1971; Gross and Underwood, 1971; Johnson, 1978; Morren, 1977; Winterhalder, 1977; Smith, 1980). Here the purpose is to describe major flows of energy through the human population and to compare the energetic consequences of rather different activities. Thus, errors introduced by this estimating technique are usually unimportant.

Standarized values for a wide variety of activities and occupations can be found in Passmore and Durnin (1955), Durnin and Passmore (1967), Godin (1972), FAO/ WHO/UNU (1985), and Astrand and Rodahl (1986). The principal limitation in using published values is that these cannot account for population differences in expenditure rates for local conditions (e.g. high altitude or disease that may cause rate variations) Furthermore, determination of values is heavily biased to activities and rates of young men so that is is difficult to estimate the expenditure rates of children through simple corrections for weight; the most recent FAO/WHO/UNU report (1985) provides techniques for addressing this problem. As we have stated before, when internal population comparisons between sex-age groups and body types are desired, more precise methods have to be used. These are most informative when presented in terms of kcal/min/kg body weight, so that the expenditure rates of individuals of different weights can be compared. As Montgomery and Johnson (1976) have argued, presentation of the degree of variation around the mean energy cost for a group may provide valuable data which is too often omitted in energy cost tables.

In figure 5, Astrand and Rodahl (1986) show the range of variability that can exist for some activities. When this range is very broad, such as working with an axe (see the last entry in the table), it suggests that more specific definitions of that activity are needed. One particularly important aspect of knowing the range of variability lies in being able analytically to overestimate or underestimate costs. Thus, in testing a hypothesis where a high energetic efficiency ratio (input to output) is expected, it would be appropriate to consistently overestimate inputs. Following this procedure, the investigator could confidently state that the energetic ratio is at least at the stated level, and most probably well above this. Frequently when one has low confidence in the accuracy of central tendency measures, or when a large range of variability exists, it is best to perform the analysis from one end of the range. Then, as we have just suggested, one can assume more accurate values would fall consistently above or below those stated.

Under some circumstances, relying upon published values can be justified. For example, in Winterhalder's study of the food-acquisition activities of Cree Indians in the boreal forest, the value of the more precise information that could have been obtained using indirect calorimetry merited neither inconveniencing the subjects nor the additional research efforts that would have been necessary. The use of more elaborate techniques of estimating energy expenditure would have precluded the simultaneous collection of other equally important data and might have deterred the study population from participating. Winterhalder states that since "the patterns and levels of various activities that go into foraging are highly variable . . . precise information on expenditure for specific acts is of reduced value" and that the increased accuracy would have been "illusory when placed in the broader context" (Winterhalder, 1977, p. 602).

Winterhalder's assessment of his own research situation appears valid, and similar realistic assessments are always necessary in deciding for or against any methodological strategy. However, strategies such as Winterhalder's must be weighed against the fact that the energy values of a specific activity may vary quite considerably among populations and under different conditions. In many studies, a combination of strategies has been utilized such that published tables are deemed adequate for some activities, while more precise indirect calorimetry techniques are adopted for measuring "critical" activities.



Fig. 5. Energy expenditure of different activities (after Astrand and Rodahl, 1986, p. 439).

Besides the use of published values, a second method of estimating non-measured energy-expenditure values can be used in conjunction with limited indirect calorimetry. This estimation involves the interpolation of results of measured activities to those not measured. Such a method may be preferable to using published values, in that the non-measured activity values are derived from information gathered from the same population and under the same general environmental conditions. In this way, at least some of the inaccuracies inherent in the use of standardized tables can be overcome.

Results derived from the use of such methods can be seen in table 6, which lists activities performed by the Nuñoa population in producing potatoes (Thomas, 1973a). Comparing this table with table 3, which lists activities actually measured, it can be noted that the values for many specific activities could be estimated with some confidence through interpolation. Ultimately, these figures were used to analyse seasonal variations in the energy costs of activities and to provide energetic efficiency ratios for various productive activities. Comparisons of heart rate or self-assessment of effort among measured activities and those to be estimated provide a basis for interpolation.