Contents - Previous - Next

This is the old United Nations University website. Visit the new site at http://unu.edu

Methods of data collection

Random or spot observations
Day-long observations

Observation, interview (respondent recall), and diary recording (respondent report) are the three basic methods by which to obtain time-activity data. Research designs may employ activity sampling techniques that are either opportunistic, random, or targeted. Whichever method or combination of methods is used, information must be recorded in standardized activity categories appropriate to the culture and research questions at hand. Whose time one follows, at what intervals and by what method depend on the initial research questions and the characteristics of the study population, particularly their "time sense" and willingness or ability to co-operate. The availability of research assistants to collect, process, and interpret the information, which can be collected by the various methods, and any additional purposes for which the data might be used, are also considerations.

The results of activity analysis can be used to supply information on the daily schedules of individuals and households, productive work inside and outside the household, how time is allocated to different work and leisure activities, and a number of other social and economic questions. Therefore, it is important to be as sure as possible of the range of one's research questions in advance. Certain methods, e.g. random sampling designs, usually preclude information on sequencing and co-ordination of activities. To chart activity profiles and levels of children, particularly very young children who may be the target of nutritional concern, probably only extensive observations or experimental observation frames can provide the information needed.

Furthermore, in all cases, research questions and settings determine whether one follows, for example (a) all members or social categories in a household or large social unit, (b) focal women or men, or (c) all children of a certain age. Also, the manner in which other types of information are collected, such as dietary intakes, energy or resource expenditures, and nutritional status - whether simultaneously with "time"activity data or subsequently - must also be carefully planned, so as to produce accurate information without upsetting normal household routines. Alternative strategies for accessing information on these and other questions will be considered in the following sections.

Preliminary Information

As mentioned above, observation and interview in the socio-cultural group should always precede the systematic investigation of time allocation among households. Preliminary study will help to establish the ecological, economic, and social context in which households operate and can contribute to determining significant criteria for classifying households and individuals for further study. In addition to background information, including demographic, health, sanitation, and socio-economic data, preliminary assessments should specify: (a) income-producing and other activities by men, women, and children of different ages or other social divisions; (b) the decision-making processes for allocating time and income for food acquisition and processing; (c) the decision-making processes for selecting food; (d) the social organization of food procurement and preparation; (e) the social organization and decision-making for partitioning food and eating; and (f) extra-household food-consumption patterns for various members. Seasonal factors in food supply, dietary patterns, activity schedules, and health should also be determined in advance to ensure adequate sampling. All these observations may form part of a more general ethnographic report on the community and contribute to related studies of physical activities, work output, social organization, and community functioning in relation to the food supply.

From these initial data should also emerge enough data on the time sense of the population, relative visibility of members performing an array of tasks at different times, culturally appropriate activity categories, and the relevant social unit or units for the study, so that one can proceed to select a workable method of data acquisition and a sampling design appropriate to the research questions and cultural setting.

Direct Observation

Direct observation is the method most commonly used by anthropologists to provide a description of the activity patterns, food habits, and social organization of a study population. If frequency, scheduling, and duration of tasks are the objects of investigation, direct observation is the only method that can be used among populations with little clock sense of time. Anthropologists often use the method of participant observation to describe the typical tasks, social relations, and nutrition and health components of a culture. More targeted observations may be needed for studies investigating the ranges of time and energy required to perform specific tasks, such as clearing gardens, weeding and harvesting, marketing, and breast-feeding. These studies may require focused, timed observations of selected individuals. Longer-term observations of particular households may be needed to answer research questions on such topics as residence rules and household dynamics that facilitate infection responsible at least in part for malnutrition.

Already cited above, the ethnographies of Lee (1969), Rappaport (1967), and Richards (1939) include examples of how time-activity data, collected by observation, can be analysed toward nutritional ends. Lee, working among the !Kung San of the Kalahari Desert, tried to demonstrate food-energy returns relative to labour-energy expenditure for food by computing the number of man-days of work as a percentage of man-days of consumption (table 1). For this calculation, he observed the daily activities of the San living at the Dobe waterhole for one season (medium dry, neither the best nor the worst) over a period of one month. Stratifying the population according to age and sex, he recorded adult activities - hunting, gathering, visiting, and resting at home - and found adults worked an average of 2.5 days out of 7, with a working day of about 6 hours in length.

While Lee's method is ingenious, he does not make clear how he counted "work" or, for that matter, "consumption." Children, for purposes of the study, are considered adults, yet time allotted by them to foraging and the food energy they acquire, which may make up a significant component of children's time and intake, are not specifically included in this calculation. Lee provides figures on how much time it takes women to crack some quantity of mongongo nuts, a major item in the diet, and how much time men spend on different types of hunting, but it is not clear whether he includes time spent in transit in this "work" figure.

Lee also notes that work patterns are erratic. For example, following a streak of bad luck, hunters may abandon the chase for a month or more and turn to visiting and dancing. He discusses the "cost" of short versus long-distance trips, but it is unclear whether this unlabelled unit of effort refers to energy expended, time, or inconvenience. Also, time spent going from camp to camp in search of sociability and food is not figured into the "work" of the food quest. Nor is there an adequate calculation of time given for and nutrient intake taken from "snacking," non-food sharing behaviour (see also Hayden, 1981). Thus, "work" in the food quest is probably underestimated. Work, in total, is underestimated still more, since it is not clear how Lee accounts for crafts or other activities. While he can discuss seasonal differences in food and water availability, and how these affect the food quest (time spent gathering food according to proximity, density, and diversity of foodstuffs), there are no systematic seasonal comparisons. Although the method of calculating how many days of work per week are needed to feed a certain number of people is suggestive, it is incomplete.

Rappaport's (1967) study of the Tsembaga sweet-potato gardeners/pighusbanders of highland New Guinea provides a second example of time-allocation calculations used in human ecological studies. Data were gathered on the entire population of 204 individuals, and the nutrient flow/ecological balance of people and the environment were analysed, in general, within a 20-year ritual cycle. To calculate dietary intake in energy and selected nutrients, Rappaport observed intensively daily consumption over a 10-month period in four households, composed of 16 individuals. To arrive at figures of energy expenditure in gardening, he observed the time it took to complete various activities, converted these figures into energy expenditure, and used the figures to calculate returns in energy per work hour.

Table 1a. Dobe work diary: a record of the activities at the Dobe camp for the 28-day period 6 July-2 August, 1964

a. Each entry in column 3 equals the sum of the entries in columns I and 2 for the given date.
b. Week 2 (13-19 July) shows an unusually low work output. The investigator contributed food on 12 and 17 July, resulting in a decreased subsistence effort for the seven-day period. Week 2 therefore has not been included in the final calculation of the S ratio.
Source: Lee,1969.

Table 1b. Summary of Dobe work diary

Source: Lee, 1969.

Time calculations were based on time-motion studies of small numbers of individuals. Two individuals were observed fence-making and seven individuals on three separate days were observed clearing underbrush for gardens. He carefully described the nature and pace of tasks, and, when performance was not uniform, he described his observations. He observed that the tempo, except in the case of the frail and elderly, was quite uniform; and the tempo of the same actors was also consistent. Men worked continuously at a set pace, with the only break a protracted stretch around midday. He then calculated the area cleared per hour.

He assumed that the more muscular men cleared more rapidly, and "eyeball impressions" confirmed this assumption (table 2). No systematic data were supplied on the time and scheduling of women's and children's activities, although he does suggest that the time is ripe for slaughtering pigs when they complicate women's activity schedules by getting into their gardens.

Audrey Richards' study (1939) of the Bemba of Northern Rhodesia provides a third example of the ways observational methods can be arranged to study native time use. Tables 3a to 3e provide an example of the methods she used to describe Bemba time allocations to different food system activities. Working in two villages over very short periods of time, she carefully planned to include data on how long it took to complete a typical task, such as building a fence or making a garden, average work days for men and women in different seasons, the amount of time it took to prepare food, including processing grain and making relishes, and the timing, structure, and content of "meals" and other food intake. She included these observations as part of a more general study of the social relations surrounding food production, distribution, and consumption. Interview and diary records were used to check her impressions based on observation.

Table 2. Clearing underbrush - time and motion study

Source: Rappaport. 1967.

Table 3b. Estimates of male and female labour required for each activity (in weeks, by calendar period)

Source: Richards, 1939, pp. 396 - 397.

Table 3c. Tasks in meal preparation for different staple foods^a

a. Comments on time-attitude requirements: Millet and maize are prepared one day at a time, and it is not necessary to plan ahead. Cassava requires a 4- to 6-day routine, and one must plan ahead to have some roots soaking, some drying, and some being pounded into flour at any one time.

Source: Richards, 1939.

Table 3d. Time required for meal preparation

Source: Richards, 1939.

Table 3e. Timed processing of staple grain by seven women in Kampamba village (mins)^a

Average per Ib.: 7.4 minutes

a. Leaving out of account the dried legumes which have to be left to stew from four to six hours.

Source. Richards, 1939, p. 104

Using a small sample of seven women she calculated in an experimental situation the average time it took to thresh, winnow, and then grind a pound of flour (about 7.4 minutes a pound), and the total amount of time it would take to prepare six pounds of flour (about 45 minutes), the usual household expenditure. To this would be added the time it took to make the relishes (vegetables, spices, occasionally fish or meat) eaten with the flour (three-fourths to twice the time of the flour) for a total food preparation time of 21/2 hours. To this she added half an hour for fetching wood, to arrive at an average of three hours for the preparation of the evening meal, the only meal of the day.

Richards also notes the problems of housewives who were too tired to cook. Women in general prepared food after a full day in their gardens. Women too tired to go to the bush to find the ingredients for the relish to eat with the porridge simply did not cook, since without relish they said they could not eat the staple. Within the constraints of this food system, she also noted how children sometimes changed household compounds, and assiduously ingratiated themselves to the new eating group by contributing labour. Thus, the potential flexibility of the social organization to provide work and food was an important factor for individual food intake. Beyond this principle, there were also social obligations to share along kin lines, although she saw these being circumvented in times of dearth.Richards also noted that the pace of work was erratic: she was unable to specify the minimum time necessary to perform particular tasks, or, for that matter, the actual time natives spent, since they simultaneously spent time socializing. Interviews confirmed that natives did not conceptualize labour/time use in the European sense; most could not plan their work to obtain maximum results, could not calculate the labour required for any particular task, nor respond to her questions of how long tasks took. Nor did people eat according to any "regular" plan. Meals were scheduled in relation to agricultural activities, usually one meal at the end of the day. Seasonal routine and individual preferences, responsibilities, and resources altered the time of the main meal and actual daily intakes in the form of snacks. Women's daily gardening work ranged from 0 to 6 hours; an average figure might be 2 3/4 hours. People might work very hard for five hours in the garden, but then "collapse." In the slack season, old men in one village worked 14 days out of 20 and young men 7, while in the next village, in the busier season, men of all ages worked an average of 8 or 9 days out of 20. Average working days for men were calculated at 2- hours, and, for women, 2 hours gardening plus 4 hours of domestic work.

The foregoing three examples provide evidence of the value of general ethnographic observation for answering both questions about the time and energy expenditure demanded by activities in different seasons, as well as questions about social organization which may contribute to the adequate nutritional provisioning of households. Such reports have been criticized, however, as too impressionistic because they were not based on random sampling but were usually focused on a few subjects chosen for intensive study. Nor did they adequately sample the full range of activities over the course of a day or a year. To meet these criticisms, the method of random or spot observation has been developed.

Contents - Previous - Next