Contents - Previous - Next

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

Random or Spot Observations

The method of "spot observations" has been used to characterize the activities of groups of specific age, sex, and occupation within particular societies. These studies have been designed primarily to document labour time use for different societies. A "spot" record captures by observation the activities of all individuals present at the moment of entry into the household, and, by subsequent interview with those present, the activities of those not in view. The data are then coded in standard activity categories, such as work inside and outside the household and leisure time. Then the data are processed to determine the percentage of time spent on any particular class of activity by any individual or class of individuals. One simply calculates the number of instances of that activity as a percentage of all activities recorded (i.e. as a percentage of the number of visits) (Rogoff, 1978).

An early example of the use of spot observations was a study of the activities of males and females of different ages in a Mayo village in Sonora, Mexico, by Charles Erasmus and his wife in 1948. They used the activity data to compare working days of Mayos to those of semi-skilled labourers near the Washington, D.C. area (Erasmus, 1955). Although they did not use a random number table to schedule visits, they found that by checking the record charts they were keeping on every man, woman, and child by hours of the day, they were able to visit those individuals at hours of the day that were being neglected and fill out profiles for the 32 households and 200 inhabitants of this community. They were able to record a total of 5,000 observations in three months.

Erasmus then tallied activity observations, in culturally appropriate categories, into three general categories: "household activities," "economic activities," and "leisure." He presented each of the specific activities as a percentage of total observations for each sex, to show sexual division of labour, and then graphed the relative distributions of the three general activity categories (5 a.m. to 6 p.m.) for each sex (table 4). From the graphs, one can see at a glance the "work day," "siesta times," and the predominance of household over remunerative labour activities for women, in contrast to men (figs. la, lb). On the basis of his observations and analysis of spots, he was able to correct his initial erroneous "impressions" that Mayos in this community were "lazy" (i.e. spent a lot of time sitting around) and did not work very much. In fact, his data showed both men and women worked about nine hours and spent four hours at leisure in the course of the thirteen-hour day covered by the observers. Comparing his results to a timeallocation study of semi-skilled residents from the Washington, D.C. area, he found patterns of time use, in terms of household, economic, and leisure activities, to be roughly the same.

Table 4. Sex division of daily activities in a Mayo village

a. Less than 0.1 per cent (only one observation).
Source: Erasmus, 1955.

Fig. 1a. Distribution of daily activities in a Mayo village: men (after Erasmus, 1955).

Fig. 1b. Distribution of daily activities in a Mayo village: women (after Erasmus, 1955).

Johnson (1975) also chose the spot observation method to document the work activities of a Machiguengo community in Peru as a more precise economical alternative to impressionistic recording. He and his wife collected spot observations over a ten-month period, while also gathering data for a broader study of the cultural ecology, social organization, and sex roles of this community. Using location and distance from a central point as a principle of selection, they visited all of the households within a 45-minute walking distance of their homestead, divided them into two groups (upriver/ downriver, to cut down on walking times), and set up a schedule of "random" visits for daylight hours, during which they recorded what everyone in the household was doing at the moment immediately before they became aware of the observers' presence.

They made spot observations in 13 households with a total of 105 members over 1,345 days, which resulted in 3,495 cases. Their hours of observation ranged from 6 a.m. to 7 p.m., during the daylight hours when it was convenient to travel, which were also the hours when the households they observed indicated they did not mind such unexpected visits. The data, coded according to person, time location, and activity, were punched directly onto IBM cards. (Probably today one would go directly from data-coding sheets to interactional data entries on computer tapes. ) After the observations were completed, the researchers simply calculated, for each social category of interest, the number of observations of each activity category as a proportion of the total, and converted these figures into the relative time spent on each activity. They were then able to compare activity rates by sex (table 5) and age class. As in the case of the Erasmus study, spot observation was combined with more traditional participant observation methods to acquire data on qualitative questions, such as cultural ecology, family organization, and sex roles.

As a third example, Minge-Klevana (1971) used spot observations to test the following hypothesis: as children allocate more of their time to education and less to the farm, parents make up for the forgone labour and offset the additional educational costs by allocating more of their own time to family-based labour. Working in a Swiss Alpine village, she studied time allocation of all family labour, the percentage performed by mother/wife, husband/father, and children, comparing those households where fathers did wage labour with those in which fathers stayed at home. From her original sample she selected 32 families for analysis because they had had more than 40 observations each. Although she used Johnson's method of "spotchecking" and accumulated 3,433 person-observations, it is not clear how she decided when to go where, nor is it clear from her descriptions how she calculated children's contributions to the household operations. Some of her difficulty in calculating children's work stems from the nature of defining "work" versus "leisure" activities for children who are ordinarily in school. For example, if children tend cows when the parents cannot both tend cows and harvest potatoes, should the anthropologist consider this "a holiday," as children themselves do?

While spot observations are more amenable to statistical analysis and can be considered less impressionistic than other forms of observation, the activity data so collected and analysed do not allow one to study how differences in household organization or individual capacity affect work effficiency. Nor is there an opportunity to observe and report on social interactions, qualitative differences, or sequencing in time use and performance. To overcome some of these deficiencies, some researchers combine spot-checking with 5-, 10-, or 15-minute narratives. This method provides them with a profile of standardized behaviours for selected social groups as well as a more qualitative record of interactions between people and their environments.

Table 5. Division of labour by sex (married adults)

a. Differences are significant at p < .01 level (t-test).

Source: Johnson, 1975.

Whiting and Whiting's (1975) research team has used the method of spot observation with narratives extensively to observe children in six cultures and to chart their development in terms of a set of standardized behavioural categories. The final cross-cultural comparison of childrearing practices was based on observations of 67 boys and 67 girls, aged 3-11. Each child was observed at least 14 times. Five-minute narratives described in detail the behaviours during the observation period, including activities coded in standardized categories, and also details of interaction, conversation, and physical movement. With one exception, the total number of behaviours recorded in each culture was at least 1,000. The end result was a comparison of behaviours of males and females of different age sets in six cultures. General terms, such as "nurturant" or "aggressive," were used to describe the behaviours, which were characterized as the outcomes of child-rearing practices and sex-role development in the different cultures.

While this method of observation and reporting potentially provides more information than simple spot-checking, the value of the data depends heavily on the scope of the activities record. For these studies, the behavioural categories, "acts sociably," "insults," "offers help," "reprimands," "seeks dominance," "seeks help," "suggests responsibly," "offers support," "seeks attention," "assaults sociably," "touches," "assaults," in which investigators were trained in advance to record, may have been useful for cross-cultural comparison of child development, but hold little meaning for analysis of other behaviours.

By contrast, studies of mothers' activities and child care carried out by the same methods in ongoing projects on cross-cultural child development do utilize categories of interest for other social scientists, including nutritional anthropologists.

Bloch (1979) collected five-minute observations on samples of Senegalese mothers involved and not involved in seasonal work in the cash economy. She stratified her sample of 50 mothers in terms of the age and sex of a mother's youngest child, using age categories of 0-6 months, 712 months, 13-18 months, 19-24 months, 2 years, 3 years, and 4-6 years. She compared differences in proportions of time allotted to personal activity, primary child care, domestic work (not primary child care), and economic activity. While she did not analyse separately the food-related activities of the mother in relation to the youngest child, number of children, or the economic organization and livelihood of the mother, her activity records of spot observations with five-minute narratives potentially provide such data.

Spot observations on the interactions between children of different ages and their environment can also potentially contribute to questions of nutritional concern, particularly when the records include information on factors such as proximity of the child to mother or mother surrogate a measure of child care. This measure of mother-child interaction can then be further tested to see how it relates to nutritional status.

In assessing the value of spot observations over more impressionistic methods of observation recording, one should note that the "spots" produce immediate quantitative "snapshot" data, remove the problem of observer bias or interference with the individuals observed (but not bias by observer), and provide reliable figures on activity profiles by different classes of individuals throughout the year. The "snapshot" will not give one a clear idea of: the duration of any particular activity; the sequencing or co-ordination of units of activity; or the time/quality of social interactions.

Nevertheless, some of these deficiencies can be made up by either combining spots with narratives or by making other kinds of longer observations. Spot observations can probably provide a reliable indication of factors related to the tasks of the elderly and of the young? e.g. the ages at which children of either sex begin to work, the patterns by which they become responsible for tasks and master them, and the hours and kinds of contributions they make to the household economy. If sufficient data are collected in different seasons and among different kinds of households, the method can also supply an accurate account of differences in activity between households with varying social structures and occupations and significant seasonal differences in work and leisure-time use. The use of spot observations to sort out differences between household work patterns, rather than "average" activity profiles, might provide the raw data from which to work towards a new measure of "household function," which can then be correlated with other factors, such as nutritional status.

None of these analyses deals specifically with the subject of time allotted to food versus other activities. Indeed, Erasmus (1955) highlights one of the problems with classifying the data in his example of eating: Is eating a "work" or "leisure" activity? For women, who fit eating in between and around food preparation and feeding the rest of the family, eating is a "work" activity, but for men, who spend more time eating, it could certainly be counted as a leisure activity.

Spot observation, nevertheless, seems to offer an ideal way to characterize the activity patterns of a culture, in particular food acquisition, food preparation, and the scheduling of meals and snacks. Significant differences in activity profiles could then be used to select households for longer observations to refine and extend the initial spot observations. Any intake data, of course, would have to be collected separately from the initial stage of spot observation study. Spot observation is also a potentially useful way to characterize labour allocations in different households. Such information can shed light on eating patterns and schedules and the relationship between women's work schedules and (a) meal preparation and (b) participation in development projects.

Spot checking could also be used to show the proportion of time allotted to different feeding behaviours and resultant food intakes. Results of spot checks on mothers' activities and patterns of child care for different ages and sexes could also be used to develop measures of nurturant behaviour. Also, spot-checking could become an efficient method of recording how family labour-time-allocation patterns relate to feeding patterns and nutritional well-being, particularly where there are significant differences in labour use among families. Where education and labour demands impinge on children's time, spot-checking can reveal conflicts in household and school schedules that interfere with children's food intake. For example, one could establish the meal-time profiles of households, and check these against children's work and school schedules.

Although random spot observations have many advantages, one should also consider some disadvantages. Spots are time-consuming to collect and may require large amounts of time for travelling from one household to the next, as many observations are needed to make the results statistically reliable. Each of the studies cited above covered a small sample of households and required substantial time commitments. Another disadvantage is that activity data, averaged across households and seasons the form in which most "spot" studies have been presented - do not allow one to analyse how differences in household organization or individual capacity affect work efficiency. Even when combined with narratives, there is insufficient opportunity to observe social interactions, work rates, co-ordination, and other qualitative differences in time use and performance. Such data must be collected separately by observation and interviews.

Since very small numbers of subjects and a larger number of spot observations are used, this method also entails special problems of sample selection. What constitutes a representative sample must be determined carefully and depends on the variance of the sample. Also, variations in the numbers of children in the household and specifics of household organization make attempts at statistical comparisons of behaviours and activity schedules of age/sex categories within and between cultures difficult, and the interpretations of such findings must be carefully qualified. Moreover, since the many spot observations or spots with narratives are potentially disruptive to the normal flow of activities, given observer presence, one can question whether the data they produce have greater scope and utility than data provided by day-long observations or some other form of short day observations. The latter two methods can provide information on sequencing and duration of acts, and, in food-related studies, can also supply data simultaneously on eating behaviours and intake by the focal personae of the study.

Contents - Previous - Next