Previous - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Gretel Pelto and Pertti Pelto
The impact or effects of food programmes on the nutritional and health status of individuals is mediated. both positively and negatively, by a series of characteristics of households. For example, sanitary practices regarding food preparation and storage in the home are well recognized as factors that affect morbidity examined in relation to food intake. Since the household, in most societies. is a primary setting for the acquisition, preparation, distribution, and consumption of food, it is apparent that household composition and organization should be regarded as an intervening or confounding condition affecting the impact of nutrition and health interventions.
At the same time, it should also be recognized that household characteristics can, themselves, be strongly affected by food and nutrition programmes. For example, food programmes may change the schedule of work activities and food preparation within households. Thus, in the analysis of any supplementary feeding programme one can approach the matter of household organization and composition from two perspectives:
- as an intervening or confounding variable that has to be taken into account in assessing the impact of the programme on nutritional status; and
- from a sociological perspective. as a dependent variable (or set of variables) that is affected by the programme directly.
In practically all human societies, the household is a primary economic and social unit. It can be defined as the "smallest coherent social unit of people who reside together and maintain collective organization of food procurement and use, as well as other joint activities." Until quite recent times, the household, in many societies. was the primary unit of production as well as consumption. Industrialization, urbanization, and the impact of these processes on rural communities is bringing about rapid changes in the production/consumption relationship in households. Nonetheless, the household remains a fundamental social unit in all societies. It is also important to note that inter-household networks can have a major impact on the organization of food procurement and food use, and these, too, should be taken into account in the assessment of household variables.
The usual method for collecting data on households is the Basic Household Interview, or Household Survey. This method, which is based on a face-to-face interview using pre-established questions, can be used to collect a great variety of data. Typically, it is conducted with the female household head when a major focus of the interview is on food and nutrition-related variables. Often, however, it is necessary to collect economic and other data from males in the household, usually the male household head. Thus, the basic household interview may be administered in several parts to different individuals.
Full-scale data-gathering on all aspects of household structure and composition would be a time consuming and costly operation. Most evaluation projects will be able to select certain portions of these research procedures, based on specific hypotheses about the effects of food programmes in particular areas and districts. Also, preliminary ethnographic reconnaissance in selected communities can often identify key features of households for research emphasis.
The types of variables that are likely to be of greatest importance to the evaluation, and that should be included in the Basic Household Interview, can be summarized under the following headings:
Within these broad categories, there is latitude for collecting a wide range of information. However, it is strongly recommended that some questions from each of these domains or headings be included in the interviews. To the extent possible, selectivity should be exercised with respect to the degree of elaboration of questions within each of the headings, rather than through the complete elimination of any one general category.
In most cases it will be useful for evaluations of projects to collect data at more than one level of intensity, drawing a fairly large sample of households for general assessment, and selecting special sub-samples within the general sample for intensive examination of special features. For example, the following strategy might be used in a situation where ethnographic reconnaissance has tentatively identified important changes in women's work allocations, with strong effects of "household type" as an intervening variable.
General sample: 480 households (basic survey data on each household: composition, MSL, ethnicity, nutrition status)
Special sample: stratified. 100 households (50 nuclear; 50 extended. multi-generation) Additional data: four randomly-allocated "spot check" observations of task performance. Focus on female tasks, plus interviews of reproductive history of female household head.
Types of Households
In different social systems, the "cultural norms" of marital residence result in different household types. Differences in household composition within and between communities can have very important consequences for food programmes and their evaluation, so effective data gathering requires careful consideration of variations in household composition. In some instances the isolation of households as distinct units will be made more difficult because they are embedded within larger kinship or family groups. However, the following types of households are generally recognized, within anthropology, as common social units:
1. Nuclear family household. This type, common in many parts of the world, consists of a husband and wife and their dependent, unmarried offspring. With urbanization, this type is becoming increasingly common in parts of the world where other household types have been the norm. A variation of nuclear households is the addition of a single adult. especially the widowed parent of husband or wife, or an unmarried adult sibling of the husband or wife. (In many areas marriage may not be clearly formalized, so the category "nuclear family" will include male-female pairs in consensual unions.)
2. Extended family household. This type consists of two or more married pairs, of different generations. The most common form is that of an older married couple. plus their married sons with spouses and children. Less often such an extended family household is "matrilocal", with inclusion of married daughters plus their husbands and children.
In those parts of the world where, in earlier decades, large extended families were considered the cultural norm, the forces of modernization have tended toward reduction of these large families into smaller units (nuclear family households). Nevertheless, there are exceptions where large extended families are still prevalent.
3. Joint family household. This type consists of two married pairs of the same generation, such as two brothers and their wives (or, less commonly, two sisters and their husbands) plus offspring.
4. Matrifocal household. This household type is composed of female head-of-family, her children, possibly other kin, but no resident spouse. Such households are common in many urban areas, as well as in rural regions where extensive labor migration has removed the males from regular permanent residence. (Note: such households should not be considered as "aberrant" or "temporary", as they may be the most numerous type, and the predominance of these has now persisted for generations in some regions.)
5. Polygynous household. This type is composed of a male plus two or more wives, and their children. In many societies the separate wives maintain separate hearths and cooking activities. However, the patterns may be variable, and the separate sub-units (women plus their children) are generally interdependent in the sense that control and maintenance of food-getting resources is generally in the hands of the male head (fatherhusband).
6. Other types of households are also found, but tend to be unusual both numerically and in their socioeconomic significance. However, in urban areas, households of unrelated adults are becoming more common.
The Significance of Household Type
There is a considerable body of evidence that differences in household type affect a number of variables that have to be assessed in evaluating programme impact, including fertility, nutritional status of children, and maternal behavior. First of all, differences in household type are likely to be associated with differences in the ratio of adults to children, as well as differences in the ratio of adult males to females. In societies where there is fairly clear differentiation of male and female tasks, differences in adult sex ratios can have important consequences for household functioning. Differences in the ratio of adults to children may have a large impact on the quality of child care, on child training practices, and on intrahousehold food distribution. When additional adults are "dependents," the effect of their greater nutrient requirements compared to children's may have negative consequences for nutritional status, especially of children. On the other hand, when the presence of more adults creates a condition of greater economic flexibility for the household, the effect may be positive.
These issues are significant for evaluating the impact of feeding programmes because most communities have several types of households, so that this is usually an important source of intra-community variation that must be taken into account in sampling design Many urban communities have high percentages of nuclear households, and matrifocal households are also common. In rural areas, there are likely to be extended. joint, and nuclear households. In societies where polygynous households are common, there will be relatively small frequencies of households with three, four, or more wives, compared to households with only two wives or with nuclear families.
Implications for Sampling
Assuming that household type may be an important variable, that the distribution of household types within a community is not random, and that the ratio of one type to another within a community may be very unbalanced, it follows that simply sampling by households can be misleading A preliminary survey to determine the rough geographic distribution and frequency of household types may be important as a basis for deciding whether a stratified sample. or perphaps quota sampling, by household type should be utilized.