Contents - Next
This is the old United Nations University website. Visit the new site at http://unu.edu
Nutrition and health
Studying health and nutrition behaviour by
examining household decision- making, fintra-household resource
distribution, and the role of women in these processes
The School as a data source for food and nutrition surveillance systems in central America and Panama
Ascariasis and digestibility: a study in cameroonian children
Feeding from the family pot for prevention of malnutrition
Breast-feeding and weaning practices for infants and young children in Rangoon, Burma
Studying health and nutrition behaviour by examining household decision- making,intra-household resource distribution, and the role of women in these processes
Ellen Gail Piwoz and Fernando E. Viteri
Food and Nutrition Program, Pan American Health Organization, Washington, D.C., USA
Underlying this article is a recognition of the relationship between poverty and poor health and nutrition and a realization that poverty does not affect al/ the members of a household uniformly. We believe that households as a whole do not operate to promote the common good of all their members. Within conditions of chronic resource scarcity, some family members consistently fare worse than others. It is, therefore, necessary to identify intra-household factors that influence health and nutrition behaviour. Given the fact that household behaviour is determined by a number of factors, several types of intervention are proposed. To improve the chances of lasting success for development programmes, we advocate designing and testing educational messages that address all aspects of household behaviour.
* * *
Recently, intra-household resource distribution has received attention as development programmers and planners have come to recognize the importance of household dynamics to their efforts. Unfortunately, little systematic research has yet been carried out to determine the functional consequences of practices related to intra-household resource distribution, and to document how changes in resource availability affect household behaviour.
The purpose of this article is to address these informational shortcomings. Its major objectives are threefold: first, to identify aspects of household dynamics that influence distribution practices; second, to describe a method for carrying out both community- and household-level surveys to identify the determinants and consequences of household behaviour; and, third, to examine decision-making and resource allocation to identify household factors that influence health and nutrition behaviour in order to design more effective educational interventions.
Household food consumption is determined by three variables: food supply, management, and distribution to individual household members. The food supply is influenced by variables affecting production and acquisition, such as the environment, access to fertile land and capital, income, food prices, and price supports. Food management (storage, processing, and preparation) for human and nonhuman use is determined by available technology and resources as well as tradition. Food distribution is thought to be based on perceptions of individual need. These perceptions are formed from assessments of size, age, physical activity, and lifecycle phase (menstruation, pregnancy, lactation, menopause) and beliefs about an individual's taste, digestive ability, and food preferences. Food distribution patterns may also be based on the perceived economic and social value of individual household members.
Most programmes aimed at bettering nutrition seek to improve health care, increase the household food supply, or modify the acquisition and preparation of food in the home. Evidence suggests, however, that improvements in the supply of food to the household alone are not enough to ensure the adequate nutrition of all its members [14, 131]. Undernutrition has been found to persist among households within communities and among individuals within households where income and reserves are both stable and sufficient to avoid such conditions. The opposite has also been found; there are well-nourished individuals among the undernourished in communities where both health care and food supply are chronically short.
Because nutritional status and income can vary independently, many development planners and programmers, as well as public health and nutrition specialists, have recently become concerned with the study of food distribution within the home [80, 86, 103, 117, 122, 149]. The focus of much of their attention has been the development of models to identify variables that influence acquisition, availability, preparation, and allocation of food to individual household members. Several authors [22, 32, 117] have specifically identified patterns of- distribution and have attempted to infer causal factors and determinants of observed nutritional behaviour. From these works, it has been concluded that many variables affect household allocation behaviour and that, in order to assess the contribution of specific nutrition-related beliefs and distribution practices to undernutrition, we must consider decision making related to resources and their distribution in the home [134, 1501.
INTRA-HOUSEHOLD RESOURCE DISTRIBUTION (IHHRD)
The study of IHHRD is important because it provides us with insight into the decision-making processes of households and, in turn, into their priorities, perceptions, and norms for survival. All too frequently in the past the efforts of development planners and programmers have not taken intra-household behaviour into consideration.
Underlying our advocacy of the study of household resource distribution is the belief that the household is both a programmatically useful and important unit of analysis [1341. The vast majority of people in the developing world live in groups (households) which are units of production, consumption, and socialization. Household composition may vary by culture, socio-economic status, and location [urban versus rural), and may change through time. It may include one or more families or members unrelated by blood.
In this study we are concerned with households that at a minimum contain mother-child pairs. We will examine how the household, as a multigenerational unit, produces, organizes, and distributes responsibilities and resources inside and outside the home, how knowledge about appropriate health and nutrition practices is transmitted, and how respect for opinions and decisions is gained by different household members.
The household as a unit has been most frequently studied by anthropologists and economists. Anthropologists have traditionally described the household in terms of genealogy and residence patterns, with particular regard for production and subsistence practices (fig. 1). Their focus has been on describing "the fabric of culture" and how different aspects of social organization and economic production interact to affect behaviour in specified population groups. These studies have recognized the role of the household as a transmitter of knowledge and socializer of expected behaviour, but their normative nature has limited their usefulness for research and programme planning [6,47,140, 157, 161].
FIG. 1. Nutritional ecology model (after Jerome, Pelto, and Kandel  )
Economists, on the other hand, have used the household unit as a whole as the subject of many micro-economic studies. Most studies of household economic behaviour have subscribed to the "new home economics theory" [10, 111. These studies have assumed the existence of a "joint utility function", or the belief that household behaviour is motivated by a "collective concern for economic efficiency" [Becker, cited in reference 50]. To apply this theory one is required to accept that the household acts as a unit to promote the common good of all members, and that household form and function remain constant over time  .
Studies assuming the existence of a joint utility function have only considered household behaviour at the aggregate level. There is ample evidence, however, that neither goods nor the benefits of development are distributed equally to all household members. The weakness of the joint utility function perspective is substantiated by evidence on intrahousehold, sex-biased differences in food consumption and undernutrition [25, 26, 68, 89, 107], literacy and level of educational attainment , and the distribution of other resources such as clothing, health care, money and time. Surveys of time use [16, 69, 82, 123] and income disaggregation [3, 62, 78, 84] further reveal that households may have separate internal economies, and that division of labour and economic responsibilities are well-defined by age (generation) and sex (gender).
The results of these studies emphasize the need for a new approach to analysing intra-household rather than aggregate household behaviour. These analyses must identify how households produce and manage resources (the division of labour and allocation of time) and on what basis household decision-making occurs. Within the home it is important to consider the division of income-generating and income saving responsibilities, and differential access to food, health resources, clothing, education, and care of different household members. The distribution of leisure time may also be an important indicator of societal values, expected sex-specific behaviour, and of interpersonal relations within the home.
Time Allocation Studies
Time studies are important for identifying the division of labour and for determining what is considered acceptable work and behaviour among different groups [56, 102]. How households and individuals manage their time, given the tasks to be accomplished, may be a very important indicator of how they will respond to new interventions. Many health and nutrition programmes require participation in new, "time-consuming" activities. An analysis of time allocation may reveal unanticipated restrictions on participation.
Interventions to save time, or labour-saving devices, may be necessary in order to provide time for programme participation [53, 65] . Such interventions may themselves indirectly result in significant nutrition and health improvements [8, 51, 52].
TABLE 1a. Time allocation of rural men and women in Upper Volta (14-hour day)
|Activity||Average time in minutes|
|A. Production, supply, distribution|
|1. Food and cash crop||178||186||0.96|
|2. Domestic food storage||4||1||-|
|3. Food processing||132||10||13.2|
|4. Water and fuel supply||44||2||-|
|B. Crafts and other professions||45||156||0.29|
|C. Community activities||27||91||0.30|
|E. Personal needs||158||269||0.59|
|F. Free time||77||118||0.65|
|Total work (A + B + C + D)||587||453||1.30|
|Total personal and free time|
|(E + F)||235||387||0.61|
|Total work (A+B+C+D)/ total personal and free time (E + F)||2.50||1.17||2.14|
Source: McSweeney .
It is assumed that observation is necessary for understanding total household production, the sexual division of labour, and the multiple factors influencing intra-household behaviour. Time is the common variable with which to identify behavioural choices and elasticity for change. Methods for the study of time allocation have been well documented [16, 31, 42, 77] . Examples of how these data may be presented are given in table 1.
Time allocation studies alone, however, may not provide us with all the key elements influencing specific actions. Other important variables include practices related to household income-generation and internal budgeting, the determinants of individual health care, and the distribution of food and childcare responsibilities within the home. Such information is important for determining how increases in earned income by different family members will affect household resource acquisition, and how socialization and division of labour affect the distribution of resources to individual family members.
Our purpose is not only to identify behaviour but to suggest how to motivate behavioural change. To realize this goal we must study the bases of decision-making power in the home, and identify those factors that influence perceived individual need and subsequent decisions about resource use.
THE STUDY OF DECISION-MAKING
TABLE 1b. Time budgets in rural households (Laguna, Philippines)
|Activity||Average hours per day|
|Father||Mother||Children (average 4 per household||Per child|
|3. Free time (does not include sleeping)||5.78||5.35||20.22||5.05|
Source: King and Evenson [ 83 ] .
TABLE 1c. Hours per week spent by family members on child care in relation to mother's work status ( Laguna, Philippines)
|Mother's participation in market production||Location of mother's job|
|Family member||None||One or more hours per week||No job or job near home||Job far from home|
|Older siblings||2.5||4 5a||4 0||3.0|
a. Paired t-test significant at 0.05 level. Source: King and Evenson .
The identification of methods for the systematic study of household decision-making processes is a subject that has received only limited attention, in part because experts have not agreed on the most relevant variables for study  . Acharya and Bennett , for example, recognized that decisionmaking was an interactive process. In their indepth studies of women's roles in eight villages in Nepal, they studied different aspects of the decision-making process, including suggesting, deciding, implementing, and disagreeing. From these studies they found that decisionmaking was dependent on the dichotomy between men's and women's spheres of activity. Where the dichotomy was greatest, men predominated in decision making. In villages where men's and women's roles were less rigidly distinguished, women had greater decision making power. Women were more likely to make suggestions than to have the responsibility of making final decisions.
In-depth studies are time-consuming and difficult to carry out on a wide scale. In recognition of these logistical problems it has been suggested that outcome, or actual behaviour, be used as the key indicator of household decision-making behaviour [ 135] . It is important to note that outcome, though easier to measure, may not reveal the elements of family interaction that are believed to be significant for understanding household distribution practices. Outcome may be the result of compromise and may not reveal what is considered rationally optimum behaviour by scientific standards or by the expectations of influential household members. Also, behaviour may be carried out on the basis of past experience or by means of interpersonal persuasion, and thus those who only observe outcome may underestimate the importance of other factors influencing behaviour.
It is unknown whether household decision-making responsibility can be attributed to individuals on the basis of outcome alone, or if household processes such as suggestion, negotiation, and persuasion must also be considered in the targeting of interventions. Special attention should be paid by programme designers to determining the role individual household members play in influencing decision-making practices. For the purposes of this article we assume that household resource distribution (outcome) is one useful indicator of intra-household decision-making behaviour for the design and targeting of health and nutrition education programmes; the approach we advocate towards the use of process and outcome variables is discussed in the next section.
HOUSEHOLD DECISION-MAKING IN THE CONTEXT OF HEALTH AND NUTRITION BEHAVIOUR
The study of IHHRD and decision-making will provide us with insight into how the benefits of development are transmitted to individual family members . We also believe that a study of household dynamics will illuminate the cultural, economic, and social factors that impede efforts to promote behavioural change  .
Most health and nutrition education programmes have traditionally sought to transmit knowledge and to influence decision-making through a didactic process, or through the use of concise and simple messages to get a specific point across. These methods tend to isolate the targeted behaviour from factors surrounding or influencing present practices ("the co-variants of behaviour"). These co-variants, or influencing factors, include: tradition and personal experience; the monitoring of expected behaviour (in the household and the community); the level and source of present knowledge; decisionmaking power; restrictions on time and on other resources; and competing needs.
Take for example a programme to encourage feeding during a child's episode of diarrhoea. Diarrhoea is considered a common occurrence among most children in the developing world. A woman's personal experience and traditional practices may dictate that she withhold fluids and food because, according to the woman, "feeding only made the diarrhoea worse in my older children." She asks her mother and even a health centre official and they concur. Her mother-in-law lives with her and will hold her responsible for any "wrong" or unusual acts. She is thus unlikely to be innovative, despite new information the dangers of dehydration. In this example, time is less important in the context of adapting new behaviours because it is usually available for the feeding of the child when he is well.
If the programme had also been directed towards other decision makers in the community, including mothers-in-law and husbands, it might have had improved results. If such a programme is also aimed at improving a mother's decision-making role (through improvements in her economic social status) it may also enhance the chances of innovation and behavioural change.
Table 2 includes examples of the co-variants of selected behaviours. This list is not exhaustive. Additional behaviour-specific research is needed to identify other relevant factors for the development of broadbased, multi-targeted educational messages.
The analysis of resource distribution and decision-making behaviour requires a new approach to the study of household dynamics. Rather than focus on either process or outcome alone, we suggest a two-stage approach. The first stage requires the identification and analysis of specific community level indicators that will indicate trends in decision making and distribution processes; these will be discussed in detail later. If analysis at this stage reveals inadequate decision-making processes or inequitable resource distribution practices, specific intra-household behavioural studies (the second stage) are recommended.
TABLE 2. Examples of the co-variants of behaviour
|Behaviour||Tradition/expected behaviour||Monitoring of tradition/ expected behaviour||Decision- making power||Time and other restrictions||Competing needs|
|Level of knowledge||Source of knowledge||Personal experience|
|1. Breast-feeding (BF)||Initiation 1 day after birth; feed from 1 breast only 3 times daily||Mother Mother-in-law (MIL) Friends Health workers||Breast-fed prior children 6 months or until next pregnancy||MIL Friends||MIL Herself cannot breast-feed||Mother works out- side the home; "modern mothers on demand||Desire to appear "modern" because do not breast-feed"|
|2. Diet during pregnancy||Knows there are additional nutri- tional needs; fears large child, difficult birth||Health worker Mother||Previous difficult birth||Husband Her mother Health worker (rare)||Husband Mother Herself||Food is scarce. Mother is ill. Fears expenses||Husband requires food for work|
|3. Women working out- side the home||Traditional Islamic culture: low-status given to women who work||Family Friends MIL||Her mother did not work; mobility restricted||MIL Husband||MIL Husband||Lack of job skills within context of deteriorating eco- nomic conditions, other household responsibilities||Mother believes she must be home to raise her children properly|
|4. Household expenditure on alcohol||Men work for wage and are free to spend their money as they choose||Family Cultural norm||All males are free to socialize and to relieve tension||Friends||Husband||Money spent on alcohol;time spent outside the home||Social demands have priority over household demands|
|5. Growth monitoring (GM)||Judges growth in comparison to other siblings. Expects dif- ferent growth for different siblings (based on pregnancy " experience, sex of child)||Family Friends Mother||Children "look" healthy||Herself Mother||Mother Husband Herself||Mother works, has little time to go to clinic for preventive care, Mother does not perceive a growth problem. Mother has not been instructed in the process of GM||Partial informa tion leads her to believe GM is un necessary in view of other health activities|
To construct intra-household behavioural profiles an analysis of power bases in the home is necessary.
POWER BASES, BARGAINING POWER, AND INTRA-HOUSEHOLD RESOURCE ALLOCATION
In considering patterns of household decision-making and resource distribution there are generational and gender issues at play. Control over resource use, including time, food, and money, is determined by power bases in the home [13, 138]. The term "power bases" refers to the relations between family members, and the relative bargaining power, influence, and respect each member has in determining the use of household resources.
Power bases reflect social status and have traditionally been determined by family structure. Productive versus reproductive roles also define household power relations.
In traditional patriarchal societies, for example, young women have the lowest social status, which preserves patrilineal transmission and lines of inheritance. Though women work inside the home their productive activities are not valued; they gain power with age and through the bearing of children (especially sons) to carry on the line
Mothers-in-law in these societies wield great amounts of power. New wives are of lowest status, and have very little say in household resource use. Abdullah and Zeidenstein [ 1 ] observed that newly-wed women in Bangladesh were expected to do all the heavy and tedious household tasks, while the mother-in-law of the house made all the decisions regarding health, the use of food, and the provision of child care.
A woman's low status in these societies is perpetuated by her economic dependence on her husband and sons, and by the low value ascribed to her time and household work. She, in turn, may be more likely to perpetuate this status ascription through the rearing and socialization of her children: boys will probably be respected for their future economic role, whereas girls will be treated as second class citizens and raised to be non-economically productive in their parental homes and bearers of children in their husband's homes. Socialization will stress compliance, obedience, and a low degree of self-assertiveness, that is, the development of qualities valued by a traditional husband. Resource distribution will favour boys over girls and possibly over women.
In less traditional, matrilineal, or transitional societies, though a woman's status is still viewed in the light of her reproductive role and responsibilities, the dependency pattern is less influential. The balance of power is shifted, for example, in the Caribbean and in many Latin American countries, where marriages are more often consensual  and property transmission is not formally carried on through men. Furthermore, in very poor families, particularly those headed by females, survival concerns may supersede the force of social restrictions on women's mobility, since women must enter the labour force for economic survival. Their productive opportunities rather than their reproductive role become a primary concern.
A woman's decision-making power within the home is traditionally defined by her reproductive role. Her power increases with age, but the norm is perpetuated among her children through family and community socialization processes. In less traditional circumstances, where consensual unions prevail, or where women are less economically dependent on their husbands, their power bases in the home will vary. Although women are the primary socializers and caretakers in matters of health and nutrition, the role and involvement of men should not be overlooked.
INTRA-HOUSEHOLD RESOURCE ALLOCATION AND WOMEN'S ROLES
We have highlighted the roles and status of women to answer the following questions:
- What are the functional consequences of women's social
status on their household decision-making powers, on their
resource distribution practices, and ultimately on their own
health and the health of their family members?
- What other variables besides social status (or societally prescribed beliefs and values) influence decision-making power and resource distribution behaviour?
- What mix or series of interventions can best strengthen or modify the co-variants of behaviour? (See table 2.)
A woman's control over household resources and decisionmaking power has been found to vary in relation to a number of factors. Two key variables are: education and income generation and control.
The role of education in improving bargaining power and intra-household resource allocation is little understood. It is very difficult to untangle the effects of education from those of other social and economic conditions that influence household interaction and decision-making processes. There is evidence, however, that maternal (not paternal) education is associated with decreases in child mortality and improvements in family health [12, 27, 159]. It is also believed that educated women make better use of their time and available resources. Caldwell  postulates that the education of women significantly alters the balance of power in the home, making women less fatalistic and giving them greater confidence to take decision-making into their own hands. An educated woman, he claims, will allocate a larger share of the household's resources when needed to feeding and caring for her children.
Ware , however, argues that there is no direct evidence to indicate a causal sequence of events.
LeVine  and Caliendo and Sanjur  believe that educated women have better attitudes about life and may thus be better child rearers. They are more likely to recognize children's developmental stages and will foresee a greater range of future roles for them. It is unknown, however, if it is education alone or the specific content of the education that is associated with observed improvements.
To assess the contribution of women's education to household decision-making and improvements in child health we must also evaluate the contribution of factors such as the income of the mother's parents, maternal health and nutritional status, and other variables that define the level of social development in a particular society [48,118,155]. There is evidence that literacy and level of maternal education have a greater beneficial impact on child health and mortality in impoverished environments compared to more affluent ones [99, 118]. Maternal education may be related to past or present personal income and socioeconomic status, or it may reflect a level of societal development that also includes improvements in health care, sanitation, social services, and social justice, which are directly responsible for hypothesized improvements in women's bargaining power and improvements in household health status [12, 48, 118, 155].
Drummond  and Freire  argue that the educative process provides important tools for reasoning and initiative-taking, and gives one a sense of power in controlling one's own life. These attributes, they feel, are critical to personal and social development. If we find that education alone, independent of parental income, present income, maternal health, and societal development, can result in improvements in women's status, their decision making power in the home, and family health, we must:
- identify whether it is education itself or the content of
education that results in improvements;
- identify what constraints there are to providing appropriate types of education to improve women's status, decisionmaking power, and the use of household resources (for example literacy training, primary education, or specific health, nutrition, and home economics education):
- take measures to overcome these constraints and ensure that these educational programmes exist and are considered socially acceptable and accessible to women and girls;
- ensure that those persons involved in household decisionmaking, particularly in the distribution of tasks and time, recognize the value of education and take the necessary steps to enable women and girls to take advantage of educational opportunities.
It is important to add that, if education operates independently to improve decision-making and resource use and is capable of generating a broadened perspective of the future, we must not limit our advocacy to educational programmes for women only. Education of boys and men is also of concern. The decision to educate males, however, must not be taken at the expense of education for females, given the observed association between women's education and household, personal, and child health. The following important research questions must be considered:
1. How does education influence a mother's use of available resources, such as food, money, time, health care, and sanitation?
2. How does education influence the way she distributes those
resources? Does education influence a father's distribution?
(a) If educated women are in fact more likely to foresee a greater range of futures for their children, how does this affect their distribution of resources to boys versus girls?
(b) How does it affect the socialization and treatment of each?
(c) Are educated women more likely to educate their daughters and to give them greater exposure to in fluences outside the home?
3. How does education influence behaviour to produce observed
health and nutrition improvements?
(a) Through delayed age of first marriage?
(b) Through improved maternal health or birth spacing?
(c) Through improved health-seeking behaviour?
(d) Through improved bargaining power?
4. Most important, does education operate through employment
and income generating ability to improve acquisition of resources
and control of their use in the home? Does this result in:
(a) A delayed age of marriage?
(b) Improvements in a woman's own health?
(c) Improvements in the health of her family?
Contents - Next