Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Intra-household food distribution among Guatemalan families in a supplementary feeding programme: Mothers' perceptions - Patrice L. Engle and Isabel Nieves


Mothers' views on the appropriate distribution of food within a household, and the relation of these views to actual food distribution, are explored. The subjects were 45 mothers with one child enrolled in a health centre's food-supplementation programme in periurban Guatemala because of low weight for height. In the interviews, most of the mothers did not report believing that more food should be given to the targeted child. The four distribution rules mentioned were giving preference to adults or workers, giving preference to children, giving preference to males, and equality. The latter two rules were significantly associated with actual food distribution within the families: mothers who stated a male preference (N = 13) were likely to give more food to males than those who did not, and mothers who stated a preference for equality (N = 29) gave a relatively higher proportion of food to children. Recommendations for food supplementation programmes are made.

Editorial introduction

This paper, investigating the relationship of mothers' attitudes regarding the appropriate distribution of food within the family and the appropriate use of supplementary food to actual distribution, illustrates the importance of evaluating the beliefs and perceptions of programme recipients and addresses a practical issue that confronts many involved in food-distribution programmes.

This is the first paper the Food and Nutrition Bulletin has received examining discrepancies between the need for dietary energy based on age, sex, and activity and actual intra-household distribution of food since the one by Chaudhury in 1988 [10(2):26-34]. Chaudhury's paper indicated that in Bangladesh the perception of preference being given to males was not borne out by observation of the actual distribution of food within the family. The present paper is particularly useful because it also addresses the issue of the actual use of supplementary food intended for the malnourished child in a family.

The findings are of importance for public health workers responsible for supplementary feeding programmes for poorly nourished children because they show that mothers are seldom informed why they should give the supplementary food to the malnourished child. Such advice may also run counter to the concept of many mothers that supplementary food should be made available equitably to all family members.


Supplementary feeding programmes often fail to achieve the goal of improving targeted children's nutrition status even though the food is distributed properly to households. Interveners worry that the food is sold, given away, or used by non-targeted family members. One of the problems may be that the programme managers have not taken into account the existing investment strategies of the families. Whereas the interveners may assume that the family will give the supplementary food to the targeted child, possibly leaving others hungry, parents may feel that it is more appropriate to distribute the food equally to all, or to those who will contribute most to the well-being of the household. This study explores mothers' perceptions of how food should be distributed and the relationship of these perceptions to their actual food distribution patterns. Because there is little information on spontaneously expressed beliefs about "just" forms of food distribution in Guatemala, the approach here is exploratory.

The subjects were 45 mothers who had one child enrolled in a CARE food-supplementation programme for malnourished children run through the health centres in two towns on the outskirts of Guatemala City. Although these are not shanty towns, poverty is common and child malnutrition is a frequent problem. A companion paper documents the intrahousehold food-distribution patterns from two days of food weighing and diet recall, and discusses the methodology for the food assessment in depth [1].

Recent work has suggested that parents in third-world settings may have different criteria for judging the way resources should be allocated from those of health professionals [2, 3]. Health professionals tend to assume that parents will direct resources to the most undernourished, sickly child because it is the most needy. Cassidy has labelled this the altruistic "activist" perspective [2]. She found that many third-world parents with scarce resources attempt to maximize the well-being of the social unit of the family and allocate resources to those who are more likely to provide a long-term benefit to the group, such as the workers or healthier children with better chances for survival. This approach has been labelled the "adapter" position [2].

These two "world-views" probably imply different perceptions of justice. Social psychologists have described several bases for conceptions of justice which influence behaviour in the United States (discussed in more detail elsewhere [1]). One common perception of justice is based on a sense of equity or deservingness: those who work hardest should receive the most. However, in certain situations, such as a family setting, a criterion of need or a norm of social responsibility may apply [4]; in this case resources would be allocated altruistically to those who are in need [5]. The former concept, called a "contribution rule" here, may be functioning for the adapter families; the latter, called a "needs rule," characterizing the behaviour of the activists, may be assumed by the health centre.

Most studies of intra-family food distribution have examined differences in the adequacy of the diets of children by age and sex, of adult males, and of pregnant and lactating mothers [6]. The contribution rule explains most findings on intra-family food distribution better than the needs rule: older age, male gender, and income-earning status have all been associated with greater food ingestion [1]. Findings for the same 45 families as in the present study suggest that, although there were no differences in food distribution by gender and age, income-earners and wives consumed relatively more of the family's calories and husbands received relatively more of the family's protein; there was no evidence that the child targeted by the health centre for the supplementary food received any more than his/her siblings [1].

Underlying these investigations of within-family variance is the assumption that someone is making decisions about food allocation. Research on decision-making strategies has shown that the process is complex, depending on the domain of the decision [3], the economic circumstances of the family, and even who reports the decisions. It is beyond the scope of this investigation to determine whether the mother was the sole decision-maker about food distribution in these Guatemalan households. Instead, the approach taken was to select only households in which the mother was present and to observe a noon meal in each home to determine who served the foods and how additional food or second helpings were obtained. If the mother was seen to be the major food distributor, it was assumed that she had a role in the food-allocation decisions. This assumption is supported by another study in the same village which indicated that mothers had responsibility for food purchases in 85% of the families [7]. If they purchase the food, they may have a role in the distribution of the food.

It was hypothesized that not all mothers have the same perception of food-distribution rules, even within this homogeneous culture, and that their perceptions are probably different from those of the health centre programme. Finding differences between families is a step toward the research agenda described by Haaga and Mason [6]7 who argued for an examination of "which types of families buffer children and which discriminate against them" (p. 155).

Statements about their principles for distributing food were elicited from the mothers by asking them to respond to a hypothetical food-distribution situation, by asking whether or not they agreed with each of ten statements about how they would distribute food, and by asking about reasons for giving the supplementary foods to the target child. Spontaneous mention of the target child's need for special feeding could also occur in responses about their reasons for serving the supplementary foods to their pre-schoolage children. Presumably, the number of mothers mentioning the special needs of their low-weight child should give an indication of the operation of a needs rule for these children.

Food distribution patterns were measured with a discrepancy score summed over two 24-hour periods. This score is the difference between how much of the family's protein or calories an individual should have received on the basis of his or her RDAs and how much he or she actually received. The measure, controlling for the actual family ingestion and age and gender differences, is described in detail elsewhere [1].



Village description

The subjects were from two Ladino (non-lndian) villages about five miles apart. Both villages lie close enough to Guatemala City that bus transportation is readily available and people can travel to the city daily for work, although the cost is relatively high. In the first village a number of factories employ some of the local men and a very few women. In the second village a number of families are supported by either the husband's or the wife's income on local fincas, or coffee plantations. During the coffee harvest, large numbers of residents of both villages, including children, work in the fields.

Each of the two communities had a centrally located public health centre, with an auxiliary nurse present daily and a physician (equivalent to an intern in the United States) several days a week. Difficult cases could be referred to the regional health network. The supplementary feeding programme was administered by CARE through the health centres.

Supplementary foods were distributed each month, the amount depending on the quantity of food sent by CARE. The foods were non-fat dry milk, cracked wheat, polenta (a corn gruel), and oil - generally two to three pounds of each of the dry foods and half a bottle of oil. Someone in the family, not necessarily the mother, had to pick up the food from the health centre. The auxiliary nurse filled plastic bags with the food with the help of a community volunteer. Mothers were asked to contribute a small amount for the purchase of the plastic bags.


The sample consisted of the mothers from 45 families receiving supplementary foods from the health centre in these two communities. All the families were Spanish-speaking.

To receive the supplement, a family had to have a child between one and five years old who had been identified as being of low weight by the physician at the health centre; the definition of "low weight" depended somewhat on the physician but tended to be based on the Nabarro Thinness Scale. Although the public health nurse attempted to monitor all children through well-baby check-ups, there was no systematic screening in the community. Families brought children they were concerned about to the health centre for care, and children who were brought in for other reasons were screened. The child who was enrolled was monitored every three months. If he or she increased sufficiently in weight, the family was dropped from the supplementary feeding programme.

To be included in the study, a family had to have received supplementary food from the health centre within the past month or to have supplementary food still available. All of those from the first village who qualified and were willing to participate were included and most of those from the second village, to bring the total sample up to the desired size of 45 families. One family in the first village refused to be in the sample and none in the second.

The 45 families included eight women (18%) not currently in a union (married or consensual), slightly higher than the rate typical of these villages. Another eight women reported that their spouses were not earning an income or not contributing to the household. In all, 35% of the households could not count on a father's earnings. Thirty-five per cent of the mothers had not attended school and were illiterate. The median years of education for the mothers were two, and for the fathers three. The average number of members of the nuclear family was 5.16 (SD=1.93), and the average income per family member was Q 32.99 per month (SD = 30.78), approximately US$15 per person per month. Almost half of the mothers (45%) were involved in some kind of income earning activity. The sample tended to be more economically disadvantaged than other residents of the villages.

Dietary patterns

The primary dietary staples were corn, in the form of tortillas, and beans. Additional foods were noodle soups and bean soup with greens. Meat was rarely served. Snack foods tended to be fruit from local trees and tortillas. Because of the poverty of the families, little food was purchased outside the home, but school-age children sometimes bought snacks from vendors on the street; one of the limitations of the study is that such foods were not measured or included in the assessment.


The potential subjects were informed by the health centre personnel that people from INCAP (the Institute of Nutrition of Central America and Panama) would be visiting their homes to see what kinds of food were being eaten. The families were visited initially to obtain their permission to be in the study and were informed that the tester would come on two separate days during one of the subsequent weeks. It was stressed that they were to eat exactly what they had intended to, that their eating habits would not be reported to the health centre. Two separate days were observed in order to eliminate variations in eating behaviours. The families were not told the specific day on which the observer would visit. Because the observers were in the home for four hours or more per day, they became familiar to the families and could clarify the fact that they were not representing the health centre. Since INCAP is well-known in these villages as a research institution, the mothers were likely to understand the distinction between the health centre and the observers.

Instruments and variables

Anthropometric measurements

Anthropometric measurements of all children five years old and under were taken during the visits. There were 18 siblings for whom both dietary and anthropometric measurements were available. Height was measured with a horizontal board for children under two, and with a vertical board for those over two. All weights were taken with a Salter scale. Weight and height measurements were converted to Z scores using the US National Center for Health Statistics (NCHS) norms [8]. Anthropologists were trained at INCAP to acceptable levels of reliability.

Diet variables

The collection of dietary data is described in detail elsewhere [1]. The two measures used here were the discrepancy score for protein and the discrepancy score for calories. These discrepancy scores are the adequacy of each individual's diet relative to the average adequacy of the diets of all family members (based on the recommended dietary allowances, RDAs, for each person according to age and gender)-that is, they indicate the discrepancy between the percentage of the total family's food the individual actually receives and what he or she would receive if all members received amounts proportionate to their relative needs according to dietary recommendations.

Mealtime behaviours

Observations of feeding behaviour were made at the noon meal in all the households in order to assess who distributed the food; whether children asked for more; if they received more whether it was requested, offered, or simply taken; and who took the initiative in beginning the meal.

Distribution rules

At the end of the second day the mother was interviewed about her attitudes and beliefs about food distribution and her opinions of the supplementary food provided at the health centre and was asked her reasons for feeding the target child the supplementary foods. The mothers' distribution rules were determined in two ways: by an open-ended question to see what kinds of distribution strategies they would mention spontaneously, and by a series of ten agree/disagree statements designed to tap attitudes about feeding. The open-ended question was: "If you were given three eggs, how would you distribute them among the members of your family?" In the series of statements (table 1), two items measured each of five hypothesized distribution rules: (1) children should get more than adults (child preference); (2) adults/workers should get more than others (adult/worker preference); (3) boys should get more than girls (male preference); (4) all family members should receive the same amount (equality preference); and (5) individuals should receive what they ask for (demand preference). The respondent was asked to either agree or disagree with each statement. The distribution rules used in the final analysis were derived from both kinds of questions.

TABLE 1. Preference statements and the percentage of mothers agreeing with each (N=45)

Child preference
Small children need more expensive food
than adults.
When there is not enough, adults can handle
being hungry better than children.



Adult/worker preference
Those who work more need more food than
those who work less.
It is more important for the adults, who sus
tain the family, to eat well than for the
small children.



Male preference
Boys need to eat better than girls of the same
If I have one piece of bread and two children
who are hungry, a boy and a girl. I prefer
to give it to the boy.



Equality preference
I want to give an equal amount of food to all
members of the family.
When someone gives me extra food, I will
give an equal piece to each member of my



Responding to demand
The one who asks for more receives more.
If I have something for another member of
the family and a child cries, I prefer to give
it to that child rather than to the other, who
remains hungry.



Data analysis

The results include a description of the major kinds of distribution rules that the 45 mothers articulated and the association between these distribution rules and actual patterns of observed food distribution within the family. T-tests compared the adequacy of protein and calorie ingestion and the family members' discrepancy scores with the mothers' distribution rules.


Observations of mealtime behaviour

The mother served the food in 87% of the families. In these cases, she was present during the entire visit, prepared the food, and supervised feeding the children. In the five cases in which the mother was not present all the time, an adult relative (grandmother, father, older sibling) or a paid helper was responsible for primary care of the children, food preparation, and food service during some period of the visit.

The families tended to eat three meals a day, with the midday meal the main source of calories. Because of limitations of seats and serving utensils and the complexity of schedules, few of the families ate together. The mother or caregiver served plates from the main pot and presented them to each family member. Fathers tended not to be present at the midday meal, but food was put aside for them when they returned. The meal was most often initiated by the server: of the lunches observed, 55% were initiated by the caregiver; in 45% of the cases the child requested food before it was served.

The general pattern was for the caregiver to serve the children separately, then continue with a chore or activity in close proximity to the child (82% ). Most of the mothers did not eat with the child (72%). For younger children, the mother fed the child (32% of meals). In only one case did a child take more food from the family pot without asking. Almost half of the children requested more food, and 91% of those received it. However, only 22% of the caregivers offered the child more food.

Reasons for feeding decisions about the target child

When the mothers were asked why they fed the target child the donated foods, the most common responses were "custom" and "the child likes the food." Only two mothers (4%) spontaneously mentioned low weight as a reason for feeding the target child a supplementary food.

All the mothers correctly reported which child was targeted. Why, then, did they mention the specific needs of the target child for extra food so infrequently? Three possible explanations were explored: (1) the target child may have been no more malnourished than other children in the family; (2) the health centre personnel who distributed the food may not have explained the special needs of the target child; or (3) some of the mothers may have had their own investment strategy which was different from the needs rule of the health centre.

The target children were significantly more wasted (weight for height) than a group of siblings under five years old but did not differ in height for age or weight for age [1]. The target children were also compared with a sample of 239 children 8-35 months old collected randomly in one of the two villages [9]. When the children were categorized into those weighing below -2 SD of the standard and those below -1 SD, it appeared that the target children were only slightly more likely to be malnourished than the sibling group or the random sample group (fig. 1). Thus one of the reasons for the mothers' failure to mention the target child's low weight may have been that the these children did not appear to be very different from other village children.

The mothers reported little instruction by health centre personnel to target the malnourished child. Almost half of the mothers (42%) reported receiving no information from the doctor about the food, and 32% reported receiving none from the nurse. Only 5% of the mothers said that the foods were "special to give to the child" or that the "food was for the malnutrition of the child." It is, of course, possible that instructions were given but not remembered because they were inconsistent with the mothers' beliefs.

The most common information received from the doctor was that the food increased weight and had vitamins. From the nurse, the most common information was how to prepare the food and that it had vitamins. (Consciousness of the value of vitamins was also seen in ethnographic observations in the villages: when an ambulatory peanut salesman hawked his product-identical with many others-as "vitamin-enhanced," it was obvious that vitamins had made it into public awareness.)

FIG. 1. Mean nutritional status of target. sibling, and comparison groups (children 1-5 years old; NCHS norms)

The instructions that were heard may have depended on the food. Milk was more likely to be perceived as special for the child: five of the mothers (11%) reported that the doctor had told them that the milk was for the malnourished child. Milk was also more likely to be given to the target child than the other two foods, polenta and cracked wheat: almost half of the mothers observed gave milk to the target child, compared to approximately 20% for the other two foods. (Ingestion of supplementary oil was not recorded, since it tended to be mixed with other oils.)

The third possibility noted above is that the mothers believed in a distribution strategy different from, and inconsistent with, that of the health centre. For example, a belief that food should be distributed equally would conflict with the strategy of giving more food to one particular child at the expense of another.

Mothers' distribution rules

Responses to the open-ended question

Four kinds of answers were given to the open-ended question: (1) distribute the food equally (16 mothers, 35%); (2) give more to the parents (13, 29%); (3) give more to the husband, less to the wife (7, 16%); and (4) give more to the children (9, 20%). Some of the responses seemed to reflect a contribution rule (give more to the husband, or give more to the parents), whereas others were closer to a needs rule (give more to the children). The highest proportion of the mothers stated a rule that had not been predicted: 35% favoured equality, giving an equal portion to every family member. Because children have less nutritional need than adults, this rule would tend to result in children receiving more food proportionally than adults, a version of a needs rule.

None of the mothers spontaneously mentioned giving more food to the target child, and none mentioned gender as the basis for distribution.

Agree/disagree statements

Table 1 shows the frequency of agreement with each of the ten agree/disagree statements. Some of the items were not used to categorize the mothers because there was insufficient variability in the responses or too many "don't know" responses. Almost all the mothers (91%) reported that they would respond to a child's request for food, and 98% agreed with the statement "Adults can handle hunger better than children." Three mothers could not answer the item "If I have something for another member of the family and a child cries, I prefer to give it to that child rather than to the other, who remains hungry," so it was not used.

Relations between the groups of items designed to be related were tested using chi-square analyses. The two equality-preference items were significantly associated (chi square = 9.38 [N = 45, df = 1], p < .002), and the two malepreference items were marginally related (chi square = 3.27 [N= 42, df = 1], p < .07). The items in the worker- and child-preference categories were not related, raising questions about the meaningfulness of these categories.

Four categories of responses were identified, roughly those that emerged spontaneously with the open-ended question. These were combined with the seven remaining questions to categorize the mothers as agreeing or disagreeing with four distribution rules: adult/worker preference, male preference. child preference, or preference for equal distribution. The first two reflect a contribution rule and the last two a needs rule. Each mother was categorized as agreeing or not agreeing with each distribution rule. The system for categorization helped to distinguish broad groups of responses but is not psychometrically adequate.

A woman was coded as expressing adult/worker preference if, on the open-ended question, she said that she would give more egg to an adult, or to her husband, and if she agreed with statement "It is more important for the adults, who sustain the family, to eat well than for the small children" (N= 23, 51%).

A woman was coded as expressing male preference if she agreed with either of the statements "Boys need to eat better than girls of the same age" or "If I have only one piece of bread and two children who are hungry, a boy and a girl, I prefer to give it to the boy" (N= 13, 29%).

A woman was coded as expressing child preference if she said that she would give more egg to children in the open-ended question, or if she agreed with the statement "Small children need more expensive food than adults" (N= 26, 58%).

If a woman reported that she would divide the egg equally among all family members in the open-ended question and agreed with the statement "I want to give an equal amount to all members of the family" or "When someone gives me extra food, I will give an equal piece to each member of my family" (N = 29, 64%), she was categorized as expressing equal preference.

FIG. 2. Mean discrepancy scores of males and husbands as a function of mothers' expression of male preference. Levels of significance of differences between preference groups: *p < .05; **p < .01.

FIG. 3. Mean discrepancy scores of adults, children under 18 years old, and target children as a function of mothers' expression of equal preference. Levels of significance of differences between preference groups: *p < .05; **p < .01.

Relation of distribution rules to family feeding

To assess the relation of the mother's distribution rules to family food distribution patterns, l-tests were performed comparing the calorie and protein discrepancy scores for particular subgroups of family members (e.g., males) as a function of the mother's stance on each of the four distribution rules. Two of the distribution rules - the male-preference and the equality-preference rules-appeared to be associated with food distribution patterns within the household as measured by the average discrepancy scores.

Male preference

To determine whether the mothers who expressed a preference for feeding males actually fed them more, the average discrepancy score for all males in their families was compared to the average discrepancy score for all males in families in which the mother did not express a male preference. Only males were included in this analysis. Mothers who expressed a male preference (N = 13) were significantly more likely to give proportionally more protein and calories to males than those who did not (t = 2.42, p < .02 for calories; t=2.85, p<.007 for protein-see fig. 2). They were also likely to give more protein (but not calories) to their husbands than were women who didn't state a male preference (t = 2.37, p < .02). The other three preferences were unrelated to the male discrepancy score.

Equality preference

To examine the relationship between equality preference and food distribution patterns, five different kinds of family members were compared: all adults, all males, husbands, all children under 18 years old, and the target child. Figure 3 shows these discrepancy scores for mothers who expressed a preference for equal distribution and those who did not.

For calories, there were no significant differences between those in families with equality-preference mothers and those in families with non-equality-preference mothers for any subgroup. For protein, however, adults in families with equality-preference mothers (N= 29) ingested a significantly smaller proportion of the family's protein than those in families with non-equality-preference mothers (At= 15) (t=3.09, p<.003). Children under 18 in families with an equality preference mother ingested a significantly larger proportion of the family's protein than those in families with a non-equality-preference mother (t = 2.97, p < .005). However, data presented in the companion paper [1] indicates that the adults in both groups were ingesting at least 100% of the protein that they theoretically required.

The husbands of mothers who expressed a preference for equal feeding also received proportionally fewer calories (t = 1.98, p < .05) and somewhat less protein (t = 1.82, p < .08) than those of mothers who did not express this preference, although it is important to note that these husbands were still receiving the proportional amount that they should (the mean discrepancy score for calories was -0.014, and for protein 1.90); it is simply that, compared to others, they were receiving their fair share, not any extra.

Other associations

Mothers who expressed an adult/worker preference did not show any indication of giving more food to adults than did those who did not have this preference. Nor was child preference associated with feeding patterns.

The final question concerns what kinds of mother preferences are associated with target children's ingestion of a relatively higher proportion of the family's calories or protein. Target children whose mothers expressed an equality preference received a higher proportion of the family's protein than those whose mothers who did not (t=2.19, p<.03). In part, this effect may be a function of the mother's tendency to give food to all of the children rather than to specifically target the malnourished child; these mothers tend to feed all of their children a higher proportion of the families' protein than the non-equality-preference group. No other distribution rules were significantly associated with feeding the target child.

The associations between socio-economic variables and the expression of the malepreference and equality rules were examined with tests. Women who expressed male preference tended to have illiterate husbands (chi square = 3.11, df = 1, p < .08) more often than women who did not. There was no relationship between the expression of the equality rule and mothers' and fathers' education, literacy, income, quality of the house, or income per person.

Discussion and conclusions

Most supplementary feeding programmes that target a specific child as malnourished assume that the food distributor, usually the mother, will use the extra food to enhance the diet of the undernourished child. They assume that mothers will behave in accordance with a needs rule, which states that the most needy (in altruistic terms) members of the family should receive a higher percentage of scarce food.

In a sample of 45 families enrolled in a supplementary feeding programme, most of the mothers did not report that they fed the "malnourished" child any extra food. This lack of targeting may have been due in part to inadequate instruction at the health centres. Only a little over half of the mothers were aware that they were supposed to give more food to the target child. This response appeared to be somewhat related to the food: the mothers were more likely to perceive milk as a special food for the malnourished child than either wheat or polenta. A second reason for this lack of targeting may be that the target children may not have looked much more malnourished than other children, since malnutrition is fairly common in these communities. A third reason may be that the mothers held different distribution rules from those assumed by the health centre personnel. Sixty-five per cent of the mothers described a norm of equality-that is, that food should be distributed equally to all family members. This concept of a just distribution would be in direct conflict with a strategy of giving more to one particular child at the expense of others.

There are several reasons for the frequency of these equality judgements. Equality as a basis for justice has been found to be the norm in some noncapitalistic cultures [10]. In more communally oriented cultures, such as China and India, people have been found to favour need or equality rather than equity as criteria for justice [11, 12]. It is also more likely to be a criterion for perceived justice in a family than in a public context [4]. Finally, women are more likely than men to use equality rather than equity or merit as the basis for allocating rewards in laboratory situations [13]. The lack of association between the equality rule and common socio-economic variables might reflect the limitation of the sample size, or the expression of this rule might be a function of unmeasured variables such as religious attitudes.

The absence of targeting might be due to the mother's perceptions of the reason for the target child's malnutrition. Research in the United States on altruism has suggested that people are much more likely to help those who they believe are in need because of circumstances outside their control (e.g.. natural disasters); they are less likely to offer help if they believe that the person's need comes from lack of effort, laziness, or lack of foresight [14-16]. In other words, the attribution of cause can influence the likelihood of an altruistic response. If mothers attribute malnutrition to a quality in the child (the child is weak, sickly, does not want to live), they may be less motivated to make the extra effort than if they believe that the child is a victim of difficult circumstances (e.g., someone took care of the child inadequately). Work in Brazilian shanty towns has illustrated cases of mothers who attributed illness and malnutrition to the child's characteristics, or to God's will, and made little effort to help the child [3]. Such attributions might differ from child to child and might be important for the health centres to investigate during the intervention.

Two of the distribution rules expressed by the mothers were significantly associated with the actual distribution of food within the household: equality and male preference. Neither adult/worker preference nor child preference was related to food distribution. Male preference was expressed by a relatively small percentage of the mothers, which is not surprising given the relative lack of explicit male reported in most Latin American settings [17]. Male preference and illiteracy among fathers were marginally associated. This effect could be a function of the greater physical activity of illiterate fathers, who would be more likely to work in agriculture. However, this feeding bias was seen among young as well as adult men, suggesting that it is not simply a reflection of higher activity levels. There may be more traditional sex-role attitudes in these less-educated families.

One implication of these findings is that it might be possible to substitute questions to the mother about how she plans to distribute food to discover intra-household processes without the extensive dietary analyses performed here. A second implication is that health centre personnel might determine the mother's investment strategy and incorporate that information into their instructions about how to use the food. They might recognize that there are individual differences in these investment strategies and adapt their information accordingly. Third, they might specifically examine the mother's beliefs about the cause of the malnourished child's condition. Since the most common response among the mothers was an equality preference, a fourth implication is that supplementary feeding programmer might target a family, rather than a child, for supplementary foods. In this way, they would build on existing belief systems rather than trying to teach the mothers a new one.


The research reported here was supported in part by a USAID grant to INCAP for the study of food supplementation programmer, and in part by a Fulbright research grant to the first author. The authors wish to express their appreciation for the support of the Guatemalan research team (Lic. Marta Amanda Barrera, Lic. Helen de Ramirez, Blanqui Suceli, and Connie Padilla) and the assistance of Dr. Martin Immick, Dr. Edmundo Alvarez, Elena Brineman, Lica. Alex Praun, and the health centre staff and families in the two research villages. The help of Shanna Kinser in all phases of editing the paper is appreciated.


1. Engle PL, Nieves 1. Intra-household food distribution among Guatemalan families in a supplementary feeding program: behavior patterns. Soc Sci Med (in press).

2. Cassidy CM. World-view conflict and toddler malnutrition: change-agent dilemmas. In: Scheper-Hughes N. ed. Child survival. Dordrecht, Netherlands: D. Reidel, 1987:293-324.

3. Scheper-Hughes N. Culture, scarcity, and maternal thinking: mother love and child death in northeast Brazil. In: Scheper Hughes N. ed. Child survival. Dordrecht, Netherlands: D. Reidel, 1987:187-210.

4. Deutsch M. Distributive justice: a social psychological perspective. New Haven, Conn, USA: Yale University Press, 1985.

5. Leventhal GS. What should be done with equity theory? new approaches to the study of fairness in social relationships. In: Gergen KJ, Greenberg MS, Willis RH, eds. Social exchange: advances in theory and research. New York: Plenum, 1980:27-35.

6. Haaga JG, Mason JB. Food distribution within the family: evidence and implications for research and programmes. Food Pol 1987;12(2):864-916.

7. Engle PL. Influences of mother's and father's income on children's nutritional status in pert-urban Guatemala. Soc Sci Med (in press).

8. Jordon M, Staehling N. Anthropometric statistical package, version 3. Atlanta, Ga, USA: CDC, 1986.

9. Engle PL. Mother's work and child care strategies in peri-urban Guatemala: nutritional effect. Child Devel 1991 ;63:95465.

10. Sampson EE. On justice as equality. J Soc Issues 1975;31(3):4564.

11. Leung K, Bond MH. The impact of cultural collectivism on reward allocation. J Pers Soc Psych 1984;47:793804.

12. Murphy-Berman V, Berman JJ, Singh P. Pachauri A. Kumar P. Factors affecting allocation to needy and meritorious recipients: a cross-cultural comparison. J Pers Soc Psych 1984;46:1267-72.

13. Major B. Adams JB. The role of gender, interpersonal orientation, and self-preservation in distributive justice behavior. J Pers Soc Psych 1985;45:598-608.

14. Barnes RD, Ickes W. Kidd RF. Effects of the perceived intentionality and stability of another's dependency on helping behavior. Pers Soc Psych Bull 1979; 5:367-72.

15. Gruder CL, Romer D, Korth B. Dependency and fault as determinants of helping. J Exper Soc Psych 1978;14:227-335.

16. Meyer JP, Mulherin A. From attribution to helping: an analysis of the mediating effects of affect and expectancy. J Pers Soc Psych 1980;39:201-10.

Contents - Previous - Next