Biocultural correlates of child nutrition and growth and development in Costa Rica

Material Information

Biocultural correlates of child nutrition and growth and development in Costa Rica
Simpson, Sharleen Hirschi ( Dissertant )
Lieberman, Leslie S. ( Thesis advisor )
Bernard, H. Russell ( Reviewer )
Doughty, Paul L. ( Reviewer )
Dinning, James S. ( Reviewer )
Safa, Helen I. ( Reviewer )
Place of Publication:
Gainesville, Fla.
University of Florida
Publication Date:
Copyright Date:
Physical Description:
vii, 182 leaves : ill., map ; 28 cm.


Subjects / Keywords:
Child psychology ( jstor )
Children ( jstor )
Diseases ( jstor )
Families ( jstor )
Food ( jstor )
Infants ( jstor )
Malnutrition ( jstor )
Mothers ( jstor )
Nutrition ( jstor )
Nutritional status ( jstor )
Anthropology thesis Ph. D
Child development -- Costa Rica -- San Jose ( lcsh )
Children -- Growth -- Costa Rica -- San Jose ( lcsh )
Dissertations, Academic -- Anthropology -- UF
Nutrition -- Psychological aspects ( lcsh )
bibliography ( marcgt )
non-fiction ( marcgt )
Spatial Coverage:
Costa Rica -- San José


Groups of low (n=19) and normal (n=25) weight for length children were selected from an initial random sample of 107 children from poor neighborhoods in San José, Costa Rica. These children were all born in 1980 and were between the ages of 1 and 2 years at initial measurement. They were than measured monthly for six additional months. Data were also collected about socioeconomic status, food habits, maternal reproductive history, medical history of the child, and food buying patterns. The Denver Developmental Screening Test (DDST) was administered and a clinical assessment of nutritional status was also made. In-home observations of child-rearing activities were made during which time nurturing or mothering behaviors were recorded. No significant differences between the two groups are reported with respect to total income, food costs, household size, environment, children under 6, birth order, or household food frequency. The matrifocal extended family was found to be an effective survival strategy for poor families. Low weight/length children were breast-fed slightly longer with later introduction of bottle feeding and solid foods and obtained slightly more abnormal or questionable scores on the DDST than normal weight/length children. Families of low weight/length children owned more material goods, spent more money on food, and tended to be less organized in their habits than were families of normal weight/ length children. By use of multivariate regression and discriminant function analyses a model was constructed to predict membership in the low or normal weight/length group. Birth weight, the number of children aged 6 years and under, and other variables representing maternal competence and economic managing abilities as well as material style of life were found to be significant predictors of low or normal weight for length status. Mothering or nurturing behaviors were found to be positively associated with height, weight, and growth in the low weight/ length group. The maternal factor, which includes adequate prenatal care and nutrition, as well as maternal competence, is suggested to be of great importance in determining whether low income families have normal or low weight/length children.
Thesis (Ph. D.)--University of Florida, 1984.
Bibliography: leaves 134-145.
General Note:
General Note:
General Note:
English and Spanish.
Statement of Responsibility:
by Sharleen Hirschi Simpson.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
000476231 ( alephbibnum )
11750803 ( oclc )
ACP2485 ( notis )


This item has the following downloads:

Full Text







The data for this dissertatation were gathered between September

1981 and December 1982 under the auspices of a grant (#BNS8104679) from

the National Science Foundation. I would like to thank all personnel

from the Costa Rican Ministry of Health who aided in the data gather-

ing. In particular, I wish to thank Dr. Carlos Diaz Amador, Dr. Enrique

Freer Miranda, Dona Evelyn Jaensthke Maglakin, and Doia Juanita

Alvarez Elizondo de Quiros. Dr. Jose Antonio Camacho was very helpful

in advising me about the urban situation. To Flory Desanti Jimenez and

Eulalia Obando Naranjo, who as former health auxiliary workers were of

invaluable assistance, I give deepest thanks. Appreciation is also

due Dr. Leslie S. Lieberman, chairman of my supervisory committee, for

her support and comments, and the other members of my committee,

Drs. H. Russell Bernard, Paul L. Doughty, James S. Dinning, and Helen I.

Safa. A special thanks goes to Ron Thomas for his patience in advis-

ing me in the statistical analysis. I wish also to thank my parents,

Donna and LaVell Hirschi, for giving me a great start in life and pro-

viding continuing emotional support. To Roderick and Randy Simpson I

give thanks for being extremely supportive and loving to a frazzled

mother trying to finish a dissertation, and for adapting so well in

Costa Rica. Most of all, I wish to thank those Costa Rican families

who with great kindness and patience allowed me to make innumerable

interviews and observe them in action. A final thanks goes to the chil-

dren; for them I have great love and respect.




ACKNOWLEDGMENTS . . . . . . . . ... ...... .ii

ABSTRACT . . . . . . . . ... . . .... .vi


THE SITUATION IN COSTA RICA . . . . . .... .. 1

Causal Factors Related to Malnutrition . . . . 4
Socioeconomic Factors . . . . . . . . 6
Changes in Child-Feeding Practices . . . . . 8
Changes in Family Structure and Women's Roles . . 9
Economic Development as a Contributing Factor .... 10
The Costa Rican Situation . . . . . .... .13

II STATEMENT OF THE PROBLEM . . . . . . .... .25

III METHODS . . . . . . . ... . . . .33

Phase I: Setup . . . . . . . .... . 33
Contacts . . . . . . . .... . .. .33
Archival Research . . . . . . . ... .33
Interviewers . . . . . . . .... . 34
Pretest . ............ 34
Phase II: Initial Survey Target Population and Samp-
ling Design . . . . . . . . .. . . 35
Anthropometry . . . . . . . .... . 38
Denver Development Screening Test ......... 39
Maternal Level of Living: Quality of Environment 40
Assessment of Clinical Signs of Malnutrition . . 41
Phase III: In-Depth Longitudinal Study . . ... 41
Long-Term Data Gathering . . . . . . . 43
Assessment of Socioeconomic Status . . . ... 44
Weekly Marketing Items and Food Budget ...... 44
Maternal Reproductive History . . . . .. 45
Dietary Data . . . . . . . . . 45
Q-Sort Techniques ..... ... . . . . . 46
Observation of the Index Child . . . . .. 47
Fees Paid to Subjects . . . . . . . . 49
Phase IV: Data Analysis . . . . . .... .50



Family Networks. .... . . . . . . . 54
Family Structure . . . . . . . . ... .55
Food-Buying Patterns . . . . . . .... .60
Child Rearing . . . . . . . . ... . 62
Family Interaction and Child Development ...... 65
Child Health . . . . . . . .. .. . 68
Diet and Food Habits as They Relate to Children . .. 70

V RESULTS . . . . . . . . ... . . . .77

Anthropometric Data . . . . . . . . . 77
Growth Patterns . . . . . . . . . . 79
Maternal Reproductive History and Birth Weights . .. 84
Illness Classification . . . . . . . . 85
Psychomotor Development . . . . . . .... .87
Socioeconomic Factors . . . . . . .... .87
Dietary Data . . . . . . . . ... . 99
Food Frequency . . . . . . . . . 99
Breast-Feeding Patterns . . . . . . .. 102
Attitudes About Food .......... ... 102
In-Home Observation of the Index Child . . ... 107
Multivariate Analyses . . . . . . . . 114
Stepwise Multivariate Regression . . . ... 116
Stepwise Discriminant Analysis . . . . .. 116


Socioeconomic Factors . . . . . . .... 122
Dietary Factors . . . . . . . . . . 125
Conceptual Model Describing Children at Nutritional
Risk in the Urban Area . . . . . . .... .126
Conclusions . . . . . . . .... .. .. 128

REFERENCES . . . . . . . . ... .... . . 134


1 FAMILY REGISTRY . . . . . . . . ... . 147








6 SYMPTOMS OF MALNUTRITION . . . . . . ... 160

7 SOCIOECONOMIC STATUS . . . . . . . . 162




COSTA RICA . . . . . . . ... .. . . 175

BIOGRAPHICAL SKETCH . . . . . . . ... .. .. . 182

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of
the Requirements for the Degree of Dcotor of Philosophy



Sharleen Hirschi Simpson

August 1984

Chairperson: Leslie S. Lieberman
Major Department: Anthropology

Groups of low (n=19) and normal (n=25) weight for length children

were selected from an initial random sample of 107 children from poor

neighborhoods in San Jos6, Costa Rica. These children were all born

in 1980 and were between the ages of 1 and 2 years at initial measure-

ment. They were than measured monthly for six additional months.

Data were also collected about socioeconomic status, food habits,

maternal reproductive history, medical history of the child, and food

buying patterns. The Denver Developmental Screening Test (DDST) was

administered and a clinical assessment of nutritional status was also

made. In-home observations of child-rearing activities were made dur-

ing which time nurturing or mothering behaviors were recorded. No

significant differences between the two groups are reported with respect

to total income, food costs, household size, environment, children

under 6, birth order,or household food frequency. The matrifocal

extended family was found to be an effective survival strategy for

poor families. Low weight/length children were breast-fed slightly

longer with later introduction of bottle feeding and solid foods and

obtained slightly more abnormal or questionable scores on the DDST than

normal weight/length children. Families of low weight/length children

owned more material goods, spent more money on food, and tended to be

less organized in their habits than were families of normal weight/

length children. By use of multivariate regression and discriminant

function analyses a model was constructed to predict membership in the

low or normal weight/length group. Birth weight, the number of chil-

dren aged 6 years and under, and other variables representing maternal

competence and economic managing abilities as well as material style of

life were found to be significant predictors of low or normal weight

for length status. Mothering or nurturing behaviors were found to be

positively associated with height,weight, and growth in the low weight/

length group. The maternal factor, which includes adequate prenatal

care and nutrition, as well as maternal competence, is suggested to be

of great importance in determining whether low income families have

normal or low weight/length children.


The incidence of undernourished children in the developing world

is known to be high. The Commission on International Relations of the

National Research Council (1977:63) cites FAO estimates that as many

as 30% of preschool-aged children in low-income countries suffer from

second- or third-degree malnutrition, and an additional 40-50% have

first-degree malnutrition based on weight for age standards such as

those set by Gdmez (1956). Jelliffe (1968) notes that the children most

at risk nutritionally are between the ages of 6 months and 3 years.

This observation has also been verified in nutritional surveys con-

ducted in Africa (Khan and Gupta 1979; Omololu 1978), in Asia (Berg 1973;

Winikoff 1978) and in Latin America (Home et al., 1977; M6nckeberg

1976; Schrimshaw and B6har 1976). At this age breast-feeding no longer

meets the child's needs. Weaning usually takes place during this time;

the diet of adults is not adequate for the child; and the infant is more

vulnerable to infections (Jelliffe 1968). Malnutrition may occur during

or immediately after episodes of infectious disease, even though the

child may have been in a state of adequate nutrition before and growing

at a normal rate (Mata 1978b). Field studies in India and Guatemala have

identified the interaction between malnutrition and infectious disease

(Mata 1977; Wyon and Gordon 1971). Mortality in nearly all types of

infectious disease is greater in nations with prevalent malnutrition


because of a decreased ability among debilitated or malnourished chil-

dren to fight off illness (Schrimshaw et al. 1968).

More recently attention is turning to what has been called

"invisible malnutrition." According to international health experts,

the image of the starving baby is too often used to represent the

developing world. Visible and obvious malnutrition occurs relatively

rarely, mostly during times of famine and war. Invisible malnutrition,

on the other hand, affects about one-fourth of the developing world's

children, stealing their energy, lowering their resistance to disease,

and thus retarding growth (International Health News 1983:1). Noting

that the child's first reaction to the lack of energy intake is to re-

duce energy output, it is observed that by conserving health and

growth at the expense of activity, the child can maintain a normal ap-

pearance. Studies have shown that "even children who are regularly

eating only three-fourths as much food as they actually need can still

maintain weight and growth by cutting out discretionary activity"

(International Health News 1983:2). The report further notes that the

mutually reinforcing relationship between invisible malnutrition and

infection is responsible for the majority of the 40,000 deaths every

day among the developing world's infants and children.

Undernutrition of the preschool child has been found to have

substantial long-term effects on the subsequent ability of the child

to develop, to be educated, and otherwise to function as a useful mem-

ber of society. Marcondes et al. (1973) found that children with his-

tories of severe malnutrition have poor performances on the Gesell even

after being rehabilitated. The Gesell test, developed by Arnold Gesell

and Catherine Amatruda (Knoblock and Pasamanick 1974) is a timetable

characterizing infant development during the first years. It includes

the categories of physical, gross motor, fine motor, sensory, vocaliza-

tion, and socialization. Richardson et al. (1972) and Cravioto and

DeLicardie (1968, 1976) found that children who showed evidence of

undernutrition in earlier years were less able to adapt and perform in

school. Cravioto and Robles (1963) found the area of hearing and speech

to be most retarded in the long run. Yatkin and McLaren (1970), compar-

ing two groups of infants recovering from severe marasmus, found that

stimulated children showed greater improvement than did the unstimulated

group. Five mental functions comprising the development quotient given

with the Griffiths Mental Development Scale were evaluated. These in-

cluded the areas of locomotor, personal-social, hearing and speech, eye

and hand, and performance. Neither group, however, attained the normal

quotients by the end of the rehabilitation.

The volume edited by Greene (1977) documents the long-term dele-

terious effects of malnutrition on the learning capacities and conse-

quent socioeconomic development of whole societies, particular socio-

economic strata (i.e. peasants), and ethnic groups (i.e. Ladinos)

throughout the world. Thus, the problem is an important one, the ef-

fects of which are potentially very harmful to the well-being of a

nation, particularly a developing nation (Aranda-Pastor 1975; Ashworth

and Picou 1976; Berlin and Markell 1977; Burgess et al. 1972; Jelliffe

1966; Seth et al. 1979; Villarejos et al. 1971; Wray and Aguirre 1969).

Although the existence of undernourished and malnourished chil-

dren has been documented in developing countries the world over, and

many important sociocultural variables have been identified, until re-

cently there has been little work done at the household level on the

behavioral, attitudinal, and habitat correlates of malnutrition. Most

of the work which has been done has taken the form of cross-sectional

surveys rather than longitudinal in-home observations. Surveys, how-

ever, often do not uncover the causal chain of events leading to con-

sumption-related malnutrition.

Causal Factors Related to Malnutrition

Nutrition workers generally agree that malnutrition in developing

countries is the result of many interrelated factors including low in-

come and underemployment, ignorance, poor sanitation, lack of access to

medical facilities, family instability (Beghin et al. 1979), changing

women's roles in relation to urbanization and participation in the work

force (Popkin 1980a; Uyanga 1980), and family structure (size, birth-

spacing, one-parent families, etc.).

Taylor and Taylor (1976) constructed a model which, according to

them, shows the interrelationships of factors that cause most of the

malnutrition in the world. They group causal factors under three head-

ings: the production of food, its distribution, and its utilization.

Under production they include factors related to agricultural labor,

distribution of land and technology which affects the level of produc-

tivity, thus determining the supply of food available. Under distribu-

tion are included economic factors (prices and income); demographic

factors (population growth and urbanization); cultural factors (beliefs

about food, childcare and feeding, social status); and health and

nutrition services (feeding programs, integrated services). Under


utilization are included physiological differences (vulnerable groups,

e.g. pregnant women and children; malabsorption, genetic adaptation),

level of activity, and infection and parasites. Taylor and Taylor

note, however, that while they regard these as being general causes

they must be adjusted to local situations. Similar patterns of mal-

nutrition may result from different combinations of causes. Thus,

they advocate nutrition surveys which go beyond standardized descriptive

data to include information which will permit the development of a local

causal profile.

Brown and Brown (1977), in their paper giving guidelines for

finding the causes of protein-calorie malnutrition in the community,

suggest that adequacy of household food supplies, availability of food

supplies (agriculture), food procurement (purchases),and food use and

feeding practices are useful categories for evaluating the local situa-


Fleuret and Fleuret (1980:250) suggest the following circumstances

leading to what they term "consumption-related malnutrition": (1) Food

production is inadequate due to lack of land, labor, capital, or any

one of these; (2) food production is adequate but some people cannot

afford enough food or the right kinds of food; (3) food production is

adequate but cultural factors (e.g. food preferences, intrahousehold

distribution) cause unhealthy consumption patterns; (4) both aggregate

food production and overall income levels are adequate; beliefs and

values are nutritionally neutral, but certain categories of the popula-

tion are constrained by other social and economic factors to make con-

sumption decisions that are inconsistent with good nutrition. Of these

approaches, the latter seems to be the most useful in terms of develop-

ing countries.

Reutlinger and Selowsky concur with this latter set of circum-

stances, noting that "at the global level malnutrition is the inequi-

table distribution of world income and not the result of an insuffi-

cient supply of food" (1979:21).

Socioeconomic Factors

Keeping in mind general guidelines and models, one can find a

number of specific studies which illustrate in detail some of the

causal factors included in the general framework already mentioned.

Rawson and Valverde (1976) in San Ram6n, Costa Rice, found that

access to land, fathers who worked as day laborers, working mothers,

physical condition of the house, and the number of children under 6

years of age negatively affected the nutritional status of children.

Wray and Aguirre (1969) in Candelaria, Colombia, found that the in-

come of the father, family size, birth spacing, and birth order were

related to malnutrition in children.

Increasing urbanization as a result of industrialization and the

migration of rural people to the cities in developing countries have

contributed to the development of malnutrition. Clark (1980) found

urban preschoolers to be smaller and less physically developed than

their rural counterparts. Villarejos et al. (1971) found the opposite

in Costa Rica where rural children exhibited more retarded growth than

did urban children. It seems that the change in environment caused by

the move to an urban area combined with existing culturally determined

food habits has created problems. In Costa Rica this is evidently

mitigated by the existence of extensive health and social welfare pro-


Specific customs of food consumption may interact with local con-

ditions to produce childhood undernutrition. For example, among the

Tongans Clark studied, the adult males began the meal,followed by adult

females and younger males, with children eating last. When meat was

consumed by the family, this was seldom reflected in the diet of the

child. Although this custom was prevalent in the rural situation, in

the city it created great problems because food items such as domestic

meats were less available. Clark also notes that in rural areas where

cash income levels were often lower, food purchases were supplemented

by family production of meats, fish, and vegetables. Urban families,

while enjoying higher cash incomes, lacked sufficient purchasing power

to replicate the food variety and quantity available to rural families.

Clark also notes that "bush foods" (wild foods which are hunted or col-

lected)were constantly present in the rural diet and almost totally

absent from urban diets. This decrease in variety of foods included

in the diet seems especially significant in view of the work of Dewalt

et al. (1980) in highland Mexico, which suggests that complex diets

contain more of the essential nutrients than do more restricted diets.

In addition to these socioeconomic factors a number of other

parameters having biological and social significance have indicated

that large families, short interpregnancy intervals, and a large num-

ber of preschool children at home are risk factors for malnutrition.

MacCorquodale and Rond6n de Nova (1977) found that in Santo Domingo

malnourished preschool children were from families averaging 4.7 chil-

dren, while well nourished children came from families with an average

of 4.0 children. In Costa Rica, Rawson (1975) found that a significant

risk factor for malnutrition was the presence of more than 2 preschool-

age children in a household. In Haiti, Ballweg (1972) reported a

greater likelihood of undernutrition associated with large families.

In addition to family size, the sex of the infant may be a corre-

late of malnutrition. Dorjahn (1976) reported that infant mortality

in Sierra Leone was higher for females in both the rural and urban

areas and suggested sex differential feeding as a causal factor. Other

researchers (Rawson 1975) have noted no sex differences in the fre-

quency of malnourished children.

Changes in Child-Feeding Practices

Another important causal factor in the development of malnutrition

is the changing pattern of infant feeding. In developing countries

this is particularly important because, as Berg (1973) notes, the move

away from traditional breast-feeding places a child in double jeopardy:

first because of the loss of a high quality food supply (formulas are

frequently mixed incorrectly) and secondly through the increased

potential for contact with infection as formulas are mixed with unclean

water and placed in dirty bottles. This gives rise to what E. F. P.

Jelliffe (1979) calls "comerciogenic" malnutrition.

Popkin et al. (1980) note that examination of worldwide data

based on large nationally representative samples suggests that, al-

though there has not been a widespread general decline in the prevalence

of breast-feeding (the percentage of women who ever breast-fed), there

has been a decline in the duration of breast-feeding, especially in

urban areas of Latin America and the Caribbean.

Changes in Family Structure and Women's Roles

Changes in family structure and women's roles have directly af-

fected child-care and thus child-feeding practices. One of the most sig-

nificant changes in family structure has been the rise of the one-

parent household, primarily with the migration of men or women to urban

areas and/or the inability of the man to obtain an adequate job. The

presence of only one parent has been found to have a negative effect

on a child's nutritional status, primarily because a woman who has a

child not supported by its father has little economic security. In

Jamaica, Desai et al. (1970) found that presence of both parents, small

family size, and birth intervals of at least 24 months were associated

with high rates of weight gain in children.

Of primary importance is the entry of women into the work force,

particularly in urban areas where work may be in factories and is not

generally compatible with child care (as selling goods in a village

market may be). Women's work hours and conditions have a great in-

fluence on child care. Uyanga (1980) notes that having a job which is

compatible with child care is positively associated with breast-feeding

behavioral changes. He also notes that in urban areas the presence of

other members of the family, e.g. daughters, other nonnuclear family

members (aunts, grandmothers, etc.), has a positive impact on child-

care time and thus a potentially positive effect on nutritional status.


Reutlinger and Selowsky (1976) have observed that infant mal-

nutrition as a result of absence from the home of lactating mothers

participating in the labor force is a case of income redistribution

within the family in the face of urbanization. Unless sufficient

food is purchased for the infant to compensate for the nutrients

lost by reduced breast-feeding, the child suffers a negative income

effect. Recent research from India according to Reutlinger and

Selowsky suggests that a woman would have to spend 76% of her earn-

ings to provide adequate quantities of formula or milk for a 4-month-

old infant. In Kenya this would require 58% of the wage, and in

Tanzania 51% (Latham 1979). Latham also points out that,since many

tropical countries have rather small dairy industries, most of the

manufactured milk formulas are imported, requiring the use of scarce

foreign exchange. This also contributes to the dependency of develop-

ing countries on foreign capitalist powers and multinational corpora-

tions like Nestl6.

Economic Development as a Contributing Factor

It is difficult to demonstrate a relationship between socio-

economic development and nutritional status because of the multi-

factoral etiology of malnutrition and because it is difficult to iso-

late the effect a specific program may have had. Commericalization

of agriculture, long a favorite goal of third world economic de-

velopment programs, is recognized as leading to a decline in

nutritional status (Fleuret and Fleuret 1980; Gross and Underwood

1971). Commercialization of agriculture usually implies introduction

of capital-intensive technology used for producing cash crops for

export, e.g. coffee, soy beans, African palm, bananas, horticultural

products. Along with this comes a decline in subsistence produc-

tion as small farmers become involved as wage laborers on large

plantations and agroindustrial plants or in producing cash crops on

contracts. Women may be left to bear the burden of the garden, or

subsistence farming, responsibilities for which they may be unprepared,

as they may lack education or experience with gardening or farming.

The result is a decline in both yields and in subsistence farming as

a whole. Staples must then be purchased with hard earned cash, which

tends to arrive in lump sums rather than being distributed throughout

the year. This has the effect of making peasant households less self-

sufficient and less able to withstand seasonal variations in the supply

of staple foods. The net result is an increased vulnerability to

malnutrition. This is especially true since private enterprise is

rarely able or willing to provide directly for the reproduction of

labor power which would entail paying wages sufficient to enable a

worker to support a family and obtain adequate housing and health ser-

vices (Roberts 1978). Certainly Gross and Underwood's (1971) study

of the introduction of sisal agriculture in northeastern Brazil is

an excellent example of this process. In this example the introduc-

tion of sisal, which employed men in the hard and exhausting labor of

harvesting and processing, resulted in such increased energy costs in

relation to the low wages that a systematic deprivation of adequate

calories to the wives and children of the sisal workers was the result.

In order to maintain physical strength, the male of the household had to

eat the lion's share of the available food. Thus in the population

studied, the children were exhibiting signs of malnutrition-retarded

growth and development. The working poor were essentially subsidizing

capitalist private enterprise.

The implication of the preceding discussion is that malnutrition,

especially infantile protein-energy malnutrition,has a complex etiology.

Solutions based on food and nutrients alone are completely inadequate

(Pellet 1977). Expecting malnutrition to disappear with economic

development and increased GNP seems unrealistic also, in view of the

past record and because evidence indicates that low-economic groups

benefit unequally in the process of capitalist-oriented economic develop-

ment (Cornelius and Trueblood 1975; Roberts 1978).

According to Reutlinger and Selowsky (1979:22), empirical evi-

dence suggests that the caloric intake will increase at approximately

half the rate at which per capital income rises. On the basis of these

assumptions it would take 30 years before the poorest 20% of the popula-

tion of these countries could reach an adequate level of caloric in-

take. In their view the high incidence of malnutrition among chil-

dren of developing countries is not likely to be reduced as part of the

development process. The only effective solution according to them

lies either in a more equitable income distribution or in supplying

food to the target population at low prices. This view is echoed by

Teller et al. (1979:21) who state that in general poor countries will

attain an adequate nutritional status only when they can provide most of

their people with a "minimum socioeconomic package" (permanent employ-

ment, adequate income, social security, housing, etc.).

The Costa Rican Situation

Costa Rica (see Figure 1) is unique in Central America because it

has long had a democratic form of government, a social conscience among

the upper classes, and a fairly large middle class. The Spanish settlers

of Costa Rica, throughout the colonial period, found themselves forced to

reside in the country where each family labored to produce its own food

(Munroe 1918:140).

As the population grew, the entire Meseta Central became divided

into small farms. There were a few wealthy and influential families who

had been given special privileges by Spain, but they never occupied the

dominant position which the aristocracy of Guatemala and Nicaragua had

been able to assume, and the land they held never amounted to more than

a small portion of the cultivated area of the colony. In 1848 through a

special decree the people were permitted to buy the land they had fenced

and were using, a measure which had the effect of greatly increasing the

number of landholders (Munro 1918:142-143).

Costa Rica achieved independence from Spain in 1821 when she joined

with seven other provinces of Central America to form the United Provinces

of Central America, which was later disbanded in 1838 because of internal

conflicts. Full national sovereignty was not established until 1848.

The period between 1938 and 1902 was largely one of dictatorships succeed-

ing one another through coups. Also, during this period Costa Rica had

to defend herself against a Nicaraguan invading force in 1836 and from

the troops of the North American adventurer William Walker in Nicaragua

in 1853-57 (Fanger 1968).

The radical transformation of Costa Rican geography since its inde-

pendence can be attributed to the development of two agricultural














S: r

products, coffee and bananas (Hall 1976:14). Coffee reached Costa Rica

some time before the end of Spanish rule in 1821. A letter from

Governor Tomds de Acosta to the king in about 1808 mentions the plant,

and by 1829 it was the most important produce of the country. A rush

to plant coffee took place in 1841, and the country's economy and

thinking became oriented around coffee growing and coffee prices

(Lundberg 1976:114). During the 1940s large uncultivated tracts of

land owned by the government were sold at low prices or given away as

premiums to encourage the planting of coffee. During the last years of

the 19th century many persons acquired land this way, and gradually a

class of large landholders developed, although the Meseta Central was

still composed largely of small farmers (Munro 1918).

The banana has been cultivated since the end of the 19th century

in the coastal regions of Costa Rica on great plantations, a system

commonly associated with the cultivation of agricultural products for

export in many developing countries. These regions have only recently

been integrated into the Costa Rican economy and then with great diffi-

culty. Communication and transport were and are major problems. Costa

Ricans from the interior (Meseta Central) had little to do with the

development of the banana regions. In the Atlantic zone, the north

American fruit companies were responsible for developing plantations with

labor provided by Negroes from the West Indies (Hall 1976:14).

Although there is at present a small class of large landholders

resulting at first from the introduction of coffee and bananas,as noted,

and later extending to the Guanacaste area with the introduction of

cattle raising for export of beef (Edelman 1983), there has always

existed an egalitarian philosophy on the part of the government and the

upper classes. Beginning, particularly in 1948, with the progressive

reforms of Calder6n Guardia, a series of enlightened social welfare pro-

grams have been enacted concerning social security, health, housing, and

protection of children. At the present time, Costa Rica, with a popula-

tion of slightly over 2 million people, has one of the most effective

an aggressive health systems in Latin America, encompassing both cura-

tive and preventive medicine. An infant mortality rate in 1980 of 19.1

per 1000 population and a life expectancy at birth of 73.4 years (Mini-

sterio de Salud 1982) reflect the general health status of the popula-

tion. This is comparable to that of the United States in 1970 when the

infant mortality rate was 20 per 1000 population. Life expectancy at

birth in the United States in 1978 was 73 years. In 1978 the United

States had an infant mortality of 12/1000 for whites, 23.1/1000 for blacks,

and 21.1/1000 for blacks and other minorities. The infant mortality rate

overall was 13.8/1000 (U.S. Dept. of Commerce, Bureau of the Census 1981).

Other countries in Central and South America had the following infant

mortality rates: El Salvador and Guatemala in 1979, 53 per 1000 and 70.1

per 1000 respectively; Colombia, 39.5 per 1000 in 1977; and Argentina,

40.8 per 1000 in 1978 (Demographic Yearbook-1980,1982). Mata et al.

(1980a, 1981) notes that the diarrhoeal disease death rate was 11 per

100,000 with about 140 infant deaths per year due to diarrhea. No cases

of poliomyelitis or diptheria have been recorded in recent years and few

cases of measles and whooping cough, indicating a favorable national

health immunization system.

The health system, as it is currently organized, dates from the

early 1970s when the National Plan of Economic and Social Development

was implemented. This plan has as its goal the improvement of the

quantity and quality of basic services available to Costa Rican citi-

zens and was begun in rural areas in 1973 and in urban areas in 1974.

In relation to health care, these basic services involve the coopera-

tion of the Ministry of Health, the Costa Rican Social Security Adminis-

tration, the Costa Rican Institute of Aqueducts and Sewers, and the

National Insurance Institute. The goals of the program are as follows:

1. To extend basic health services to poor rural and urban com-

munities using health auxiliary workers.

2. To control and reduce infectious diseases through immunization


3. Referral of patients to either preventive or curative facili-

ties, depending on medical diagnosis.

4. Planned visits to homes and businesses to evaluate health


5. Emphasis on environmental health, health education, and com-

munity organization. (Freer Miranda 1980)

In 1966 a national nutrition survey was conducted in Costa Rica by

the Instituto de Nutricidn de Centro America y Panam--INCAP (1969).

Rawson (1975), Rawson and Valverde (1976), and Valverde et al. (1975)

investigated the nutritional status of preschool children in the com-

munity of Concepcidn de San Ram6n, Costa Rica. In general these studies

indicated nutritional deficiencies in preschool children which were

moderate but chronic. Deficiencies were more pronounced in low socio-

economic groups. According to the INCAP study (INCAP 1969), the

four types of malnutrition which most frequently occur in Costa Rica

are protein-energy malnutrition (PEM), nutritional anemias, vitamin A

deficiency, and goiter (iodine deficiency). Mata (1979), however, notes

that of these four the single most important deficiency is that of

protein-energy malnutrition, or PEM. It is the most common and has

the most complex etiology which makes its study and control more diffi-


Results of the San Ram6n study, in which the G6mez scale (Gdmez

et al. 1956) was used, revealed that 38.9% of 149 children aged 0-5 years

were of normal weight, whereas 46.3% exhibited first degree malnutrition,

13.4% second degree malnutrition, and 1.3% third degree malnutrition.

Therefore, more than half of the children (61%) were found to be below

the normal weight range. Fifteen percent were in the lowest two groups

(below 75% of normal weight). Contrary to findings already cited for

other areas, the effects of malnutrition in San Ram6n apparently become

worse as the children grow older. This pattern of chronic rather than

acute malnutrition in neonates and preschool-age children points to fac-

tors or interrelationships of causal factors different from those found

in other developing countries.

Rawson (1975) and Rawson and Valverde (1976) note a number of

socioeconomic and cultural variables associated with nutritional status.

Variables significantly (p 5 0.02) associated with childhood malnutrition

were access to less than 1.4 hectares of land, more than one sibling under

6 years of age, father works as a day laborer, mother works outside the

home, house in poor physical condition, poor stove quality, and little

access to fresh milk. Variables not significantly associated with child-

hood nutritional status included education level of the parents, type of

family unit (nuclear, extended, or single parent), sex of the child, age

at weaning, and salary income.

A more recent National Nutrition Survey (Ministerio de Salud 1978)

also noted relationships between house conditions and nutritional

status. Poor housing conditions were associated with poor nutritional

status among preschool children (Table 1). The condition of the house

was assessed on type of house construction, number of rooms, presence

and condition of bathrooms and toilets, water supply, presence and con-

dition of windows, quality of cooking facilities, etc.

Table 1 indicates that the highest prevalence of malnutrition

(G6mez's first, second, and third degree combined) in children aged 0-5

years occurs among children in poor housing in the urban areas-63.1% com-

pared to 51.3% in the dispersed rural areas and 55.5% in concentrated

rural areas. The overall national percentage of children and infants in

poor housing with first degree malnutrition is 45.1%. As housing condi-

tions improve, the proportion of children underweight for age decreases


Some studies have suggested that malnutrition in Costa Rica is fre-

quently the result of factors other than lack of food, such as passive

child abuse manifested as neglect. Stunting and wasting were found to be

much higher among abused children (Mata et al. 1980b). Mata, in another

paper (1980), cites decreased maternal competence, failure of bonding,

and infections as factors which are just as important as available food

in the development of PEM.

Salazar and Cervantes (1979),in a preliminary report based on the

national survey, also note positive relationships between education of the

Table 1. National Distribution of Children 0-5 Years by Weight for Age
According to Condition of Housing and Degree of Urbanization

Weight for age

Condition Second and
of housing Overweight Normal First degree third degree













11.4 42.7

14.2 47.7

13.0 44.5

5.6 34.2

















Dispersed rural 7.7 40.8

Good 8.4 44.5

Fair 11.5 46.2

Poor 3.8 34.4

Source: National Nutrition Survey
aG6mez (1956) classification.

37.3 8.6

31.6 6.5

36.2 6.3

45.1 15.1


























de Salud 1978).

head of the family and spouse and nutritional status of preschool chil-

dren. In households where the head of the household is illiterate,

57.9% of the children 0-5 years of age are malnourished based on the

G6mez scale. As the educational level increases, the proportion of mal-

nourished children decreases: secondary-level education-37% mal-

nourished; university-level education-20.2% malnourished.

The 1978 survey also revealed a downward trend in the incidence

of breast-feeding. Thirty-seven percent of infants in rural areas were

weaned in the first month of life, often at birth, and by 5-6 months

approximately 60% had been weaned to formula milks. Also, infants who

were breast-fed frequently received supplements; only 19% were exclus-

sively breast-fed at 4 months of age. In response to this problem

Dr. Leonardo Mata and his team of workers from the National Institute

for Health Research (INISA) have begun an innovative and very interest-

ing nutrition-related project. The primary site of this work is the

area of Puriscal, primarily involving dispersed and concentrated rural

settlements. The Puriscal study was begun in September 1980 in conjunc-

tion with an intervention aimed at increasing breast-feeding. All the

mothers in this area are included through the prenatal clinics. Eighty-

four percent of deliveries are in the San Juan de Dios Hospital, 13% in

maternities and clinics of San Jose,and 3 percent in the home. The

objective of the project is to increase breast-feeding by increasing

the opportunities for maternal/infant bonding in the hospital situation

after birth. Attempts were made to change hospital policy to foster

rooming-in to some degree. A milk bank was created to provide milk and

colostrum for high-risk neonates. The project includes in-home sur-

veillance as well as the hospital stimulation. The result has been a

dramatic increase in the incidence of breast-feeding in the study popula-

tion (Mata et al. 1981).

Another program of great interest is that of the canasta basica

alimentaria, or basic food basket program. This has been in the process

of development since the survey of 1966. A basic diet was elaborated

which as of 1981 included 12 basic food items considered necessary for

an adequate diet (see Table 2). This list is used to determine the im-

pact of cost-of-living increases and also to plan national food produc-

tion with the aim of ensuring adequate supplies of these items. Another

objective of the basic food basket concept is to control prices so that

these items remain within the reach of the poor (Murillo et al. 1981).

Until 1982 most of the items on this list were subject to controlled

prices. At that time, inflation forced prices to be raised. The con-

cept is still very useful, however, because it tends to focus national

attention on the basic requirements for maintaining good nutrition for

all population groups.

Costa Rica also has a nutrition rehabilitation center housed in

the facilities of the Costa Rican Institute of Investigation and Teach-

ing in Nutrition and Health (INCIENSA) located in Tres RTos. In addi-

tion to ongoing research projects, this institution is a referral center

for severely undernourished children (3rd degree by G6mez standards).

These children come from all over Costa Rica. Their numbers are rela-

tively small, however, because of the ongoing surveillance program con-

ducted by the community health workers.

As part of the national nutrition surveillance program,the System

of Nutrition Information (SIN) was set up to utilize the data which are

Table 2. Structure of the Basic Costa Rican Food Basket in Grams per
Person per Day (average for the country)

Quantity Percent of
Food (grams) Calories total calories

Milk (liquid) 445 289 9.95

Eggs 31 46 1.60

Meats 84 206 7.10

Beans 74 255 8.80

Vegetables 105 34 1.17

Fruits 44 17 0.60

Bananas and
plantains 49 56 1.95

Roots and tubers 69 78 2.70

Rice 187 682 23.50

Wheat bread 78 218 7.50

Tortilla 68 136 4.70

Sugar 102 377 13.00

shortening 51 447 15.40

Ground coffee 13 29 1.00

Soft drinks 96 30 1.05

Total 1496 2900 100.00

Source: Murillo and Mata (1980:101-104).

constantly being collected locally by health workers to define and de-

scribe target groups of households at risk for undernutrition on a

national level. The goal of SIN has been to develop a functional classi-

fication which involves describing categories of poor families and relat-

ing the prevalence of malnutrition with relevant specific social,

economic, and cultural factors. A functional group is defined as a set

of families sharing a similar pattern of living, that is, the same type

of production, the same type of social and economic constraints and

access to existing resources, and similar cultural practices (Valverde

et al. 1981).

Thus, while nutrition problems exist in Costa Rica, a number of

very effective and sophisticated programs exist for the purpose of evalu-

ation and surveillance of nutritional status. Nutrition problems in

Costa Rica, while not severe, are complex, requiring subtle and inno-

vative ways of measuring the extent of the problem and evaluating the

progress already made. Subsequent chapters will deal in more depth

with the statement of the problem and methods used in looking at the

status of child nutrition among the urban poor in San Jose.


The major purpose of this research was to define biocultural

causal factors in undernutrition and retarded growth and development in

a sample of infants aged 12 to 24 months in the poor urban barrios of

San Jos6, Costa Rica. A holistic anthropological approach was used to

elucidate the diverse strategies employed at the household level which

lead to well or poorly nourished children.

Undernutrition is most prevalent among poor people; yet, among

the poor there are those who maintain good or excellent health and

nutritional status. The underlying hypothesis of this research is that

strategies of resource accrual and dispersion as well as absolute levels

of nonshelter-related real income within a household are significant

factors in the determination of nutritional status. It was anticipated

that below a certain level of income, the proportion of undernourished

children would increase dramatically with relatively small decreases in

real income. The object of the research was to do in-depth, cross-

household evaluation of resource utilization in poor households, compar-

ing those with well nourished children to those with undernourished chil-


Nutritionally related high-risk factors and strategies have been

identified for rural populations (Desai et al. 1970; Rawson and Valverde

1976). These factors tend to be related to ownership and access to farm

lands. In the urban context, however, little work has been done to


elucidate those factors which contribute to well or poorly nourished

children. The urban populations are of special concern because as

migration continues from rural to urban areas, the increased popula-

tion size puts pressure on the economic system to provide more jobs and

wage-related income for recent migrants. These migrants, usually young

and within the growth phase of family development, are suddenly con-

fronted with new economic and environmental stressors which require the

development of new resource-related strategies and the development of

new skills to enhance their wage-earning abilities. As a result, the

more resourceful have developed coping mechanisms which enable them to

survive on very little, primarily within the informal economy or small-

scale sector, e.g. lottery ticket sellers, street venders (Roberts 1978).

This "tertiarization" of the economy-the expansion of employment in the

commercial and service sectors rather than in manufacturing-has been

associated with urban poverty as an adaptive mechanism allowing migrants

to the city to eke out an existence. Peattie (1975) provides an excel-

lent description of how this phenomenon works in a Latin American city.

The resourcefulness of the poor is further emphasized by such works as

that of Lomnitz (1977) describing mutual reciprocity networks and Arizpe

(1980) who investigated relay migration (family members taking turns go-

ing to the city to live and work). Both of these works were carried

out in Mexico and illustrate very well that if anyone can squeeze better

nutrition out of a low income, these people can. One of the objectives

of this research was to become more familiar with these strategies as

they exist in Costa Rica and to glean knowledge which would be helpful

in understanding this complex problem.

In particular, the research focused on strategies of food acquisi-

tion and allocation within the household context and on the decision-

making associated with these strategies. As noted earlier, malnutrition

has a multifactorial etiology; nutritional and health surveys often do

not uncover the underlying and/or intervening variables which are the

essential links between access to food and the nutritional status of

the child. In addition, survey data tend to obscure intracultural di-

versity which accounts for the observation that among the poor some

children are well nourished while others are not. In short, we know

that malnutrition exists. We have many supposed causes. We do not

understand very well how these causes relate to specific cases at the

family and household level. In other words, we do not know much about

the target group, particularly in the urban areas where most of the

poor are now concentrating in developing countries. It may be that

below a certain level of income the only way to improve nutrition is to

increase income. Above this income level, however, there is a tremen-

dous gray area which still includes many families in which there are

undernourished children. This gray zone may be amenable to many kinds

of programs, such as education, food subsidies, etc. The problem occurs

when a program is aimed at those people below the crucial economic

level but uses methods which are meant for the people in the gray zone.

The program may then be evaluated as a failure when very few people in

the target group respond to it. Or it may be judged as a success by

the number of people participating but yet not reach those who are

critically in need of assistance. This research, therefore, was aimed

at investigating intracultural diversity at the household level by use of

a variety of research techniques to obtain a clearer picture of causal

factors involved in undernutrition of specific groups. This research

contributes methods for characterizing children at risk nutritionally

which allows more effective targeting of nutrition intervention pro-


Data indicating the importance of family structure to the sur-

vival of low income families came from an in-depth multidisciplinary

study done in the urban barrios of San Jos6 which focused on marijuana

use by adult males (Carter et al. 1976). This study contains some in-

formation on the nutritional status of adult males. More important

for the purposes of this research was the information on family dynam-

ics, kinship-based networks, and the general environment of lower-class

neighborhoods. As described by the adult male informants in the study,

a typical family of orientation has about five members, including both

parents, two or three children, and occasionally a bilateral relative

such as a sibling or an aging parent. The range in family size, how-

ever, is reported to be wide. A frequent pattern encountered was a

periodic doubling up on space within an extended family group to reduce

costs during periods of economic stress, which frequently occurs be-

cause many children and adults are seasonal laborers. In this study

20% of the marihuana users and 5% of the nonusers came from single

parent households headed by the mother. Often surrogate parents, e.g.

grandmothers, aunts, or uncles, were child caretakers. In contrast to

other areas in Latin America, fictive kinship, godparents, and com-

padrazgo are relatively unimportant (True 1976:69).

Although the households were found to be largely restricted to the

immediate nuclear family, both marihuana users and nonusers were almost

always in constant contact with extended kindred networks. These net-

works are potentially very important to the nutritional status of the

child since they tend to be the source of most surrogate parents and

kin may share food, material goods, and money.

Costan Rican anthropologist Eugenia Lopez de Piza (1979:4-5) has

described four family types which she had found to be common in Costa

Rica: the nuclear family, consisting of father or father substitute,

mother, and children; the extended family, consisting of related nuclear

families; the matrifocal family centered around the mother and her chil-

dren; and what she has termed the "Queen Bee" family, which is a vari-

ant of the matrifocal type and consists of the grandmother, her daugh-

ters, and their childern. In this latter family type, the daughters

work and the grandmother cares for the children and controls the money.

This extended matrifocal family is similar in structure and function to

the three-generation family with a female head described by R. T. Smith

(1956:106) for British Guiana. It is also similar to the "grandmother

families" described by M. G. Smith (1962) for the West Indies. Lopez

de Piza (1979:15) calls the matrifocal family in Costa Rica, particu-

larly the "Queen Bee" variant, the best adaptive mechanism enabling

poor women to obtain economic necessities and care for and educate chil-


Mothering ability or maternal competence is another factor which

has been implicated as a risk factor for malnutrition. Mata (1979) in

Costa Rica cites deficiencies in maternal technology and social path-

ology as becoming increasingly important causal factors in child mal-

nutrition in modern times. Alvarez et al. (1977) in Chile found that

maternal education levels were important in influencing the development

of healthful food habits among children.

Data from the study on marihauna consumption in San Jos6 indi-

cated that while most subjects reported that meals were prepared and

served at regular hours in their homes, because of employment hours

at least half of the males did not eat at the same time as the rest of

the family. The rest of the family (mother/wife, children, and other),

however, generally followed a regular schedule for meals (Rawson and

Phillips Arizmendi 1976:589-590).

Alcohol use by a parent was cited as a cause for going hungry and

other unhappy childhood experiences which could directly affect the

nutritional status of young children. Slightly less than half of mari-

huana users and nonusers had fathers who drank heavily (True et al.

1976; True 1976).

A model was developed based on prior work done in San Ram6n, a

small village, and the surrounding rural area by Rawson (1975). This

model was modified to make it more appropriate for testing in the urban

situation and to include most of the risk factors already mentioned.

Figure 2 illustrates this model. The model was to be tested through

the collection of extensive biocultural data and by use of a mixed

cross-sectional and longitudinal design. Both the methodologies and

the model were to be evaluated for applicability to other urban set-


Unlike Rawson's study (1975) in which the basic independent

variable was access to an adequate amount of agricultural land (i.e.

two manzanas, or 1.4 hectares), this study used a factor labeled "real
























income." "Real income" is composed of wages, credit, savings, monetary

gifts, inheritance, etc., averaged weekly to allow for variations such

as seasonal ones. There were a number of intervening variables involv-

ing the allocation of resources, the adequacy of the physical environ-

ment, and the quality of child care. These variables were directly

observable and measurable and are described in Chapter III in the sec-

tion on methods.

One of the objectives of the study was to augment the community-

based nutritional research completed by the Ministry of Health, the

National Institute for Health Research (INISA), and the Costa Rican

Institute for Research and Training in Nutrition and Health (INCIENSA).

The project emphasized a longitudinal, in-depth behavioral approach

which was used to delineate the variables and diversity of coping

strategies leading to poor and well nourished children among the urban

poor. This information will be used to characterize that portion of

the population which is at greatest risk for malnutrition and to assist

in the development of intervention programs which are targeted specifi-

cally at the high risk group. This approach has as its goal the con-

servation of scarce resources by facilitating directed programs which

provide the greatest return on invested materials and manpower. The

identification of risk factors and strategies which lead to malnutri-

tion on the basis of household data can be extrapolated for use on the

community level.


Phase I: Setup


In my position as field director of the project,I traveled to

Costa Rica in August 1981. The appropriate contacts were made with the

project consultants,Dr. Carlos Diaz Amador, director of the Department

of Nutrition of the Ministry of Health, and Dr. Jos6 A. Camacho, profes-

sor of anthropology at the University of Costa Rica. Through these

individuals, contacts were made with the director of the Community Health

Program, Dr. Enrique Freer Miranda, and Dr. J. Manual Alvarado Obando,

director of Health Region 1, which has metropolitan San Jos6 in its juris-

diction. After obtaining the appropriate authorization, contacts were

made with the directors and personnel of all the health centers from

which the samples of children were to be drawn. This selection was made

with the aid of Dr. Diaz Amador and Dr. Camacho and by use of the criteria

of a large population of poor people and, where possible, areas which

had not been overstudied by various investigators in the past.

Spanish translations of the proposed project and the permission

form to be used for inclusion of each child in the study were presented

to the Ministry of Health and were approved.

Archival Research

Archival research was carried out in the libraries of the Univer-

sity of Costa Rica, the National Children's Hospital, the Costa Rican

Institute for Research and Teaching in Nutrition (INCIENSA) at Tres Rfos,

the National Museum, and the Ministry of Health. Documents were also ob-

tained from the National Office of Planning (OFIPLAN), the National

Nutrition Information System (SIN), the Ministry of Health, and the

National Institute for Research in Health (INISA).


Through the auspices of Dr. Enrique Freer Miranda, head of the

Community Health Program at the time of this study, two women who had

worked for four years as health auxiliary workers in the Hatillo Health

Center were hired. These two women, Flory Desanti Jimenez and Eulalia

Obando Naranjo, worked with me to develop the survey instruments,

assisting particularly with translation of the language to the idiomatic

Spanish of the urban lower class. They were also trained to conduct

structured interviews, do anthropometric measurements, and give the

Denver Developmental Screening Test (DDST).


With the help of the two interviewers, 10 households were chosen

from various neighborhoods which had people similar to the sample popu-

lation but which were not to be included in the survey. In these house-

holds children were weighed and measured, interview schedules were filled

out, and the Denver Developmental Screening Test (DDST) was given.

Polaroid photos were taken and left with each family.

In addition to this pretest situation, prior to the beginning of

the formal study, about 70 children from a squatter settlement were

measured and weighed to provide additional training in measurement tech-

niques for the research assistants. These children are not included in


this study, however. The DDST was also given to 30 children with myself,

Flory, and Eulalia scoring each test simultaneously. By the time the

formal investigation was begun,everyone was comfortable with all instru-

ments and procedures. There was also very little variance in the manner

with which each procedure was carried out.

Phase II: Initial Survey Target Population
and Sampling Design

Greater San Jos--San Josd and surrounding urban areas--had a

population in 1973 of 837,000 which is projected to reach 1.8 million

people by the year 2000, increasing the present labor force from 260,000

to 500,000 persons. In 1976 there were more than 240 slums in the

metropolitan area in which there were one-room shacks occupied by

several families, according to a USAID report (1976). According to this

report, there were an estimated 110,986 families in Costa Rice (average

family size-6 persons) which were considered mas pobres ("more poor" or

"poor poor") with incomes of less than $825 per year. Of these families

40.6% (45,000) lived in greater San Jose. In addition, there were

197,268 other families who were considered pobres ("poor" or "rich poor")

with incomes below $1,395 per year. Of these families 53% (104,610)

lived in the greater San Jose area. The income figures are based on cen-

sus data and primarily refer to salary income. More recent data was

difficult to obtain because similar USAID reports and Costa Rican

government reports are now considered classified material and I was not

allowed access to them. It is doubtful if the situation described has

changed much for the better, however. If anything, things are prob-

ably worse because Costa Rica has suffered an economic crisis and in-

flation has changed the value of the colon from (12.50 per $1 to

(40-60 per $1 in 1981-82. Times are tougher and Costa Rica has been

hard pressed to try to pay off its foreign debt. It has also had a

tremendous influx of refugees from Nicaragua, El Salvador, and Guate-

mata, an additional drain on the economy.

Seven health centers were selected as sites for obtaining the

sample population. They were Hatillo, Alajuelita, Pavas, Tibas, Paso

Ancho, Guadalupe, and Cristo Rey. The areas selected were known to

have a high percentage of poor people. Figure 3 shows the distribution

of this population, Although each of these health center regions has

some more well-to-do neighborhoods included in its jurisdiction,

only areas noted as very poor were selected, e.g. Calle Blancos, Cinco

Esquinas, Concepci6n de Alajuelita, La Esperanza in Pavas. A sampling

frame was constructed from birth records kept by each health auxiliary

worker for his/her district. All the children born in 1980 in each of

the areas selected were placed on a list and given a number. A table

of random numbers was used in selecting over 200 children. After each

child was selected from the birth registry, his or her family record

was located and the address and other social data noted. About 50 to

60 of these could not be located in the files of the health auxiliary

workers. Another 25 to 30 were located in the file but when the health

auxiliary worker arrived in the neighborhood, they had moved and left

no forwarding address. The final large sample contained 107 children.

In each of these 107 families, a family registry, or listing, of all

members of the household, their ages, sex, education,and migration

history was taken. In addition to these data a medical history of the

child during the past year, an assessment of material level of living


and quality of the environment, an examination of the child for clinical

symptoms of malnutrition, and anthropometric measurements of the index

child were obtained. The Denver Developmental Screening Test (DDST)

was also performed on this initial visit. (See Appendices 1-5 for

examples of the forms used.)


The following measurements were carried out according to the IBP

Handbook protocols (Weiner and Lourie 1969).

Length. This was taken in centimeters by use of a headboard with a

metal tape measure attached which could be taped to a hard surface such

as the floor or a table. Another board was then held up to the bottom

of the feet to determine the length. This instrument was adapted to fit

into a shoulder bag because bus transportation was used at all times.

Weight. In the interest of making the instruments portable, it

was decided to use a bathroom-type scale. Most of the children being

measured could walk; therefore this was not a problem. Three scales were

obtained; each of these was tested with standard weights up to 30 kilos

and found to be accurate. Thereafter they were tested every 6 weeks.

Before use the scales were placed on zero. The weight was obtained with

the child wearing light clothing and without shoes.

Triceps and subscapular skinfolds. These measurements were ob-

tained in millimeters with Lange skinfold calipers. These measurements

were taken on the large sample at the beginning of the in-depth study

and at the end of the in-depth study. It was decided not to take them

each time because of the difficulty of doing this accurately with chil-

dren and because of the need to have only one person take the measure-

ments. Thus, I took all the skinfold measurements.

Arm and head circumference. These measurements were taken in cen-

timeters with a fiberglass nonstretchable measuring tape.

Denver Development Screening Test (DDST)

The Denver Developmental Screening Test is a simple and effective

way of assessing the developmental status of children during the first

6 years of life. It evaluates the following aspects of the child's

functioning: Gross motor, fine motor-adaptive, language, and personal-

social areas. The test relies on observations of what the child can do

and on reports by a parent or parent surrogate who knows the child.

Direct observation is used whenever possible. The younger child may be

tested while sitting on the mother's lap, which is the procedure we used.

The test is easy to administer, score, and interpret and is useful for

repeated evaluations of the same child. The test has a high rate of

validity and reliability (Camp et al. 1977; Frankenburg and Dodds 1967;

Frankenburg et al. 1971a, 1971b, 1971c). It is not an intelligence

test but is intended as a screening instrument for use in clinical prac-

tice to note whether the growth and development of a particular child

is within the normal range.

Because the DDST was standardized on the Denver population, it has

been noted that some of the norms are not applicable to children in other

cultural areas (Bryant et al. 1974; Solomons and Solomons 1975). It has.

been successfully adapted and standardized for use in Japan and Okinawa

(Ueda 1977a, 1977b). It has been translated into Spanish and has been used

to some extent in Guatemala, Chile (Fandal 1980), and Yucatan, Mexico

(Solomons and Solomons 1975 ). The test has also been used cross culturally

in the United States (Frankenburg etal. 1975). Upon arriving in Costa Rica,

I discovered that research conducted by the University of Costa Rica

had been carried out to standardize the DDST for use in Costa Rica and

Central America (Howard and Nieto de Salazar 1982). However, because

this standardization was not yet complete and because of what I per-

ceived as problems with the design of the standardization research, I

decided to use the official Spanish version of the DDST,which had been

obtained from the La Doca Foundation in Denver, Colorado. Because the

comparisons to be made were to be within a cultural group rather than

between cultural groups, I felt that the results from this point of view

would be valid in spite of any possible cultural biases inherent in the

test. However, many of the suggestions developed by the Costa Rican

group to facilitate administration of the test to the local population

were incorporated.

The DDST was given on the first visit to all 107 children included

in the sample. After it was explained and demonstrated, it was found to

be of great interest to the people and proved to be an excellent entree,

providing parents with graphic demonstrations of their children's

abilities in various areas. It was repeated once more,after at least

three months had passed,on the smaller group of 44 children included in

the in-depth study.

Material Level of Living:
Quality of Environment

A checklist was used to assess the material level of living, e.g.

presence or absence of radios, TV's, sewing machines, and other indica-

tors of quality of life. Quality of the living environment was also

assessed with an evaluation of the condition of the house, type of

construction, number of rooms, number of people per room, number of beds,

type of cooking facilities, disposition of excreta and garbage, light-

ing, ventilation, and other aspects of the environment (see Appendix 4).

Assessment of Clinical Signs
of Malnutrition

A list of symptoms of primary clinical malnutrition was developed

by use of the form suggested by Rawson (1975) which was then adapted

for use in San Jose with the aid of Dr. Carlos Diaz Amador, director of

the Department of Nutrition of the Ministry of Health. (See Appendix 6

for a copy of this list). I performed this examination on all the chil-


Phase III: In-Depth Longitudinal Study

Because the initial large sample showed very few children at or

below the fifth percentile of weight for height when the NCHS statistics

were used (probably because of the excellent national program of nutri-

tion, surveillance, and intervention), it was decided to focus on the

mildly to moderately undernourished child, victim of invisible malnutri-

tion,as it has been termed (International Health News 1983), currently

the major problem for Costa Rica and many other developing countries.

The criteria for severely undernourished children are quite clear,

e.g. fifth percentile or below on the growth charts of the National

Center for Health Statistics (NCHS), 80% of the median or below, again

on the NCHS growth charts (Mata 1978a; Waterlow and Rutishauser 1974),

or second and third degree malnutrition by the classification of G6mez

(1956). On the G6mez scale, which is based on weight for age, normal

is 91-110% of standard weight for age, first degree malnutrition is 79-

90% of standard weight for age, second degree malnutrition is 61-75% of

the standard, and third degree malnutrition is 60% or less of the stan-

dard (G6mez el al. 1956:77). In recent years the classification of

G6mez et al. has come under criticism because it is based on standards

developed among much heavier children than is the norm in Latin

America and other parts of the third world. Mata (1978a) notes that it

was developed more than 20 years ago as a means of categorizing children

admitted to the hospital for treatment of malnutrition and was not in-

tended to apply to the general population. Generally the trend has been

to use weight for height as the criterion for determining acute malnutri-

tion, and height for age as the indicator of long-term nutritional status.

Recently the concepts of wasting(deficit in weight/height) and stunting

(deficit in height/age) have been introduced to allow populations to be

classified in different groups for intervention purposes (Mata 1978b;

Waterlow and Rutishauser 1974).

As mentioned already, the criteria for severely undernourished

children are clear; however, the cut-off point for delineation of mildly

undernourished children with subclinical malnutrition is not so easily

determined. The Costa Rican Ministry of Health uses growth charts based

on NCHS standards and considers children to be moderately under-

nourished if they are at or below 90% of the median. Because of a

decision to focus on the mildly to moderately undernourished child, the

in-depth sample included all those children who were at or below 90%

of the median of NCHS standards and/or had at least first degree malnutri-

tion by the scale of G6mez et al. Originally there were 26children falling

into this category. Two were eliminated because they had had ill-

nesses shortly after birth (meningitis and encephalitis) which had left

such effects that it would have been impossible to say what retarda-

tion in growth and development was due to undernutrition and what was

due to disease processes and their residuals. Of the 24 left, only

19 agreed to participate. Another group of children was selected, by

use of the table of random numbers, from the remaining 81 normal weight/

height children. The families of 25 of these children agreed to let

them participate in the in-depth study. In the end there were 26 males

and 18 females participating. Ten of the females and 9 of the males

were in the low weight/height group. Eight of the females and 17 males

were in the normal weight/height group. Among the normal group, the ran-

dom selection yielded very nearly an equal number of boys and girls.

The difference apparently results from the fact that more of the parents

of boys were willing to let them participate then were parents of

girls. It is not certain whether this is a chance occurrence or some

cultural bias having to do with boys and their roles and position in


Long-Term Data Gathering

The plan had originally been to pick up the children to be fol-

lowed in groups of 10, adding another increment of 10 each month until

all were being followed and then gradually tapering off as the 6 months

of surveillance were reached. At the time the in-depth portion of the

research was to begin, however, the political situation in Costa Rica

and all Central America seemed quite tense. It was just prior to the

elections, there had been some terrorist activity, arms caches had been

found, and there had been border incidents with Nicaragua. Because of

these factors and because the interviewers proved to be very capable

and easily trained, a decision was made to carry all the families simul-

taneously, visiting each family every month for measurements and addi-

tional data gathering. Each visit had to be made within six days

either before or after the date which equaled one month exactly from the

last visit. The object was to make the data-gathering period as compact

as possible in case there developed a situation in which the ability to

visit families living in slum areas would be interrupted. This arrange-

ment worked out well, proving to be a better plan than the original.

In addition to the family registry, anthropometry, medical history

of the child, checklist of material living conditions and quality of the

environment, and assessment of clinical malnutrition already mentioned,

several other types of data were obtained.

Assessment of Socioeconomic Status

This interview schedule included data about the occupations, in-

come, expenditures, gifts, etc., of all members of the household. Be-

cause of the sensitive nature of these data and our wish to keep every-

one in the study for all of the 6 months, this schedule was not adminis-

tered until we had been visiting the family about 4 or 5 months. Only

after that time did we feel enough rapport had been established to enable

us to attempt to elicit information about socioeconomic status with any

accuracy. (See Appendix 7 for a copy of this schedule.)

Weekly Marketing Items and Food Budget

With the help of the two interviewers, a schedule was constructed

which elicited the frequency, place of purchase, and cost of the foods

commonly purchased. This information was used to construct a monthly

food budget and was generally elicited at the same time as the socio-

economic data, both being rather sensitive in nature. In addition to

this information, shopping lists were obtained from several families

and other families were accompanied on food-buying trips by myself.

The object of this data was to determine what kind of access people had

to food supplies and how they allocated their resources with respect to

food purchasing (see Appendix 8).

Maternal Reproductive History

This schedule included data about parity, number of pregnancies,

complications of pregnancy, miscarriages, length of pregnancy, and pat-

terns of breast-feeding. This material was not deemed to be sensitive

because women tend to talk with each other and other people about their

ills and pregnancies, thus it was administered early in the study to the

mother of the index child. Generally, the women seemed to enjoy reminisc-

ing about their pregnancies and problems associated with them. (Appendix

9 contains a copy of this schedule.)

Dietary Data

In addition to buying patterns relative to food, dietary data were

elicited in two ways. The first was a 24-hour recall of what the index

child had eaten. Four of these were obtained, one for the weekend and

one for the weekday during the dry season (December to April) and one for

each weekend and weekday during the rainy season (May to November). Be-

cause of the difficulty of being able to arrive on the appropriate day

for the recall, some were done by us as 24-hour recalls and the rest

were left with the mother or mother substitute to be filled out as a

diary for the day needed, e.g. a Saturday or Sunday. We did not have

high hopes for the success of this method and only resorted to it

several months into the study when we lacked needed data. Much to our

surprise, the mothers were highly motivated and we received all but two

reports back. These data were not used as a way of determining indi-

vidual intake because of the errors inherent to recall data (Acheson

1980; Bernard et al. in press; Morrison et al. 1959) (e.g. 20-50%

rate of error in recording intake), but rather as a way to get an idea

of the typical diet of the child in relation to family food consump-

tion and customs associated with mealtimes.

Q-Sort Techniques

A Q-sort of 100 cards of local foods most frequently consumed was

devised. The Q-sort techniques have been adapted from Stephenson (1953).

Correlational methods for analysis of ordinal and nominal data were em-

ployed as suggested by Cohen (1957) and Wittenborn (1961). The Q-sort

used in this research was adapted from that described by Fitzgerald

(1977), who used it to develop a food choices game. The same type of

game was developed for this study and was adapted to local food habits.

Fitgerald noted that the advantage of using this type of approach rather

than just an interview to elicit recall data is that it is non-

threatening and involves active participation. We found this to be

true. Even though the mother, or the principal caretaker, of the index

child was the one asked to perform the sort with the food cards, many

times it ended being a family consensus, which is probably a more accu-

rate reflection of the food habits of the family anyway according to the

findings of Bernard et al. (in press) regarding the increased accuracy

of recall data among groups as compared to individuals. Five different

kinds of information about food habits were elicited by use of this

procedure (not all at the same time). First a food frequency test was

done. Then the mothers were asked to sort those foods most liked and

least liked, those considered most nutritious and least nutritious,

those most appropriate for the index child, and the informant's own

classification of the 100 cards by whatever criteria she cared to devise.

Observation of the Index Child

In addition to the data mentioned already, each child was observed

for 4-5 hours during the morning of one day. Procedures described by

Spradley (1980) and Wilson (1977) were incorporated into an adaptation of

child-following as described by Wilson (1974). The behavior of the child

and the family in general were the target of this observation. Although

what was eaten by the child during the time the observer was there was

recorded, these data were not used to determine the intake of the child

but rather to learn something about mealtimes, intrafamily food distri-

bution, how the child was cared for, how he/she ate, and family inter-

actions. Because of the large number of children to be observed in a

relatively short time, it was considered unfeasible to carry out child-

following as Wilson had described it. Instead an attempt was made to

observe during a similar period of time in each household in the same

way so that the observer bias at least would be the same and, it was

hoped, the data would provide a sample of what happened during the same pe-

riod of time in each child's day. This observation was done only after at

least 3 or 4 visits monthly had been carried out. Because it is rather

inconvenient for a family to have a stranger in their house for half a

day, appointments were set up in advance. The observations were car-

ried out only by myself. The people were very cooperative and, al-

though undoubtedly some things were changed because of prior knowledge

of the visits, i.e. the house was made cleaner and perhaps a special

effort with the food served was made, the observations were believed

to be of great value in understanding the family environment. The

cooperation of the family seemed to be enhanced also by the fact that

we made appointments with them. Although the sampling procedures may

have suffered because people knew an observer was coming, rapport with

and confidence in the researchers were increased.

The total observation period was usually 4-6 hours in length;

however, the child's behavior was recorded every 15 minutes for 3

minutes on a checklist during only 3 hours of that time. Ethnographic

notes were recorded every 15 minutes during the entire observation

period, and included information about behavior, food eaten, and any

other activities which were of interest (see Appendix 10 for a copy of

the checklist of behaviors). The family was informed before the obser-

vation was set up that I would be making notes of the child's behavior.

Because the focus of attention was on the index child, the pressure

was taken off the other family members to a large extent, thus making

them more at ease. A short, informal history of the family was also in-

cluded at the end of the observation because by that time I had been able

to elicit that kind of information. Photos were also taken to illustrate

some of the child's behaviors. Because photos are expensive, they were

taken only when the child changed behavior rather than every 15 minutes.

A copy of the photos was given to each family. This gesture was very

well received.

One of the interesting offshoots of this in-home observation was

a first-hand acquaintance with Costa Rican cuisine because I was expec-

ted to eat with each family. Because among Costa Ricans sharing food

is an integral part of social behavior, these invitations were not to

be taken lightly, even though most families could ill afford another

mouth to feed. I used the opportunity dining with each family pres-

ented to gather data about typical menus served. A wide range in the

quality of the cooking was noted; some food was excellent and some was

terrible. I found that good cooks can make tasty dishes with very

little money to spend.

Fees Paid to Subjects

In order to provide an incentive for each family to remain in the

study for the entire time, the project had been set up to pay a certain

fee each month. Originally this sum had been set at $5; inflation,

however, had so affected the exchange rate that this sum had become a

fairly large amount in colones. The project advisors, Dr. Diaz Amador

and Dr. Camacho, did not believe that it would be wise to pay that sum

because it would set a precedent which could make data gathering harder

for other individuals and organizations whose budgets did not include

money for fees to be paid to subjects. It was also feared that too

high an amount would affect socioeconomic status unduly. The sum

finally agreed upon was (125, which was at the time the equivalent of

$2 and was approximately the amount a semiskilled worker would receive

for a day's wages. The extra money was used to give each family the


photos mentioned earlier and for a small gift for each child at the end

of the study.

I believe that the monthly sum did help keep a number of the

families in the study, particularly those who were extremely poor. I

provided each family with my telephone number and instructions to call

me if they were going to move so that we could get directions to the

new home. Also they were to call if for some reason they could not be

there on the day we were to visit,because we were trying very hand to

be as precise as possible in getting measurements at one-month intervals.

I really had no great hopes that people would call because almost all

had to use public telephones, but much to my surprise I regularly got

calls advising me of changes needed in scheduling,and of changing

addresses from those three families who proved to be nomadic. Also a

number of the mothers began to call periodically to ask advice about

many things. One of the ideas I tried to impress upon each family was

how important the information they could give me was. Because I re-

garded their input as important, they too seemed to take things more

seriously. At any rate we lost none of the 44 families during the 6

months of observation.

Table 3 presents a summary of the various methods used, along with

their frequency, who performed them, and which informant was used.

Phase IV: Data Analysis

The data were coded and entered into the computer for data manipu-

lation and analysis. The Statistical Analysis System (SAS) was used for

descriptive and inferential statistical analyses, including multivariate

techniques for correlation, analysis of variance, and various





(D I0


c 7

() -
w w








4JU 4J U mD

* E 0n > 0
S0)3 O E (3 -m
S- 0 : c-0 U S- S- -
i0 *r- 0 : o s-,-
>, 4i 4-'10 04-' -- ( = 3
S Lv, Sj -o 0o -o IS-
a0 0 0 ( U 0 Si
u- Co m )O u c
L V) M: c: u



4-' -C


4- ,-

> 3 4-1

U) 0 0

4-- S-
it en 0
* r- c s

r >
+- r



t1 1




41 4-

C >,

i 4"


C 3

C -'
a 4-



0a '







( o

S- a)

o -

4 -P

o* 10

0- 0
c O
0. 0


4- 0
3 4-'

0 S-
4.) 0
- O

0 LC
40 -




R C-
c L
Q J.






nonparametric statistical techniques. These statistical analyses, along

with a discussion of the results, will be discussed in more detail in

Chapters IV and V.


Family Networks

Social interaction in Costa Rica is based on kinship. At all

levels of social class, the persons with whom an individual is most

likely to celebrate New Year's Eve or Christmas, or upon whom he or she

depends in times of emergency, are family members. First impressions

may lead the outsider to assume that there exists a great deal of

"neighboring," i.e. interaction between nonrelated individuals who live

close to one another. A closer look, however, will show that in most

cases where neighbors interact they are also related. The barrios in

which this study was carried out were dotted with clusters of two or

more related families. Over 50% of the 44 families in the in-depth

study had relatives as neighbors.

Although the scope of this research did not include a rigorous

description of the networks which do exist, an acquaintance of any depth

and over a length of time will automatically bring one in contact with

these kinship-based networks. Usually there are two major networks

within each household: one with the wife as a link to her family and

the other linking the husband to his family. Sometimes one of these two

networks may be nonfunctional, depending on the physical distance be-

tween family members and the strength of their relationship with one


The basic importance of the family within the overall Costa Rican

social structure has ramifications for the nutritional status of the

child. Among the families followed, I found this kin-based network to

be the primary source of financial assistance in times of economic

depression, not unlike the situation observed by Safa (1974) in the

Shantytown families of Puerto Rico. It is also the main source of

mother substitutes or child caretakers should the mother of a child

need to work or should something happen to her. There is a lot of food

sharing, not only of foodstuffs but also of meals eaten at one another's

house. The latter is particularly important for children.

True (1976), in the study of marihuana users, mentions that

among the poor there are two social spheres, that of the woman and that

of her husband or companion. My research did not contradict this find-

ing but rather found the same kind of division. My experience, how-

ever, put me in contact primarily with the people included in the mater-

nal networks.

Family Structure

Of the 44 families in the in-depth study, 11, or 25%, were matri-

focal families. Only one did not fit the definition of the "Queen Bee,"

or "grandmother,"matrifocal family described earlier. The one which did

not fit was a "Queen Bee" type which had broken down; i.e. the grand-

mother and great grandmother had refused to let a granddaughter and her

children live with them any longer. However, she still spent most of

the day with them. It was also noted that "Queen Bee" families could

include grown sons who were not married.

Of the 44 families, 11, or 25%, were extended families. The

extended family as I found it included various persons from different

nuclear families, although these might not be complete families, e.g.

daughters or sons and their children and possibly a spouse. The

extended family always had a male as its head, however.

Of the 44 families, 22, or 50%, were nuclear families. In San

Jos6, however, the designation nuclear family can be misleading. For

example, I followed two children from nuclear families who lived in

one of the barrios of Alajuelita. The first family consisted of Miguel

and Antonia and their two children. They lived in a complex of three

other separate houses on the same block. Antonia's parents lived in

front; her sister, at one side; and her brother,in a house behind their

house. They all shopped together on Saturdays (at the farmer's market

or the central market) and the children were in and out of all the

houses and could eat at any of them. Goods, food, and services were

exchanged freely. For all practical purposes, they functioned as an

extended family but had separate dwellings.

Not far from Miguel and Antonia, another family lived in similar

circumstances. Jose and Sara Lived as a nuclear family with their three

children; however, they too lived in a family complex consisting of

several of Josd's brothers, a sister, and his parents. Figure 4 shows

this layout-a pattern frequently found among the families in this

study and also among many middle and upper class families who were

social acquaintances. Thus it seems to be that the term nuclear

family has a very different connotation in Costa Rica from that exist-

ing elsewhere. The main difference between the nuclear family and the


Jose Empty lot
and Sara proposed site
of house for
another brother

Sister Courtyard

Jose's Jose's
Parents older Brother

Main Street

Figure 4. Diagram of the Family Complex of Jos6 and Sara.

extended family in many cases is that in the former the family members

have separate roofs and may cook separately; otherwise they act much

like a traditional extended family.

R. T. Smith (1970) had observed almost 30 years earlier that matri-

focality is associated with a class position at the lowest rank of

society. Vertical mobility is limited and the men are usually unskilled

laborers holding jobs which have little prestige. He sees matrifocality

as being the final phase of a cyclical process which includes a period

of sex experimentation and spouse selection, the nuclear family, and

finally the matrifocal household, which occurs when the role of wife-

mother gradually gains dominance over the role of husband-father.

In Costa Rica, as greater job opportunities have become available

for women in the textile and garment industries, women have become more

independent, a phenomenon Safa (1974) also found in Puerto Rico. With

this independence has come a decreased need for marriage to obtain

financial security and thus leaving little, in many cases, to hold a

man and his wife together. Here as in Puerto Rico, each is bound to

his/her own kin group, blood ties being more important than marriage.

Given this situation, marital problems such as alcoholism or unemploy-

ment on the part of the husband have tended to result in the formation

of the matrifocal family. The "Queen Bee," or "Grandmother," family

seems to result when daughters, able to obtain employment, find formal

marriage unnecessary or impractical because they can remain in their

mother's house and raise their children there. Increased availability

of employment for women may be encouraging a trend toward elimination

of the nuclear family phase of family evolution described by Smith.

The Program of Social Welfare and Family Aid (Programa de Desarrollo

Social y Asignaciones Familiares), which has been created to aid low-

income children and pregnant and lactating mothers, may have a ten-

dency to make poor families, particularly female-headed families, more

dependent on public aid, not unlike the situation described by Safa

(1974:46) for urban Puerto Rico.

As noted earlier, in Costa Rica the matrifocal family particularly

the "Queen Bee," or "Grandmother" variant, has been considered an excel-

lent adaptive mechanism enabling poor women to obtain economic necessi-

ties and care for their children and to educate them (Lopez de Piza

1979:15). The data gathered in this study tended to verify this find-

ing. Only 3, or 27%, of the matrifocal families had children in the

low weight/height group. Although income on the family level did not

differ significantly between male-headed and female-headed households,

per capital income did. Female-headed households had a per capital income

of t580/month ($14/month based on an exchange rate of $1/45) whereas

male-headed households had a per capital income of 821/month ($18/month).

The general linear model (GLM) (SAS, 1982) was used to run an analysis

of variance and a Duncan test of the means. This analysis indicated

that female-headed households were significantly larger, with a mean of

8.1 persons as opposed to 5.7 in male-headed households (F=7.34, pO.01).

In general, then, female-headed households had smaller individual

salaries but there were more persons working and pooling resources,

which compensated for the low salaries.

An example of this phenomenon is the household of Doia Pilar,

who left her husband years ago because of his drinking. She lives with

her daughters and three grandchildren. Three of her daughters work in

the garment industry, making clothing. Sara and Reina work in a factory;

Luz works in a smaller establishment. Yolanda, the youngest, has been

under treatment for psychiatric problems. Dona Pilar cares for the

children and also sews at home. (Figure 5 shows a diagram of this

family.) The three working daughters earn g2400, 02000, and (1200

respectively for a combined income of (5600 ($124) per month, which

puts them slightly above the mean for female-headed families (X=04697,

or $104) and for male-headed families (X=C4719, or $105). A total of

03742 ($83) is spent on food each month and (650 ($14) for rent and

utilities. This leaves &1208 ($27) to be spent on clothes and other

incidentals. They receive medical care through the social security

clinics and the neighborhood health center. They speak of wanting to

move to a new housing project but remain where they are because rent

is cheap, allowing them to spend more on food, clothing, and other


Food-Buying Patterns

Relating family structure to food-purchasing patterns, I found

that in the female-headed household the person in charge of purchasing

is usually also the principal child caretaker. This means that the

buyer for a female-headed household has much less mobility and greater

dependency on local corner grocery stores (pulperfas) and street vendors

whose products are usually more expensive. She has the alternative of

taking small children along, leaving them alone for a period of time,

or sending an older child to the store with a list of groceries.

In male-headed families, in contrast, 36% of the time the man is

the main buyer and 39% of the time both the husband and wife do the shop-

ping. Only 12% of the time is the wife the principal buyer.

Doa Pilar
Doi~a Pilarl

Sara Reina Luz Yolanda

Juan Maria Jose

(Index Child)

A Absent

Figure 5. Kinship Chart of the Family of Doia Pilar. This is an
Example of the Extended Matrifocal, or "Queen Bee," Family.


-- --


In general, male-headed households were better able to patronize

the subsidized state-run food outlets, the central market, larger

grocery stores (almacenes), and other places which tend to offer better

prices, whereas female-headed families were dependent on what could be

bought within walking distance. Male-headed households spent an

average of 43884 ($86) per month on food, which was 82% of the total

family income. Female-headed families spent an average of t3790 ($84)

per month, or 81%, of their total family income. Per capital expenditure

on food, however, differed more. Male-headed families spent /670 per

capital ($14), whereas famale-headed families spent (468 per capital ($10)

per month (see Table 4). All families used bus transportation and most

bought groceries with cash (over 90%) rather than on credit. Family

members tended to get food or money to buy food from their kin networks

rather than having the local corner grocery store extend credit.

Child Rearing

Children occupy a favorable position in Costa Rica. The birth of

a child is usually a welcome event. Lomnitz (1977) notes that in

Mexico this attitude is common among members of marginal populations

because children represent a potential cheap source of income. Their

cooperation is dependable and the cost of their maintenance insignifi-

cant. The same is apparently true in Costa Rica among the poor, al-

though children also seem to be valued for themselves, as family mem-

bers and as sources of pride and prestige. Positive attitudes toward

children are evident throughout Costa Rican society. At the national

level the Patronato Nacional de la Infancia has been formed to serve

as the watchdog of children's rights. Over 60% of community health

00 00 00 00
00 00 00 00


A'nq jou pLg


jonbt a~ePS







c 0 Co


i i M Ln


0I I I I II o
I I I I 1 I 3 J i

a .-- to to u'-o c< a co n
0. 0 CJ- C, 01 cn

i- ai
c -j c- J mm -U

C- *r 01 (Cl -

00 00 00 00
00 00 00 00
oo oo oo oo

co CIA



o I I

I I 0 IC

M 2 m ID

I I I I-


a a

U Cl


Rnq ;ou pLa


AonbIt[ @a2S


s WIl I ej

S1 B





pooJ a@e;S





00 00
00 00


CD n 1 1

E-- F --- F E- il n E ) s- E--
- :LL C U. = C UL- N U- OO U. -" 3Z u 0-
SL ] *r-. \- 00 0C
0 0 C U*r- S_ 4- *
co L, cl

00 00 00 00 00
00 00 00 00 00

CO)- CO0 C O0 C0 CD D

In -I- r- CI M I -

I I II | | 10 |
I I I 1 I I




co co 711

I I Min ? I I I I C

a U
_E 4-)

+ 0
UO 0
) O4-

U] -
0 O

c c
s- U

0 0

0 (O
*- U



0) -
0 In


0 0

-0 4-
om aa

0 0
4- *-
a a)

S-e 0
U]IL 0

=) 0)
00 U '

a, s- a,

L>'1 0)

a, u -a

&_ 0
-) I + ll

OCL> O 4-

ajuQJ a
Q. l0E
Pr- !
=1 t S

a-<:u- Q
10 <> -

programs concern maternal and child health (Freer Miranda 1980). There

is a children's library in the main plaza of San Jos6 and a fine chil-

dren's hospital, as well as many other services for children.

Aside from these indications, I found the attitudes toward chil-

dren in most homes to be those of patience and warmth. During the

course of my observation I noted very little corporal punishment. In

only one case was discipline carried out in this manner. In general,

the behavior of children was controlled by diverting their attention or

by removing them from the immediate area. In response to my questions

mothers and caregivers told me that it was not customary to beat or

spank children. This contrasted with what I had observed in the United


Family Interaction and Child Development

I spent time in each home observing the index child in the manner

described previously. Generally, after the first few minutes the child

ignored me and went on about his/her usual activities. In the physical

environment I noted a lack of high chairs or raised stools or other

equipment enabling the child to eat comfortably at the table. Children

ate sitting on their mother's lap, sitting on a low chair reaching up

to the table, standing on a chair, or sitting on top of the table or

on the floor. Only 11 families,or 25%,of those observed had special

furniture or made any provision for the child's comfort and ease while

eating. Most children were encouraged to eat by themselves although

some were completely fed and in the case of some a combination was

employed-the child eating alone and then the mother helping to make

sure that he/she had eaten enough. The lack of children's furniture

was probably the result of both tradition and low income. Locally made

wooden chairs and other furniture pieces were available and were rela-

tively inexpensive.

An analysis of variance obtained by use of the GLM procedure of

the Statistical Analysis System showed some significant differences in

certain activities observed in the home, based on the sex of the child.

Girls were observed more frequently than boys playing by themselves

(F=4.71, pO0.04), playing with toys which stimulate fine motor develop-

ment (F=9.47, p'O.004), and engaging in imitative play, e.g. playing

house, pretending to be an adult, etc. (F=9.96, plO.003). Boys were

more frequently observed running (F=4.44, pO0.04) and playing with toys

which stimulate gross motor development (F=8.43, pO.O06) than were

girls. Mothers were observed stimulating language development with

girls significantly more often than with boys (F=6.64, pO.01).

It is not surprising that boys were more often observed running

than were girls. Soccer is the national sport, and one of the first

skills a boy learns is to run and kick a ball. Girls in Costa Rica are

expected to be homemakers; therefore, it is not unexpected to find them

engaged in such activities as playing house or playing with dolls or

other toys requiring the use of fine motor coordination and imagination.

Also, some of the differences observed may result from the fact that boys

between the ages of 1 and 2 years may be a little behind girls in their

development. Observations at an older age might yield different results,

because playing with some toys requires more coordination which boys may

not develop until later. Distinguishing physiological causes from

culturally induced differences would require much more in-depth observa-

tion, however, and was not within the scope of this study.

Most children had access to toys and played with them, although

the variety was not as great as in the United States. Most of the chil-

dren customarily played with other children, usually brothers and sis-

ters or cousins. Twenty-five percent of children were observed playing

with an adult male at least once, usually another family member, i.e.

father, older brother, or uncle. Forty-five percent of mothers were

observed playing with the child. Although almost every family had a

TV, very few of the children watched. Because electricity is expensive,

the TV was usually turned on only at night and on special occasions when

a number of adults were present. It was not at all unusual to find a

family living in a shack but owning a television set.

Prior to beginning the research, I had been told by middle class

Costa Rican informants that children of poor people were not at all stimu-

lated and that it was common practice to leave babies in drawers or cribs

without attention. I had also observed during my time in South America

that, at least among the middle and upper classes, children were kept

dependent for a much longer period of time than in the United States,

probably because of the availability of servants. For example, they

were fed rather than taught to eat with spoons. The children I observed

in San Jose, however, were encouraged to be independent and to learn to

take care of themselves. Although some children customarily drink a baby

bottle of milk each night before bedtime until they are 5 or 6 years

old, they still know how to drink from a glass and to eat using uten-

sils. The baby bottle is a habit or treat. I also noted that the

index child received a good deal of physical and cognitive stimulation.

Sixty-eight percent of the mothers were observed cuddling their child

at least once during the visit. Forty-five percent were observed en-

couraging language development by talking with the child. Sixty-three

percent of the children were observed conversing with other children or

adults. In short, the majority of homes provided an environment that was

full of potential stimuli, certainly not what could be called a severely

deprived atmosphere.

Child Health

Based on my sample, I found children living in San Jos6 to have

good access to primary health care through the community health program

operating in each health center. As an indicator of the quality of

health care, over 90% of the initial 107 children were found to have had

their BCG, DPT, polio, and measles (rubella and rubeola) immunizations

completed for their age. In comparison, records show that in Florida

93.7% of all school entrants had been immunized. This higher value,

however, is undoubtedly influenced by the fact that obligatory immuniza-

tion begins to function at school entrance age in Florida. Nationally,

only 60% of the children aged 1-4 had been immunized. Although spe-

cific figures for children aged 1-4 or 1-3 were not available, the

Flroida State Health Plan indicated that immunization rates for this

age group are considerably below that of school entrants, probably more

in line with the national rate noted (Florida State Health Coordinat-

ing Council 1981). (Table 5 gives a breakdown of immunization figures

for the group followed in depth.)

Nine percent of the children in the longitudinal study were ob-

served to have dental caries upon a simple inspection. Other than this,


no clinical symptoms of malnutrition were observed. (Table 6 gives a

breakdown for the group as a whole.)

Table 5. Immunization Rates for Costa Rican Children from Poor Urban

Type of Basic
immunization complete

Basic No
incomplete immunizations

--------------------.------- % .-------------------.__ .--------
BCG 95 5

DPT 93 7

Polio 98 2

Rubella 96 2

Rubeola 96 2

Note: Percentages are based on a sample of 44 children selected
at random from health center birth registers.

Table 6. Clinical Symptoms of Malnutrition in a Sample Population of
Children in San Jose

study group
Symptom (N=44)

Group excluded
from longitudinal Total group
study (N=63) (N=107)

--------------------------- % -.----- .--------------------
Dental caries 9 11a 10

Apathy 2 1

Irritability 2 1

Lethargy 3 2

Dry hair 3 2

Scant hair 2 1

Diarrhea 3 2

aMost of the symptoms other than dental caries were observed in
two undernourished children eliminated from the study because of other
problems complicating their nutritional status.

The mother is the person who usually cares for the child when he/

she is ill, but grandmothers and aunts may also help. Fathers were

never indicated as being directly involved in child care. Most of the

families in the sample went to health centers, with a slightly smaller

number going to social security clinics for medical care. Private

physicians were used by a few and about 5% claimed to have no access

to medical care. (See Table 7 for a breakdown of these figures.)

Diet and Food Habits as They Relate to Children

Although breast-feeding was once the norm in many developing coun-

tries, including costa Rica, recent work has indicated that both the

incidence and the duration of maternal lactation are declining (Jelliffe,

E. F. P. 1979; Popkin et al. 1980). This trend was also noted in Costa

Rica (Mata et al. 1981), and programs have been begun to encourage in-

creased maternal lactation.

Data about maternal lactation were gathered on the 44 children in

the longitudinal study. In an effort to compare these data with earlier

information, Table 8 was constructed to show our results from 1982 as

compared to results for Region I (Metropolitan San Jos6) and the country

as a whole in the national survey of 1975 (Diaz Amador 1975). This in-

formation appears to indicate a decline in numbers of children never

breast-fed and a general trend toward longer periods of maternal lacta-

tion, although the percentage of children breast-fed more than one year

is reported to be the same.

Table 9 shows comparative data regarding the age at which bottle

feeding is begun and indicates a decline in babies never bottle fed and

more bottle feeding started in the first month of life. However, a

Table 7. Who Cares for the Sick and What Facilities Are Used

Longitudinal Group excluded
study group from longitudinal Total group
(N=44) study (N=63) (N=107)

------------------.- -.-.---- % ------- -.----_--------------


Mother 82 86 84

Grandmother 9 1 5

Aunt 2 5 4

Mother and
grandmother 5 5 4

Others 2 3 3

Total 100 100 100

Facility used

Health center 34 46 41

Social security 43 35 38

physician 7 5 6

Both health
center and
CCSS 5 5 5

None 7 3 4

Other 4 6 6

Total 100 100 100

Table 8. Comparative Data on Duration of Maternal Lactation

Length 1982 1975 total
of time study Region I country

--------------- % ---------------

Not breast-fed
at all 9.1 16.0 13.3

Less than 1
month 6.8 12.2 13.3

1-3 months 25.0 30.2 20.3

4-6 months 13.6 3.7 4.3

7 months to
1 year 18.2 1.8 5.1

More than 1
year 27.3 27.4 34.1

Do not know 8.5 9.5

Note: Time periods given on the 1975 study have been combined so
that they are equivalent to the time periods used in this study.
aRegion I includes metropolitan San Jose.

bDiaz Amador, 1975.

Table 9. Age at Which Bottle-Feeding Is Introduced Among Costa Rican
Children-Comparative Data

Length 1982 Region I country
of time study 1975a 1975b

---------.------- % -----------------

Not breast-fed
at allc 4.5 7.5 5.9

Less than 1
month 61.4 59.5 55.0

1-3 months 11.4 18.8 17.8

4-6 months 11.4 1.9

7 months to
1 year 6.8 2.8 3.1

More than one
year 4.5 -

Do not know 8.5 16.1

aRegion includes Metropolitan San Josd.

bDiaz Amador, 1975.

CTime periods have been combined in the 1975 study to make them
equivalent to the time periods used in the current study.

large number of mothers are apparently introducing bottle feeding later

as well. The changes shown in both Tables 8 and 9 may be the result of

Ministry of Health-sponsored campaigns to promote breast-feeding. It is

probably too soon, however, to say definitively that this is the case.

Data were also gathered about why mothers discontinued breast-

feeding. Table 10 shows a summary of these data compared to reasons

given in 1975. The basic reasons a mother decides not to breast-feed

are insufficient milk supply and child rejection of breast-feeding.

My research also indicated that 23% of the mothers had intro-

duced solid food within the first month of the child's life and by three

months of age 77% of the infants had been introduced to regular food.

By the end of the first 6 months, 93% of mothers had introduced their

infants to regular food. In fact, looking at the incidence of breast-

feeding without taking any of these other factors into account is mis-

leading and can give a false impression. What actually happens is that

most mothers are breast-feeding and bottle feeding simultaneously,

starting when the infant is very young. Regular food is added as soon

as the infant can tolerate it, beginning in the first month of life.

Infants are first fed bits of food from the table and then gradually

given their own portions.

Information about whether a child is being breast-fed is not as

important as estimating the amount of breast milk received. The latter

is difficult to do. There are mothers who report breast-feeding for

6 to 12 months, but after the first 2 or 3 months only once or twice

a day; in this way breast-feeding only makes a small contribution to

Table 10. Reasons Given for Discontinuing Breast-Feeding-Comparative

1982 Region I country
Reason study 1975a 1975b

------------------ % --------

Milk dried up or
did not have
sufficient supply 41.2 46.8 42.4

Child rejected
breast-feeding 14.7 15.6 14.5

Mother pregnant 11.8 -

Mother's illness 8.8 6.5 6.8

Mother's work 5.9 7.8 5.1

Physician's orders 5.9 1.3 2.9

Child biting 5.9 5.2

Child hospitalized 2.9 2.3

Child ill at home 2.9 0.6

Advice from friends
and relatives 1.0

Believe cow's milk
is better -- 0.6

Other -16.9 23.8

aRegion I includes Metropolitan San Jos6.

bDiaz Amador, 1975.

the child's diet. The scope of this study did not include estimating

the amount of breast milk received by the child.

Other information from 24-hour recalls for the index child show

two facts of special interest. First, milk, whether breast or bottle,

is the primary source of protein for the child under 3 years of age.

The second item of interest is that, although rice and beans are staples

of the Costan Rican diet and good sources of protein, the child under

3 years of age is by tradition fed rice and caldo de frijol, or the

juice in which the beans have been cooked. Although caldo de frijol

is a relatively good food from a nutrition point of view, it is not

equivalent in protein to the whole bean. Government policies which

seek to decrease milk available to small children from poor families

and substitute rice and beans could be detrimental for children under

3 years of age.

A brief discussion on some of the more pertinent social and cul-

tural factors which may have important influences on the nutritional

status of the urban poor child in San Josd, Costa Rica, has been pre-

sented. The next chapter will deal with the actual comparison in

growth and development between children of low weight/length and those

of normal weight/length.


Anthropometric Data

A total of 107 children, 54 males and 53 females, born in 1980

were chosen at random from birth registers in low income areas asso-

ciated with the health centers of Alajuelita, Hatillo, Pavas, Tibas,

Paso Ancho, Guadalupe, and Cristo Rey. In the initial survey, measure-

ments of length, weight, head circumference, arm circumference, tri-

ceps, and subscapular skinfolds were taken. In addition to these measure-

ments, the Denver Developmental Screening Test (DDST) was given to each

child and a family registry and assessment of the environment wenr done.

Table 11 shows the distribution of the children surveyed with respect

Table 11. Distribution of Weight/Length Percentile Rankings by Sex
of the Initial Sample of 107 Children from Poor Urban
Households in San Jose, Costa Rica


Sex <5th 5th-10th >10th-25th >25th-50th >50th

Male 5 5 7 14 16

Female 3 3 8 19 18

Total 8 8 15 33 43

Percent 7.5 7.5 14 31 40

Note: Percentiles are.based on NCHS standards (National Center
for Health Statistics 1977).

to percentile rankings. As indicated, only 7.5% (n=8) of the children

were found to be below the 5th percentile. Another 7.5% (n=8) fell be-

tween the 5th and 10th percentiles. These findings indicate a low per-

centage of severely undernourished children. It was decided therefore

to focus on the moderately undernourished child. The ministry of

Health criterion of 90% of the median of National Center for Health

Statistics (NCHS) weight/height standards and/or first degree malnutri-

tion by Gomez's criterion was used to delineate a group of mildly under-

weight children to be followed and compared with normal weight/length

children. These criteria were used even though they included more chil-

dren in the low weight/length group than would occur by use of the 5th

percentile category. This was felt to be a reasonable division given the

fact that less is known about problems involving invisible or mild under-

nutrition in children, particularly the boundaries which define it.

The families of 19 of the children in the low-weight/length group

agreed to let them participate. Twenty-five normal weight/length chil-

dren were chosen at random and also were asked to participate. Although

the drawing was done randomly, more girls than boys of both the low

weight/length and the normal group were not allowed to participate. In

the case of the 25 normal weight children, substitutions were made by

drawing another random number. The result was 18 females and 26 males

in the final study, even though the initial group was evenly divided.

After the group for the longitudinal study was selected, six

additional monthly measurements of length, weight, head circumference,

and arm circumference were taken. Triceps and subscapular skinfold

measurements were taken at the beginning and at the end of the 6-month

longitudinal study. For the purposes of this particular study, only the

weight/length figures are analyzed. Table 12 shows the distribution of

the smaller group of 44 children with respect to percentile distribution

of the initial and final measurements.

Table 12. Distribution of Weight/Length Percentile Rankings of Initial
and Final Measurements by Sex in a Sample of 44 Children
from Poor Urban Households in San Jos6, Costa Rica, Who Were
Followed for 7 Months


Sex <5th 5th-10th >10th-25th >25th-50th >50th

Initial measurement

Male 3 5 3 8 7

Female 1 3 5 7 2

Total 4 8 8 15 9

Percent 9 18 18 31 21

Final measurement

Male 3 5 6 5 7

Female 2 7 5 4

Total 5 5 13 10 11

Percent 11 11 30 23 25

Note: Percentiles are based on NCHS (National Center for Health
Statistics 1977).

Growth Patterns

As Table 12 indicates, there was some shifting in the percentile

rankings over time. Although there was a slight increase in the per-

centage of children under the 5th percentile at the end of the study,

there was also a decrease in children in the 5th-10th percentile cate-

gory and an increase in the number of children in the 10th-25th and

50th percentile categories. The overall impression is that growth

shows a great deal of individual variation and tends to occur as a

series of mini-catch-up spurts for both height and weight. Specific

illness episodes or other events (e.g. weaning) only partially explain

the observed variation in growth magnitude and velocity. Figure 6

illustrates this pattern well. The child is a female of normal

weight/length with no reported illnesses or unusual family circum-

stances, but whose growth velocity varies a great deal from month to

month. Figure 7 illustrates the classic pattern of the mother becom-

ing pregnant and weaning the child. The child, a male, was normal in

weight/length but dropped 0.5 kg. (4% of body weight) during the ad-

justment period. In this case the child suffered no severe effects

because he had reserved body fat. The underweight child in this same

situation would not be so fortunate. Figure 8 illustrates another

typical problem related to growth. The child, a male of normal weight/

length, was cared for by the grandmother while his mother worked. The

grandmother went to work, leaving a rather emotionally unstable aunt

in charge of the children. The index child of this family lost weight

(1.5 kg., 13% of body weight) rather abruptly and did not begin to re-

gain it until the grandmother returned to stay with the children.

These examples have been chosen as representative of typical

growth patterns found among the children followed and as such demon-

strate the need for many measurements over time in order to determine

the state of a child's health and nutritional status. Cross-

sectional studies do not necessarily reveal this individual variation.

In w


C o



o r

CO a


L 5



I I I I | I I
0 0 0 0 0 0 0 0 0
oLO 0 L 0 0 n nO
Ic\j T d L r O

( ) 0 0H0)03M







0 c

r-r 1-


So o o o

m 4 Cj 04 0Q C- C5 6
\ -01
\ CO C'1 C0

e S-

(5i iM
| I -| ---- [ --- r -- i i i -- -- i --- -- -^ ---- e <
S ~ ~ ~ ~ ~ C IfEl 001 OOl
o ^ o < o r- i cM o r- io cM
CT C C'J N cM T- T- T- 0 d d d r
^- *^ ^ *^ ^ ^ r- r ^ ^ r- ^ r-

( B~i) 1H93M I

w: =
0. uj<
) 20>-crU
co ) (0 zg 0-1
m) < Zo=
Qc n- ~h.23









l I I I I i i I I I L I

0 (O N 0 00 o o o o
-N 0- dd

( By ) LHDF3M

Maternal Reproductive History and Birth Weights

A SAS general linear model (GLM) analysis of variance showed that

no significant differences existed between mothers of low weight/length

children with respect to total number of pregnancies, age, the order of

the index child in terms of total pregnancies, or the order of the

index child in relation to living children.

Birth weight was found to be lower among the low weight/length

group as compared to the normal weight/length group (F=5.94, pCO.02).

Because of the large number of boys in the group, particularly the nor-

mal group, analyses were also done by sex. Low weight/length boys were

found to have significantly lower birth weights than normal weight/

length boys (F=4.64, p50.04). Low weight/length girls, however, were

not found to differ significantly from normal weight girls with respect

to birth weight, even though their average birth weight was lower

(2818 grams as compared to 3060 grams). In this study birth weight was

not controlled but rather was one of the parameters allowed to vary. In

an effort to determine to what extent birth weight was influenced by

genetic factors, the height and weight of each child's mother were also

obtained. An analysis of variance indicated no significant difference

between the height and weight of the mothers of low weight/length chil-

dren and those of the mothers of the normal children. In the former

group both weight and height tended to be lower but not significantly

so. This suggests that the low birth weight associated with the mildly

underweight child may be more related to maternal undernutrition during

pregnancy than to other factors.

Illness Classification

In addition to the health history obtained initially for the year

prior to the beginning of the study, health histories were elicited

monthly from primary caretakers. An illness classification was devel-

oped based on the work of Martorell and Yarborough (1983). Diarrheal

illnesses include diarrhea and gastroenteritis, including pega. Respira-

tory illnesses include bronchitis, cough, cold, asthma, flu, and grippe.

Selected common symptoms include all of the above diseases and symptoms

plus fever, vomiting, skin infections, tonsilitis, parasites, inflama-

tions, ear infections, allergy, kidney or bladder infection, loss of

appetite, and dyhydration.

There were not significant sex differences in the average number

of illnesses, although males had more diarrheal illnesses and illnesses

included in the selected common symptoms category. Slightly more

females were, on the average, reported as having respiratory illnesses.

A large number of children (23, or 52%) were reported as having no

diarrheal illnesses.

A total of 160 illnesses, including 74 respiratory illnesses and

30 dirrheal illnesses, were recorded for the sample over a consecutive

7-month period (Table 13); 50% of the illnesses occurred during the

rainy season and 50% during the dry season. There was no significant

difference in the occurrence of respiratory illnesses by season.

Diarrheal illnesses showed a significant seasonal trend with a mean of

X= 3.00 cases in the dry season and X = 6.00 cases in the rainy season

(t=2.60, pO0.05). Select common symptoms showed no significant

seasonal variation with means of X = 20.00 and X = 26.67 for the dry

season and rainy season respectively (t=0.544 N.S.). The mean number

of illnesses per child was X= 3.64.

Table 13. Seasonal Frequencies of Illness Among Low-Income Urban
Children in San Jos6, Costa Rica

Month and Selected common
season Diarrheal Respiratory symptoms

Month 1
Dry 1 6 7

Month 2
Dry 5 12 24

Month 3
Dry 1 15 26

Month 4
Dry 5 8 23

Month 5
Rainy 7 6 22

Month 6
Rainy 7 13 33

Month 7
Rainy 4 14 25

Total 30 74 160

With respect to elicited health histories for the year prior to

the initiation of the study, significantly more children in the low
weight/length group were reported as having fever ( 2=4.62, pO.03)

than in the normal weight/length group. Other than this there was no

appreciable difference between the two groups in the reported incidence

of illness according to the health histories. Table 14 gives a summary

of the illness episodes reported in the histories.

Psychomotor Development

Table 15 gives a summary of the group as a whole with respect to

the results of the Denver Development Screening Test (DDST). Twelve

percent (n=8) of the normal weight/length group and 17% (n=4) of the

low weight/length group had abnormal or questionable results on the

DDST. It is suggested that a more discriminating instrument may be

necessary to detect differences between mildly underweight children

and normal weight children. However, more interesting is the fact

that 50% (n=5) of the children who were short for their age but normal

in weight/length had abnormal or questionable results on the DDST.

Unfortunately, these children, because they were considered normal in

weight/length and because they represented a different nutritional

problem, were not included in the in-depth study. These results sug-

gest that an additional follow-up of short-for-age children may be war-

ranted, even though they may not be technically considered under-

nourished by weight/length criteria.

Socioeconomic Factors

There were slightly more (X=2.5) children under 6 per household

in the low weight/length group than in the normal group (X=2.0); this

difference, however, was not statistically significant (see Table 16).

Household size tended to be larger, mother and father older, household

income less, and education of both mother and father less in the low

weight/length group. Total household income was calculated by summing



5- a 0 0U
EU E 3

U ) U

5- a I I 0 I I I I II I I

c/ .

0 o a)

0 0S
=-3 a o

0C C I ) c IN O I I I I I II I l

* o O-

0 s-

0 *l
00 0

33 M- 0I C

4- 5
0 In W

00- u) v1
+- ct c L3 r- Iu I (J i U I I I I I I I I C I

0 S-' 0 C- N0
0 4' 4-

20 0 S- D-

I- >u-u5-. uztj c nt)('=:0)
E^ '~ '

I- ~oau~~ Q>Q--(_U-(_cto QQQ S -[/3e O



I 1


"' CI CO

0 ~On

o0 lo 0

I I I 1
I I I 0

c0 I I


0 >
C-- r--

0. Ci

Lz1 00

c coO

CO'O, I fl


Lc No



^. 6


J 0

- C
s i

-I 0


0 o
z z


I- S
L3 0

r-- 03

"- CO
crC U

r- 0 a
m --E

nL 0 1

-r- 0 fO
- i-- "E LL..

I n m O

all salaries of family members plus the value of food or other items

which were gifts, and any other support, i.e. government aid. Since

this information was collected after about 5 months of visits, it is

felt to be a reasonably accurate estimate. The exchange rate from

colones to dollars was calculated at 45/1 which was approximately what

it was when the questionnaires were answered.

Table 16. Means of Selected Social Characteristics of the Families
of Low and Normal Weight/Length Poor Children in San Jose,
Costa Rica

Low Normal
weight/length weight/length
Characteristic (n=19) (n=25)

Household size 6.6 6.2

Children 6 and
under 2.5 2.0

Father's age 32.4 31.1

Father's education
(no. of years
completed) 6.5 7.0

Mother's age 28.2 25.3

Mother's education
(no. of years
completed) 6.3 7.0

Total household 4505 04872
income per month ($100) ($108)

Total food cost 03991 l3761
per month ($89) ($84)

The cost of food was calculated on a per month basis according

to the frequency of purchase of food items as reported to the investi-

gators. The prices were as accurate as possible given the frequent

fluctuations caused by inflation during 1982. A greater proportion of

income was spent on food in the households of the low weight/length

group (89%) than in the households of the normal group (77%).

Fifty-three percent of the families of the low weight/length

group and 36% of those of the normal weight/length group either owned

or were buying their homes. Table 17 gives a summary of the type of

tenure of the dwelling. More families among the normal weight/length

group were renting, and fewer owned their own homes.

Table 17. Type of Tenure of Dwelling in a Sample of Low Weight/Length
and Normal Weight/Length Children from Low Income Families
in San Jose, Costa Rica

Low Normal
weight/length weight/length
Type of Tenure (n=19) (n=25)

------------- -------------
Own home 32 16

Buying home 21 20

Rent home 42 52

Borrowed home 5 12

Total 100 100

aIn this case families were usually living in a house belonging to
some other family member, i.e. mother, father. Rent was paid when funds
were available.

More of the parents of children in the low weight/length group

were married than in the normal weight/length group (68% as compared to

56%). Forty percent of the parents of the normal weight/length group

and 16% of the parents of the low weight/length group were single.

Thirty-two percent of the households of the normal weight/length group

were female headed, whereas 16% of the households in the low weight/

length group were of this type.

The sample was slightly biased toward families with a more

stable residence because of the method of selection. Those who had

moved within the previous year were not usually included in the birth

registers we used to select the sample. Eighty-four percent of the

families of children in the low weight/length group and 72% of those

of children in the normal weight/length group had lived in their present

residence for three years or more. Fifty-eight percent of the families

of the low weight/length group and 56% of the families of the normal

weight/length group had lived 5 or more years at the same residence.

Thirty-eight percent of the fathers of the low-weight/length group

and 11% of the fathers of the normal weight/length group were unemployed

or worked only occasionally or seasonally. Seventy-nine percent of the

mothers of the low weight/length group and 67% of those of the normal

weight/length group did not work outside the home. Thus, not only was

employment less for men but also for women in the families of the low

weight/length group.

Data collected about buying patterns indicated that most families

did not buy on time; however, more of the families of the low weight/

length group did-32% compared to 12% of the normal weight/length

group. Thirty-two percent of the families of the normal weight/length

group also reported having some savings as compared to 5% of the

families of the low weight/length group.

With respect to environmental conditions, wooden houses predomi-

nated in both groups (58% and 64% in the low and normal groups respec-

tively). A few were made of brick or cement block (42% and 32% res-

pectively) (see Table 18). Most had cement floors and zinc roofs.

Condition of housing was generally good. Approximately 26% of the

families of low weight/length children and 24% of the families of normal

weight/length children lived in government housing projects. The

houses had 1 to 5 rooms with an average number of 2 rooms in both

groups. The number of beds per family ranged from 1 to 6, with a mean

of 3.5 beds per household.

Most families in both groups cooked with electricity although gas

wood and/or charcoal, and kerosene were also used by a small number

(25%). During the course of the study a number of people began to use

wood and/or charcoal again because the cost of electricity had risen

as a result of inflation and the increased cost of importing petro-

leum to Costa Rica. A total of 15% of the families of the low weight/

length group and 24% of the families of the normal weight/length group

were either using charcoal exclusively or a combination of charcoal

and electricity and/or gas. Charcoal and wood are local products and

relatively cheap. When a combination of cooking methods was used,

usually for items requiring long cooking periods such as beans, rice,

soups, and stews, charcoal was used, whereas the electric stove or hot

plate was used for foods which could be cooked rapidly, i.e. eggs,