Contents - Previous - Next

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

Human nutrition

The deprived family: Different approaches to its analysis

María de la Luz Alvarez and Fanny Wurgaft


A "deprived family" is here defined as one that is deficient in more than one of the following areas: social, economic, cultural, and psycho-social. There is observable external poverty (housing), and there are other internal deprivations that cannot be observed as easily, such as poor personal hygiene and insufficient knowledge of child rearing and child socialization. A family that has such a background is more likely to produce a malnourished infant, because the sum of these factors creates a milieu of great deprivation [1; 2].

The purpose of this paper is to present various approaches to the study of deprived families in developing countries. The better we understand the problem, the better we can make policies to help solve it. A holistic view is required.

Deprived families are usually at the lower end of the socio-economic strata and are easily identified within this context. They share some of the values? beliefs, and knowledge of their society as a whole, but the differences they exhibit are sufficient to place them in a subgroup within the low socioeconomic class.

The analytical diagram (fig. 1) shows the aspects that should be considered in the study of the deprived family.

First, the family should be seen as a social unit with at least some structured organization among its members. The structure reveals a hierarchy of family members, each with a different role. The father is generally more active, the mother more socially expressive, and the children are more or less passive responders to both. Each role is based on behavioural patterns that are guided by societal norms and, more specifically, by those norms that govern the social group as a whole. The norms originate from within the culture and are also stimulated by individual aspirations.

Furthermore, a family is affected by both an internal and an external milieu. The family is located in a physical place (housing) with objects (basic goods) and stimuli (outside influences). The internal milieu provides a series of relationships: couple, parent-child, siblings, relatives, and non-family members. At the same time, to satisfy its needs, the family maintains contacts with the community. The external milieu then begins to play an important role by providing such services as health care, education, sanitation, recreation, and community organizations such as parents' centres (at school), mothers' centres, and sports clubs.



Figure 1 (see FIG. 1. Diagram of analysis) was compiled using the three different methodological designs described below. The results provide a better understanding of more than a thousand families in greater Santiago.


An experimental design uses two groups, experimental and control, both of which are given post-intervention tests. The cases are matched for different parameters (e.g. age, sex, socio-economic stratum, placement, outpatient clinic attendance).

Our experiment using this design started with the fact of frank infant malnutrition. In this case, the experimental group was represented by the existence of infant malnutrition in the family, a condition that is impossible to produce "voluntarily." The control group consisted of families with normal infants.


A quasi-experimental design uses a pre- and post-intervention test without use of a control group. Thus, the group studied serves as its own control. It is necessary to measure the variables before and after intervention in order to quantify the impact of the intervention (e.g. changes in housing, social treatment).


A descriptive design requires a representative sample from which it is possible to generalize to the population studied. A comparative design compares two or more samples with similar parameters but with one different factor, for example, socio-economic stratum.


The techniques used for gathering data under the three designs varied depending on the design, the information to be obtained, and the type of informant interviewed. These techniques may be summarized under the two following categories.


Interviews were structured and semi-structured. The structured interviews used both simple open and closed questions (e.g., in the socio-economic survey: activity, schooling, social security of the head of the household, housing conditions) and complex ones (e.g. past and current aspects of the mother's life, assignment of priorities for the children's future, subjective questions about values).

The semi-structured interviews used a guide for all very personal questions, and their sequence depended on the person being interviewed (e.g. narration of earlier events in life, expression of feelings) [3; 4].


Observations were of a structured, naturalistic kind as well as more formal in nature. The structured observations were used to describe the natural milieu of the family. These observations were either simple (e.g. noting the presence or absence of more than essential objects and colours in the physical environment) or complex (e.g. recording behaviours concerned with the gestural language used by the mother with her infant and with food preparation) [5; 6].

Observations were videotaped to provide an experimental situation that was identical for all participants. Recording the observations was complex. One observer prepared a structured observation report by watching the video. Afterward, each observer described the observations separately, and the results of both were compared. The results with the greatest degree of agreement between observers (87%) were accepted as valid.

Below, some studies are described that show the design and techniques employed. In some cases more than one technique may have been used within the same study but with different methods of data analysis.

Experimental design

The experimental design was the one most used. Samples were matched for between 40 and 200 mother-infant or father-infant dyads. For statistical analysis, the minimum acceptable number was 40 (20 in each group).

In a pilot study on the non-verbal language of mothers with malnourished infants [5], the sample consisted of 40 mother-infant dyads. 20 with malnourished infants (experimental group) and 20 with healthy infants (control group). All the families participating in the study had two to six children and a mother who did not have a steady job and who had no other female relatives living in the same house. The infants were between five and eleven months old, born by normal delivery after a full-term (38-40 weeks) pregnancy with a birth weight exceeding 2,500 grams. The dyads attended the same health centre and were from a similar socio-economic stratum.

The techniques used were (I) a socio-economic survey, (2) a semi-structured interview about the mother's past history (family structure, peer group, family relationships, and jobs held) and current situation (education level, duration of infant's breastfeeding, personal satisfaction in her family life), (3) structured naturalistic observations, which included a list of gestures that have a wide range of positive or negative emotional connotations, and (4) a Wechsler Adult Intelligence Scale used to measure the mother's IQ. The gestures observed in 3 above were classified as facial (ocular or labial) and gestural (hands); the observation was made by two investigators who sat in the same room watching the mother-infant dyed while the child was being fed.

A general non-verbal language index derived from the observations showed a significant difference in expressiveness between the mothers with malnourished infants and those with healthy infants (see FIG. 2. General non-verbal language index (x20, = 6.442, p < .01 , df = 1 )), and a sub-index for non-verbal language use during feeding showed all the mothers of malnourished infants displaying low expressiveness (see FIG. 3. Sub-index of non-verbal language used during feeding (x20 = 7.059, p < .01, df = 1 )). Finally, 70% of the mothers with malnourished infants indicated that they felt little or no satisfaction in their family life, whereas 85% of the mothers in the healthy-infant group considered themselves moderately or completely satisfied (see FIG. 4. Index of mothers' personal satisfaction (x20 = 8.7447, p < .01, df = 1)).

Another study, on "parental models" [3; 7; 8], used a sample made up of 202 mother-infant dyads: 50 with severely malnourished infants (< 70% adequate nutritional status), 51 with moderately malnourished infants (c 80% adequate nutritional status), 51 with slightly malnourished infants (< 90% adequate nutritional status), and 50 with normal infants (> 90% adequate nutritional status). The precentages of nutritional status were based on the weight-for-age ratios given by the US National Center for Health Statistics table [9]. The three groups with malnourished infants constituted the experimental group, and the group with normal infants was the control. The dyads were matched for socio-economic status, maternal age, outpatient clinic attendance, and age of the infant (c 24 months). The techniques used were (1) a simple structured interview to measure socio-economic status, aspects of pregnancy, and aspects of infant development; (2) a complex structured interview to measure socio-emotional relationships of parents and the mother-infant dyed; and (3) a semi-structured, in-depth interview to measure the impact of the parental models on the infant.

TABLE 1. Factors used in the calculation of the work-history index



Age when started to work


Reason for early start


Number of job changes while living with the child's mother


Time in a given labour sector


Time with social security


Duration of jobs


Minimum score


Maximum score


Evaluative scale

Points Evaluation
0- 10 positive
10-20 intermediate
20-30 negative

The data were analysed through self-evaluation scales and parental-image indices that included the way the mother saw her own father and mother, and family rejection of her measured through her own perceptions, It was found that mothers with severely malnourished infants gave less stimulation when interacting verbally with their infants (see FIG. 5. Mothers' verbal interaction with infants of different nutritional status (x20 = 10.3677, p < .01, gL = 3)), were dissatisfied with their own family life, and had a negative parental image.

A study of fathers with malnourished infants [10-13] applied the same experimental design. The sample included 212 father-infant dyads: 107 with severely malnourished infants (the experimental group) and 105 with healthy infants (the control group). Simple and complex structured interviews were carried out over a five-year period.

The data were analysed through indices such as work history, including the age when the father began working, the number of job changes, periods of unemployment, and time spent in the same job (table 1). The experimental group showed a more negative work history (table 2). Multiple correlations of the father's variables and the nutritional status of the infant are shown in table 3 . The work history index is associated with the total number of jobs, the average number of occupations (as an employed worker and/ or self-employed), and schooling.

TABLE 2. Work history of fathers with malnourished and healthy infants Infant's status

Work history

Infant's status







































x20 = 16-5462. p < .0002. df = 2.

TABLE 3. Multiple correlation of infants' nutritional status and fathers' variables (N = 212)


Variables (as listed at left)

1 2 3 4 5 6 7 8 9 10
1. X weight 1.00                  
2. Work history index 0.13 1.00                
3. Total number of jobs 0.07 0.34** 1.00              
4. X self employed 0.05 0.30** 0.35** 1.00            
5. X employed 0.05 0.26* 0.60*** 0.15 1.00          
6. Age 0.06 0.11 -0.24* -0.21 -0.36** 1.00        
7. Socio economic stratum 0.00 0.22 -0.18 0.06 -0.21 0.07 1.00      
8. Schooling 0.16 0.29* -0.02 0.14 0.11 -0.49** 0.34** 1.00    
9. Alcoholism 0.02 0.07 - 0.09 -0.03 0.01 -0.11 -0.04 -0.03 1.00  
10. Total unemplovment 0.12 0.06 0.04 -0.01 0.37** -0.14 -0.08 0.15 0.05 1.00


TABLE 4. Representative indicators of variables used for defining social problems

Variable Indicators
Employment Absolute unemployment
Looking for a job (person of active age, has not worked previously)
Unstable job (temporary or independent job)
No social security
Works for board only
Physical inability
Income Does not cover minimal needs
Temporary problems derived from salary schedule
Low income is not used efficiently
Housing Living in someone else's house
Poor environmental sanitation
Poor conditions of the house (floor, ceiling, roof, windows, . . .)
Legal problems in the land tenancy (should leave house immediately)

Quasi-experimental design

A quasi-experimental design using pre- and post-intervention tests without a control group, is difficult to apply because it requires an intervention to provoke change. Furthermore, it is very difficult to control the variables within social situations.

A study evaluating changes in hygiene in food preparation that occurred when families living in slums were relocated to adequate, or so-called "basic," housing used this design [5; 14; 15]. It measured hygiene practices in food preparation, especially the use of water, studying the same variables in the same families both before and after the housing relocation. The subjects were a random sample of 52 families with a child under seven years of age, taken from a total population of 135 families.

The techniques used were (1) a complex structured interview conducted in the home, covering such questions as whether the people were of urban or rural origin and the length of time they had lived in Santiago altogether and in the slum in particular, and (2) a structured naturalistic observation of meal preparation in the home before and after the move. This observation was identical in both stages and focused on physical hygiene (the presence of flies on pots and pans, the cleanliness of food containers, the cleanliness of the table before the meal was prepared) and on the preparation and handling of food (washing of greens and vegetables, washing and drying of hands, cleaning and storage of utensils after use).

Four individuals were trained to carry out the interviews and to observe the meal preparation. Special sheets were prepared for recording the information gathered. The data collected for each item of the observation were quantified through numerical indices, which were calculated by taking the percentage of an item's components registered during an observation period. For example, the item "material hygiene" (physical hygiene) included 15 categories that could have been recorded, but only 10 of these were used. If the 10 categories recorded were considered positive for that particular item, the index given was 70%. The index for each item was built on the basis of positive findings only.

It was found that the use of water was significantly greater after the change in housing (see FIG. 6. Washing of vegetables under running water before and after housing relocation (x20 = 24.413, p < .001, gL = 1)). This change took place without any special educational intervention.

An evaluation model for family problems was applied to a population of 259 families who had a malnourished infant under treatment at a nutritional rehabilitation centre of the Corporación pare la Nutrición Infantil (CONIN) [16]. The effectiveness of social work was assessed by comparing diagnoses of the families' situations before and after treatment. A standardized code of the problems diagnosed was designed, with 11 variables to describe the current conditions of the families. For, each variable there were a number of indicators; for example, table 4 shows the "employment" variable with such indicators as unemployment, job hunting, temporary job, and absence of social security. To obtain the data, a chart was filled out with the following information for each family: dates, diagnosis of problems(s), activities (for diagnosis and treatment), and time spent with the family. Diagnosis and treatment were recorded on this chart by the social worker using the respective code numbers.

Figure 7 (see FIG. 7. Diagnosis of the situations of families with children in nutritional rehabilitation centres, at the time of the child's arrival (N = 259: average problem load = 57)) shows the diagnosis of the 259 families on the basis of information obtained at the time the child was brought to the rehabilitation centre; unemployment was the first family problem. Figure 8 (see FIG. 8. Comparison between the situations of families at the time of their child's arrival at the rehabilitation centre (vertical bar on the left for each variable) and on second diagnosis (bar on the right), after an average of five months at the centre (N= 110; Sd=4.25; S2= 18.1)) presents a comparison of the original situations of 110 families with their situations after an average of five months of treatment. All problems decreased significantly except for the "no legal identification" and "parent-child relationship" variables.

Descriptive/comparative design

Descriptive and comparative designs were often used when it was necessary to generalize to the population as a whole.

A study of the acceptability of a new programme for supplementary feeding [17-19] analyses the consumption and distribution of foods. The programme, carried out by the Ministry of Health at no cost to the beneficiaries, included the provision of whole powdered milk (26% fat) and fortified food for preschool children in families of low socio-economic status. The survey was conducted by means of a complex structured interview with the mothers in a representative random sample of 366 families in the National Health Service areas of Santiago.

From the results, shown in table 5 , it can be seen that the milk was shared by the entire family, thereby substantially reducing the amount received by the infant.

A study of values held by mothers with a malnourished infant was both descriptive and comparative [1]. A sample made up of 144 mothers of low socio-economic status with a malnourished infant who had been admitted to a CONIN nutritional rehabilitation centre was studied. A complex structured interview with subjective questions was carried out at the time of the infant's admission to the centre.

As shown in table 6 , a majority of the mothers assigned great importance to the following factors: communication (84%), nursing-bottle hygiene (80.6%), training courses (77.5%), health (71.5%), and education (66.7%). They attributed a more limited degree of importance to housing (68%), parental roles (67.4%), the mother's ideal role (66.8%), and work (59%). The majority did not consider religion important.

In a second study, 50 mothers, randomly selected from the original sample of 144 mothers described above, were matched with 100 mothers with healthy infants, 50 of low and 50 of middle socio-economic status, who were given the same interview as the first group [1].

It was found that health, parental roles, education, and time-space orientation indices were valued significantly more by the mothers of the middle than by those of the low socio-economic stratum (table 7). There were no significant differences in the values given to work, housing, and communication.

TABLE 5. Milk consumption the family

Family member (age in years)

Consumes milk










Infant (0-2) 211 29.0 10 1.2 221 14.4
Pre-school (2-6) 163 22.3 94 11.7 257 16.7
Mother 149 20.5 211 26.2 360 23.5
Father 88 12.0 256 31.9 344 22.4
School-age (6-12) 76 10.5 123 15.3 199 13.0
Adolescent (12-18) 42 5.7 110 13.7 152 10.0
Total 729 100.0 804 100.0 1,533 100.0

p < .001.

TABLE 6. Areas considered important by mothers with a malnourished infant



Total Partial None
Health 71.5 23.6 4.9 < .001
Bottle hygiene 80.6 19.4 0 < .001
Education 66.7 31.2 2.1 < .001
Training courses 77.5 6.5 16.0 < .001
Work 39.6 59.0 1.4 < .01
Communication 84.0 16.0 0 < .001
Parental roles 20.1 67.4 12.5 < .001
Ideal father's role 48.0 52.0 0 NS
Ideal mother's role 33.2 66.8 0 < .001
Housing 28.5 68.0 3.5 < .001
Neighbourhood 54.0 46.0 0 NS
Time-space orientation 2.8 50.0 47.2 NS
Religion 25.0 75.0 < .001  

TABLE 7. Valuation indexes of mothers with malnourished and healthy infants in various socio-economic strata (N= 150)





Contingency coefficient




< .01


Parental roles



< .01





< .001


Time-space orientation



< .001





< .30





< .30





< .30


Discussion and conclusion

Quasi-experimental designs are recommended for situations where there is direct action introducing variables that encourage a change in behaviour. This change may be the result of social treatment or government actions (relocation from a slum to adequate housing) or the application of a controlled educational programme. This type of design includes evaluation.

Descriptive designs are useful to determine general features of the general population under study. They also point out important areas for further, in-depth, studies. Finally, comparative designs for studying the effects of differing socio-economic strata enable us to put the deprived family within its real context.

What do deprived families share with the rest of society, and where do the differences lie? The answers to these questions will make it possible to deal with the deprived family according to its own patterns, instead of using norms that pertain to other groups.

The identification and study of deprived families as a subgroup within a society is of considerable importance for the implementation of social and health policies, because such families should be treated appropriately. This may be done by direct interventions, which may be either big-medical or social. They include efforts to prevent increases in the number of deprived families through the implementation of such policies as the provision of supplementary food, health control for pregnant mothers and infants at risk of malnutrition, the establishment of state or local programmes to provide work for unemployed fathers with malnourished children, priority for malnourished children and their siblings in nurseries, and family allowances for abandoned mothers who have malnourished infants.


  1. Alvarez ML. Deprivación y familia. Santiago, Chile: Editorial Universitaria, 1982.
  2. Alvarez ML, Wurgaft F, Salazar ME. Mediciones de nivel socioeconómico bajo urbano en familias con lactante desnutrido. Arch Latinoam Nutr 1982;32(3):650-62.
  3. Alvarez ML. Deprivación y modelos parentales. Santiago, Chile: Editorial Universitaria, 1986.
  4. Alvarez ML, Miskasic D, Ottenberger A, Salazar ME. Características de familias urbanas con lactante desnutrido. Arch Latinoam Nutr 1979;29(2):220-30.
  5. Alvarez ML, Wurgaft F, Wilder H. Non verbal language in mothers with malnourished infants: a pilot study. Soc Sci Med 1982;16:1365-69.
  6. Alvarez ML, Wurgaft F. Changes in hygienic habits in food preparation: facilitates to use water. Canadian J Public Health 1987;78(Jan-Feb):43-46.
  7. Alvarez ML, Wurgaft F. Conductas socioafectivas de madres con lactante desnutrido. Rev Chil Nutr 1981 ;9(2):91-99.
  8. Wurgaft F, Carrasco D, Alvarez ML. Vida pasada y presente: su influencia en madres con lactante desnutrido en Santiago, Chile. Rev Chil Nutr 1984;12(1):29-34.
  9. Hamill PVV, Drizd TA, Johnson CL, Reed RB, Roche AF, Moore WM. Physical growth: National Center for Health Statistics percentiles. Am J Clin Nutr 1979; 32:607-29.
  10. Alvarez ML, Wurgaft F. Historia laboral del padre conlactante desnutrido. Rev Chil Nutr 1981;9(2):91-99.
  11. Alvarez ML, Wurgaft F. Pérdida de status del padre a nivel familiar y desnutrición infantil en Santiago, Chile. Rev Chil Nutr 1982;10(2):155-66.
  12. Alvarez ML, Wurgaft F. Barahona S, Castillo M. Factores asociados al padre y desnutrición infantil en Santiago (Chile). Rev Chil Nutr 1984;12(1):35-40.
  13. Wurgaft F, Alvarez ML. Conducta paterna y desnutrición infantil. Rev Chil Nutr 1985;13(3):170-74.
  14. Alvarez ML, Wurgaft F. Presencia de un equipo médico y practicas de salud en una comunidad urbane de bajo nivel socioeconómico. Rev Chil Pediatría 1985; 56(6):456-68.
  15. Alvarez ML, Wurgaft F. Erradicación de un campamento: un análisis critico. Boletin Insta Vivienda. 1988;3(Sep):6-18.
  16. Alvarez ML. Evaluación en trabajo social: una experiencia metodológica. Trab Social 1979;28:29-35.
  17. Alvarez ML, Guzmán MT, Jaque G, Bell K, Gattás V. Estudio descriptivo sobre Fortesán: alimento pare preescolares. Arch Latinoam Nutr 1977;27(3):521-29.
  18. Alvarez ML, Ségure T, Miranda M, Vial M. Evaluación de la distribución y consumo de leche. Rev Chil Pediatria 1978;49(1-6):247-52.
  19. Alvarez ML, Guzmán MT. Jaque G, Bell K, Gattás V. Estudio de aceptabilidad de madres y niños de una mezcla proteica para preescolares . Rev Chi Pediatría 1979;50(1):89-94.

Contents - Previous - Next