Contents - Previous - Next

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


Changes in SES indicators over time

Water and sanitation

The census forms reported water sources, methods of grey-water disposal, and methods of faeces disposal. The percentages of families with some means of disposal of human waste and with a source of water at home in 1967, 1974, and 1987 are presented in table 3. No method of grey-water drainage was available in any of the villages in either 1967 or 1974; in 1987 only 5% of the families in any village had any such drainage.

TABLE 3. Characteristics of houses by village and year

  Large fresco (SD) Large atole (C) Small fresco (ES) Small atole (SJ) Total
Percentage of houses with facilities
Human-waste disposal  
1967 7 8 1 5 8
1974 15 19 24 12 18
1987 46 65 54 66 62
Water sourcea  
1967 0 5 9 0 4
1974 < 1 21 15 < 1 10
1987 7 82 71 9 44
Non-dirt floor  
1967 5 5 12 2 6
1974 13 13 16 10 13
1987 62 27 38 28 42
Mean house factor score
1967 2.08 1.76 1.64 2.15 1.87
1974 3.18 2.64 2.38 2.97 2.80
1987 4.01 3.52 3.27 3.45 3.60

Sample sizes are given in table 1.
a. A well or connection to a public water system.

In 1967 most families in the large fresco and the small atole village obtained water from a public well or spigot. In the small fresco village water was obtained from open, hand-dug wells, and in the large atole village from a public water hole. By 1974 half of the families in the large atole and the small fresco village and virtually all families in the other two villages obtained water from a public spigot. By 1987 the main source of water continued to be a public spigot in the large fresco and the small atole village, whereas in the small fresco and the large atole village, over half had a hook-up to a public network. Overall, only 44% of the families had water in the home by 1987.

In 1967 few households had any method for faeces disposal. By 1974, 18%, and by 1987, 62% had some form of disposal of human waste, such as a latrine, in or near their homes.

Housing and possessions

The typical house in all the villages except the small fresco village has one or two rooms with adobe walls, a dirt floor, and a tile or metal roof. After a major earthquake in 1976 a number of tile roofs were replaced with sheet metal, which can be considered comparable to tile in quality. In the small fresco village, which is located in a lower, warmer area, most houses have thatched roofs and walls made of reeds and mod. Most homes have a kitchen located in a separate room or area outside the house, and most families own the land on which their homes are constructed.

Dirt floors present a potential health risk, particularly for young children. Over 85% of the homes had dirt floors in 1967 and 1974. By 1987 the figure was down to 58%; in the large fresco village only 38% had dirt floors. The means for the house factor scores increased in all the villages over the 20 years but were consistently lowest in the small fresco village. No houses were equipped with electricity in 1967. In 1974 electricity was reported for 60% of the families in the small fresco village. By 1987 two-thirds to three-fourths of the families in all the villages had electricity in their homes.

Families generally have few possessions. The most common possession is a radio; in 1967, about one-third of families per village had one compared with about one-half in 1974 and 1987. In the large fresco village two-thirds owned radios by 1987. Both in 1974 and 1987, sewing machines were owned by less than 10% in all but the large fresco village, where the figure was 20%. Less than 5% owned large items such as a television, record player, refrigerator, bicycle, or vehicle in either 1967 or 1974. There was little change in the purchase of any of these items by 1987 except for televisions; 20%-30% owned televisions in all the villages except the small fresco village, where the figure was 10%.


Fathers' occupations are shown in table 4. In 1967 and 1974 more than 90% of the male heads of household were primarily engaged in agriculture. By 1987 the figure had dropped to 65%, and the proportion of men working at a skilled trade had increased to 29%. However, the proportion of fathers who were farming land that they owned dropped from 41% in 1974 to only 32% in 1987. The percentages varied by village.

TABLE 4. Fathers' primary occupation by village and year (percentages)

  Large fresco (SD) Large atole (C) Small fresco (ES) Small atole (SJ) Total
Agricultural wage labour  
1967 6 3 21 15 11
1974 5 7 29 4 10
1987 5 5 38 2 10
Tenant farmer  
1967 58 28 42 26 39
1974 29 19 22 28 39
1987 54 33 30 34 23
Small landowner  
1967 26 68 28 56 44
1974 27 53 28 55 41
1987 17 52 15 40 32
Skilled trade  
1967 8 <1 7 0 4
1974 14 7 11 4 9
1987 47 21 21 28 29
1967 1 >1 7 0 1
1974 1 0 2 3 1
1987 2 2 4 3 3

Sample sizes are given in table 1.
In 1987 unemployed were coded separately. and 3% were coded as unemployed.

In 1967 fewer than 5% of the women reported an occupation. Table 5 presents the percentages of women who reported an occupation in 1974 and 1987, categorized into the same groups as the men. In 1974, 32% of the women reported an occupation, compared with 23% in 1987. The rates vary considerably by village, with high rates in the small fresco village, where women wove baskets for sale. Agricultural work was rarely reported (< 4%); the most common occupations were skilled trade, including artisan work (21% in 1974 and 15% in 1987) and commercial work (6% in 1974 and 7% in 1987).

TABLE 5. Mothers' primary occupation by village and year (percentages)

  Large fresco (SD) Large atole (C) Small fresco (ES) Small atole (SJ) Total
Agricultural worka  
1974 1 9 3 2 4
1987 >1 2 3 1 2
Skilled trade  
1974 13 15 76 11 28
1987 15 14 48 14 22
None reported  
1974 86 76 20 86 68
1987 84 82 49 85 77

Sample sizes are given in table 1.
One worker was listed as commercial/selling.
a Includes wage labour and small farming on rented or own land.

Literacy and schooling

Literacy rates and years of school passed are shown in table 6. Mothers at least partially literate ranged from 25% to 40% in 1967 and from 57% to 63% in 1987. For fathers the range changed from 38%-60% in 1967 to 61%-76% in 1987. As with literacy, the years of schooling completed increased, but by 1987 still averaged only two years. Both schooling and literacy were higher in the fresco villages for men in both 1967 and 1987, and for women in 1967.

TABLE 6. Parents' literacy and schooling by village and year

  Large fresco (SD) Large atole (C) Small fresco (ES) Small atole (SJ) Total
Families with parent at least somewhat literate (%)
1967 40 25 26 29 30
1974 51 37 35 39 41
1987 63 57 58 63 60
1967 47 38 60 38 46
1974 56 49 67 44 54
1987 65 65 76 61 66
Mean years of school passed
1974 1.2 0.9 1.0 0.8 1.0
1987 2.3 1.8 2.1 2.1 2.1
1974 1.2 1.1 1.9 0.8 1.2
1987 2.5 2.1 3.2 2.1 2.4

Sample sizes are given in table 1.
Data on schooling were not collected in 1967.

Marital status

The percentage of women in each marital status category remained constant over time (table 7). At each census, approximately 25% of the women were unmarried (mostly widows or divorced or separated; few unmarried mothers). About 40% were in consensual unions, and 36% were legally married. These rates differed by village, with women in the small fresco village least likely to be in a consensual or legal union.

TABLE 7. Marital status by village and year (percentages)

  Large fresco (SD) Large atole (C) Small fresco (ES) Small atole (SJ) Total
Single, widowed, separated/divorced          
1967 26 23 35 20 26
1974 27 24 31 14 25
1987 28 24 29 20 25
Consensual union          
1967 26 55 53 17 39
1974 28 49 55 22 39
1987 31 52 39 27 38
Legally married          
1967 48 22 12 63 35
1974 45 27 15 63 36
1987 40 25 31 53 36

Sample sizes are given in table 1.

Comparison of villages on SES indicators

Initial comparison

To determine whether there was initial equivalence between the atole and the fresco villages in SES, the continuous measures (house quality, father's occupation, number of pregnancies, fertility rate, number of children who had died) were analysed using a treatment (2) analysis of variance with unbalanced cells (general linear model, SAS, 1988). For parents' literacy, a treatment (2) by literacy (2) chi-square analysis was calculated (tables 8 and 9).

TABLE 8. Analysis of variance of SES and demographic indicators by village type, 1967

  N F Fresco
House factor score 577 0.03 1.89 1.90
Father's occupation 509 5.61* 2.42 2.62
No. of pregnancies 577 1.98 5.19 5.61
Fertility rate 519 25.74** 0.22 0.27
Child deaths 526 7.49** 1.43 1.87

* p < .05. ** p < .01.

TABLE 9. Comparison of parents' literacy by village type, 1967

  % literate X2 N
Fresco Atole
Mothers 34 26 3.72* 532
Fathers 53 38 12.49** 512

*p<.05. **p<.01.

In 1967 the fathers in the atole villages had significantly higher-ranked occupations than those in the fresco villages. House quality did not differ with village type. The number of pregnancies per woman also did not differ by village type, but many women had not completed their families (mean age of the mothers sampled 39.1 yrs, SD = 14.6). However, the fertility rate was significantly higher in the atole villages, as was the number of children who had died. Finally, both mothers and fathers in the fresco villages were more likely to be somewhat literate than those in the atole villages in 1967.

Comparison in 1987

For literacy, which in 1987 was coded on a three-point scale, a 2 (village type) x 3 (literacy code) chi-square analysis was performed. The fresco village fathers (but not mothers) were more likely to be literate (X2 (2) = 6.91, p < .05 for fathers; x2 (2) = 0.35 for mothers).

Table 10 presents results of an analysis of variance by treatment (2) for the continuous variables house quality, years of schooling, occupation, and demographic variables in 1987. Parents continued to have significantly more years of schooling in the fresco villages in 1987. Fertility rates were also lower as in 1967. No differences in the number of pregnancies were found in 1987, as was the case earlier. However, some differences favoured the fresco villages later that were not seen in 1967; fresco village houses were rated as of higher quality, and there was no difference in the occupations of the fathers, which had been significantly higher in the atole villages 20 years previously. On the other hand, whereas in 1967 more children had died in the atole villages, no differences were seen in 1987.

TABLE 10. Analysis of variance of SES and demographic indicators by village type, 1987

  N F Fresco
Mother's schooling 1,092 5.07* 2.22 1.90
Father's schooling 967 17.91** 2.77 2.07
House factor score 1, 157 12.22** 3.71 3.49
Father's occupation 966 0.81 6.11 6.40
No. of pregnancies 1,163 1.92 5.22 5.56
Fertility rate 1,091 12.09** 0.24 0.27
Child deaths 1,163 2.90 0.81 0.97

*p < .05. **p < .01.


Despite the apparent poverty of the area, these villages experienced progress over this 20-year period. Literacy and schooling increased, especially for women. Houses improved, people owned more possessions (notably television sets), and they were more likely to be working in skilled trades. Women were more likely to report an occupation, and probably even more worked than reported it, since occupation is often underreported [6]. On the other hand, adequate water still was not available in the homes, and faeces disposal was lacking for 40% of the families. Almost half of the families continued to have dirt floors in their houses. Finally, although literacy rates improved markedly, the average school grade passed by parents was only second grade.

The absence of change in demographic indicators such as the number of children and the fertility rate may threaten the well-being of the residents. The significant increase in the population, despite continued reliance on agriculture, suggests that the scarce resources of this arid region may not be able to support the rate of growth that current patterns indicate. These topics are addressed elsewhere in this issue [3].

In terms of the experimental design, it appears that neither the assumption of initial equivalence on SES indicators nor of equivalence prior to the follow-up study is justified. In 1967 the fathers in the atole villages were on the average higher in occupational status than those in the fresco villages. An examination of the frequencies of the specific codes reveals that few fathers were commercial workers or skilled tradesmen; rather, the major difference between the villages was that the atole fathers were far more likely to own their own land than the fresco fathers. Conversely, the parents in the fresco villages in 1967 were better off educationally than those in the atole villages. House quality did not differ overall among the villages, but the large fresco and the small atole village had higher house quality than the other two. Thus, adjustment for initial differences seems to be necessary for an interpretation of results.

The implications of the socio-economic differences observed in the 1987 data for the experimental design depend in part on the direction of the differences. Those favouring the atole villages would have made interpretation of any treatment differences on adolescent test performance more difficult. However, they tended to favour the fresco villages. Both mothers and fathers in the fresco villages had more years of school, and fathers, but not mothers, were more likely to be literate. Houses were, on average, of better quality in the fresco villages. No differences in fathers' occupations were seen. There were no differences in the number of pregnancies, but overall the atole mothers had more children per year. There were differences in the children's exposure to education [3]. Children reared in these conditions would be likely to have higher performance on test batteries, particularly those reflecting school-related knowledge and experience. Thus an appropriate test of the hypothesis that supplementation has an effect on cognitive performance in adolescence and young adulthood should control for differences in SES that might tend to reduce the effects of the supplementation on test performance.

It is difficult to explain the different trajectories of the atole and fresco villages. One might have expected that the atole villages would have advanced more rapidly. Not only did they benefit from the protein-energy supplement, but also the large atole village received several years of an intensive development effort by the Institute for Cultural Affairs, which brought electricity, a drip irrigation system, a preschool, and an improved water supply. Also, in 1967 a much larger percentage of families in the atole villages owned their own land than in the fresco villages.

One clear difference was the greater level of education both of parents and of children in the fresco villages. One can speculate about other differences, such as the presence of a local artisan industry and high rates of employment among the women in the small fresco village, or the proximity of the large fresco village to Guatemala City. Further research examining not only mean differences but also variability in these indicators would help to answer these questions.

A direct implication for analyses of the effects of the food supplementation programme on growth and development outcomes is the need to control for education and socio-economic status. This is precisely the approach followed in a study contrasting cognitive development in the atole and fresco villages [4]; the authors show, for example, that failure to control for education yields biased results.


Data collection and analyses were supported by NIH grant HD22440. The study was a collaborative effort involving investigators at several institutions: R. Martorell (principal investigator, originally at Stanford University, now at Cornell University), J. Rivera (INCAP, Guatemala), E. Pollitt (University of California at Davis), and J. Haas (Cornell University).


  1. Bradley RH, Caldwell BM, Rock SL. Home environment and school performance: a ten-year follow-up and examination of three models of environmental action. Child Dev 1988;59:852-67.
  2. Martorell R. Habicht J-P, Klein RE. Anthropometric indicators of changes in nutritional status in malnourished populations. In: Underwood BA, ed. Methodologies for human population studies in nutrition related to health. Washington, DC: US Government Printing Office, 1982:99-110.
  3. Bergeron G. Social and economic development in four Ladino communities of eastern Guatemala: a comparative description. Food Nutr Bull 1992;14(3):22136.
  4. Pollitt E, Gorman K, Engle PL, Martorell R. Rivera J. Early supplementary feeding and cognition: effects over two decades. Chicago, 111, USA: Society for Research in Child Development (in press).
  5. Johnston FE, Low SM, De Baessa Y. MacVean RB. Interaction of nutritional and socioeconomic status as determinants of cognitive development in disadvantaged urban Guatemalan children. Am J Phys Anthropol 1987;51:501-06.
  6. Engle PL. Maternal employment and child mortality in rural Guatemala. In: Wagner D, ed. Child development and international development: research-policy interfaces. New Directions in Child Development, vol. 20. San Francisco: Jossey Bass, 1983:57-75.

Contents - Previous - Next