Contents - Previous - Next

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

Women's work and children's nutrition in south-western Kenya

Deborah S. Rubin

This paper reports a time-allocation study conducted in south Nyanza, Kenya. Data on household composition, identifying each household member, age, and relationship to each other were collected from 75 households. The data describe the internal household division of labour, especially how men and women differently allocate productive and reproductive work time, and how patterns of labour allocation differ between households that have entered into contracted sugar cane production and those that have not. While the sugar cane scheme has resulted in higher aggregate household incomes for its participants, nutritional status has not improved measurably. Other environmental factors such as water quality, sanitation, and access to medical care may play stronger roles in child nutrition status than women's time-allocation and participation in the sugar cane scheme. The results reveal gender-specific patterns of labour that may have consequences over the long run for children's nutrition status.


The labour allocation among households involved in or surrounded by a sugar cane outgrowers scheme in southwestern Kenya has implications for preschooler feeding practices. Historically, sugar cane cultivation has had a bad reputation for bringing hunger and hardship to its workers and their families. The labour patterns of cane production have changed over time, from the use of slaves to the employment of rural proletarians, and, increasingly in Africa today, to the use of an outgrower model in which an estate is supplied by contracted farmers [1, 2]. We can attribute this shift, as has been noted with respect to the Caribbean, "to parallel movement in the stages of the development of overseas agricultural capitalism" [3] Reliance on outgrower systems reconstitutes the working or productive unit from one of individuals to one of farming households, with potentially radical consequences for gender relations within the household. Writing about such a contracting set-up in Gambia, Watts [4] noted:

Household labor continues to be the dominant social form in which labor power is mobilized but under conditions directly determined and shaped by production management.... In conditions in which peasant households are contracted externally as the basic unit of labor supply, the consequences for the labor process (the social organization of work) and property (relations of production) are always domestic in character, reflecting the dominance of kinship and gender in the access to and control over resources.

These statements remind us of the need to institute changes in the domestic sphere in which food consumption resides, to effect changes in global economic systems, and to reflect on the nuances of their expression in gender relations.

Results of a time-allocation study in south Nyanza, Kenya, reveal gender-specific patterns of labour that may have consequences over the long run for children's nutrition status. The time-allocation data described the internal household division of labour, especially how men and women differently allocate productive and reproductive work, and how patterns of labour allocation differ between households that have entered into contracted sugar cane production and those that have not. These data provide both a snapshot of the ethnographic context of the internal organization of household production, as well as a montage suggestive of more dynamic processes linking household production to larger socio-economic systems with the increased pressure of greater agricultural commercialization.

Participants in the outgrowers scheme, and possibly designers of the scheme. imply that farmers in the project area are similar except that some grow sugar cane and some do not. Choosing to grow cane, however, is not simply a substitution or addition of a single crop; it reflects different productive arrangements. These arrangements, in turn, are influenced by culturally defined, gender-specific patterns of work. Thus, the choice to grow cane influences the shape of other elements in the production process, eventually influencing choices about crop mixes, food preferences, and feeding patterns.

Analysis of the time-allocation data shows that participation in household labour differs according to members' involvement in sugar cane production, gender, and social position in the household. Briefly, among all the agricultural households, men appear to contribute slightly more or equal time to agricultural tasks than do women, with their involvement concentrated in land preparation. Men also have more leisure time than women, who of course have a heavier load of domestic work. Women with small children spend less time in farming than do other women in these households.

Both men and women in families who do not grow cane put more labour into food production because they hire out to perform agricultural wage labour on other farms. Much of this work entails weeding contracted cane fields. This suggests that movement into sugar cane growing is not equally accessible to all households, and that participation in the sugar cane scheme within the project area is at least partially supported by an unequal system of labour recruitment.

In sugar-cane-growing households, women spend less time in agricultural work than their counterparts in other households, and more time in domestic work. Similarly, men in these households spend less time in agricultural work than their non-cane-growing neighbours. Differences in time use were paralleled by other differences between the two groups. Sugar cane farmers have larger farms, own more large livestock, and earn higher incomes [5, 6], implying that the introduction of cane cultivation has affected the dynamics of both internal household production and community production structures.

Materials and methods

The study area

Research was carried out in south Nyanza, on the eastern side of Lake Victoria in south-western Kenya. Sugar cane was introduced as a commercial crop through an outgrowers scheme in 1978 at the South Nyanza (SONY) Sugar Company, although local varieties of cane had long been grown on a small scale for jaggery production. Both in Kenya and elsewhere, the introduction of inedible cash crops has sometimes been associated with an increase in pre-schooler malnutrition [7, 8]. The precise mechanisms by which changes in farming patterns influence nutrition status remain unclear, however.

Research conducted by the International Food Policy Research Institute (IFPRI) in this area in 1983-1984 and again in 1985-1986 found only small differences in the nutrition status of pre-schoolers in agricultural households as measured by caloric intake and anthropometric indicators [5, 6]. A number of differences in farming patterns between households involved in sugar cane cultivation and those growing other food and cash crops were noted, but there was no obvious connection between farming patterns and pre-schooler nutrition status over the short term.

Over a longer time, however, these shifts in farming patterns represent significant differences in the processes by which households provision themselves, and could result in consumption differences. Details of an ethnographic study [6] showed, for example, that non-sugar farmers exhibit a greater diversity in cropping patterns, such as increased use of intercropping and drought-resistant crops, as well as a different pattern of gender control over food plots than the sugar farmers. Men in households not growing sugar cane are more involved in food-crop farming than those in sugar.

In many African areas, women's primary role in food production is often the first sphere to be affected by new crop mixes and crop-production technologies. The conceptual framework of the IFPRI study [5, 7] proposed that women's time spent in both productive and reproductive activities is linked to their control over income as well as to the nutrition and morbidity patterns of their children. By shifting the work load within the household, increasing commercialization of agriculture consequently influences the time women allocate to child care and other activities that have either a direct or an indirect bearing on health and nutrition status. Although the first phase of the investigation into the shifting labour patterns associated with sugar cane cultivation [8] did not find preschooler mothers' labour patterns to be significantly associated with consumption differences, more detailed scrutiny of the time-use patterns among all household members, including men and other women, raises interesting possible parallels between differences in cropping patterns and changing feeding practices among pre-schoolers.


Data gathering

Data on household composition, identifying each household member, his or her age, and relationships to each other, were collected from 75 households. These data proved important in desegregating the time-allocation material. Sugar-cane-farming house holds showed a higher rate of polygynous marriages, both historically and during the study period, than did the households not farming sugar. This suggests that cane farmers were wealthier than their neighbours even prior to their participation in the scheme.

After participants were observed performing their household activities, a time-allocation questionnaire was designed to capture micro-level differences in time-use. Over the course of 12 months, more than 800 interviews were conducted about respondents' activities during the previous 24 hours.

Problems exist in this form of data collection. An extensive review reported that time-allocation studies are generally less accurate than those based on direct observation [10]. There is also an unavoidable bias towards household members who are found in and around the compound. Accurate weighting of simultaneous activities is also difficult. The results should thus be interpreted in light of these limitations.


Interesting differences appeared between men's and women's activities, in the use of time by different groups of women, and in the use of time between households engaged in cane farming and other agricultural work. The study first revealed clear segregation of tasks by gender. Women's work centred around the daily performance of household duties as well as food crop farming. Men rarely assisted in domestic and food-related tasks.

Both the observations and the time-allocation reporting concurred that men and women took part in agricultural work and animal care, but their tasks and crop assignments differed. Men cleared, ploughed, and prepared the fields to a greater extent than women. Both men and women planted, weeded, and harvested staple grain crops, but women did a greater share of this work. Women cultivated vegetable fields; men cultivated tree crops. Marketing and processing were the provinces of women.

This categorization by gender, however, glosses over the differences in patterns of labour allocation that are influenced by economic status or position within the household. In cane-farming households, less time was spent per person per day on agricultural work than in non-cane-farming households, and women spent on average less time per day in agricultural work than men.

Women also differed regarding their involvement in hired agricultural labour. No women from cane households reported having worked on another person's fields for pay. In contrast, those from non-sugar households hired themselves out 8% of the total days spent in agricultural activities. Among men, the difference was also striking: only 2.2% in sugar house holds worked as hired labour, compared to 15% in non-sugar households.

Women's time-use also differed in other ways. Women in cane households spent more time in domestic tasks, and in craft work (primarily basket making), which was carried out in the home while supervising children. They also had more leisure time. Women in households not growing sugar spent slightly more time not only in agricultural work on food crops but also in marketing and transport as well as in hired agricultural labour. These latter activities took place away from the home and were more difficult to combine with child care, but they contributed to female-controlled income. Since non-sugar households were also less likely to be polygynous, these data suggest that children in these households might be spending less time either with their mothers or in an adult's care. Children cared for by an adult were found to have a slightly better short-term nutritional status [8].

Other differences appeared when women were divided according to whether or not they had preschoolers. In both types of farming households, women with pre-schoolers spent more time in domestic work and less time in agricultural work than did their co-wives with older or no children. This finding confirms that the household division of labour is based not only on gender but on social position in the household. It is not only important that one is a woman, but which woman-first wife or fourth wife, a daughter-in-law or a daughter, a divorcee or a widow, an old woman or a young one.

The larger IFPRI study concluded that the incomes of sugar farmers in the outgrowers scheme grew as a result of the project and were generally higher than those of their non-participating neighbours. More money, however, did not translate into significant health or nutrition benefits for pre-schoolers [9]. Neither, however, did cane growing seem to have a negative effect on pre-schoolers, as was anticipated.

The differences in women's time-use by activity found in this study did not correlate with the lack of difference in children's nutrition status found in the IFPRI survey. If women's time-use is a critical factor in children's health and nutrition status, greater differences in the anthropometric data and the levels of children's caloric intake would be expected. Several interpretations are possible. Assuming the accuracy of the measurements of pre-schoolers' nutrition status, it may be that women's time-use is not a major factor influencing children's nutrition status. Other environmental factors such as water quality, sanitation, and access to medical care may play stronger roles. The project area suffers from one of the highest levels of child mortality and malnutrition in Kenya. Since the population is ethnically homogeneous, culturally mediated beliefs about child feeding and health standards are also important in defining consumption standards. These factors are equally problematic for all rural households, thus potentially negating the positive effects on household income resulting from participation in the sugar scheme.

Data from observations of food-preparation and feeding patterns in the household are nonetheless suggestive when examined beside the time-allocation material. Two points deserve mention.

Porridge, a staple food in south Nyanza, is made from one or a combination of flours, and is the primary food given to infants and toddlers, beginning at the weaning stage. Some households prepare a sour porridge by allowing the soaked grain to sit overnight and begin to ferment before it is cooked. Sour porridges have several advantages for pre-schoolers over plain porridges. According to Kenyan research [11], they become less acidic and less susceptible to the bacterial growth that causes diarrhoea in small children. The sour porridges are able to sit longer than plain porridges after preparation without special preservation measures. In addition, the souring reduces the final viscosity of the porridge. This means that to get a sour porridge of a desired thickness, more cereal flour has to be used than if the porridge was not soured. For any given viscosity or thickness, sour porridge therefore contains more nutrients per unit volume than plain porridge. Thus for every cup of porridge eaten, sour porridge will provide more nutrients than plain porridge if the porridges are prepared by the mother so as to be of the same thickness, a thickness that she judges best suitable for the child. [11]

In a collection of household cooking recipes, research assistants discovered that some study households were using sour porridge. In six (67%) of nine households, the porridge was prepared by mixing flour and water the evening before and letting it ferment before cooking it the next morning. In all instances a mixture of flours was used: cassava and finger millet flours in five, and maize and finger millet flours in one. The three recipes that did not include souring used only a single flour, either maize or cassava alone. More careful re-analysis of this point is needed, but it seems that the non-sugar households are among those that more frequently made sour porridge with mixed flours, while sugar households tended toward single-flour porridges.

A second item concerns the frequency of child feeding. Children in households growing sugar cane were fed somewhat more often during the day than those in households not growing sugar. This seemed to be linked to both the greater number of women present in the sugar households and the greater amount of time they spent working around the house.

These two examples provide contradictory but suggestive possibilities for linking time-allocation feeding practices. Women in non-sugar households, who have less income, spend more time away from the home, and cultivate a greater variety of grains, may be more likely to prepare the multi-grain porridge and let it sit before cooking. As a result, their pre-schoolers receive a more nutritious, healthier porridge than those in sugar households. Conversely, pre-schoolers in sugar households may be fed more often and with a greater variety of foods. These examples show that the consequences of both income and time use may have conflicting effects on health and nutrition. The conflicts average out in the survey data to draw similar portraits of the affected populations, but tracing out the pathways provides keys to magnifying the positive choices and minimizing the negative.

Further analysis of data on feeding observations is underway to identify more clearly the nutrition consequences for pre-schoolers of women's changes in agriculture production patterns. Thus, while participation in the cane outgrowers scheme may not now be measurably influencing children's nutrition status, it clearly has other effects on people in the project area and on household practices. The time-allocation results give us some insights into these changes. Documenting differences of time-use related to food production and agricultural wage labour shows that the production patterns of sugar farmers and non-sugar farmers differ not only in their patterns of sugar cane cropping but also in food growing. Cane cultivation ties the farmer into a complex network of labour relations both with the sugar factory and with hired labour from within and outside the community. The income benefits from sugar cane result from these changing labour relations between farming groups, and are, in part, supported by the labour of those not growing sugar.

Just as labour relations between farming households are affected, so are gender relations regarding food production within the household. The move into cane farming has resulted in a decreased input of male time into food production. Women in cane-farming households also spend less time in food production than women in non-sugar-farming households, but their responsibility for food plots is greater. Greater use of hired labour in cane households parallels these shifts.

While the sugar cane scheme has resulted in higher aggregate household incomes for its participants, nutrition status has not improved measurably [5, 9]. Men and women, adults and children are being affected in different ways. The changes in time-use across and within agricultural households indicate that important shifts in production and consumption practices are occurring in areas other than nutrition that may have less favourable effects on the welfare of some of the project-area population.


  1. Dinham B, Hines C. Agribusiness in Africa. Trenton, NJ, USA: Africa World Press, 1984.
  2. Williams S. The Mumias Sugar Company: a nuclear estate in Kenya. In: Williams S, Karen R, eds. Agribusiness and the small scale farmer. Boulder, Colo, USA: Westview Press, 1985:35-56.
  3. Mintz S. The rural proletariat and the problem of rural proletarian consciousness. J Peasant Stud 1974;1(3):291325.
  4. Watts M. Idioms of land and labor: producing politics and rice in Senegambia. Paper presented at a conference on Popular Culture and Popular Protest in Africa. Stanford, Calif, USA: Stanford University, 1989.
  5. Kennedy ET, Cogill, B. Income and nutritional effects on the commercialization of agriculture in southwestern Kenya. Research report 63. Washington, DC: International Food Policy Research Institute, 1987.
  6. Rubin D. Changing production practices in a sugarcane-growing community. Report to the USAID Office of Policy Coordination. Washington, DC: International Food Policy Research Institute, 1988.
  7. von Braun J, Kennedy E. Commercialization of subsistence agriculture: income and nutritional effects in developing countries. Working paper on commercialization of agriculture and nutrition no. 1. Washington, DC: International Food Policy Research Institute, 1986.
  8. Cogill B. The effects on income, health, and nutritional status of increasing agricultural commercialization in south-west Kenya. Doctoral dissertation. Ann Arbor, Mich, USA: University Microfilms, 1987.
  9. Kennedy ET. Effects of sugarcane production in southwestern Kenya on income and nutrition. Final report to USAID Office of Policy and Program Coordination. Washington, DC: International Food Policy Research Institute, 1988.
  10. Gross D. Time allocation: a tool for the study of cultural behavior. Ann Rev Anthropol 1984;13:519-58.
  11. Alnwick D. Sour sorghum porridges-are they good weaning foods? Nairobi, Kenya: UNICEF Regional Office, 1987.

Contents - Previous - Next