Contents - Previous - Next

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

Hunger, technology, and society

On-farm research and applied nutrition: some suggestions for collaboration between national institutes of nutrition and agricultural research
Economic aspects of food protein supplies in the world

On-farm research and applied nutrition: some suggestions for collaboration between national institutes of nutrition and agricultural research

Robert Tripp
Centro Internacional de Mejoramiento de Maz y Trigo (CIMMYT), Londres, Mexico


There is a growing concern that nutritional considerations should play a larger role in determining the direction of agricultural research. This paper attempts to address one part of that issue by examining ways in which the increasing interest in farming systems research might contribute to resolving nutritional problems. The following discussion looks only at possible ways in which agricultural change might improve the nutrition of rural people, and thus ignores a whole series of larger policy issues linking agricultural production and nutrition.

The paper is based on two assumptions: It is hopelessly optimistic to believe that increases in aggregate production and income are sufficient to assure nutritional improvement; and it is hopelessly naive to believe that agricultural technologies can be planned around farmers' diets. A compromise must be sought. This paper develops a series of suggestions for activities involving national institutes of nutrition and agricultural research in order to develop such a compromise.

The connection between agricultural change and nutrition has long been recognized (1), but there has been difficulty in establishing a nutritional direction for applied agricultural research. Most work has been done in connection with rural development projects, but even here the results are unclear. In some cases, nutritional analysis proceeds independently of agricultural research within a given project (2), and in others it proposes comprehensive approaches that provide no direction for agricultural research carried out with limited resources (3) Only recently has a concise set of guidelines been developed for including nutritional concerns in agricultural projects (4).

Although the suggestions presented in this paper are relevant to the project approach, the focus is on the more modest, everyday activities of national researchers and extension agents who attempt to bring new ideas and practices to their clientele, the majority of whom are resource-poor farmers. What are the possibilities and limitations for including nutritional concerns in this type of work? At what stages should nutritionists participate in the process? What analytical tools are appropriate to this participation?

This paper discusses a set of research procedures used by CIMMYT, known as on-farm research (OFR), that derive from a concern about the effectiveness of traditional agricultural research and development. Many efforts at agricultural development have failed because researchers have not taken full account of the circumstances under which farmers operate, nor have they developed and tested their recommendations under local conditions. It would not be unfair to say that much applied nutrition suffers from these same deficiencies of a "top-down" approach. The success of OFR in developing recommendations for farmers suggests that this type of approach may also prove useful for applied nutrition. The emphasis here will be on conceptual issues in OFR of common interest to nutrition and agriculture, procedural implications, and suggestions for possible courses of action by national agricultural research and nutrition institutions.


On-farm research is part of a broad array of agricultural research strategies that have been developed over the past decade, often referred to collectively as "farming systems research" (FSR). Although there is interest in the relationship of nutrition to FSR (5), the definition of FSR itself remains unclear. An examination of the literature reveals a welter of different methods and viewpoints (6). It will be best to concentrate here on one specific research approach, which we call on-farm research. The place of OFR within the farming systems movement is outlined elsewhere (7) and will not be described here. A clearer idea of OFR, and its possible relation to nutritional concerns, should emerge from the discussion of conceptual issues that follows.

Two sets of issues of particular relevance to nutrition are considered. The first concerns area selection and the location-specificity of OFR. These determine the ability to target agricultural research for nutritional impact The second set of issues is related to the selection of research priorities. OFR generally focuses on a few crops and problems at a time within a given system, in an attempt to improve productivity and income. It will be argued that this is an appropriate nutritional strategy as well, even though it may seem to conflict with concerns regarding dietary diversity and commercial agriculture.

Area Selection and Location-Specific Research

The issue of targeting is an important one for both agricultural development and applied nutrition. Many agricultural development programmes have not had a positive nutritional impact simply because the circumstances of small farmers have not been well understood, nor has the farming population been disaggregated so that efforts can be aimed at those most in need (8).

The choice of targets for development programmes is a policy issue that is often managed by decision-makers at levels above the institutions that are the subject of this paper. But if these institutions have at their disposal precise methods for identifying target groups, then they can be more articulate in their interactions with policy-makers. In the case of OFR, the concept of the recommendation domain (9) is used for identifying roughly homogeneous groups of farmers whose circumstances are similar enough so that they are all eligible for the same recommendation. Very often policy mandates are stated in terms of target groups whose definition (e.g., "small farmers"} masks considerable variation in circumstances and potential. The use of recommendation domains makes agricultural research efforts much more precise in their response to policy objectives. It offers the same opportunity for sharpening the focus of applied nutrition.

Both farming systems and nutritional systems are distinguished by their location specificity. The concept of recommendation domain is one expression of the location-specific nature of OFR. Problems are identified, experiments are carried out, and recommendations are made, all with reference to a well-defined set of farmer circumstances. The possibility of including nutritional problems in that set of circumstances allows a clear point of entry for applied nutrition.

There is no need to be labour the inefficiency of a "top-down" approach to agricultural research. It is now generally recognized that there are few "appropriate technologies" specific to the poor or malnourished. Solutions must instead be worked out within the local context. To cite one example, chemical weed control appears to be a labour- and cash-saving technology for small farmers in southern Honduras (10), while in central India herbicide use may be both uneconomical for farmers (11) and a threat to the livelihood of landless female labourers (12). The interplay of agronomic, socio-economic, and nutritional considerations can only be examined and resolved at the local level.

Farming System Characteristics and Nutritional Targeting

The preceding section outlined the importance of precise targeting and the need for location-specific strategies in both agricultural and nutritional programmes. What is the degree of congruence between agricultural and nutritional problems that would allow joint efforts at targeting? The answer to this question lies in the epidemiology of malnutrition among rural populations; the particular patterns of malnutrition must be considered.

In the initial targeting of areas for joint efforts between agricultural research and applied nutrition, the obvious relationship between poverty and malnutrition (13) will play an important role. It is likely that a nutrition focus will imply consideration of the relatively poorer rural areas of the country, where farm incomes are lowest. But the decision to direct attention to such areas is not straightforward. It requires both judgements as to the technical feasibility of bringing about improvements in a given area and policy considerations on the trade-offs between possible gains for disadvantaged groups and productivity returns to a focus on more favoured areas. Once these considerations have been taken into account, however, OFR can certainly accommodate this type of target.

Parameters such as income level are often difficult to measure and sometimes not the best indicators of nutritional status, however. At times it may be a particular resource base or cropping system (2, 14) that is related to a high incidence of malnutrition. Again, OFR would have no problem in adopting this as a target.

Once an area has been identified, the challenge is to see how precisely the agricultural research can be directed to the malnourished. In the context of OFR, this is a question of the congruence between recommendation domains and sectors at high risk of malnutrition.

Often there is more malnutrition in a given area among farm families with smaller holdings (15, 16). To the extent that size of holding is correlated with other circumstances that determine farmer practices (17), farm size may be included in the definition of the recommendation domain. In areas where there is a higher incidence of malnutrition among the smallest farmers, research in a recommendation domain defined by small farm size would be particularly relevant. The choice to work in such a domain would, of course, be conditioned by the technical and policy considerations mentioned above.

In other instances, opportunities for agricultural research are not determined by size of holding, and particular technologies are appropriate over a relatively wide range of farm sizes (18). In these cases nutritional relevance for agricultural research is best achieved by simply assuring that the smallest farmers are well represented in both the research and extension activities of OFR.

But malnutrition in rural areas is not always associated with size of holding (19, 20) or even income (21, 22). In such cases, what prospects are there for a precise nutritional targeting of agricultural research? It will depend on the degree to which the causal factors related to malnutrition are agricultural in character. For example, in parts of southern Africa, nutritional status may be related to migratory wage labour and the resulting patterns of household composition and labour allocation (3). These same factors are correlated with household production strategies (23) and may form the basis for determining recommendation domains. In this case there may be a possibility for parallel efforts in nutrition and agriculture.

But in many other instances the epidemiology of rural malnutrition is either not clear or identifies causal factors such as domestic organization, family size, or access to sanitary facilities that bear no direct relationship to the targeting of agricultural technology. In these instances it may be that the best strategy is simply to make sure that OFR is aimed at the majority of farmers in an area, and leave more specific nutritional activities to other agencies.

The Design of an Experimental Programme: Research Opportunities

One way of distinguishing OFR from many other farming systems research activities is its focus on a few experimental variables at a time. OFR has been described as "farming systems research in-the-small," in which the ratio of experimental variables to parameters is quite low. OFR thus takes account of the elements and interactions in a farming system without feeling obligated to do research on all of these variables simultaneously. OFR proceeds by selecting a few priority enterprises and research themes within a given system.

There are several reasons for choosing this course of action. The first is economic. The systems perspective has at times led researchers to consider projects that are much too ambitious. National programmes rarely have the personnel, experience, or funds to meet the requirements of a research project that includes a large number of variables. The time horizon of such projects usually strains not only programme budgets but also farmers" patience and ignores opportunities for short-term gains.

In addition, there is good evidence that farmers are much more likely to adopt technologies in small steps rather than in complete packages (24, 25). Research that respects such adoption behaviour and offers farmers incremental changes easily accommodated to their own production system stands a much better chance of success.

This incremental, well-focused approach that characterizes OFR has implications for several themes important to current debate regarding nutrition and agricultural change. These include the issues of dietary diversity and commercial agriculture.

Dietary Diversity and the Selection of Research Priorities

A strategy that directs agricultural research to a few crops and problems may seem to run counter to the nutritionist's concern for promoting dietary diversity. There are two points that need to be examined here.

The first consideration is the nature of the malnutrition problem. The majority of malnutrition is of the calorie-protein type. Although previous analyses of the relationship of agriculture to nutrition gave priority to more specific problems, such as protein production (26), current thinking emphasizes the importance of improving the quantity of local diets (27-29). This reinforces a strategy such as OFR that usually begins with local staples. There are certainly instances, however, where serious micronutrient deficiency would cause researchers to consider working with other crops.

A second point concerns the feasibility of working with so-called "minor" crops. There is much interest in developing some of the crops that have been overlooked by agricultural research (30, 31). Farmers try to plant a wide variety of crops for reasons that include efficient use of available land and labour, dietary preferences, nutritional value, and risk avoidance. Production programmes want to preserve this diversity. But priorities for short-term research must be chosen, and the opportunity cost of dividing research among a number of crops must be considered. The agronomic and economic potential of these minor crops must be examined along with their actual importance in the diet. In some cases (e.g., certain garden vegetables) there may be considerable potential, but in many other cases attention to minor crops may offer neither economic nor nutritional benefits.

An analysis in northern Ecuador, for instance, led to the conclusion that a research effort into the traditional (and highly nutritious) crops quinoa and lupine would be counterproductive (32). In this case research began with the principal staples, maize and beans. But as the Ecuadorian programme evolved, research soon included broad beans, potatoes, and peas. Work on these crops would have been much less feasible, however, without initial progress with the principal staples.

Commercial versus Subsistence Agriculture

Means for improving nutrition through agriculture must be linked to innovations that offer farmers gains in productivity. In many cases this implies a focus on dietary staples for home consumption, but in other instances the research may be directed towards increasing farmers' capacities for marketing these or other crops.

The current debate regarding commercial versus subsistence agriculture (33, 34) may be misdirected and is certainly poorly documented. Although this is not the place to review all of the evidence related to this complicated issue, at least two points should be made. First, there are few instances of truly subsistence agriculture to be found in the world (35). Most farmers are involved to some extent in markets, and analyses that try to ignore this by referring to a pristine, self-sufficient system are inappropriate. Second, in many cases where peasant producers have suffered (nutritionally or otherwise) from the introduction of a development scheme, it is not the fault of commercial agriculture per se, but rather a lack of attention to targeting, on-farm experimentation, and economic feasibility.

It is well known that improper targeting has allowed a privileged minority to capture the bulk of the benefits of many agricultural development projects (36, 37). Implications for nutritional impact are obvious and require no further comment. A careful identification of targets is essential before work begins.

Equally important is location-specific research that include! on-farm testing before recommendations are made. Many examples of the nutritional failure of agricultural development projects are testimony to the general lack of competence with which these projects are planned and executed, rather than to a lack of specific nutrition I, planning. Dewey (35), for instance, has described the negative nutritional impact of a project in Tabasco, Mexico, in which participating farmers were turned into day labourers, and the project appears to be failing from the agricultural standpoint. The nutritional result is thus a symptom of a larger problem. More effective projects do not necessarily require extensive nutritional analysis, but rather more attention to targeting and technology design.

A further problem is the adequacy of markets for proposed production increases and purchasing demands. More than one development project has learned the importance of assuring local market capacity for increased production, especially for minor crops (38). Equally important, local markets are often unable to provide foodstuffs at reasonable prices to those who have switched to commercial crops (39). When these factors are not taken into account, the nutritional impact of a production increase may be disappointing.

However, there are many examples of commercial agriculture functioning for small farmers and providing nutritional benefits as well. The improvement in nutritional status for children of Jamaican farmers that came during the adoption of a national "self-sufficiency" policy (40) seems to have come at least in part from increasing opportunities for marketing produce and earning higher incomes. In a recent study in eastern Kenya, Fleuret and Fleuret (41) report evidence of the positive impact of commercial agriculture on nutritional status among peasant farmers.

The option of commercial agriculture must thus be approached with an open mind. The challenge for OFR is to select crops that most benefit the rural poor. In many cases it will be a food crop that poor people grow for their own consumption. But it may also be a food crop that the poor have an advantage in growing for sale (42, 43) or a commercial crop that gives farmers more flexibility in their cropping patterns (44).

Additional Considerations for a Productivity Strategy

It must be conceded that attention to increasing the productivity and incomes of target farmers is a necessary, but not always sufficient, condition for ensuring the nutritional impact of agricultural research. As mentioned previously, there are a number of examples of increases in productivity or incomes among small farmers without attendant nutritional gains (45-47). Sometimes these can be explained by failures in targeting, but in other cases there are additional factors that modify the nutritional impact of agricultural development.

In many instances increased productivity would be more effective nutritionally if it did not have to compete with the "background noise" of inadequate sanitation and health care (43). Sometimes problems in the management of added income (e.g., lump sum payments for commercial crops} and inexperience with (or lack of access to) markets for additional dietary items also limit the nutritional impact of agricultural development. These considerations imply a need for nutrition education or even a more integrated approach to rural development. But the success of these latter efforts will largely depend on the establishment of income and productivity gains that farmers can appreciate. An association with OFR activities would give nutritionists the opportunity to plan educational and other activities in support of agricultural change.

Another complicating issue in determining the nutritional impact of agricultural change is women's role in this process. The nutritional implications of women's participation in agriculture has been the subject of a number of analyses 148, 49). There are at least two concerns. The first is the fact that in many parts of the world women are farmers, and research and extension should be directed towards them. This is a question of targeting and is relatively straightforward.

The second concern is more subtle, and involves the possible effects of technological change on women's time and income. The basic dilemma is the balance between the positive nutritional impact of increasing women's incomes (20) and the deleterious effects of a shift in time allocation from child-rearing to income-earning activities (50). Changes in household time allocation and income are certainly important for OFR, but when these have nutritional implications as well, OFR will have to modify its data collection and analysis procedures.

The emphasis on productivity also overlooks the fact that a strategy that simply provides farmers with more of the same diet may not be sufficient. Although calorie-protein malnutrition is the most common problem, there are other nutrient deficiencies of importance. The use of added income to diversify the diet is one way of attacking these problems, but more specific agricultural solutions may be possible in some cases.

The introduction of a new variety of the local staple with improved nutrient content is one possibility; quality protein maize is certainly the most prominent example here (51). In other cases, a change in cropping patterns may be indicated, as with the introduction of vegetables to improve vitamin A availability. There are several prerequisites for such a change however: Nutritionists must demonstrate the seriousness of the particular deficiency in the local setting. Evidence must be available that the increased nutrient supply will reach those in need, either directly or through some sort of educational programme. And, as always, agricultural scientists must provide agronomic and economic evidence of the attractiveness of the new varieties or crops to farmers. Again, participation in the OFR process will allow nutritionists to help identify those situations where a specific nutrient may be provided through this type of agricultural change.

Agricultural development may at times be responsible for other nutritional changes as well. One such problem is a shift in cropping patterns that increases the production of certain crops at the expense of others. It must be assured that the resulting production pattern provides an adequate supply of nutrients or that these are available through the market. Because of the incrementalist approach of OFR, drastic changes of this type are unlikely, but they still must be guarded against.

Another nutritional effect of agricultural change may be the impact on farm labourers, who often represent the most nutritionally vulnerable section of the rural population. Whether new technologies increase or decrease opportunities for labourers is also an important issue for OFR. The problem is a difficult one, however, and may be related to policy issues that go well beyond the concerns and expertise of both location-specific agricultural research and applied nutrition.

These cautions regarding a strategy that focuses on increasing farmer's productivity and income provide additional justification for the inclusion of nutritionists in OFR. First, they present opportunities for linking applied nutrition activities, such as educational programmes, to technological change in agriculture. But they also demand the ability to analyse special situations in which new technologies may affect women's time or income, cropping patterns, specific nutrient availability, or rural labour markets. The thesis presented here is that in the majority of cases of OFR these factors will not be relevant, but the exceptions are important enough to require development of skills to accomodate them when they do arise.


Contents - Previous - Next