GLOBAL RESEARCH CHALLENGES: INCLUDING SMALL HOLDERS IN RURAL DEVELOPMENT'
Peter E. Flildebrand2
Agricultural research has been "globalized" since at least the 1960s when the International Agricultural Research Center system was inaugurated, the Green Revolution emerged, and the agricultural researchers and administrators from the foundations, universities, USAID, the IARCs, and some national agricultural research programs began globetrotting. In the 1 970s, multiple cropping research, as an example, became international, if not global, in scope. By 1972 in Central America it involved an IARC (IRRI), a regional research organization (CATIE), a national agricultural research institute (CENTA in El Salvador), and a U.S. university (Florida). And at least in El Salvador, this work was also multidisciplinary. It was also during this period of time that the term "farming systems"' began to be applied by the globetrotters to activities around the world, including Guatemala and Colombia. These were all multidisciplinary activities oriented specifically toward small holders. Also in the 1 970s, animals began to be incorporated into what were previously crop-biased "farming systems."
Despite this long history of a globalized research effort, a number of factors exist that have made ineffective our quest to include small holders in rural development. Chief among these is their great diversity. We tend to work where we can see broadly adoptable results and these efforts are supported by industry. Land is not necessarily the most limiting resource on small farms, yet we tend to look mostly at "yield" increasing technology measured in output per unit land area. "Our" crops are not necessarily the priority crops of the farmers, yet relatively little effort has been put in minor crops and in livestock. Average farms do not exist, yet we frequently work with averages. Thus our technology tends not to be appropriate for the poorer half of the population and they are not able to benefit from existing rural development efforts.
But small farms are not going away. Even though as a percent of the population, rural numbers are decreasing, farm populations in most countries in Latin America are still increasing. Therefore, it is time to take the challenge of including them in rural development. We need to work with the diversity that is both a characteristic of these small farms and a critical need of their livelihood systems. A multidisciplinary methodology that is broadly adaptable, but that conserves this diversity and can lead to different technologies for diverse groups of the poorest farmers is described in the paper.
Keywords: Diversity, limited resources, multidisciplinary, participatory
Invited keynote address at the lst Henry A. Wallace Inter-American Scientific Conference on Globalization of Agricultural Research. CATIE, Turrialba, Costa Rica February 25-27, 2002. 2Professor, Food and Resource Economics Department, University of Florida, Gainesville FL 32611-0240. email@example.com
GLOBAL RESEARCH CHALLENGES: INCLUDING SMALL HOLDERS IN RURAL DEVELOPMENT'
Peter E. Llildebrand2
Agricultural research has been "globalized" since at least the 1960s when the International Agricultural Research Center system was inaugurated, the Green Revolution emerged, and agricultural researchers and administrators from the foundations, universities, USAID, the IARCs, and some national agricultural research programs began globetrotting. In the 1970s I had foundation directors, university presidents, USAID officers and IARC scientists in my field plots in El Salvador and Guatemala. In 1972, Richard Bradfield, a Rockefeller Foundation scientist from UMR in the Philippines who had started working with multiple cropping, came to see our field plots in El Salvador on his way to visit the CATIE multiple cropping experiments in Costa Rica. In the Agricultural Economics Department of CENTA, we were already orienting our work toward practical applications for small holders and his visit inspired us to begin what became the national multicultivos program for CENTA. Multiple cropping research had become international, if not global, in scope. It involved an IARC, a regional research organization, a national agricultural research institute, and a U.S. university (Florida). And at least in El Salvador, this work was also multidisciplinary.
The Green Revolution, with its roots even before this time, was widely heralded, but slowly it was recognized that it was not benefiting many of the world's small farmers, including most of those on marginal lands, and particularly those with very limited resources. In the 1970s, new efforts to provide for the 'poorest of the poor' became common. It was slowly being realized that more than agronomy was required to find technologies that the poorest of the poor farmers could utilize and benefit from. In El Salvador, the multiple cropping program was headquartered in agricultural economics. Perhaps because of the farm management orientation in this department, our multiple cropping was oriented from the beginning to work under poor farm conditions as opposed to the work at IRRI and CATIE, which were more focused on the scientific interactions of different species. In 1973 we established our first on-farm trial even as we were still experimenting with new ideas on the station.
In El Salvador there was intense pressure on the land because of the high population density. Farms were small but there was relatively abundant labor. Seed and chemical inputs were generally available, and in the specific area where we were working (Zapotithn), there was irrigation. We were looking at ways to help farmers produce more cash crops on their limited land without decreasing the amount of maize and beans. By late 1973 we began to be visited by extension agents, small farmers and even some of the wealthy farmers. In December alone, 100 small farmers visited our plots. It was
' Invited keynote address at the I' Henry A. Wallace Inter-American Scientific Conference on Globalization of Agricultural Research. CATIE, Turrialba, Costa Rica February 25-27, 2002.
2 Professor, Food and Resource Economics and Director of International Program, Institute of Food and Agricultural Sciences, University of Florida, Gainesville FL 32611-0240. peh(~ufl.edu
interesting to note that after looking over what we were doing, the small farmers often said that if they could do it, they would not need more land. The wealthy farmers often said that if the small farmers could do what we were doing they wouldn't need more of the wealthy farmers' land.
The multicultivos project created a lot of excitement. We were visited in January 1974 by the President of the Agricultural Development Bank (Banco de Fomnento Agropecuario) of El Salvador, the Director General of CENTA and the Sub-secretary (vice minister) of the Ministry of Agriculture. Farmers also continued to visit in groups. In March our visitors to the multicultivos included the Minister of Agriculture and his Sub-secretary, the Director General of CENTA, the commanding Colonel of the Cavalry who was interested in putting his soldiers to work producing some of their own food, the Director of Planning of the Ministry of Agriculture, the owners of a food processing company, a large number of farmers from many areas of the country, some extension agents, and a group of managers from cotton farms who were interested in ways their workers might be able to produce more of their own food. In April, among other visitors, was a group of reporters for radio, TV, newspapers and magazines. As a result of their visit there were five articles on the mullicultivos in the newspapers, and reports on both radio and TV. We were also visited by the Board of Directors of the Agrarian Reform Institute who were interested in the implications of this intensive system on the amount of land small farmers might need and the potential income they might be able to expect. It was obvious that research directly focusing on the conditions and needs of small farmers was seen as unique and potentially very beneficial for the country. By 1976 there were nearly 600 on-farm trials of the multi cultivos technology throughout the country.
Beginning in 1974 in ICTA in Guatemala where I had gone, we added anthropologists to the mix of agronomists and economists in the "Rural Socioeconomics" unit and increased our efforts at working for the small farmers. This was when the Sondeo (Hildebrand 198 1) methodology was formulated and where we began keeping records with the small farmers (Hildebrand 1982). Rather than study the "adoption" of technology to find the characteristics of farmers were "innovative" and "early adopters" (Rogers) and thus worth working with, we assessed the acceptability of the technology for the conditions and capabilities of the small farmers. We were more concerned about why some farmers were unwilling or unable to adopt the new technologies being promoted by ICTA. What were the characteristics of the technology that made it unacceptable to the small farmers who were the clients? We were looking at the farm as a whole unit, not just at one crop at a time.
The Rural Socioeconomics unit also had its own on-farm trials and, again, these were visited by many of the "rural development" globetrotters. It was during this period of time that the term "farming systems"' began to be applied by the globetrotters, to our activities and to the activities of two other agricultural econom-ists, David Norman in Nigeria and Mike Collinson in Tanzania, as well as to other activities such as the DPI projects in Colombia in which Hubert Zandstra, among others was involved. These were all multidisciplinary activities oriented specifically toward small holders. Also in the
1970s, animals began to be incorporated into what were previously crop-biased "farming systems" (McDowell and Hildebrand).
By 1980, "farming system' programs were being created in many developing countries, and in 1982 the Farming Systems Support Project, FSSP, was created at the University of Florida by USAID. Its purpose was to provide technical assistance, training and networking for its widely scattered farming systems projects in Afiica, Asia and Latin America. "Farming Systems" was the golden buzzword during the 1980s. But many of those projects were not based on farming systems methodology nor oriented specifically toward small holders. USAID gradually became disenchanted with farming systems. A new buzzword began to emerge - sustainable - and it began to be substituted for the words farming systems in project titles. The bureaucracy ignored the fact that "sustainable" technology or sustainable7 agriculture were states of being and that "farming systems" as the term was being used was a methodology. It was even more curious that farming systems methodology was necessary in proposed sustainable projects in order to get approved.
Gender became incorporated in farming systems methodologies in an important way in the mid 1980s partly as a result of a conference held at the University of Florida in 1986 (Poats et al., 1988). This conference brought together the Women in Development and the Farming Systems communities from around the world. The emphasis was not on the potential confrontation of "Why aren't women taken into accountT' but rather on "How to better take gender into account" in agricultural and rural development activities. Gender as a concern has persevered.
The new buzzword is "participation." It was thought that previous efforts at reaching small holders were not being as effective as anticipated because the stakeholders - the small farmers - were not sufficiently being taken into account (Chambers). One frequently referenced successful participation project is that of Jackie Ashby's work with bean breeding at CIAT. But more purist participation projects tend to wander away from agricultural technology development and into such things as potable water, schools, roads, etc., so have not been popular with agricultural research organizations, national or international.
Constraints to including small holders
As promising as each of these new waves of rural development practices seemed to be at the time, they still often failed to reach most of the world's poorest farmers. A number of factors exist that have made ineffective our quest to include small holders in rural development. Chief among these is their great diversity. Even in areas where we used to define "homogeneous systems' (Hildebrand, 198 1) or "recommendation domains" (Byerlee et al. 1982) we now know that there is also tremendous diversity among households because of the composition making up those households. Our scientific and professional baggage is another.
Land is the most limiting resource on small farms
To many of us, increased production means increasing the amount of product produced per unit land area, because arable land is a globally limited resource. Thus we tend to look at technologies that increase yield per unit land area even if they require other resources such as more female or male labor or cash that many small holders have little of. Examples using cash include mechanization, inorganic fertilizer or pesticides. Yet, if we measured the productivity (yield) of any of these in terms of product per unit of cash, it might well be lower than what the farmers are already doing. Because labor, cash or seed are often more limiting than land on small-scale, family farms, we first need to consider increasing the productivity of these resources. Increased productivity of land may follow. In crops like potatoes or beans which are staple foods and can readily be sold, yet somewhat difficult to store under rustic conditions, amount of seed can be much more limiting than land. In Nariflo, Colombia, a well-known minifundio area, farmers planted potatoes in low densities to maximize the productivity of each potato planted even though it reduced yield per unit land area. When they ran out of seed, that was what detennined the size of the field. Even on those very small farms, seed was the limiting resource for potatoes, not land (Andrew 1970). In the Pakistani Punjab in the 1960s, farmers with limited water, spread it out on as much land as possible to maximize the productivity of the water. This, of course, reduced the yield of the crop per unit land area. But if they had increased the per ha application of water, it would have reduced the productivity of the water, resulting in less product. (Andrew and Hildebrand 1982).
"Our" crops are the priority crops
We also tend to put priority only on the animal, crop or crops in which we are interested, forgetting that the farmer must allocate resources among all the activities and needs of the household. Ignoring the seasonal needs of farmers can lead to inappropriate technologies. For instance, early (short season) crops usually have lower yields than later (longer season) crops. Yet farmers grow them to provide food or cash at a time when they are needed even if they yield less. Also, "late" planting is often the result of a difference between our priorities and those of the farmers. The farmer may put a higher priority on a different crop than the one in which we are interested so plants the other one first and ours later than we "know" is optimal.
We need to see results
We tend to define "results" in different ways. Researchers usually define results as statistically significant differences among treatments. If our trials do not produce statistically significant differences, we do not see "results." This leads to tight control of non-test variables in experiments to reduce unexplained variance, thus increasing the probability of achieving statistically significant differences among test treatments. When
3 Parts of the following are from a previous paper presented at the Workshop on No-till Farming in South Asia's Rice-Wheat System: Experiences from the Rice-Whcat Consortium and the USA, 21 February 2002, Ohio State University.
this practice is followed in on-farm trials, it creates an artificially superior environment that the farmer is unable to provide over time on a field basis (ffildebrand and Russell). Another method of helping assure statistically significant results is to have higher yields, to achieving a lower coefficient of variance, thus increasing the probability of significant differences among treatments. In on-farm trials this leads to selecting the best farms or the best fields or even the best spots in fields on which to conduct trials. The DG of one national agricultural research organization for which I worked said, "Pete, if you work on those hillsides (where the small farms were) you won't get any response." By response, he meant results as measured by statistically significant differences.
But results also are measured by breadth of adoption. Whether it is a cultivar, another input or a cultivation practice, broader adoption is always better than limited adoption. This is also appreciated by industry. A limited number of broadly adoptable products is better than a larger number of narrowly applicable items. This leads us toward practices and materials that require the modification of the field environment so that they can be productive, similar to what is done in experiments. 4 This practice has been very effective in areas such as the corn belt of the United States. It can also be effective in flooded rice paddies. But most small farmers with limited resources cannot modify the highly diverse environments of their farms and fields to suit the requirements of new technologies.
Looking for results can also lead us into working with the "innovators" or "early adopters" as Rogers (1962) would call them. These inevitably are the farmers with more resources, those who have the time to devote to meetings and in working with us, and those who can accept risk of experimentation. But orienting our work toward them so we can see "results" also leads us to those who are best able to modify their environments to meet the requirements of the technology. They inevitably are not the poorest of the poor small farmers,
Average farms do not exist
An "average' farm household in an area might have 1.2 adult males, 1.4 adult females, 1.7 adolescent males, 0.6 adolescent females and 2.4 children. It would probably have more land than at least half of the households in the area. It could have from 0.4 to 1.6 cattle or horses. Other averages could be added, but nowhere could such a farm household be found. Yet we inevitably tend to use averages when we discuss yields, farm size, available resources or capabilities to adopt new technologies. We ignore the fact that somewhere around half of the farms do not achieve the average yield, nor are that large, nor have other resources or capabilities to adopt new technologies. And technology mostly is not scale neutral. It is not the size of the field nor farm, but the resources that small farms do not have compared with large farms that make the difference. I once heard the research dean of a U.S. Land Grant university declare, "Certainly our technology works on small farms as well as large ones. We test it on small plots, don't we?"
4 Conway, p. 39, expresses a similar concern in a slightly different way,
Small farms are not going away.
Although the percentage of people living in rural, as opposed to urban areas is declining, in many countries the number of people living in rural areas and the agricultural population continues to increase. In Central America, between 1970 and 2000, the agricultural population increased by 2.6 million people (FAO). In Guatemala, there are as many people in agriculture now as there were total people in the country in 1970. Although there are no statistics that I know of, the majority of these people have to be living on small farms. This could be as many as 400,000 new small farms in Central America over this 30-year period.
It is time to take the challenge
Over this same 30 year period, even after realizing that Green Revolution technology has had a very limited effect on limited resource farmers on marginal lands, we have avoided the challenge of working with the great biophysical and socioeconomic diversity in which the world's small farmers struggle to survive. Perhaps we thought small farms would go away. "Get bigger or get out." Perhaps we thought that our technology was scale neutral and that it was just a matter of trickling down to the late adopters and laggards that were the small farmers who did not adopt it. Perhaps we thought that average farms represented all farms. Perhaps we feel we have to work where we can easily measure results the way we are traditionally accustomed to do. Perhaps we are convinced that land is the most limiting resource on small farms and that yield is measured only per unit land area. Perhaps we have forgotten, or do not know that farmers have many priorities and theirs may not coincide with ours.
Not all farms will adopt new technology
For whatever reasons, over this period of time, only a small fraction of our research and technology development efforts have been oriented directly and adequately at intensification of the still increasing numbers of small-scale, family farms with limited or marginal resources. For forty years we have been convinced by Everett M. Rogers (Diffusion of Innovations, 1962) that many farmers are slow adopters or even laggards or non-adopters of new technological innovations. Because we believe our technology is good we believe what Rogers tells us. We lean toward working with the "innovators" or "early adopters" who are those who are most apt to adopt our technology. We feel it is not our fault that some farmers do not adopt what we think is good technology. Furthermore, it is not our fault that some farmers do not have access to credit nor the cash resources to acquire our good technology or that it is not available in local markets. We know that it takes time for the benefits of our technology to be understood and for farmers to be "motivated" to adopt it (ffildebrand 1980). As a result, the research impacts the relatively small number of larger farmers, or those small farmers in better environments or with more resources, Figure 1, but does little or nothing for the very large number of limited resource farmers at the bottom of the pyramid. We have taken the easy road of asking those farmers with sufficient resources to change the environment to suit the technology. Now it is time to accept the challenge of creating technology to
suit the diverse biophysical and socioeconomic environments of most of the world's small farmers.
Environmental 7 7eso u -irc e--sQual ity
Lower No. of Farmers Less
Figure 1. Farm resource diversity and relative numbers of farmers. (Hildebrand 1993)
Agricultural intensification on small farms
To effectively help intensify highly diverse small farm agriculture it is necessary to comprehend the livelihood systems of these small farmers. A livelihood system is comprised of all the on- and off-farm activities available to farmers in an area from which they can select their strategies to survive and thrive. This includes not only all the crops and livestock they raise, but different ways or times of raising them. Besides production activities, it is also important to understand reproduction and community activities as well because they also use scarce farm resources. Production activities are those that result in the production of goods such as food (for consumption or sale), or cash. Farming, fishing, carpentry, cottage industry, migrant work, paid labor, civil service, etc. can be considered production activities whether on or off the farm. Reproduction activities are those like maintenance and care of the family unit that result in the survival and succession of the family or household. Meal preparation, hauling water or fuel, childcare, laundry, house cleaning, re-roofing, house building (for family), or caring for elderly or disabled are among reproduction activities. Community activities are more difficult to quantify in terms of inputs and outputs, however they play a key role in understanding how households and communities function. Community activities might include: attending or organizing meetings; forming or participating in women's groups, men's groups, children's groups, or producer groups; acting as part of a village or
community council, household food sharing; or the like. Seasonality of activities and periods of cash or labor scarcities are important to understand as well as which of the household members is involved in each activity.
Different households do not all adopt the same strategies. Livelihood strategies are a function of the characteristics of the households such as wealth, sex of the household head, relative age of the household and household composition (sex, age and relationship of household members). Even though all households in a livelihood system have access to all activities, the constraints and resources reflected in these characteristics cause the members to choose different subsets of the activities as strategies. To effectively help all these diverse households intensify production, it is critical to assess the capabilities of each type of household in order to mold the technology to the needs and constraints of each type.
But this sounds like anthropology. And everyone knows that anthropologists take years to do their ethnographic studies in remote villages and then don't really want anyone or anything to change "their" village. This is no longer the case. Many anthropologists with solid agricultural backgrounds are productive members of multidisciplinary teams, Also, many agronomists now have solid anthropological training. Incorporating these kinds of scientists in teams working to intensify diverse small farm agriculture is highly productive. Economists (heaven forbid!) with agronomic and/or anthropological training can also be useful members of the team.
Modeling small-scale, limited-resource family farm livelihood systems, such as by ethnographic linear programming (Bastidas; Breuer; Cabrera; Grier; Gough; Kaya, Kaya et al.; Litow; Mudhara; Pomeroy; Sullivan; Thangata), is one effective way to integrate crop, animal, anthropologic and economic knowledge gained through fanner participation to help predict which households may be able to adopt different kinds of new technologies even prior to their being offered to farmers in the community. 5 These models help us understand the kinds of technologies that are needed by the different types of households in an area and thus can guide innovative thought into unique approaches to agricultural intensification that can help even the poorest kinds of small farm households survive and thrive.
Figure 2 represents a highly efficient and very effective methodology for incorporating these different kinds of scientists in a participatory process with farmers that incorporates diversity, both in problem or constraint assessment and in recommendations. With the availability of laptop computers, it is now feasible for modelers to work in the field with farmers in the process of creating, validating and using their models. When these models are validated (adequately simulate the existing livelihood system), alternatives can be pre-tested in the models, even while on-farm trials are being conducted, both to help researchers better understand how the alternatives would fit into the strategies of the different kinds of households, and to help characterize the recommendation domains for
5 See Hazell and Norton for details on linear progranuning. See Hildebrand, 2001, for details on ethnographic linear programming. For examples of ethnographic linear programming see the other authors listed above.
which the technologies are appropriate. The results of the on-farm trials and knowledge gained from continuous contact with the farmers can be used to improve the ethnographic linear programn-fing models which should constantly be modified to make them even more useful.
Figure 2. Schematic representation of the methodology for ex-ante evaluation of potential technology, infrastructure or policy changes. (based on Bastidas, 2001)
It can be done
Mniaturizations of computer hardware, and advances in software have generated the potential to create and use sophisticated models in the field while working with farmers in their diverse environments in a participatory, ethnographic mode. We know these methods work. Now is the time to put them all to work together. This will require concerted, multidisciplinary efforts and the will to shed many approaches to which we tend to cling. The remaining challenge is for research and extension personnel, infrastructure managers and politicians to become innovative in their search for technologies, infrastructure and policies specifically oriented to the still increasing number of highly diverse, small-scale, limited resource, family farms in many countries of the world including those in Latin America and the Caribbean.
Success (reaching poor farmers) will not be achieved either by applying modem science and
technology, on the one hand, or by implementing economic and social reform on the other, but
through a combination of these that is innovative and imaginative. (Conway p. 42).
American Society of Agronomy. 1976. Multiple cropping. ASA Special Publication Number 27. Madison.
Bastidas, E.P. 2001. Assessing potential response to changes in the livelihood system of diverse, limited-resource farm households in Carchi, Ecuador: modeling livelihood strategies using participatory methods and linear programming. PhD dissertation, Food and Resource Economics, University of Florida.
Breuer, N.E. 2000. The role of medicinal plants in rural Paraguayan livelihoods. M.A. thesis, Latin American Studies, University of Florida.
Byerlee, D., L. Harrington and D.L. Winkelman. 1982. Farming systems research strategy and technology design. American Journal of Agricultural Economics 64:899904.
Cabrera, V.E. 1999. Farm problems, solutions, and extension programs for small farmers in Cafiete, Lima, Peru. M.S. thesis, Agricultural Education and Communication, Farming Systems Concentration, University of Florida.
Conway, G. 1997. The doubly green revolution. Cornell University Press, Ithaca.
FAOSTAT Agricultural Data. http://apps.fao.orgipage/collections?subset=agriculture
Grier, C.E. 2002. Potential Impact of Improved Fallows on Small Farm Livelihoods, Eastern Province, Zambia. M.S. thesis, Food and Resource Economics, University of Florida.
Gough, A.E. 2002. The Starter Pack Program in Malawi: Implications for Household Food Security. M.S. thesis, Agricultural Education and Communication, Farming Systems Concentration, University of Florida.
Hazell, P.B.R. and R.D. Norton, 1986. Mathematical programming for economic analysis. Macmillan, New York.
Hildebrand, P.E. 1979. The ICTA farm record project with small farmers: four years of experience. ICTA, Guatemala.6
Hildebrand, P.E. 1980. Motivating small farmers, scientists and technicians to accept change. Agricultural Administration 8:375-383.
Hildebrand, P.E. 1981. Combining disciplines in rapid appraisal: The sondeo approach. Agricultural Administration 8:423-432.
6 Also pp. 309-314 In: Shaner, W.W., Philipp, P.F. and W.R Schmet. 1982. Farming systems research and development: guidelines for developing countries. Westview Press, Boulder.
Hildebrand, P.E. 1993. Targeting technology diffusion through coordinated on-farm research. Presented at the Association for Farming Systems Research-Extension North American Symposium on Systems Approaches in North American Agriculture and Natural Resources: Broadening the Scope-of FSRE. University of Florida.
Hildebrand, P.E. 2001. Ethnographic linear programming: limited-resource, family-farm household models. Class handout. AEB 5167, Economic analysis in small farm livelihood systems. Food and Resource Economics Department, University of Florida.
Kaya, B. 2000. Soil fertility regeneration through improved fallow systems in southern Mali. PhD dissertation, Forest Resources and Conservation, University of Florida.
Kaya, B., P.E. Hildebrand and P.K. Nair, 2000. Modeling changes in farming systems with adoption of improved fallows in southern Mali. Agricultural Systems 66: 51 -68.
Litow, P.A. 2000. Food security and household livelihood strategies in the Maya Biosphere Reserve: The importance of milpa in the community of Uaxactiin, Pet6n, Guatemala. M.S. thesis, Agricultural Education and Communication, Farming Systems Concentration, University of Florida.
McDowell, R.E. and P.E. Hlildebrand. 1980. Integrated crop and animal production: Making the most of resources available to small farms in developing countries. A Bellagio conference. The Rockefeller Foundation, New York.
Mudhara, M. 2002. Assessing the livelihood system of diverse smallholder farm households: potential adoption of improved fallows in Zimbabwe. PhD dissertation, Food and Resource Economics, University of Florida.
Pomeroy, C. 2000. An evaluation of a crop diversification project for low resource hillside farmers in the Dominican Republic. M.S. thesis, Agricultural Education and Communication, Farming Systems Concentration, University of Florida.
Poats, S.V., M. Schmink and A. Spring. 1988. Gender issues in farming systems research and extension. Westview Press, Boulder.
Rogers, E.M. 1962. Diffusion of innovations. The Free Press, A Division of Macmillan Publishing Co, Inc. New York.
Sullivan A.J. 2000. Decoding diversity: mitigating household stress. M.S. thesis, Agricultural Education and Communication, Farming Systems Concentration, University of Florida.
Thangata, P. 2002. The potential for agroforestry adoption and carbon sequestration in smallholder agroecosystems of Malawi: an ethnographic linear programming approach. PhD dissertation, Interdisciplinary Ecology, Natural Resources and Environment, University of Florida.