Getting a new idea adopted can be very difficult. This is all the more frustrating when it seems to the proponents of the new idea that it has very obvious advantages. It can be a challenge to try to introduce new ideas in rural areas, particularly in less-developed societies, where people are somewhat set in their ways—ways that have evolved slowly, through trial and error. It’s all the more difficult when those introducing new ideas don’t understand why people follow traditional practices. Rural sociologists and agricultural extension researchers who have studied the diffusion of agricultural innovations have traditionally been oriented toward speeding up the diffusion process (Rogers 1983). Pro-innovation bias has sometimes led sociologists to forget that ”changing people’s customs is an even more delicate responsibility than surgery” (Spicer 1952).
Although innovation relies on invention, and although considerable creativity often accompanies the discovery of how to use an invention, innovation and invention are not the same thing. Innovation does, however, involve more than a change from one well-established way of doing things to another well-established practice. As with all innovations, those in agriculture involve a change that requires significant imagination, break with established ways of doing things, and create new production capacity. Of course, these criteria are not exact, and it is often difficult to tell where one innovation stops and another starts. The easiest way out of this is to rely on potential adopters of an innovation to define ideas that they perceive to be new.
Innovations are not all alike. New ways of doing things may be more or less compatible with prevalent norms and values. Some innovations may be perceived as relatively difficult to use and understand (i.e., complex), while others are a good deal simpler. Some can be experimented with in limited trials that reduce the risks of adoption (i.e., divisible). Innovations also vary in the costs and advantages they offer in both economic and social terms (e.g., prestige, convenience, satisfaction). In the economists’ terms, innovation introduces a new production function that changes the set of possibilities which define what can be produced (Schumpeter 1950). Rural sociologists have studied the adoption of such agricultural innovations as specially bred crops (e.g., hybrid corn and high-yield wheat and rice); many kinds of machines (e.g., tractors, harvesters, pumps); chemical and biological fertilizers, pesticides, and insecticides; cropping practices (e.g., soil and water conservation); and techniques related to animal husbandry (e.g., new feeds, disease control, breeding). Often they have relied upon government agencies such as the U.S. Department of Agriculture to tell them what the recommended new practices are.
The diffusion of agricultural innovations is a process whereby new ways of doing things are spread within and between agrarian communities. Newness implies a degree of uncertainty both because there are a variable number of alternatives and because there is usually some range of relative probability of outcomes associated with the actions involved. Rogers (1983) stresses that the diffusion of innovations includes the communication of information, by various means, about these sets of alternative actions and their possible outcomes. Information about innovations may come via impersonal channels, such as the mass media, or it may pass through social networks. From an individual’s point of view, the process of innovation is usually conceived to start with initial awareness of the innovation and how it functions. It ends with adoption or nonadoption. In between these end points is an interactive, iterative process of attitude formation, decision making, and action. The cumulative frequency of adopters over time describes an S-shaped (logistic) curve. The frequency distribution over time is often bellshaped and approximately normal.
Individual innovativeness has been characterized in five ideal-type adopter categories (Rogers 1983). The first 2 to 3 percent to adopt an innovation, the ”innovators,” are characterized as venturesome. The next 10 to 15 percent, the ”early adopters,” are characterized as responsible, solid, local opinion leaders. The next 30 to 35 percent are the ”early majority,” who are seen as being deliberate. They are followed by the ”late majority” (30 to 35 percent), who are cautious and skeptical, and innovate under social and economic pressures. Finally, there are the ”laggards,” who comprise the bottom 15 percent. They are characterized as ”traditional,” although they are often simply in a precarious economic position. Earlier adopters are likely to have higher social status and better education, and to be upwardly mobile. They tend to have larger farms, more favorable attitudes toward modern business practices (e.g., credit), and more specialized operations. Earlier adopters are also argued to have greater empathy, rationality, and ability to deal with abstractions. They are less fatalistic and dogmatic, and have both positive attitudes toward change and science, and higher achievement motivation and aspirations. Early adopters report more social participation and network connections, particularly to change agents, and greater exposure to both mass media and interpersonal communication networks.
Although Rogers (1983) provides dozens of such generalizations about the characteristics of early and late adopters, he admits that the evidence on many of these propositions is somewhat mixed (Downs and Mohr 1976). Even the frequently researched proposition that those with higher social status and greater resources are likely to innovate earlier and more often has garnered far less than unanimous support (Cancian 1967, 1979; Gartrell 1977). Cancian argues that this is a result of ”upper middle-class conservatism,” but subsequent meta-analysis has clearly demonstrated that the relationship between status and innovation is indeed linear (Gartrell and Gartrell 1985; Lewis et al. 1989). If anything, those with very high status or resources show a marked tendency to turn their awareness of innovations into trial at a very high rate (Gartrell and Gartrell 1979).
Ryan and Gross (1942) provide a classic example of diffusion research. Hybrid corn seed, developed by Iowa State and other land-grant university researchers, increased yields 20 percent over those of open-pollinated varieties. Hybrid corn also was more drought-resistant and was better suited to mechanical harvesting. Agricultural extension agents and seed company salesmen promoted it heavily. Its drawback was that it lost its hybrid vigor after only one generation, so farmers could not save the seed from the best-looking plants. (Of course, this was not at all a drawback to the seed companies!)
Based on a retrospective survey of 259 farmers in two small communities, Ryan and Gross found that 10 percent had adopted hybrid corn after five years (by 1933). Between 1933 and 1936 an additional 30 percent adopted, and by the time of the study (1941) only two farmers did not use the hybrid. Early adopters were more cosmopolitan, and had higher social and economic status. The average respondent took nine years to go from first knowledge to adoption, and interpersonal networks and modeling were judged to be critical to adoption. In other cases diffusion time has been much shorter. Beginning in 1944, the average diffusion time for a weed spray (also in Iowa) was between 1.7 years for innovators and 3.1 years for laggards (Rogers 1983, p. 204). Having adopted many innovations, farmers are likely to adopt others more quickly.
Adoption-diffusion research in rural sociology has dominated all research traditions studying innovation. Rural sociology produced 791 (26 percent) of 3,085 studies up to 1981 (Rogers 1983, p. 52). Most of this research relied upon correlational analysis of survey data based on farmers’ recall of past behaviors. This kind of study reached its peak in the mid 1960s. By the mid 1970s the farm crisis in the United States and the global depression spurred rural sociologists to begin to reevaluate this tradition. By the 1980s global export markets had shrunk, farm commodity prices had fallen, net farm incomes had declined, and high interest rates had resulted in poor debt-to-asset ratios. What followed was a massive (50 percent) decapitalization of agriculture, particularly in the Midwest and Great Plains.
Criticisms of adoption-diffusion research include (1) pro-innovation bias; (2) a lack of consideration of all the consequences of innovation; (3) an individual bias; (4) methods problems; (5) American ethnocentric biases; (6) the passing of the dominant modernization-development paradigm. The pro-innovation bias of researchers has led them to ignore the negative consequences of innovation (van Es 1983). Indeed, innovativeness itself is positively valued (Downs and Mohr 1976). The agencies that fund research and the commercial organizations (e.g., seed companies) that support it have strong vested interests in promoting diffusion. Furthermore, successful innovations leave visible traces and can be more easily studied using retrospective social surveys, so researchers are more likely to focus on successful innovations.
Since most researchers are well aware of this problem, it can be addressed by deliberately focusing on unsuccessful innovations, and by studying discontinuance and reinvention. It can also be avoided by the use of prospective research designs, including qualitative comparative case studies, that track potential innovation and innovators’ perceptions and experiences. This should facilitate the investigation of noncommercial innovations and should result in a better understanding of the reasons why people and organizations decide to use new ideas. Moreover, these methods will likely lead to a better understanding of the system context in which innovations diffuse.
One of the most strident critiques of the pro-innovation bias of the ”land-grant college complex” was voiced by Hightower (1972). Agricultural scientists at Davis, California, worked on the development and diffusion of hard tomatoes and mechanized pickers (Friedland and Barton 1975). They ignored the effects of these innovations on small farms and farm labor, except in the sense that they designed both innovations to solve labor problems expected when the U.S. Congress ended the bracero program through which Mexican workers were brought in to harvest the crops. In the six years after that program ended (1964 to 1970) the mechanical harvester took over the industry. About thirty-two thousand former hand pickers were out of work. They were replaced by eighteen thousand workers who rode machines and sorted tomatoes. Of the four thousand farmers who produced tomatoes in California in 1962, only six hundred were still in business in 1971. The tomato industry honored the inventor for saving the tomato for California, and consumers got cheaper, harder tomatoes—even if they preferred softer ones.
Several other classic examples of agricultural innovation illustrate problems that result from not fully considering the consequences of innovation (Fliegel and van Es 1983). Until the late 1970s rural sociologists, among others, studiously ignored Walter Goldschmidt’s 1940s study (republished in 1978) of the effects of irrigation on two communities in California’s San Joaquin Valley. Dinuba had large family farms, and it also had more local business, greater retail sales, and a greater diversity of social, educational, recreational, and cultural organizations. Arrin was surrounded by large industrial corporate farms supported by irrigation. These farms had absentee owners and Mexican labor. This produced a much lower quality of life that was confirmed three decades later (Buttel et al. 1990, p. 147).
The enforced ban on earlier chemical innovations in agriculture by the U.S. Food and Drug Administration provides another interesting example. Chemical innovations such as DDT insecticide, 2,4-D weed spray, and DES cattle feed revolutionized farm production in the 1950s and 1960s. In 1972, DDT was banned because it constituted a health threat (Dunlap 1981), and 2,4-D, DES, and similar products were banned soon afterward. Finally, in 1980 the U.S. Department of Agriculture reversed its policy and began to advise farmers and gardeners to consider alternative, organic methods that used fewer chemicals.
The impact of technical changes in U.S. agriculture, particularly the rapid mechanization begun in the Great Depression, put farmers on the ”treadmill of technology” (Cochran 1979; LeVeen 1978). Larger farmers who are less risk-aversive adopt early, reap an ”innovation rent,” reduce their per-unit costs, and increase profits. After the innovation spreads to the early majority, aggregate output increases dramatically. Prices then fall disproportionately, since agricultural products have low elasticity of demand. Lower, declining prices force the late majority to adopt, but they gain little. They have to adopt to stay in business, and some late adopters may be forced out because they cannot compete. This treadmill increases concentration of agricultural production and benefits large farmers, the suppliers of innovations, and consumers. Indeed, it helps to create and to subsidize cheap urban labor. When it comes to environmental practices, however, large farms are not early innovators (Pampel and van Es 1977; Buttel et al. 1990).
The individual bias of adoption-diffusion research is evident in its almost exclusive focus upon individual farmers rather than upon industrial farms or other agribusiness. There is also a tendency to blame the victim if anything goes wrong (Rogers 1983). Change agents are too rarely criticized for providing incomplete or inaccurate information, and governments and corporations are too infrequently criticized for promoting inappropriate or harmful innovations. Empirical surveys of individual farmers also lead to a number of methodological problems. As noted above, if surveys are retrospective, recall relies on fallible memory and renders unsuccessful innovations difficult to study. These surveys are commonly combined with correlational analysis that makes it difficult to address issues of causality. After all, the farmer’s attitudes and personality are measured at the time of the interview, and the innovation probably occurred some time before. As we have pointed out, these issues can be addressed by prospective designs that incorporate other methods, such as qualitative case studies and available records data, and focus on the social context of innovation.
Taking into account the social context of innovation involves shifting levels of analysis from individual farmers to the social, economic, and political structures in which they are embedded. Contextual analysis of social structures has evolved in two directions, both of which have been inspired by the adoption-diffusion paradigm. The first considers social structure as a set of social relations among farmers, that is, a social network. Typically, social network structure has been studied from the point of view of individual farmers. For instance, farmers are more likely to innovate if they are connected to others with whom they can discuss new farming ideas (Rogers and Kincaid 1981; Rogers 1983; Warriner and Moul 1992). In this type of analysis, ”connectedness” becomes a variable property of individual farmers that is correlated with their innovativeness. It is much less common to find studies that consider how agricultural innovation is influenced by structural properties of entire networks, such as the presence of subgroups or cliques, although other types of innovation have been studied within complete networks (see, e.g., Rogers and Kincaid 1981 on the diffusion of family planning in Korean villages). Field studies of subcultural differences in orientations to innovation report what amount to network effects, though networks are rarely measured directly. For instance, studies of Amish farmers have revealed that members of this sect restrict certain kinds of social contacts with outsiders in order to preserve their beliefs, which include environmental orientations based on religious beliefs (for a review, see Sommers and Napier 1993). Given the growing importance of social network analysis in contemporary sociology (Wasserman and Faust 1995), and the demonstrated importance of networks of communication and influence in innovation research (Rogers and Kincaid 1981), future studies of agricultural innovation could profitably incorporate network models and data in their research designs.
A second type of social structural analysis has considered how agricultural innovation is influenced by distributions of resources within farming communities. Much of this research has focused on the so-called ”Green Revolution.” This term refers to the increases in cereal-grain production in the Third World, particularly India, Pakistan, and the Philippines, in the late 1960s, through the use of hybrid seeds and chemical fertilizer. In Indian villages where knowledge of new farming technology and agricultural capital were highly concentrated, the rate at which individual farmers translated their knowledge into trial was higher (Gartrell and Gartrell 1979). Yet overall levels of innovation tended to be lower in such high-inequality villages. Had the primary goal of India’s development programs been to maximize the rate at which knowledge of new farming practices is turned into innovation, then these results could have been seen as a vindication of a development strategy that concentrated on well-to-do cultivators and high-inequality villages. Yet, this and other assessments of the Green Revolution in the 1970s suggested that development would likely exacerbate rural inequality; that most of the benefits of innovation would accrue to farmers who were wealthy enough to afford the new inputs (Frankel 1971; Poleman and Freebairn 1973); and that the rural poor would be further marginalized and forced to seek employment in the increasingly capital-intensive industries developing in cities. The Green Revolution, so the thinking went, contained the seeds of civil unrest in the cities—an urban Red Revolution (Sharma and Poleman 1993).
Recent research in agricultural economics paints a more optimistic portrait of the long-run distributional effects of the Green Revolution. Once small farmers were given the necessary infrastructural support, their productivity and incomes increased. Growing rural incomes and the resulting growth in consumption demand stimulates the development of a wide variety of off-farm and noncrop employment opportunities. Through participation in these ”second generation” effects of the Green Revolution, the incomes of landless and near-landless households have increased dramatically (Sharma and Poleman 1993). These indirect consequences of agricultural innovation are not limited to the Green Revolution. Innovation in agriculture and the expansion of rural, nonagricul-tural manufacturing were strongly associated in the development of Western Europe and East Asia as far back as the eighteenth century (Grabowski 1995). We tend to think of the Industrial Revolution as an urban phenomenon in which agricultural surplus labor was transferred from occupations of low productivity in agriculture to those of high productivity in urban manufacturing. Yet in this early phase of industrialization, the flow of people and economic activity went the other way—from town to country. Then, as now, agricultural innovation influenced and was influenced by the growth of rural manufacturing. Expanding commercialization of rural areas fosters innovation by providing many of the inputs needed by agriculture, as well as sources of credit for farm operations and alternative sources of income to buffer the risks associated with innovations. These reciprocal effects have been observed in the experience of China in the 1990s (Islam and Hehui 1994).
Economists have argued that agricultural innovation is induced by changes in the availability and cost of major factors of production, particularly land and labor (Binswanger and Ruttan 1978). Historical, cross-national studies show that countries differently endowed with land and labor have followed distinct paths of technological change in agriculture. Population pressures on land resources have impelled technological change and development (Boserup 1965, 1981; Binswanger 1986). Rather than focusing on individual farmers’ adoption decisions, research in this tradition has examined variation in innovation by region according to demographic and other conditions.
The social context of innovation also has an important political dimension. Political structures powerfully influence the path of innovation. When family farms were turned over to collective management by the Democratic Republic of Vietnam, local farmers in both the lowland and upland areas were unable to use their own knowledge of farm management (Jamieson et al. 1998). Traditional knowledge passed down over many generations became lost to new generations of farmers. Yet local knowledge systems are often crucial to the successful implementation of modern farming technologies brought in from the outside. Such problems are compounded when the power to make decisions about the course of development is centralized in national agencies, as is the case in Vietnam.
Innovation can also call forth new political structures. India’s Green Revolution became politicized as the terms and prices at which agricultural inputs could be obtained and the price at which agricultural products could be sold were determined by government and its local agencies. Peasant movements arose as a response to concern about access to the new farming technology. The incidence of improved agricultural practices has been associated with the rise of political parties such as the Lok Dal of Uttar Pradesh, which most clearly articulated rural interests (Duncan 1997). By asserting new identities and interests created in the changed circumstances brought about by the Green Revolution, the Lok Dal was able to mobilize across traditional lines of caste and locality.
The social, economic, and political structures of the social context of innovation do not exist in isolation from one another. In any development setting, a contextually informed understanding of agricultural innovation must consider the relationships among these different types of structures (Jamieson et al. 1998). While it may no longer be as fashionable as it once was, the adoption-diffusion model still has much to offer in such efforts. The model refers implicitly to structural effects of socioeconomic status and communication behavior, though these are conceptualized at an individual level (Black and Reeve 1992). Structural analysis has recently moved more firmly into this interdisciplinary realm, particularly in economics (see ”Economic Sociology”). With the appropriate structural tools, rural sociologists could make notable contributions to our understanding of how the social structures of markets influence innovation.
Technological change in agriculture is still vitally important throughout the world and, correctly applied, diffusion research can assist in its investigation. It is important to consider the consequences of technological change as well as the determinants of adoption of innovation. It is critical to apply the model to environmental practices and other ”noncommercial” innovations in agriculture. In-depth case studies over time are needed to further our understanding of how and why individuals and agricultural social collectives adopt technological change. Above all, the social, economic, and political contexts of innovation must be studied with the models and methods of modern structural analysis. All this provides a basis for continuing to build on a wealth of research materials.
