Supplements, not entry points: the stratification of agroecological knowledge access among smallholder farmers in western Kenya

In the fields of western Kenya, where small farms feed families and communities, a quiet revolution is taking place. It is not driven by new seeds or expensive machinery, but by a shift in how farmers learn to grow food. This approach, known as agroecology, treats farming as a complex web of life rather than a simple factory line. Unlike traditional methods that often rely on buying specific inputs like chemical fertilizers, agroecology asks farmers to observe their land, adapt practices to local conditions, and share knowledge with one another. Because these methods are so specific to each farm, they cannot be learned from a single pamphlet or a radio broadcast; they require observation, trial, and feedback. For this system to work, farmers need access to a wide variety of information sources, from neighbors and local groups to government experts and scientific institutions. The question facing policymakers is whether these different channels of knowledge are actually reaching the farmers who need them, and if so, how they fit together in a farmer's daily life.

A team of researchers from the University of Eldoret and the Regional Universities Forum for Capacity Building in Agriculture set out to answer this question by listening to 386 smallholder farmers in the Busia and Siaya counties. These two regions, located near Lake Victoria, are home to thousands of families who grow crops like maize, beans, and sweet potatoes. The researchers did not just ask which sources were most popular; they asked every farmer to list every single channel they used to learn about sustainable farming. They wanted to see the full picture of a farmer's "knowledge toolkit"—the complete set of sources they trusted and turned to. By mapping these toolkits, the team hoped to understand if the system was broken, or if it was working in a way that simply looked different from what experts expected.

What the researchers found was a landscape of knowledge that is deeply unequal, not because some farmers are left out entirely, but because the path to the most advanced information is blocked by a steep hill. The study revealed that the most common way farmers learn is by talking to each other. Nearly 70 percent of the farmers surveyed said they learned from fellow farmers. Government extension officers were the next most common source, reaching about 54 percent, followed by non-governmental organizations at 45 percent. These sources form the foundation of the system. However, the channels that connect farmers directly to formal science—such as Farmer Field Schools, where groups learn through guided experiments, and academic research institutions—reached very few people. Only about 11 percent of farmers had attended a Farmer Field School, and a mere 6 percent had turned to a research institution for advice.

The crucial discovery, however, was not just that these scientific channels were rare, but who was using them. The researchers expected that perhaps these schools and institutions served a specific, isolated group of farmers who relied on them exclusively. Instead, they found the opposite. Not a single farmer in the entire study named a Farmer Field School or a research institution as their only source of knowledge. In fact, the farmers who did use these advanced channels were already the most connected people in the system. On average, a farmer who used a Farmer Field School had access to five different sources of information, while the average farmer in the study used fewer than three. These scientific channels were not entry points for the unconnected; they were supplements for the already well-connected.

This pattern suggests that the system is stratified, meaning it is layered like a pyramid, rather than segmented, where different groups would use completely different tools. The farmers at the bottom of the pyramid rely on a narrow set of sources, usually just talking to neighbors or a single government officer. As farmers gain more resources, such as higher income, more education, or greater comfort with digital tools, they climb the pyramid. They add more sources to their toolkit, eventually reaching the top where the formal scientific channels sit. The study showed that for every step up in income, a farmer was significantly more likely to have a broader range of information sources. Similarly, farmers who felt comfortable using digital platforms were more likely to have a wider toolkit.

The researchers also discovered a surprising twist in how these networks function. It is often assumed that farmer groups act as a bridge, taking local knowledge and connecting it to scientific experts. The data, however, suggested that for many, these groups act as a closed loop. Farmers who had strong access to local groups were much more likely to learn from their peers, but they were actually less likely to reach out to the formal scientific channels. The peer networks were satisfying their immediate needs so well that they did not feel the need to look further, or perhaps the structure of the groups kept them from looking up. This means that simply training a farmer group does not automatically bring scientific knowledge into that group; the connection must be made deliberately.

Perhaps the most striking gap the study uncovered was between what farmers wanted to learn and what they were getting. When asked how they preferred to learn, nearly 78 percent of farmers said they wanted hands-on field demonstrations. They wanted to see the techniques in action, to touch the soil, and to try things out themselves. Yet, the channels that provide this kind of learning—Farmer Field Schools and agricultural fairs—were the very ones that reached the fewest people. More than 72 percent of the farmers who wanted these demonstrations had no access to them. They were left with radio programs, mobile apps, or written guides, which could not provide the physical experience they craved. This mismatch was not due to a lack of interest in technology; many farmers used social media and radio, but they found that these tools could not replace the need to see and do.

The study also highlighted that the two counties, while similar in their farming, had different paths to the same result. In Busia, the system was driven heavily by external non-governmental organizations, which reached a large portion of the farmers there. In Siaya, the system relied more on local community groups and had slightly better connections to research institutions. Despite these different intermediaries, the underlying structure remained the same: a broad base of peer-to-peer learning, with the scientific channels sitting at the very top, accessible only to those who had already climbed the ladder.

The implications of these findings are clear for anyone trying to improve farming in the region. If the goal is to bring advanced agroecological knowledge to the farmers who need it most, simply building more Farmer Field Schools or launching more research programs will not work if they remain at the top of the pyramid. These channels cannot serve as the first point of contact for a farmer who has no other sources. Instead, programs must use the channels that farmers already trust and use—neighbors, local groups, and government officers—as the entry points to guide farmers toward the more advanced knowledge. The study suggests that the horizontal sharing of knowledge among farmers is already working well, but the vertical connection to scientific expertise is missing. Without a deliberate effort to bridge that gap, the farmers who are already struggling the most will remain on the bottom rung, unable to reach the tools that could help them thrive.