The purpose of this paper is to document smallholder experiences during a
participatory experimental initiative and draw useful lessons for field practitioners. The main
methods used to collect data were participant observation, in-depth interviews among 40 farmers,
and analyses of notes taken during participatory monitoring and evaluation. These farmers were
regular participants in four collective trials*designed by scientists, managed by smallholders*
that were run for over five years in western Kenya. This paper shows that scientific trial designs
enhanced ‘conviction’ among smallholders in new technologies that were being experimented. The
study further reveals that (1) collective trials must last long enough for interactive learning to
effectively occur, (2) farmers’ selected aspects of experiments and did not generally treat the
whole trial as one concept, (3) field trials were forms of evidence of research or project activity,
and (4) trials were seen as like ‘churches’ and multipurpose congregating sites, especially because
these collective experiments had roots in local traditions.
The practical implication is that because interactive trials are social events, like other learning
platforms, they must be appropriately situated within local contexts to enhance their relevance.
This paper also shows that experiments are only a first step in learning, which needs to be
supported by in-depth research and support for smallholder experiments. Collaborative
experimentation is an effective tool for improving smallholder knowledge. Experimentation is a
hands-on tool; it strengthens interactive learning, enhances understanding of a concept and
improves ownership of the process or research among farmers. However, longer-term collective
experiments develop a strong underlying human dimension. Understanding the social dynamics
that influence the value of collective experimentation is critical for the process of scaling out